Citation
Evolution of forest fragmentation in temperate and tropical forests in Mexico for 2002, 2008, and 2013

Material Information

Title:
Evolution of forest fragmentation in temperate and tropical forests in Mexico for 2002, 2008, and 2013
Creator:
Clay, Elizabeth ( author )
Language:
English
Physical Description:
1 electronic file (77 pages) : ;

Subjects

Subjects / Keywords:
Forest biodiversity -- Mexico ( lcsh )
Forest conservation ( lcsh )
Fragmented landscapes ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Review:
Mexico is considered to be “megadiverse” due to its rich biodiversity, therefore studying forest fragmentation in Mexico is fundamental in protecting the country’s rich biodiversity as well as the forest ecosystem functions and services. This thesis develops a national level assessment of the fragmentation of the temperate and tropical forests in Mexico for three dates 2002, 2008 and 2013 (corresponding to the land use/cover layers known as Series III, IV and V created by the Instituto Nacional de Estadística y Geografía INEGI in Mexico). Then, the forest fragmentation classes identified for a date (e.g. 2002) are cross-referenced with the forest fragmentation classes identified for the next date (e.g. 2008) with the purpose of assessing how the forest area that falls in each of the fragmentation classes identified in this study are evolving in time. The same cross-reference process is carried out for 2008-2013 to analyze the changes during this period. This analytical process explores the trends in the relationship between fragmentation levels and changes in fragmentation of the forests that occur in different forest fragmentation classes over time. The Morphological Spatial Pattern Analysis (MSPA) method and the GUIDOS Toolbox will be used to identify the forest fragmentation classes. The INEGI Series III, IV and V land cover layers are used to extract the areas covered by temperate and tropical forests. A raster overlay with unique identifiers technique and the Tabulate Areas Tool in ArcGIS were used to carry out the cross-reference analysis. Understanding forest fragmentation is integral for biodiversity conservation. The results from this study can enhance the prioritization of conservation efforts of the remaining forests in Mexico.
Thesis:
Thesis (M.S.)- University of Colorado Denver.
Bibliography:
Includes bibliographic references
System Details:
System requirements: Adobe Reader.
General Note:
Department of Geography and Environmental Sciences
Statement of Responsibility:
by Elizabeth Clay.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
945638408 ( OCLC )
ocn945638408
Classification:
LD1193.L547 2015m C53 ( lcc )

Downloads

This item has the following downloads:


Full Text
EVOLUTION OF FOREST FRAGMENTATION IN TEMPERATE AND TROPICAL
FORESTS IN MEXICO FOR 2002, 2008, AND 2013
by
ELIZABETH CLAY
B.S., University of North Carolina Wilmington, 2012
A thesis submitted to the
Faculty of the Graduate School of the
University of Colorado in partial fulfillment
of the requirements for the degree of
Master of Science
Environmental Science
2016


This thesis for the Master of Science degree by
Elizabeth Clay
has been approved for the
Environmental Science Program
by
Rafael Moreno, Chair
Peter Anthamatten
Juan Manuel Torres-Rojo
November 18, 2015
n


Clay, Elizabeth (MS, Environmental Science)
Evolution of Forest Fragmentation in Temperate and Tropical Forests in Mexico for 2002,
2008, and 2013
Thesis directed by Associate Professor Rafael Moreno
ABSTRACT
Mexico is considered to be megadiverse due to its rich biodiversity, therefore
studying forest fragmentation in Mexico is fundamental in protecting the countrys rich
biodiversity as well as the forest ecosystem functions and services. This thesis develops a
national level assessment of the fragmentation of the temperate and tropical forests in
Mexico for three dates 2002, 2008 and 2013 (corresponding to the land use/cover layers
known as Series III, IV and V created by the Instituto Nacional de Estadistica y Geografia
INEGI in Mexico). Then, the forest fragmentation classes identified for a date (e.g. 2002) are
cross-referenced with the forest fragmentation classes identified for the next date (e.g. 2008)
with the purpose of assessing how the forest area that falls in each of the fragmentation
classes identified in this study are evolving in time. The same cross-reference process is
carried out for 2008-2013 to analyze the changes during this period. This analytical process
explores the trends in the relationship between fragmentation levels and changes in
fragmentation of the forests that occur in different forest fragmentation classes over time.
The Morphological Spatial Pattern Analysis (MSPA) method and the GUIDOS Toolbox will
be used to identify the forest fragmentation classes. The INEGI Series III, IV and V land
cover layers are used to extract the areas covered by temperate and tropical forests. A raster
overlay with unique identifiers technique and the Tabulate Areas Tool in ArcGIS were used
to carry out the cross-reference analysis. Understanding forest fragmentation is integral for
m


biodiversity conservation. The results from this study can enhance the prioritization of
conservation efforts of the remaining forests in Mexico.
The form and content of this abstract are approved. I recommend its publication.
Approved: Rafael Moreno
IV


ACKNOWLEDGMENTS
I would like to express my gratitude to my advisor and thesis chair, Dr. Rafael
Moreno who provided me endless support throughout my research through his immense
knowledge, patience, and passion for the subject. His guidance helped me develop my
research from the start to the finished product.
I would also like to thank the rest of my thesis committee; Dr. Peter Anthamatten and
Dr. Juan Manuel Torres Rojo for taking their time to contribute their individual expertise to
better my thesis.
v


TABLE OF CONTENTS
CHAPTER
I. INTRODUCTION............................................................1
II. REVIEW 01 LITERATURE....................................................3
Forest Fragmentation....................................................3
Characterizing Spatial Patterns of Forests..............................7
Effects of Geographic Scale on Forest Fragmentation....................11
Conclusions............................................................12
III. METHODOLOGY........................................................... 13
Introduction...........................................................13
Data Sets..............................................................13
Morphological Spatial Pattern Analysis.................................15
Analyzing the Transitions Between Fragmentation Classes During the 2002-
2008 and 2008-2013 Periods.............................................20
IV. RESULTS................................................................23
Fragmentation Levels in 2002, 2008, and 2013 with Intext Parameter Off... 23
Fragmentation Levels in 2002, 2008, and 2013 with Intext Parameter On.... 26
Transitions Between Fragmentation Classes During 2002-2008 and 2008-
2013 Periods Without Differentiating Between Internal and External
Fragmentation Classes..................................................30
Transitions Between Fragmentation Classes During 2002-2008 and 2008-
2013 Periods Differentiating Between Internal and External Fragmentation
Classes................................................................39
V. DISCUSSION.............................................................47
Defining Fragmentation Classes of Temperate and Tropical Forests.47
vi


Transitions Between Fragmentation Classes During 2002-2008 and 2008-
2013 Periods Without Differentiating Between Internal and External
Fragmentation Classes.......................................................
49
Transitions Between Fragmentation Classes During 2002-2008 and 2008-
2013 Periods Differentiating Between Internal and External Fragmentation
Classes..................................................55
Limitations and Sources of Error.........................55
VI. CONCLUSIONS AND RECOMMENDATIONS..........................57
REFERENCES...............................................62
vii


LIST OF TABLES
TABLE
1. Vegetation Types from INEGI's Series III, IV, and V included in the definition of
temperate forests and tropical forests............................................14
2. Summary table of input parameters for MSPA.........................................19
3. MSPA outputs with unique values for each fragmentation class...........................20
4. Areas of each fragmentation class for temperate and tropical forests in 2002..24
5. Areas of each fragmentation class for temperate and tropical forests in 2008. 24
6. Areas of each fragmentation class for temperate and tropical forests in 2013..25
7. Areas of each external and internal fragmentation class for temperate and tropical forests
in 2002...............................................................................27
8. Areas of each external and internal fragmentation class for temperate and tropical forests
in 2008.............................................................................. 28
9. Areas of each external and internal fragmentation class for temperate and tropical forests
in 2013...............................................................................29
10. Fragmentation classes transition matrix for tropical forests in the 2002-2008 period
without differentiating between internal and external classes.....................31
11. Fragmentation classes transition matrix for temperate and forests in the 2002-2008 period
without differentiating between internal and external classes.....................32
12. Fragmentation classes transition matrix for tropical forests in the 2008-2013 period
without differentiating between internal and external classes.........................32
13. Fragmentation classes transition matrix for temperate forests in the 2008-2013 period
without differentiating between internal and external classes.....................33
14. Fragmentation classes transition matrix for tropical forests in the 2002-2008 period
differentiating between internal and external classes.................................40
15. Fragmentation classes transition matrix for temperate forests in the 2002-2008 period
differentiating between internal and external classes.................................42
viii


16. Fragmentation classes transition matrix for tropical forests in the 2008-2013 period
differentiating between internal and external classes...............................43
17. Fragmentation classes transition matrix for temperate forests in the 2008-2013 period
differentiating between internal and external classes...............................45
IX


LIST OF FIGURES
FIGURES
1. Seven basic output classes from the MSPA analysis..................................15
2. Foreground connectivity of eight versus four......................................16
3. Effects of increased edgewidth on MSPA results.....................................17
4. Effects of transition pixels on (left) or off (right) on MSPA results..............18
5. Effects of intext on (left) and intext off (right) on MSPA results.................19
6. MSPA output for tropical forests in 2002. Similar cartographic outputs were generated
for 2008 and 2013 temperate and tropical forests..................................26
7. MSPA cartographic output for fragmentation classes of the tropical forests in 2002
differentiating between external and internal fragmentation classes. Similar outputs were
created for the 2008 and 2013 temperate and tropical forests....................30
8. Example of fragmentation classes in 2008 that were edge forest in 2002 for tropical
forests. Similar results were generated for temperate and tropical forests for all dates and
all fragmentation classes.......................................................34
9. Example of fragmentation classes in 2008 that were external edge in 2002 for tropical
forests. Similar results were generated for temperate and tropical forests for all dates and
all fragmentation classes.......................................................46
x


CHAPTER I
INTRODUCTION
The effects of forest fragmentation in Mexico have been an extensively researched topic,
particularly for specific plants or animals at a local or regional scale. However, forest
fragmentation research in Mexico at the national scale focuses only on specific types of
forests (Trejo & Dirzo 2000). Additionally, local and national studies do not use a common
methodology to compare results over time. Lastly, there is a need for spatial reference when
analyzing forest fragmentation to better understand influences on fragmentation and changes
over time. This analysis will address these shortcomings by providing a spatially referenced
national level assessment analyzing the changes in forest fragmentation over time for both
temperate and tropical forests in Mexico. The study will use the most up to date land cover
data for Mexico, and identify forest fragmentation classes using the latest methodology and
tool that has been sanctioned and agreed by a large number of national and international
forestry agencies. The purpose of this study is to provide a national level assessment of the
state of forest fragmentation for temperate and tropical forests in Mexico for 2002, 2008, and
2013, as well as analyze changes in fragmentation classes of temperate and tropical forests in
Mexico from 2002-2008 and again from 2008-2013 in relation to the fragmentation class of
the previous date, and to determine the area of each of these new forest fragmentation
classes. This study will answer the following research questions:
1. What are the levels of forest fragmentation for both temperate and tropical forests in
2002, 2008 and 2013 in Mexico?
1


2. How has each fragmentation class changed from 2002-2008 and 2008-2013 in
temperate and tropical forests in relation to the fragmentation class of the previous
date?
The results of this analysis will include maps of forest fragmentation for each date (2002,
2008, and 2013), as well as maps of the changes in fragmentation classes from one date to
the next. Also, tabular data will be provided to accompany each of these cartographic
products in order to visualize and quantify changes in forest fragmentation in Mexico over
time. The rest of this thesis is organized as follows:
Review of Literature
Methodology
Results
Discussion
Conclusions and Recommendations
2


CHAPTER II
REVIEW OF LITERATURE
Forest Fragmentation
Introduction to forest fragmentation
Forest fragmentation, or habitat fragmentation, is a very broad term given to many
different processes and effects of landscape change (Lindenmayer & Fischer 2007). It is
used as an umbrella term for ecological processes, patterns of vegetation cover, and biotic
responses that result from changes in landscape patterns (Lindenmayer & Fischer 2007).
There are many ways to define forest fragmentation, but it can be generally defined as a
process during which, a large expanse of habitat is transformed into a number of smaller
patches of smaller total area, isolated from each other by a matrix of habitats unlike the
original (Wilcove et al. 1986). However, there are numerous publications on the efforts to
define fragmentation and on the theoretical approaches that have been taken to understand
and assess it (Harrison & Bruna 1999, Haila 1999, Kupfer 2006, Lindenmayer & Fischer
2007, and Kupfer 2012). Forest fragmentation is one of the most researched topics by
conservation biologists (Fazey et al. 2005). There is a substantial body of literature studying
the effects of fragmentation on the functioning of ecosystems and conservation of diverse
flora, fauna, and ecological processes at different temporal and spatial scales (Laurance et al.
2011, Harrison & Bruna 1999, Turner 1996, Donovan & Flather 2002, Schmiegelow &
Monkkonen 2002, Thompson et al. 2002, Tscharntke et al. 2002, Fahrig 2003, Groom et al.
2005, and Schleuning et al. 2011). Forests can become fragmented due to a variety of
reasons such as expanded agricultural land, urban settlements, transportation infrastructure,
and fire occurrence (Estreguil et al. 2012). Landscape modification and habitat (including
3


forests) fragmentation are key drivers of global species and biodiversity loss, and are
believed to negatively affect virtually all taxonomic groups of animals and plants, as well as
key ecosystem components and functions for long periods of time (Tabarelli et al. 2004,
Vellend et al. 2006, and Fischer & Lindenmayer 2007). Increase in fragmentation has also
been identified as a major threat to the conservation of forest ecosystems as a whole (Murcia
1995, Laurance et al. 2000, Ribeiro et al. 2009, Harper et al. 2005, and Laurance et al. 2011).
Additionally, deforestation and forest fragmentation contribute to processes such as the
greenhouse effect and climate change, and result in negative changes in regional hydrology
and biogeochemical cycles (Mas et al. 2004).
Extent of forest fragmentation in Mexico
There have been a variety of studies examining the extent of forest fragmentation in
Mexico, particularly at local and regional scales. Many of these studies are focused on
Mexicos southern tropical forests. In Mexicos southern tropical forests, the expansion of
agriculture and urban developments have contributed to increased forest fragmentation
(Ochoa-Gaona & Gonzalez-Espinosa 2000). Landsat satellite imagery has been used to
determine annual deforestation rates and spatial patterns in the Chiapas highlands in Chiapas,
Mexico. Using Landsat imagery to determine deforestation rates in the Chiapas highlands, it
was determined that this area of Mexico is experiencing deforestation at a faster rate than the
national average (Ochoa-Gaona & Gonzalez-Espinosa 2000). Converting forests to
agricultural uses has resulted in the amount of primary forest to decrease while the area of
disturbed forest, secondary vegetation, and developed areas have increased (Ochoa-Gaona et
al. 2006). Landsat imagery analyzed during the period of 1975-2000 in the Chiapas
highlands determined that the amount of continuous forest cover has decreased, while the
4


amount of fragmented forest patches increased (Cayuela et al. 2006). Additionally, the
deforestation rates continuously increased over time in the Chiapas Highlands (Cayuela et al.
2006). These studies using satellite imagery can estimate the deforestation rates and general
spatial patterns of forests (such as patch sizes and degree of isolation), but the specific degree
of fragmentation of the forests cannot be analyzed using these methods.
Trejo & Dirzo (2000) conducted a national level assessment of changes in seasonally
tropical dry forests in Mexico in the early 1990s using land cover maps classified as: intact
seasonally dry tropical forest, or forest under good conservation status, altered seasonally
dry tropical forests (forest fragments), degraded seasonally dry tropical forests (smaller and
more isolated fragments), or converted seasonally dry tropical forests. Although these results
provide a national level assessment, they focus only on one forest type (seasonally dry
tropical forests) and do not specify the specific types of spatial patterns of the fragmented
forest, only whether or not the forest was fragmented or not.
National level assessments on patterns of forest fragmentation in Mexico have been
performed in the past (Moreno-Sanchez et al. 2014, 2012, and 2011). Moreno-Sanchez
(2011) provides tabular data on the change in forest fragmentation in Mexico over time, but
no spatial reference. Moreno-Sanchez (2012) provided a national level assessment on
changes forest fragmentation in Mexico in relation to anthropogenic pressure. Updated land
cover data has been released since the previous papers on forest fragmentation in Mexico
have been published. Additionally, a new methodology for defining the spatial pattern of the
forests will be used in this study in which more specific fragmentation classes can be
determined. This methodology has been sanctioned and agreed upon by a large number of
national and international forestry agencies.
5


Overall, there is much more extensive research regarding forest fragmentation and
deforestation in the southern tropical forests of Mexico than the rest of the nation.
Additionally, there is much more research quantifying the deforestation rates rather than
providing quantifiable metrics regarding spatial patterns of forests in Mexico.
Effects of forest fragmentation in Mexico
Due to its rich biodiversity, Mexico is considered to be megadiverse and is one of
the five biologically richest countries (Groombridge and Jenkins 2000), therefore, studying
forest fragmentation in Mexico is immensely important in protecting the countrys rich
biodiversity and ecosystem services. There have been many studies examining the effects of
forest fragmentation on specific species or ecosystems in Mexico. For example, the tropical
rainforests in Los Tuxtlas, Veracruz are heavily fragmented due to anthropogenic activities,
particularly agriculture and the conversion of the forest to pasture for cattle grazing (Myers
1988). This fragmentation and disappearance of rain forest in Los Tuxtlas has resulted in 80-
90 percent reduction of the area of natural habitat for howler monkeys (Estrada & Coates-
Estrada 1996). Dung beetles in Los Tuxtlas are dependent on howler monkey dung, which
contribute to nutrient cycling in the soil. With a decrease in howler monkeys due to reduced
habitat, dung beetle populations could experience a decrease and alter the biogeochemical
cycles in the rainforest (de Mexico 1999). Additionally, species richness of bat species in the
Los Tuxtlas region is more likely to decline as forest fragments become more isolated
(Estrada et al. 1993). Since bat species in the order of Chiroptera constitute 40-50 percent of
the mammal species in the region, they are vital for species richness and biodiversity of
mammals in these ecosystems (Estrada et al. 1993). Fragmented forest landscapes in
Mexicos southern tropical forests also effects seedling dispersal. Melo et al. (2010) found
6


that most larger mammals were eliminated from forest fragments less than 30 hectares in size
therefore influencing seed dispersal of larger seedlings and effecting the regeneration of the
forests (Melo et al. 2010). Again, there is an abundance of research describing the patterns
and effects of forest fragmentation in the specific regions and forest types in Mexico, and a
lack of literature on forest fragmentation on the national scale.
Characterizing Spatial Patterns of Forests
Introduction
It is important to not only consider the size of forest patches when researching forest
fragmentation, but to quantify the degree of fragmentation by measuring the spatial pattern of
the forest on the landscape (Fahrig 2003). The spatial pattern of the forest can be defined as
the spatial distribution of the forest across the landscape (Estreguil et al. 2012). As
mentioned previously, many studies on forest fragmentation examine patch size and degree
of isolation as determinates of forest fragmentation. Examining only forest patch size as a
measure of fragmentation does not always accurately portray the effects of forest
fragmentation on the ecosystem (McAlpine & Eyre 2002). When analyzing the effects of
forest fragmentation on woody vegetation in the Chiapas highlands, it was found that
fragmentation size, shape of the forest fragments, and possibly the forest matrix all
contributed to the species richness within forest fragments (Ochoa-Gaona et al. 2006).
Ochoa-Gaona et al. (2006) suggested that there may be too much emphasis on the size and
isolation of forest patches in forest conservation plans, and the need of more emphasis on the
effects of the shape of forests fragments on biodiversity.
7


Comparison of methods
Two major concepts of characterizing spatial patterns of forest have been
implemented through different methodologies in different areas of the world. These
applications do not only determine fragmentation based on patch size and isolation of forest
patches, but they determine the different spatial patterns of patches of forests as well as
differentiate between external and internal fragmentation. Differentiating between external
and internal forest fragmentation is an important aspect of describing fragmentation because
internal forest fragmentation increases the amount of edge within the forest, rather than on
the boundary of the external forest patch (Vogt et al. 2007). These interior edge effects may
impact species differently than edge effects on exterior forests (Zipperer 1993).
The first method is a landscape level estimation of spatial patterns of fragmentation.
Metzger and Muller (1996) used Landsat Thematic Mapper (TM) images to create indices
that reflect the typology and complexity of forest boundaries in southeastern Brazil. This
methodology was able to determine the complexity of elongated and fragmented forest
patterns in Brazil (Metzger & Muller 1996). Metzger and Decamps (1997) used similar
methodology to derive landscape level estimations of the spatial patterns of fragmentation to
determine connectivity of the forests. Their model was based on three components of the
landscape; percolation measurement, the complexity and quality of corridor networks, and
the permeability of the habitat matrix (Metzeger & Decamps 1997). Incorporating the
complexity of the forest fragments allows for a better understanding of how forest
fragmentation effects species survival and biodiversity. Bogaert et al. (2004) created an
algorithm to define fragmentation classes differentiating between internal and external
fragmentation at the landscape level through observed changes in patch area, number of
8


patches, and patch perimeter in different landscapes. However, using patches to define
fragmentation across large geographic areas (typically landscape level) may produce
inaccurate estimates of patch size and shape due to the fact that there will be a large amount
of patches in the spatial extent, which may require altering of the data and losing information
about the small forest patches (Vogt et al. 2007). Additionally these methods of defining
forest fragmentation can assign landscapes with different arrangements of forest the same
value of fragmentation (Neel et al. 2004).
An alternative to landscape level estimation of spatial patterns of forest fragmentation
is pixel level mapping of internal and external forest fragments. Pixel level mapping of
forest fragmentation offers higher sensitivity to forest shapes than landscape level analysis
and can more accurately portray changes over time (Vogt et al. 2007). One application of
pixel level mapping is through a method developed by Riitters et al. (2000). This method is
based on image convolution and uses windows surrounding each forest pixel to measure the
amount of forest and its occurrence as adjacent forest pixels (Riitters et al. 2000). The model
classifies each forest location according to the type of fragmentation that exists in the
surrounding landscape that can be defined at different scales (different window sizes). This
defines forest fragmentation as a property of the landscape that contains the forest, rather as a
property of the forest itself (Riitters et al. 2002). This methodology results in specific forest
fragmentation classes such as interior, perforated, edge, transitional, and patch (Riitters et al.
2000). The moving windows methodology has been used in several different landscape scale
studies. First, this method was used to analyze global forest fragmentation (Riitters et al.
2000). It was also used to determine forest fragmentation in the continental United States
9


(Riitters et al. 2002) as well as to analyze forest fragmentation changes over time in Mexico
(Moreno-Sanchez et al. 2014).
Another primary method used for determining the spatial patterns of forests using
pixel level mapping is through morphological image processing. Morphological image
processing analyzes the shape and form of objects in an image at the pixel level of a binary
map, offering the same advantages of mapping at the pixel level as image convolution in a
moving window (Soille & Vogt 2008). However, morphological image processing is not
dependent on the size of window used and therefore offers more accuracy across all
geographic scales at the pixel level than the image convolution technique (Vogt et al. 2007).
When comparing image convolution versus morphological image processing when
evaluating forest fragmentation in a national park in Italy, the morphological image
processing remained more accurate with changes in scale of the input map than the image
convolution. For example, small patches of forest remained small patches, regardless of the
scale allowing for accurate landscape level analysis at the pixel scale (Vogt et al. 2007).
Morphological image processing has been used to evaluate forest fragmentation and
connectivity in Europe (Estreguil et al. 2012) and by the United States Department of
Agriculture (USDA) to evaluate forest fragmentation across the United States as well as in
select cities (available online at http://www.forestthreats.org/research/tools/landcover-
maps/mspa). It is important to note that with pixel level mapping, the resolution (pixel size)
used for analysis is important in influencing the results, and input maps at different
resolutions should not be used to directly compare results (Ostapowicz et al. 2008). For
example, an input map with a pixel size of 250 meters by 250 meters is going to define the
spatial patterns of the forest differently than an input map of the same geographic area with a
10


pixel size of 100 meters by 100 meters. The 250 meter resolution will capture more forest
area per pixel than the 100 meter resolution, so therefore these two results should not be used
to make comparisons of the spatial patterns of the same geographic forest area.
Based on previous studies and available literature, pixel level mapping of forest
fragmentation using morphological image processing is the most accurate at the landscape
scale. Therefore, the Morphological Spatial Pattern Analysis (MSPA) contained in the
GUIDOS Toolbox (Soille & Vogt 2008) will be used in this study to define forest
fragmentation in Mexico.
Effects of Geographic Scale on Forest Fragmentation
The geographic scale at which forest fragmentation studies are performed vary
greatly, which impacts the conclusions on the degree of fragmentation and its effects on the
ecosystem (Fahrig 2003). Typically, studies performed at a smaller geographic scale indicate
more interior forest than studies performed at a larger scale (Riitters et al. 2000). For
example, when Riitters et al. (2000) used 81 square kilometer windows for a global
assessment on forest fragmentation, about one-third of the total global forest is characterized
as interior, whereas forests characterized as an edge or patch dominate when using 59,049
square kilometer windows. It is important to analyze forest fragmentation at large scales in
order to define the level of fragmentation at the landscape level (Riitters et al. 2000). With
the increased availability of satellite data, it is possible to examine forest patterns at larger
spatial and temporal scales (Hall et al. 1991). Landscape scale analysis of forest patterns can
relate fragmentation to the reduction of the area of the remaining forest patches, increased
isolation of the fragments, loss of overall connectivity, and increased edge effect, and
disturbances from surrounding pressures (Garcia-Gigorro & Saura 2005). The resulting
11


spatial patterns of the forest on a landscape scale can be used to determine how changes in
the spatial pattern impacts ecosystem processes such as habitat provision, gene flow,
pollination, wildlife dispersal, or pest propagation (Estreguil et al. 2012).
Conclusions
Overall, there have been numerous studies describing the effects of forest
fragmentation in Mexico on specific species or ecosystems. These studies provide insight on
the importance of habitat conservation, but do not provide a national level analysis on the
state of forest fragmentation in Mexico. Many studies in Mexico do not quantify forest
fragmentation in relation to the spatial patterns of the forest. Rather, forest fragmentation is
commonly described as the size, amount of core area, and degree of isolation of forest
patches. Previous studies examining forest fragmentation in Mexico on a national scale do
not provide spatial reference, only describe one forest type, or do not use the most updated
MSPA methodology. Additionally, there is no literature relating the changes in area in each
fragmentation class to all other fragmentation classes over time to get a more specific picture
of how the fragmentation of the forests are changing in Mexico. This study will fill these
gaps in the literature by providing a national level assessment of forest fragmentation in
Mexico using the most updated MSPA methodology in order to describe forest fragmentation
in terms of the spatial patterns of the forest. Additionally, this study will relate each
fragmentation class in one date to the fragmentation classes of the next date during the two
time periods of 2002-2008 and 2008-2013 in order to determine specifically how forest
fragmentation is evolving over time in Mexico. Lastly, this study will provide cartographic
results for the classes of forest fragmentation as well as for the transitions of each forest
fragmentation class from one date to the next.
12


CHAPTER III
METHODOLOGY
Introduction
The research questions in this study are answered using land cover/land use data for
Mexico. Several tools are used to carry out the two-fold research questions. First, the
geographic information system ArcGIS 10.2 (ESRI, Redlands CA) was used to prepare the
land cover/land use data to be able to define the levels of fragmentation. Next, the levels of
forest fragmentation are defined using the freely available Morphological Spatial Pattern
Analysis (MSPA) contained in the GUIDOS Toolbox (Soille & Vogt 2008). ArcGIS was
then used again to perform the cross referencing of the fragmentation class from one date to
the fragmentation classes of the following date. The specific processes are detailed below.
Data Sets
Land use and land cover data from the National Institute of Geography and
Informatics (INEGI) in Mexico (INEGI2014, 2012, 2009). The most recent versions of land
use and land cover data from INEGI will be used for the analysis. The original data are in
vector format and consists of all vegetation types and land uses in Mexico for the years 2002,
2008, and 2013, known as Series III, Series IV, and Series V respectively. INEGI has
homogenized the land use/ land cover classes in these layers, hence temporal changes can be
evaluated. The INEGI land cover vector data are at a scale of 1: 250,000 and provided in
Lambert Conformal Conic GRS 1980 projection with linear units of meters. For the purpose
of a national level assessment, the specific land use/land cover classes were reclassified to
the two forest types of temperate and tropical forests for each date. Table 1 specifies the
original vegetation types included in the temperate and tropical forests classes defined in this
13


study. The temperate and tropical forests vector data sets were converted to raster format
using a cell size of 250 x 250 meters for each date. The cell size was chosen based on the
scale of the original data layers (1:250,000), the level of locational certainty of the features in
the original layers, and the consideration that, although responses by plants, animals and
ecosystem functions to edge effects vary widely, 250 meters represents a conservative
estimate of the level of penetration of edge effect for many species and processes (Harper et
al. 2005).
Table 1. Vegetation Types from INEGI's Series III, IV, and V included in the definition of
_________________________temperate forests and tropical forests_____________________
Temperate Forests
INEGI's CVE UNION code Description
BA Bosque de Oyamel
BB Bosque de Cedro
BC Bosque Cultivado
BG Bosque de Galeria
BI Bosque inducido
BJ Bosque de Tascate
BM Bosque Mesofilo de Montana
BP Bosque de Pino
BPQ Bosque de Pino-Encino
BO Bosque de Encino
BQP Bosque de Encino-Pino
BS Bosque de Ayarin
Tropical Forests
SAP Selva Alta Perennifolia
SAQ Selva Alta Subperennifolia
SBC Selva Baja Caducifolia
SBK Selva Baja Espinosa
SBP Selva Baja Perennifolia
SBQ Selva Baja Subperennifolia
SBS Selva Baja Subcaducifolia
SG Selva de Galeria
SMC Selva Mediana Caducifolia
INEGI's CVE UNION code Description
SMP Selva Mediana Perennifolia
SMQ Selva Mediana Subperennifolia
SMS Selva Mediana Subcaducifolia
VSA/PT Vegetacion Secundaria de Selvas Arborea/Vegetacion de Peten
VSA/SAP Vegetacion Secundaria de Selvas Arborea/Selva Alta Perennifolia
VSA/SAQ Vegetacion Secundaria de Selvas Arborea/Selva Alta Subperennifolia
VSA/SBK Vegetacion Secundaria de Selvas Arborea/Selva Baja Espinosa
VSA/SBQ Vegetacion Secundaria de Selvas Arborea/Selva Baja Subperennifolia
VSA/SG Vegetacion Secundaria de Selvas Arborea/Selva de Galeria
VSA/SMQ Vegetacion Secundaria de Selvas Arborea/Selva Mediana Subperennifolia
VSA/SMS Vegetacion Secundaria de Selvas Arborea/Selva Mediana Subcaducifolia
VSA/BS Vegetacion Secundaria de Selvas Arborea/Bosque de Ayarin
14


Morphological Spatial Pattern Analysis
The European Union FORESTMOD Joint Research Centers Graphical User
Interface for the Description of image Objects and their Shapes (GUIDOS) Toolbox was
used to answer the first research question, What are the levels of forest fragmentation for
both tropical and temperate forests in 2002, 2008, and 2013 in Mexico. The Morphological
Spatial Pattern Analysis (MSPA) (Soille & Vogt 2008) within the GUIDOS Toolbox was
used to define the levels of forest fragmentation for each date in Mexico. The MSPA
determines forests fragmentation classes based on the geometry and connectivity of a binary
input image as well as user-specified input parameters (Soille and Vogt, 2008). The MSPA
results in seven basic classes based on pixel value. The seven classes include:
1. Core: Interior foreground area excluding foreground perimeter
2. Islet: Disjoint foreground object and too small to contain core
3. Loop: Connected at more than one end to the same Core area
4. Bridge: Connected at more than one end to different Core areas
5. Perforation: Internal foreground object perimeter
6. Edge: External foreground object perimeter
7. Branch: Connected at one end to Edge, Perforation, Bridge, or Loop
Figure 1. Seven basic output classes from the MSPA analysis (Soille & Vogt 2008)
15


The MSPA requires a specific input format in order to run the analysis. Using
ArcGIS, the temperate and tropical forest layers for each date were formatted to input into
the MSPA. First, each raster was reclassified to a binary image containing a background (no
forest, value = 1) and a foreground (forest, value = 2). Each raster was then be exported as
an 8-bit GeoTiff with no compression to be imported into the GUIDOS Toolbox. In the
GUIDOS Toolbox, an MSPA batch process was performed for temperate and tropical forests
for each date. User-defined input parameters must be selected when running the MSPA.
The first parameter option is the foreground connectivity, options of eight or four.
This option determines how the center pixel, in sets of three by three pixels, is connected to
its adjacent neighboring pixels (Vogt, n.d.). A foreground connectivity of four results in
neighboring pixels sharing a common pixel border, and a foreground connectivity of eight
results in neighboring pixels sharing a common pixel border and comer (Vogt, n.d.). Figure
2 shows the difference between a foreground connectivity of type eight and four cells. For
this analysis, a foreground connectivity of eight was used.
The second parameter option is the edgewidth. The edgewidth defines the thickness
of the non-core classes in pixels in the analysis, to define what is considered an edge
fragmentation class.
#1 Zoom 1 ___ ^ = 1 #2 Zoom 1 a |core m Edgv | Perforation | Bridge toop 1 | Branch | Islet f | Background | | No data
nfoai iXrrnjflfl.

8-connectivity 4-connectivity MSPA segmentation: 8100 <-> 4100
Parameter 1: Foreground Connectivity ( Figure 2. Foreground connectivity of eight versus four (Vogt n.d.).
16


The actual distance of the edgewidth is the number of defined edge pixels multiplied by the
resolution of the input data. Increased edgewidth results in a larger edge area and smaller
core area. An edgewidth of one was used for this analysis, meaning the edge will have a
thickness of 250 meters since the input raster resolution is 250 x 250 meters. Figure 3
displays the effects of edgewidth on the MSPA results.
| Core
| Edge
| Perforation
| Bridge
n Loop
| Branch
| Islet
I I Background
| | No data
EdgeWldth: 1 EdgeWidth: 3 Edge Width: 9
Figure 3. Effects of increased edgewidth on MSPA results (Vogt n.d.).
The third parameter option is transition. Transition pixels are the pixels of an edge or
perforation where the core area intersects with a loop or a bridge. Turning the transition
pixels off hides the transition pixels and the perforation and edge will be closed core
boundaries. Turning the transition pixels on illustrates all detected connections, but may
result in a more confusing image (Vogt, n.d.). Transition pixels were left on for this analysis
in order to display all connections. Figure 4 displays the effects of leaving transition pixels
on and turning the transition pixels off.
The fourth parameter option is intext. The intext option distinguishes internal
features from external fragmentation classes. Internal classes are features enclosed by a
perforation. When intext is enabled, it adds a value of 100 to all outputs to separate internal
and external features. For example, an external core would be assigned a value of
seventeen, while an internal core would be assigned a value of 117.
17


' V #L Zoom I
| Cora
| Edge
| PeUwaUon
| Bridge
| |Loop
| Branch
| Islei
| | Background
| | No data
Figure 4. Effects of transition pixels on (left) or off (right) on MSPA results (Vogt n.d.).
This analysis was performed with both intext on and intext off to compare the results of
distinguishing and not distinguishing between internal and external forest classes. It is
important to note that even with the intext parameter off, the MSPA still distinguishes
perforations within the interior of a forest and forest fragments within a perforation, just not
as detailed as with the intext parameter set to on. This still allows perforations within the
interior of a forest to be differentiated than forest fragments separate from interior forests,
which introduce different edge effects and ecological implications than forest patches outside
of interior forests (Zipperer 1993). Figure 5 displays the effects of the intext parameter.
The fifth and final parameter option is statistics. The statistics option produces a text
file of summary statistics from the MSPA. Statistics was turned on for this analysis. Table 2
provides a summary of all input parameters that will be used for the MSPA in this analysis.
The MSPA results in GeoTiff raster images for temperate and tropical forests for all
dates. Once complete, the output GeoTiffs were imported back into ArcGIS for analysis.
Each value in the MSPA output represents a unique identifier for the different forest
fragmentation classes, and these unique values were used to perform analyses.
18


Parameter 4: Imexi (8,1,1, ). Left; Imexi on (8111). Right; Imexi off (8110).
Figure 5. Effects of intext on (left) and intext off (right) on MSPA results (Vogt n.d.).
Table 2, Summary table of input parameters for MSPA.
Parameter Input for Analysis
Foreground Connectivity 8
Edge Width 1
Transition On
Intext On & Off
Statistics On
Table 3 displays the different fragmentation classes, their output color, and their
unique values. Note the added value of 100 for internal versus external fragmentation
classes. The colors may be manipulated once imported into ArcGIS, but the values will
remain the same for each fragmentation class.
19


Table 3. MSPA outputs with unique values for each fragmentation class (Vogt n.d.).
Class Color RGB Value [byte] internal/external
1) Core 000/200/000 117/17
2) Islet 160/060/000 109/9
3) Perforation 000/000/255 105/5
000/000/000 103/3
5 a) Loop 255/255/000 165/65
5b) Loop in Edge 255/255/000 167/67
5c) Loop in Perforation 255/255/000 169/69
6a) Bridge 255/000/000 133/33
6b) Bridge in Edge 255/000/000 135/35
6c) Bridge in Perforation 255/000/000 137/37
7) Branch 255/140/000 101/1
Background 220/220/220 100/0
Missing 255/255/255 129/129
Analyzing the Transitions Between Fragmentation Classes During the 2002-2008 and
2008 -2013 Periods
The MSPA results and ArcGIS were used to answer the second research question,
How has each fragmentation class changed from 2002 to 2008 and from 2008 to 2013 in
temperate and tropical forests in relation to the fragmentation class of the previous date?
First, each forest fragmentation class from the ArcGIS rasters created from the MSPA results
with the intext parameter set to off for both temperate and tropical forests in 2002 were
isolated by reclassifying the fragmentation raster to give the desired isolated fragmentation
class a value of one and all other classes a value of NoData. This was performed for
each fragmentation class for temperate and tropical forests in 2002. For each fragmentation
class, this process will result in a raster containing only the desired fragmentation class from
2002. For example, the edge fragmentation class has a unique value of three (Table 3), so
using the 2002 rasters, the value of three was reclassified to one and all other values were
20


reclassified to NoData for tropical and temperate forests. Therefore, the output raster only
contained the edge fragmentation class from 2002.
Next, the 2002 isolated fragmentation classes were clipped to the MSPA results for
temperate and tropical forests in 2008. This was done by extracting the values from the 2008
MSPA results using the 2002 isolated fragmentation classes as a mask. This results in the
2002 fragmentation areas with the 2008 fragmentation values in order to visualize what level
of fragmentation the 2002 isolated classes turned into in 2008. From these results, the area of
each 2008 fragmentation class in the 2002 masks were calculated using the unique values for
each class using the Zonal Geometry tool in ArcGIS. The resulting areas were in units of
square meters, and were divided by 10,000 in order to get the results in units of hectares.
This procedure was repeated by isolating all fragmentation classes for temperate and tropical
forests in 2008 and clipping them to the fragmentation classes in 2013, and then calculating
the areas in 2013.
Additionally, these same procedures were repeated with the ArcGIS rasters generated
from the MSPA results with the intext parameter set to on in order to differentiate between
internal and external forests.
In order to evaluate how the temperate and tropical forests have evolved from one
date to the next, transition matrices were created for temperate forests from 2002-2008,
tropical forests from 2002-2008, temperate forests from 2008-2013, and tropical forests from
2008-2013. This was accomplished by first using the Zonal Geometry tool in ArcGIS to
determine the area (in hectares) of each fragmentation class for the first date in each time
period (i.e. 2002 for the 2002-2008 period and 2008 for the 2008-2013 period). Next, the
percent change of each fragmentation class from the first date to each fragmentation class of
21


the next date was calculated. This results in a transition matrix describing what percent of
each fragmentation class in the first date changed into each fragmentation class from the next
date for temperate and tropical forests for both time periods (for example, eight percent of the
branch fragmentation class in 2002 turned into edge in 2008).
These processes result in a comprehensive, national level assessment of what level of
fragmentation the temperate and tropical forests of Mexico have changed into from 2002-
2008 and 2008-2013.
22


CHAPTER IV
RESULTS
The results first discuss results generated from the MSPA for both temperate and
tropical forests for 2002, 2008, and 2013 with the intext parameter set to off. Then the
results from the MSPA for both forest types and all dates are discussed with the intext
parameter set to on. Next, the results from the transitions between fragmentation classes for
both time periods with the intext parameter set to off are discussed, organized by major
trends in the data. Lastly, the results from the transitions between fragmentation classes for
both time periods with the intext parameter set to on are discussed.
Fragmentation Levels in 2002, 2008, and 2013 with Intext Parameter Off
Tables 4, 5 and 6 present the forest fragmentation classes resulting from the MSPA
method for temperate and tropical forests in 2002, 2008 and 2013 respectively. These results
were generated with the MSPA intext parameter set to off (meaning that fragmentation
classes were not differentiated between internal and external). This results in 11
fragmentation classes with unique identifiers. Each table shows the area in hectares of each
fragmentation class and the percent of the total forest area that falls within each
fragmentation class for the temperate and tropical forests.
There were a total 18,070,262 hectares of tropical forests and 22,126,451 hectares of
temperate forests in 2002 (Table 4). Approximately 73 percent of the total tropical forest
area was classified as core (interior) forests, 17.2 percent were classified as edge, and about
10 percent fell in the fragmented classes (all other fragmentation classes except core and
edge). Approximately 70 percent of the temperate forest areas were classified as core, 18.4
percent were classified as edge, and about 11 percent fell in the fragmented categories.
23


Table 4. Areas of each fragmentation class for temperate and tropical forests in 2002.
2002 Tropical Forests 2002 Temperate Forests
Fragmentation Class Area (ha) Percent of Total Forest Area (ha) Percent of Total Forest
Branch 781,794 4.33 1,234,488 5.58
Edge 3,108,631 17.20 4,074,182 18.41
Perforation 531,331 2.94 602,025 2.72
Islet 67,088 0.37 134,444 0.61
Core 13,240,831 73.27 15,549,201 70.27
Bridge 137,556 0.76 223,613 1.01
Bridge in Edge 113,981 0.63 167,731 0.76
Bridge in Perforation 1,425 0.01 3,269 0.01
Loop 46,131 0.26 76,556 0.35
Loop in Edge 32,150 0.18 42,394 0.19
Loop in Perforation 9,344 0.05 18,550 0.08
Total Forest Area 18,070,262 22,126,451
There was a total of 17,957,907 hectares of tropical forests and 21,222,037 hectares
of temperate forest in 2008 (Table 5). Approximately 73 percent of the tropical forests were
classified as core forests, 17.2 percent were edge, and about 10 percent fell in the fragmented
classes. Approximately 69 percent of the temperate forests were classified as core, 19.4
percent as edge, and approximately 11 percent of fell in the fragmented classes.
Table 5, Areas of each fragmentation class for temperate and tropical forests in 2008.
2008 Tropical Forests 2008 Temperate Forests
Fragmentation Class Area (ha) Percent of Total Forest Area (ha) Percent of Total Forest
Branch 827,375 4.61 1,217,513 5.74
Edge 3,087,438 17.19 4,117,194 19.40
Perforation 529,050 2.95 511,944 2.41
Islet 80,475 0.45 138,550 0.65
Core 13,030,538 72.56 14,686,731 69.21
Bridge 163,400 0.91 225,575 1.06
Bridge in Edge 130,438 0.73 170,938 0.81
Bridge in Perforation 2,744 0.02 3,025 0.01
Loop 56,825 0.32 83,356 0.39
Loop in Edge 37,013 0.21 49,019 0.23
Loop in Perforation 12,613 0.07 18,194 0.09
Total Forest Area 17,957,907 21,222,037
24


There was a total of 18,425,013 hectares of tropical forests and 21,271,030 hectares
of temperate forests in 2013 (Table 6). Approximately 73 percent of the tropical forests were
classified as core forest, 16.7 percent as edge, and about 10 percent fell in the fragmented
classes. Approximately 69 percent of the temperate forests were classified as core, 19.5
percent as edge, and approximately 11 percent fell in the fragmented classes.
Ta
ole 6, Areas of each fragmentation class for temperate and tropical forests in 2013.
2013 Tropical Forests 2013 Temperate Forests
Fragmentation Class Area (ha) Percent of Total Forest Area (ha) Percent of Total Forest
Branch 847,975 4.60 1,222,613 5.75
Edge 3,078,619 16.71 4,147,906 19.50
Perforation 545,919 2.96 499,644 2.35
Islet 79,594 0.43 138,288 0.65
Core 13,447,188 72.98 14,716,174 69.18
Bridge 170,400 0.92 223,944 1.05
Bridge in Edge 135,581 0.74 170,694 0.80
Bridge in Perforation 2,813 0.02 2,825 0.01
Loop 61,756 0.34 82,881 0.39
Loop in Edge 39,650 0.22 48,931 0.23
Loop in Perforation 15,519 0.08 17,131 0.08
Total Forest Area 18,425,013 21,271,030
Overall, there was a loss of 112,355 hectares of tropical forests between 2002 and
2008, and a reported gain of 467,106 hectares between 2008 and 2013. There was a loss of
904,414 hectares of temperate forests between 2002 and 2008, and a reported gain of 48,993
hectares between 2008 and 2013.
In addition to tabular data, the MSPA results in spatially referenced fragmentation
classes for the entirety of Mexico. Figure 6 shows a zoomed-in sample area of the MSPA
output for tropical forests in 2002. Similar cartographic outputs were generated for the
temperate and tropical forests for 2002, 2008, and 2013.
25


Figure 6. MSPA output for tropical forests in 2002. Similar cartographic outputs were
generated for 2008 and 2013 temperate and tropical forests.
Fragmentation Levels in 2002, 2008, and 2013 with Intext Parameter On
Tables 7, 8, and 9 present forest fragmentation classes resulting from the MSPA for
temperate and tropical forests in 2002, 2008, and 2013 respectively differentiating between
external and internal fragmentation classes (MSPA parameter intext set to on). This results in
a total of 19 fragmentation classes with unique identifiers. Each table shows the area in
hectares of each fragmentation class and the percent of the total forest area that falls within
each fragmentation class for the temperate or tropical forests.
In 2002, approximately 73 percent of the tropical forests were core forests (Table 7),
of which almost all were external core forests (73.15 percent external versus 0.13 percent
internal core). Approximately 17 percent of the tropical forests were external edge and 0.13
percent were internal edge (meaning it was an edge forest within a perforation).
Approximately 10 percent of the tropical forests were classified in the fragmented classes (96
percent in the external classes and 4 percent in the internal).


Table 7. Areas of each external and internal fragmentation class for temperate and tropical
forests in 2002.
2002 Tropical Forests 2002 Temperate Forests
Fragmentation Class Area (ha) Percent of Total Forest Area (ha) Percent of Total Forest
External Branch 742,894 4.11 1,171,469 5.29
External Edge 3,085,494 17.07 4,046,650 18.29
External Islet 66,313 0.37 132,863 0.60
External Core 13,217,810 73.15 15,526,460 70.17
External Bridge 133,350 0.74 213,194 0.96
External Bridge in Edge 111,344 0.62 161,906 0.73
External Loop 39,013 0.22 59,050 0.27
External Loop in Edge 32,038 0.18 42,269 0.19
Internal Branch 38,900 0.22 63,019 0.28
Internal Edge 23,138 0.13 27,531 0.12
Internal Perforation 531,331 2.94 602,025 2.72
Internal Islet 775 0.00 1,581 0.01
Internal Core 23,019 0.13 22,738 0.10
Internal Bridge 4,206 0.02 10,419 0.05
Internal Bridge in Edge 2,638 0.01 5,825 0.03
Internal Bridge in Perforation 1,425 0.01 3,269 0.01
Internal Loop 7,119 0.04 17,506 0.08
Internal Loop in Edge 113 0.00 125 0.00
Internal Loop in Perforation 9,344 0.05 18,550 0.08
Approximately 70 percent of temperate forests in 2002 were external core and only 0.1
percent internal core, and 18.2 percent were external and 0.1 percent were internal edge.
Approximately 11 percent of the temperate forests were classified in the fragmented classes
(97 percent in the external classes and 3 percent in the internal).
In 2008 (Table 8), approximately 73 percent of the tropical forests were core forests,
of which almost all were external core forests (72.46 percent external versus 0.11 percent
internal core). Approximately 17 percent of the tropical forests were external edge and 0.13
percent were internal edge (meaning it was an edge forest within a perforation).
27


Table 8. Areas of each external and internal fragmentation class for temperate and tropical
forests in 2008.
2008 Tropical Forests 2008 Temperate Forests
Fragmentation Class Area (ha) Percent of Total Forest Area (ha) Percent of Total Forest
External Branch 779,788 4.34 1,165,412 5.49
External Edge 3,064,719 17.07 4,097,744 19.31
External Islet 79,688 0.44 137,381 0.65
External Core 13,011,530 72.46 14,671,090 69.13
External Bridge 156,419 0.87 218,488 1.03
External Bridge in Edge 126,856 0.71 166,675 0.79
External Loop 45,450 0.25 67,663 0.32
External Loop in Edge 36,869 0.21 48,981 0.23
Internal Branch 47,588 0.26 52,100 0.25
Internal Edge 22,719 0.13 19,450 0.09
Internal Perforation 529,050 2.95 511,944 2.41
Internal Islet 788 0.00 1,169 0.01
Internal Core 19,006 0.11 15,644 0.07
Internal Bridge 6,981 0.04 7,088 0.03
Internal Bridge in Edge 3,581 0.02 4,263 0.02
Internal Bridge in Perforation 2,744 0.02 3,025 0.01
Internal Loop 11,375 0.06 15,694 0.07
Internal Loop in Edge 144 0.00 38 0.00
Internal Loop in Perforation 12,613 0.07 18,194 0.09
Approximately 10 percent of the tropical forests were classified in the fragmented
classes (96 percent in the external classes and 4 percent in the internal). Approximately 69
percent of temperate forests in 2008 were external core and only 0.07 percent internal core,
and 19.31 percent were external edge and 0.09 percent were internal edge. Approximately
11 percent of the temperate forests were classified in the fragmented classes (97 percent in
the external classes and 3 percent in the internal).
In 2013 (Table 9), approximately 73 percent of the tropical forests were core forests,
of which almost all were external core forests (72.86 percent external versus 0.12 percent
internal core). Approximately 17 percent of the tropical forests were external edge and 0.14
28


percent were internal edge (meaning it was an edge forest within a perforation).
Approximately 10 percent of the tropical forests were classified in the fragmented classes (96
percent in the external classes and 4 percent in the internal). Approximately 69 percent of
temperate forests in 2013 were external core and only 0.07 percent internal core, and 19.42
percent were external edge and 0.08 percent were internal edge. Approximately 11 percent
of the temperate forests were classified in the fragmented classes (97 percent in the external
classes and 3 percent in the internal).
Table 9. Areas of each external and internal fragmentation class for temperate and tropical
forests in 2013.
2013 Tropical Forests 2013 Temperate Forests
Fragmentation Class Area (ha) Percent of Total Forest Area (ha) Percent of Total Forest
External Branch 796,106 4.32 1,172,244 5.51
External Edge 3,053,500 16.57 4,130,775 19.42
External Islet 79,069 0.43 137,244 0.65
External Core 13,424,288 72.86 14,702,307 69.12
External Bridge 162,800 0.88 217,200 1.02
External Bridge in Edge 131,506 0.71 166,844 0.78
External Loop 48,775 0.26 67,506 0.32
External Loop in Edge 39,438 0.21 48,894 0.23
Internal Branch 51,869 0.28 50,369 0.24
Internal Edge 25,119 0.14 17,131 0.08
Internal Perforation 545,919 2.96 499,644 2.35
Internal Islet 525 0.00 1,044 0.00
Internal Core 22,900 0.12 13,869 0.07
Internal Bridge 7,600 0.04 6,744 0.03
Internal Bridge in Edge 4,075 0.02 3,850 0.02
Internal Bridge in Perforation 2,813 0.02 2,825 0.01
Internal Loop 12,981 0.07 15,375 0.07
Internal Loop in Edge 213 0.00 38 0.00
Internal Loop in Perforation 15,519 0.08 17,131 0.08
29


Overall, across all dates, a very small percentage of the tropical and temperate forests
fall in the interior fragmentation classes. Figure 2 shows a zoomed-in example of the MSPA
cartographic output for the tropical forests in 2002 differentiating between external and
internal fragmentation classes. Similar cartographic outputs were generated for temperate
and tropical forests for 2002, 2008, and 2013.
Fragmentation Classes
| | No Forest
External Branch
External Edge
| | External Islet
External Core
External Bridge
External Bridge in Edge
External Loop
| 1 External Loop in Edge
| 1 No Forest
Internal Branch
Internal Edge
| | Internal Perforation
| | Internal Islet
| | Internal Core
| | Internal Bridge
m Internal Bridge in Edge
Internal Bridge in Perforation
Internal Loop
Internal Loop in Edge
Internal Loop in Perforation
Figure 7. MSPA cartographic output for fragmentation classes of the tropical forests in 2002
differentiating between external and internal fragmentation classes. Similar outputs were
created for the 2008 and 2013 temperate and tropical forests.
Transitions Between Fragmentation Classes During 2002-2008 and 2008-2013 Periods
Without Differentiating Between Internal and External Fragmentation Classes
Results from analyzing the transitions between fragmentation classes without
differentiating between internal and external classes are organized into three major trends in
the data; the percent change from any of the fragmentation classes to no forest for both dates
and time periods, differences in transitions between temperate and tropical forests, and
30


differences in transitions between the two time periods. Additionally, results from transitions
between fragmented classes (all classes aside from core and edge) will be described. First,
the transition matrices are displayed for first tropical and temperate forests in the 2002-2008
time period and next for tropical and temperate forests in the 2008 to 2013 time period in
Tables 10, 11, 12, and 13. These transition matrices display the percent of the fragmentation
class in the first date than turned into each fragmentation class for the next date (i.e. what
percent of each fragmentation class in 2002 turned into each fragmentation class in 2008, and
again for the 2008 to 2013 period).
Table 10. Fragmentation classes transition matrix for tropical forests in the 2002-2008 period
without differentiating between internal and external classes.
% Changed to each 2008 fragmentation class
Fragmentation Class in 2002 Area in 2002 (ha) No Forest Branch Edge Perforation Islet Core Bridge Bridge in Edge Bridge in Perforation Loop Loop in Edge Loop in Perforation
Branch 781,794 25 59 8 1 1 4 2 1 0 1 0 0
Edge 3,108,631 17 3 68 2 0 9 1 1 0 0 0 0
Perforation 531,331 13 2 12 53 0 18 0 0 0 0 0 0
Islet 67,088 26 9 4 0 57 1 1 0 0 1 0 0
Core 13,240,831 7 0 3 1 0 89 0 0 0 0 0 0
Bridge 137,556 21 9 7 0 0 6 51 3 0 2 0 0
Bridge in Edge 113,981 17 4 15 0 0 11 5 45 0 0 1 0
Bridge in Perforation 1,425 8 4 5 12 0 16 6 10 30 2 1 6
Loop 46,131 22 11 7 1 1 5 9 1 0 42 2 0
Loop in Edge 32,150 16 3 18 2 0 10 2 8 0 2 37 2
Loop in Perforation 9,344 13 2 5 13 0 11 2 3 1 4 10 36
31


Table 11. Fragmentation classes transition matrix for temperate and forests in the 2002-2008
_____________period without differentiating between internal and external classes.________________
% Changed to each 2008 fragmentation class
Fragmentation Class 2002 Area in 2002 (ha) No Forest Branch Edge Perforation Islet Core Bridge Bridge in Edge Bridge in Perforation Loop Loop in Edge Loop in Perforation
Branch 1,234,488 16 76 4 0 1 2 1 0 0 0 0 0
Edge 4,074,182 12 2 80 0 0 5 0 0 0 0 0 0
Perforation 602,025 13 2 13 63 0 8 0 0 0 0 0 0
Islet 134,444 12 3 2 0 81 1 0 0 0 0 0 0
Core 15,549,201 5 0 3 1 0 91 0 0 0 0 0 0
Bridge 223,613 12 5 4 0 0 3 73 2 0 1 0 0
Bridge in Edge 167,731 10 2 8 0 0 6 3 69 0 0 1 0
Bridge in Perforation 3,269 16 2 7 6 0 7 2 9 47 1 0 3
Loop 76,556 14 5 4 1 0 2 5 1 0 67 1 0
Loop in Edge 42,394 11 2 10 1 0 5 1 3 0 1 65 0
Loop in Perforation 18,550 12 2 4 10 0 7 1 2 1 3 7 50
Table 12. Fragmentation classes transition matrix for tropical forests in the 2008-2013 period
_________________without differentiating between internal and external classes._________________
% Changed to each 2013 fragmentation class
Fragmentation Class 2008 Area in 2008 (ha) No Forest Branch Edge Perforation Islet Core Bridge Bridge in Edge Bridge in Perforation Loop Loop in Edge Loop in Perforation
Branch 827,375 6 89 2 0 0 2 1 0 0 0 0 0
Edge 3,087,438 4 1 89 1 0 4 0 0 0 0 0 0
Perforation 529,050 8 1 10 62 0 17 0 0 0 0 0 0
Islet 80,475 10 4 2 0 82 1 0 0 0 0 0 0
Core 13,030,53 8 3 0 1 1 0 95 0 0 0 0 0 0
Bridge 163,400 5 3 3 1 0 2 84 1 0 2 0 0
Bridge in Edge 130,438 4 1 5 1 0 3 2 83 0 0 1 0
Bridge in Perforation 2,744 8 3 2 8 0 8 4 18 41 1 1 6
Loop 56,825 8 3 3 2 0 3 4 0 0 75 1 0
Loop in Edge 37,013 4 1 5 1 0 4 1 5 0 1 76 1
Loop in Perforation 12,613 6 1 3 7 0 10 0 2 1 2 9 58
32


Table 13. Fragmentation classes transition matrix for temperate forests in the 2008-2013
___________period without differentiating between internal and external classes.______________
Fragmentation classes transition matrix temperate forests 2008-2013
% Changed to each 2013 fragmentation class
Fragmentation Class 2008 Area in 2008 (ha) No Forest Branch Edge Perforation Islet Core Bridge Bridge in Edge Bridge in Perforation Loop Loop in Edge Loop in Perforation
Branch 1,217,513 1 98 1 0 0 0 0 0 0 0 0 0
Edge 4,117,194 1 0 98 0 0 1 0 0 0 0 0 0
Perforation 511,944 1 0 3 94 0 2 0 0 0 0 0 0
Islet 138,550 3 1 0 0 95 0 0 0 0 0 0 0
Core 14,686,731 0 0 0 0 0 99 0 0 0 0 0 0
Bridge 225,575 1 1 1 0 0 1 96 0 0 0 0 0
Bridge in Edge 170,938 1 0 2 0 0 1 0 96 0 0 0 0
Bridge in Perforation 3,025 0 0 0 1 0 3 0 6 89 0 0 1
Loop 83,356 1 1 2 0 0 1 1 0 0 94 0 0
Loop in Edge 49,019 1 0 3 0 0 2 0 1 0 0 93 0
Loop in Perforation 18,194 1 0 0 2 0 4 0 0 0 1 3 89
In addition to the transition matrices created for temperate and tropical forests,
cartographic outputs were generated for each cross referenced fragmentation class for both
time periods. The resulting maps each contain the area of the isolated fragmentation class
from the first date with the fragmentation classes from the next date. Figure 8 is an example
of the cartographic results showing the 2008 fragmentation classes that were edge forest in
2002. The lines on the map were edge forest in 2002 and the different colors represent the
fragmentation classes in 2008, therefore displaying what the 2002 edge forest turned into in
2008. Similar results were generated for each fragmentation class for both temperate and
tropical forests for both time periods.
33


Figure 8. Example of fragmentation classes in 2008 that were edge forest in 2002 for tropical
forests. Similar results were generated for temperate and tropical forests for all dates and all
fragmentation classes.
Transition to no forest
From 2002 to 2008 for both temperate and tropical forests, the percent of each
fragmentation class changing to no forest is much higher than the percent change to another
fragmentation class. In tropical forests, islets experienced the most percent change to no
forest (26 percent) and core experienced the least percent change to no forest (seven percent).
In temperate forests, branch and bridge in perforation forests experienced the most change to
no forest (both at 16 percent) and core forests experienced the least change to no forest (five
percent).
From 2008 to 2013, temperate and tropical forests experienced similar transitions in
that a relatively high percent of each fragmentation class transitioned to no forest in 2013, but
the changes are not as drastic as from 2002 to 2008 due to much less overall change in
34


fragmentation classes. In tropical forests, islets again experienced to most change to no
forest (10 percent) and core forest experiencing the least percent change to no forest (three
percent). In temperate forests islets also experienced the most transition to no forest,
although much lower overall percent change (three percent). Temperate core forests
experienced less than one percent change to no forest during this time period. Isolated
fragmentation classes (islets) and narrow, elongated fragmentation classes (branch, bridge,
loop) were the most vulnerable to transition to no forest, while core forests were consistently
the least likely to change to no forest.
Differences in forest type
Although both tropical and temperate forests experienced a higher percent change to
no forest than other fragmentation classes, tropical forests experienced much more transition
to no forest than temperate forests during both time periods. From 2002-2008, tropical
forests fragmentation classes experienced a minimum of seven percent to a maximum of 26
percent transition to no forest, whereas temperate forests fragmentation classes experienced
minimum of five percent to a maximum of 16 percent transition to no forest. From 2008-
2013, tropical forests fragmentation classes experienced a minimum of three percent to a
maximum of 10 percent transition to no forest, whereas temperate forests experienced zero to
three percent transition to no forest. The results from the MSPA indicated that temperate
forests had more total forest area but higher percentages of that forest were edge and
fragmented than in tropical forests. When analyzing the transitions of between fragmentation
classes from one date to the next, it is evident that a higher percent of remaining tropical
forests are being fragmented and turning into no forest than the remaining temperate forests.
35


Additionally, temperate forests experienced less overall change in fragmentation classes
across both time periods than tropical forests. For example, 59 percent of tropical branch
forest in 2002 remained branch in 2008, whereas 76 percent of temperate branch forest in
2002 remained branch in 2008. Likewise, 89 percent of tropical branch forest in 2008
remained branch in 2013, and 98 percent of temperate branch forest in 2008 remained branch
in 2013. These findings are consistent across all fragmentation classes. This indicates less
change in temperate forests than in tropical forests across both time periods.
Differences between time periods
When analyzing the transitions between fragmentation classes, it is evident that there
is significant more change from 2002-2008 than there is from 2008-2013. Across all
fragmentation classes, higher percentages remained the same fragmentation class from 2008
to 2013 than from 2002 to 2008. For example, using the example from above, from 59
percent of tropical branch forest remained branch from 2002-2008, whereas 89 percent of
tropical branch forest remained branch from 2008-2013. Similarly, 76 percent of temperate
branch forest remained branch from 2002-2008, whereas 98 percent of temperate branch
forest remained branch from 2008-2013. These results were consistent across all
fragmentation classes. Overall, core forests consistently changed the least in both tropical
and temperate forests during both periods. This result is to be expected as these forests are
buffered from edge effects and represent large compacted forest areas that are less subject to
fragmentation and deforestation.
Fragmentation class to fragmentation class transitions
Analyzing the changes from fragmentation class to fragmentation class is looking at
how each fragmentation class has evolved in relation to the other fragmentation classes, not
36


determining how they changed in relation to turning into no forest. These results describe
what the majority of each fragmentation class transitioned to aside from remaining the same
class or turning into no forest. Therefore, aside from remaining the same class or turning into
no forest, in tropical forests from 2002-2008, the majority of:
Branch turned into edge
Edge turned into core
Perforation turned into core
Islet turned into branch
Core turned into edge
Bridge turned into branch
Bridge in edge turned into edge
Bridge in perforation turned into core (much of it also turning into perforation)
Loop turned into branch
Loop in edge turned into edge
Loop in perforation turned into perforation
Aside from remaining the same class or turning into no forest, in temperate forests from
2002-2008, the majority of:
Branch turned into edge
Edge turned into core
Perforation turned into edge
Islet turned into branch
Core turned into edge
Bridge turned into branch
37


Bridge in edge turned into edge
Bridge in perforation turned into bridge in edge
Loop turned into bridge or branch
Loop in edge turned into edge
Loop in perforation turned into perforation
From 2008-2013, aside from remaining the same class or turning into no forest, in tropical
forests the majority of:
Branch turned into edge or core
Edge turned into core
Perforation turned into core
Islet turned into branch
Core turned into edge or perforation
Bridge turned into branch or edge
Bridge in edge turned into edge
Bridge in perforation turned into perforation or core
Loop turned into bridge
Loop in edge turned into edge or bridge in edge
Loop in perforation turned into core
From 2008-2013, aside from remaining the same class or turning into no forest, in temperate
forests the majority of:
Branch turned into edge
Edge turned into core
Perforation turned into edge
38


Islet turned into branch
Less than one percent of core changed into a different fragmentation class
Bridge turned into branch, edge, or core
Bridge in edge turned into edge
Bridge in perforation turned into bridge in edge
Loop turned into edge
Loop in edge turned into edge
Loop in perforation turned into core
Transitions Between Fragmentation Classes During 2002-2008 and 2008-2013 Periods
Differentiating Between Internal and External Fragmentation Classes
Similar results were generated with internal and external fragmentation classes
differentiated as were generated when internal and external classes were not differentiated.
The same major trends of transitioning to no forest, differences between forest types, and
differences in time periods were identified. Since the results of the transitions between
fragmentation classes differentiating between internal and external classes are similar to the
results without differentiating between external and internal classes, they are summarized
briefly by time period.
2002-2008 time period
Table 14 presents the results from cross referencing each fragmentation class in 2002
to the fragmentation classes in 2008 for tropical forests with the intext parameter set to on,
differentiating between internal and external fragmentation classes. These results are
presented in the form of a transition matrix, indicating the percent of the fragmentation class
in 2002 that turned into each fragmentation class in 2008.
39


Table 14. Fragmentation classes transition matrix for tropical forests in the 2002-2008 period
______________________differentiating between internal and external classes._______________________
% Changed to each 2008 fragmentation class
External Classes Internal Classes
Fragmentation Class 2002 Area 2002 (ha) No Forest Branch Edge Islet Core Bridge Bridge in Edge Loop Loop in Edge Branch Edge Perforation Islet Core Bridge Bridge in Edge Bridge in Perforation Loop Loop in Edge Loop in Perforation
External Branch 742,894 25 58 8 1 3 2 1 1 0 1 0 0 0 0 0 0 0 0 0 0
External Edge 3,085,494 17 3 68 0 9 1 1 0 0 0 0 2 0 0 0 0 0 0 0 0
External Islet 66,313 26 9 4 57 1 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0
External Core 13,217,810 7 0 3 0 88 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0
External Bridge 133,350 22 9 7 0 5 51 3 2 0 0 0 0 0 0 0 0 0 0 0 0
External Bridge in Edge 111,344 18 4 15 0 11 5 45 0 1 0 0 0 0 0 0 0 0 0 0 0
External Loop 39,013 23 11 7 1 5 9 1 38 2 0 0 1 0 0 0 0 0 2 0 0
External Loop in Edge 32,038 16 3 18 0 10 2 8 2 37 0 0 2 0 0 0 0 0 0 0 2
Internal Branch 38,900 19 12 2 0 12 0 0 0 0 45 1 5 0 0 1 0 0 1 0 0
Internal Edge 23,138 13 1 13 0 26 0 0 0 0 3 33 7 0 2 1 1 0 0 0 0
Internal Perforation 531,331 13 1 12 0 18 0 0 0 0 1 1 53 0 0 0 0 0 0 0 0
Internal Islet 775 26 0 0 12 0 0 0 0 0 5 0 0 52 0 0 0 0 5 0 0
Internal Core 23,019 13 0 2 0 52 0 0 0 0 1 2 2 0 27 1 1 0 0 0 0
Internal Bridge 4,206 14 1 1 0 14 7 0 1 0 6 1 3 0 1 41 2 1 5 0 1
Internal Bridge in Edge 2,638 8 0 4 0 18 1 7 0 0 5 7 5 0 5 4 31 1 1 0 2
Internal Bridge in Perforation 1,425 8 1 4 0 14 2 7 0 1 3 1 12 0 1 4 3 30 2 0 6
Internal Loop 7,119 17 2 3 0 5 3 1 10 1 7 0 5 0 0 2 0 0 41 0 1
Internal Loop in Edge 113 33 0 0 0 0 6 0 0 0 6 0 6 0 0 0 0 0 11 33 6
Internal Loop in Perforation 9,344 13 1 5 0 11 1 2 1 10 2 0 13 0 0 1 0 1 3 0 36
Similar results were generated with internal and external fragmentation classes
differentiated as were generated when internal and external classes were not differentiated.
From 2002 to 2008 in tropical forests, a high percent of fragmentation classes changed into
no forest in 2008. From the external classes, external islet forests experienced the most
change to no forest (26 percent) and external core experienced the least change to no forest
40


(seven percent). From the internal classes, internal loop in edge forests experienced the most
change to no forest (33 percent), and both internal bridge in edge and internal bridge in
perforation experienced the least (both eight percent). Of the external fragmentation classes,
external core changed the least, with 88 percent remaining external core. External loop in
edge forests changed the most, with 37 percent remaining external loop in edge and 63
percent changing to a different fragmentation class. Of the internal fragmentation classes,
internal perforation forests changed the least, with 53 percent remaining internal perforation
in 2008. Internal core forests changed the most, with only 27 percent remaining internal
core.
Table 15 presents the results from cross referencing each fragmentation class in 2002
to the fragmentation classes in 2008 for temperate forests with the intext parameter set to on,
differentiating between internal and external fragmentation classes. These results are
presented in the form of a transition matrix, indicating the percent of the fragmentation class
in 2002 that turned into each fragmentation class in 2008.
For temperate forests from 2002 to 2008, in the external fragmentation classes,
external branch forests experienced the most change to no forests (15 percent) and external
core the least (five percent). From the internal fragmentation classes, internal islet forests
experienced the most change to no forest (39 percent), and internal loop in perforation the
least (12 percent). Of the external fragmentation classes, external core changed the least with
89 percent remaining external core, and external loop in edge forests changing the most with
65 percent remaining external loop in edge. Of the internal fragmentation classes, internal
perforations changed the least (63 percent remaining internal perforation) and internal loop in
edge forests changed the most (only 20 percent remaining internal loop in edge).
41


Table 15. Fragmentation classes transition matrix for temperate forests in the 2002-2008
period differentiating between internal and external classes.
% Changed to each 2008 fragmentation class
External Classes Internal Classes
Fragmentation Class 2002 Area 2002 (ha) No Forest Branch Edge Islet Core Bridge Bridge in Edge O- o o h-l Loop in Edge Branch Edge Perforation Islet Core Bridge Bridge in Edge Bridge in Perforation O- o o Loop in Edge Loop in Perforation
External Branch 1,171,469 15 77 5 1 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0
External Edge 4,046,650 11 2 80 0 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
External Islet 132,863 12 3 2 81 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
External Core 15,526,460 5 0 3 0 91 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0
External Bridge 213,194 11 5 4 0 3 74 2 l 0 0 0 0 0 0 0 0 0 0 0 0
External Bridge in Edge 161,906 10 2 8 0 6 3 70 0 1 0 0 0 0 0 0 0 0 0 0 0
External Loop 59,050 12 5 4 0 2 5 1 70 1 0 0 0 0 0 0 0 0 0 0 0
External Loop in Edge 42,269 11 2 10 0 5 1 3 1 65 0 0 1 0 0 0 0 0 0 0 0
Internal Branch 63,019 25 10 3 0 2 0 0 0 0 52 1 3 0 0 0 0 0 1 0 0
Internal Edge 27,531 25 2 15 0 6 1 1 0 0 2 41 4 0 2 0 1 0 0 0 0
Internal Perforation 602,025 13 i 12 0 8 0 0 0 0 i 0 63 0 0 0 0 0 0 0 0
Internal Islet 1,581 39 i 0 2 0 0 0 0 0 4 3 0 52 0 0 0 0 0 0 0
Internal Core 22,738 17 i 3 0 29 1 1 0 0 1 2 2 0 43 0 1 0 0 0 0
Internal Bridge 10,419 32 3 2 0 4 7 1 1 0 3 1 2 0 1 37 1 1 3 0 0
Internal Bridge in Edge 5,825 27 1 2 0 7 2 7 0 0 1 2 3 0 2 1 40 0 2 0 1
Internal Bridge in Perforation 3,269 16 2 6 0 7 1 8 1 0 0 1 6 0 0 1 1 47 0 0 3
Internal Loop 17,506 19 1 2 0 4 3 1 6 1 3 0 3 0 0 2 0 0 52 0 1
Internal Loop in Edge 125 25 0 5 0 0 5 5 0 5 5 10 5 0 5 5 5 0 0 20 0
Internal Loop in Perforation 18,550 12 1 4 0 7 1 2 1 7 1 0 10 0 0 1 1 1 2 0 50
2008-2013 time period
Table 16 presents the results from cross referencing each fragmentation class in 2008
to the fragmentation classes in 2013 for tropical forests with the intext parameter set to on,
differentiating between internal and external fragmentation classes. These results are
42


presented in the form of a transition matrix, indicating the percent of each fragmentation
class in 2008 that turned into each fragmentation class in 2013.
Table 16. Fragmentation classes transition matrix for tropical forests in the 2008-2013 period
______________________differentiating between internal and external classes._______________________
% Changed to each 2013 fragmentation class
External Classes Internal Classes
Fragmentation Class 2008 Area 2008 (ha) No Forest Branch Edge Islet Core Bridge Bridge in Edge Loop Loop in Edge Branch Edge Perforation Islet Core Bridge Bridge in Edge Bridge in Perforation Loop Loop in Edge Loop in Perforation
External Branch 779,788 6 89 2 1 1 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0
External Edge 3,064,719 4 1 89 0 4 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0
External Is let 79,688 10 4 2 82 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
External Core 13,011,530 3 0 1 0 95 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0
External Bridge 156,419 5 3 2 0 2 85 1 2 0 0 0 0 0 0 0 0 0 0 0 0
External Bridge in Edge 126,856 4 1 4 0 3 1 84 0 1 0 0 0 0 0 0 0 0 0 0 0
External Loop 45,450 7 3 3 0 2 5 0 77 1 0 0 1 0 0 0 0 0 1 0 0
External Loop in Edge 36,869 4 1 5 0 4 1 5 1 76 0 0 1 0 0 0 0 0 0 0 1
Internal Branch 47,588 12 12 2 0 9 1 0 0 0 58 1 4 0 1 0 0 0 1 0 0
Internal Edge 22,719 12 1 13 0 13 0 1 0 0 2 41 10 0 3 1 1 0 0 0 0
Internal Perforation 529,050 8 0 9 0 17 0 0 0 0 1 1 62 0 0 0 0 0 0 0 0
Internal Islet 788 44 1 0 6 2 1 0 0 0 9 0 2 34 0 0 0 0 0 0 0
Internal Core 19,006 7 0 2 0 36 0 0 0 0 1 4 2 0 46 1 1 0 0 0 0
Internal Bridge 6,981 10 1 1 0 5 17 1 0 0 4 4 6 0 3 44 1 0 0 3 0
Internal Bridge in Edge 3,581 10 1 3 0 11 2 12 0 0 3 5 6 0 2 2 38 1 1 0 3
Internal Bridge in Perforation 2,744 8 0 2 0 8 1 15 0 1 2 1 8 0 1 3 3 41 0 1 6
Internal Loop 11,375 12 2 2 0 6 0 0 9 0 3 0 7 0 0 2 0 0 53 0 2
Internal Loop in Edge 144 17 0 0 0 0 0 4 0 0 0 0 0 0 0 0 0 0 0 70 9
Internal Loop in Perforation 12,613 6 0 3 0 9 0 1 1 9 0 0 7 0 0 0 1 1 1 0 58
From 2008 to 2013, of the tropical external fragmentation classes, external islets
experienced the most change to no forest (10 percent) and external core experienced the least
(three percent). Of the internal fragmentation classes, internal islets experienced the most
43


change to no forest (44 percent) and internal loop in perforations the least (six percent).
Within each external fragmentation class, external core forests changed the least (95 percent
remaining external core) and external loop in edge forests changed the most (76 percent
remaining the same class). For the internal fragmentation classes, internal loop in edge
forests changed the least (70 percent remaining internal loop in edge) and internal islets
changed the most (only 34 percent remaining internal islets).
Table 17 presents the results from cross referencing each fragmentation class in 2008
to the fragmentation classes in 2013 for temperate forests with the intext parameter set to on,
differentiating between internal and external fragmentation classes. These results are
presented in the form of a transition matrix, indicating the percent of the fragmentation class
in 2008 that turned into each fragmentation class in 2013.
Of the temperate forest fragmentation classes from 2008 to 2013, similar to 2002 to
2008, there was much less overall change of fragmentation classes. Of the external forests,
external islets experienced the most change to no forest (three percent) and external core the
least (approximately zero percent). The percent change to no forest for internal
fragmentation classes were overall very low, with most classes experiencing zero to one
percent change. However, internal islets experienced a 10 percent change to no forest. Of
the external forests, external core experienced the least change (99 percent remaining
external core) and external loop in edge forests experiencing the most (93 percent remained
the same class).
Of the internal forests, internal loop in edge forests experienced the least change
(approximately 100 percent remained the same class) and internal core experienced the most
change (84 percent remained internal core).
44


Table 17. Fragmentation classes transition matrix for temperate forests in the 2008-2013
period differentiating between internal and external classes.
% Changed to each 2013 fragmentation class
External Classes Internal Classes
Fragmentation Class 2008 Area 2008 (ha) No Forest Branch Edge Islet Core Bridge Bridge in Edge Loop Loop in Edge Branch Edge Perforation Islet Core Bridge Bridge in Edge Bridge in Perforation O- o o Loop in Edge Loop in Perforation
External Branch 1,165,412 1 98 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
External Edge 4,097,744 1 0 98 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
External Islet 137,381 3 1 0 95 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
External Core 14,671,090 0 0 0 0 99 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
External Bridge 218,488 1 1 1 0 1 96 0 0 0 0 0 0 0 0 0 0 0 0 0 0
External Bridge in Edge 166,675 1 0 2 0 1 0 96 0 0 0 0 0 0 0 0 0 0 0 0 0
External Loop 67,663 1 1 2 0 1 1 0 94 0 0 0 0 0 0 0 0 0 0 0 0
External Loop in Edge 48,981 1 0 3 0 2 0 1 0 93 0 0 0 0 0 0 0 0 0 0 0
Internal Branch 52,100 0 5 0 0 0 0 0 0 0 94 0 0 0 0 0 0 0 0 0 0
Internal Edge 19,450 0 0 14 0 0 0 0 0 0 0 85 0 0 0 0 0 0 0 0 0
Internal Perforation 511,944 1 0 3 0 2 0 0 0 0 0 0 94 0 0 0 0 0 0 0 0
Internal Islet 1,169 10 0 0 1 0 0 0 0 0 0 0 0 89 0 0 0 0 0 0 0
Internal Core 15,644 0 0 0 0 15 0 0 0 0 0 0 0 0 84 0 0 0 0 0 0
Internal Bridge 7,088 0 0 0 0 1 5 0 0 0 0 0 1 0 0 92 0 0 1 0 0
Internal Bridge in Edge 4,263 0 0 0 0 0 0 10 0 0 0 0 0 0 0 1 87 0 0 0 1
Internal Bridge in Perforation 3,025 0 0 0 0 3 0 6 0 0 0 0 1 0 0 0 0 89 0 0 1
Internal Loop 15,694 1 0 0 0 1 0 0 3 0 0 0 2 0 0 1 0 0 92 0 0
Internal Loop in Edge 38 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 100 0
Internal Loop in Perforation 18,194 1 0 0 0 4 0 0 0 3 0 0 2 0 0 0 0 0 0 0 89
Cartographic results
In addition to the transition matrices created for temperate and tropical forests,
cartographic outputs were generated for each cross referenced fragmentation class for both
time periods, differentiating between internal and external forests. The resulting maps each
contain the area of the isolated fragmentation class from the first date with the fragmentation
classes from the next date. Figure 9 is an example of the cartographic results showing the
45


2008 fragmentation classes that were external edge forest in 2002. The lines on the map
were external edge forest in 2002 and the different colors represent the fragmentation classes
in 2008, therefore displaying what the 2002 edge forest turned into in 2008. Similar results
were generated for each fragmentation class for both temperate and tropical forests for both
time periods.
Fragmentation Classes
I | No Forest
External Branch
External Edge
I I External Islet
External Core
External Bridge
External Bridge in Edge
External Loop
I | External Loop in Edge
[ I No Forest
Internal Branch
H Internal Edge
I I Internal Perforation
I I Internal Islet
I I Internal Core
I I Internal Bridge
internal Bridge in Edge
internal Bridge in Perforation
Internal Loop
Internal Loop in Edge
Internal Loop in Perforation
Figure 9. Example of fragmentation classes in 2008 that were external edge in 2002 for
tropical forests. Similar results were generated for temperate and tropical forests for all dates
and all fragmentation classes.
46


CHAPTER V
DISCUSSION
Defining Fragmentation Classes of Temperate and Tropical Forests
Across all dates, there is overall less area of tropical forests than there are temperate
forests. When using the MSPA to define the fragmentation of these forests, although there
are more overall temperate forests, they are more fragmented than tropical forests. There are
less temperate core forests across all dates than tropical core forests. In 2002, 73.27 percent
of tropical forests were core and 70.27 percent of temperate forests were core. In 2008,
72.56 percent of tropical forests were core and 69.21 percent of temperate forests were core.
In 2013, 72.98 percent of tropical forests were core and 69.18 percent of temperate forests
were core. Additionally, across all dates temperate forests contain a higher percent of the
forest as edge than tropical forests. In 2002, 17.2 percent of tropical forests were edge and
18.41 percent of temperate forests were edge. In 2008, 17.19 percent of tropical forests were
edge and 19.4 percent of temperate forests were edge. In 2013, 16.71 percent of tropical
forests were edge and 19.5 percent of temperate forests were edge. With the exception of the
perforation and bridge in perforation fragmentation classes, temperate forests also contain a
higher percent of each fragmented forest class (all classes except core and edge) than tropical
forests. Therefore, although there is more total forest area of temperate forests, less of the
temperate forests are interior core and more are edge and fragmented forest classes.
The MSPA results generated with the intext parameter set to on, differentiating
between internal and external fragmentation classes, are very similar to the results with intext
set to off. Across all dates temperate forests contained less percent of the total forest area as
core forest, more percent of forest as edge, and higher percentages of fragmented forest
47


classes. The percent of total forest that were internal fragmentation classes were very small,
many less than 0.1 percent. This result is in part explained by the scale of the original data
(1:250,000), the generalization effects created when grouping several forest types into the
broader tropical and temperate forest classes used in this study, and the coarse resolution
used in the analysis (cell size of 250x250 meters). Under these conditions, small perforations
and fine details in the internal edges of the forest areas are difficult to detect.
Quantifying the degree of fragmentation in remaining forests is important for
biodiversity conservation. Although there is more forest area of temperate forests in Mexico,
they are more fragmented than tropical forests. Species are less able to adapt to changing
spatial patterns of the forest and a more fragmented landscape may reduce their ability to
survive over time (O'neill et al. 1988). This is an important factor to consider when
managing forests for conservation. The overall amount of forest is important for species
survival, but even if there are larger amounts of forest area, the fragmented spatial patterns
may affect the ability for different species to survive. For example, there may be more area
of small, isolated forest patches that have a large proportion of edge forests compared to
interior forests. This may increase predator activity and introduce invasive species, which
can influence species survival and decrease biodiversity. Therefore, understanding that there
is more total forest area of temperate forests than tropical forests in Mexico, but temperate
forests have a higher percent of that forest area as fragmented can aid in the conservation
practices of remaining forests as well as aid in understanding how the composition of the
forest may be effecting biodiversity. Conservation measures can be targeted to improve the
degree of forest fragmentation based on the combination of the tabular data describing the
percent of the total forest of each fragmentation class with the spatial reference generated
48


from the cartographic results generated from this study to identify explicit areas in Mexico
with a high degree of fragmentation.
Transitions Between Fragmentation Classes During 2002-2008 and 2008-2013 Periods
Without Differentiating Between Internal and External Fragmentation Classes
When analyzing the transitions between fragmentation classes without differentiating
between internal and external classes, clear trends can be identified. Notably, for both time
periods and forest types analyzed, the percent change from any of the fragmentation classes
to no-forest is high compared to the change from one fragmentation class to other
fragmentation class. Additionally, there were distinct differences between transitions in
temperate versus tropical forests as well as between the two time periods. Lastly, there were
important trends when analyzing the transitions between fragmentation classes, aside from
the changes to no forest or remaining the same fragmentation class from one date to the next.
Transition to no forest
Fragmentation classes in both temperate and tropical forests during both time periods
experienced a higher percent change to no forest than to other fragmentation classes. Overall,
isolated fragmentation classes (islets) and narrow, elongated fragmentation classes (branch,
bridge, loop) were the most vulnerable to transition to no forest, while core forests were
consistently the least likely to change to no forest. This is to be expected, as smaller isolated
forest patches, as well as elongated narrow forest patches, are more likely to change to other
land uses than forests that are farther from edges and span large compacted areas (i.e. core
class). Generally, the edge class shows less tendency to change to no forest than the
elongated forest patches.
49


This information relates the degree of fragmentation to the likelihood of transitioning
to no forest. It is clear through this analysis that the smaller, isolated and narrow, elongated
fragmentation classes are more likely to transition to no forest over time. Targeted
conservation efforts can be implemented to preserve these forest types. With the
cartographic results, regions with increased amounts of islet, bridge, branch, and loop
fragmentation classes can be identified in order to focus these kinds of conservation efforts.
Although these fragmentation classes are typically smaller in area than core or edge forests,
they provide important corridors and connections from larger, more intact forest patches.
These connections and corridors add to the overall landscape structure which is an important
component to species survival (Fahrig & Merriam 1994). Therefore, although they are not
contributing as much to the overall area of forest cover as larger core forests, smaller
fragmentation classes such as islets, bridge, branch, and loop are important to conserve since
they provide important connections and corridors that are lost once these areas are
deforested. Additionally, it is imperative to preserve intact interior forest areas so they do not
become smaller, isolated fragmentation classes that are more susceptible to becoming no
forest. Incorporating land use information to the cartographic results provide additional
information to determine what pressures are causing the transition to no forest in different
areas across Mexico.
Differences in forest type
Tropical forests experienced a higher percent change to no forest as well as to
different fragmentation classes during both time periods. The results from the MSPA
indicated that temperate forests had more total forest area but higher percentages of that
forest were edge and fragmented than in tropical forests. When analyzing the transitions of
50


between fragmentation classes from one date to the next, it is evident that a higher percent of
remaining tropical forests are being fragmented and turning into no forest than the remaining
temperate forests. The differences in transitions of fragmentation classes in temperate and
tropical forests reinforces empirical knowledge and results of studies stating that tropical
forests, more than temperate forests, are subjected to higher anthropogenic pressures and
risks of deforestation (Moreno-Sanchez et al. 2012). Moreno-Sanchez et al. (2012) found
that although temperate forests have high amounts of fragmentation, they typically contain a
larger proportion of forest farther away from anthropogenic influences and therefore
experience lower levels of anthropogenic pressure that is associated with a high risk of
deforestation.
These results reinforce the fact that tropical forests are at a higher risk to becoming
deforested as well as more fragmented than temperate forests. However, the results from the
MSPA also reinforce that temperate forests are currently more fragmented than tropical
forests. Therefore, policy and conservation efforts need to target both types of forests, but
may need to be implemented differently. For example, since temperate forests are currently
more fragmented but are experiencing less change to no forest and other fragmentation
classes than tropical forests, conservation efforts should be concentrated to rehabilitate the
already fragmented forest and to mitigate further fragmentation. For tropical forests, it is
essential to realize what anthropogenic pressures are causing increased fragmentation. Land
use and socioeconomic data can be incorporated with the results from this study to obtain a
thorough understanding of the pressures driving fragmentation in tropical forests.
51


Differences between time periods
Overall, there was much more change in forest fragmentation during the 2002-2008
time period than the 2008-2013 time period. This could reflect successful policy measures to
improve forest conservation in Mexico. Two major payment for ecosystem services (PES)
initiatives were undertaken by the National Forestry Commission (CONAFOR) in 2003 and
2004. One was program for hydrological services and the other a program for carbon
sequestration and biodiversity (FAO 2013). These programs were merged into one PES
National Program in 2006. The PES compensates landowners for implementing land
management and conservation practices. CONAFOR pays landowners a fixed compensation
per hectare of forest for five years. Sustainable forest practices include preserving land cover
and avoiding land use change, developing a Best Management Practices Program, using
surveillance in order to reduce illegal logging and hunting in forests, and providing signage
to inform neighbors about what is permitted on the land. The amount of compensation is
determined based on the economic value of converting forests to com production, as well as
on the level of risk of the type of forest to deforestation. For example, conserving cloud
forests provides the highest compensation rate per hectare. Since these programs began in
2003, more than 3.2 million hectares of forests have been conserved (FAO 2013). The PES
programs may have contributed to the decreased amount of change in forest fragmentation
during the 2008-2013 period. Additionally, the Inter-Ministerial Climate Change
Commission (CICC) was formed in 2005 to coordinate the development of Mexicos climate
change policies, programs, and strategies (UN-REDD 2013). In 2009, a United Nations
Reducing Emissions from Deforestation and Forest Degradation (UN-REDD+) working
group was established within the CICC (UN-REDD 2013). The UN-REDD+ program aims
52


to create financial value for preserving carbon stocks in forests through reduction of
deforestation and forest degradation (fragmentation), as well as integrate forest conservation
and sustainable management strategies for countries. Efforts to preserve forests in Mexico
for carbon storage starting with the creation of the CICC in 2005 and the integration of the
UN-REDD+ program in 2009 may have also contributed to the decrease in overall change in
fragmentation levels during the 2008-2013 time period in this study. It is also known that the
overall rate of deforestation has decreased by 55 percent from 2000-2010 (FAO 2010). All
of this information reflects successful policy measures in forest conservation, which reflect in
the findings of this study that the time period of 2008-2013 experienced much less overall
change in fragmentation than from 2002-2008 in both temperate and tropical forests.
Fragmentation class to fragmentation class transitions
Although the overall percentages of the total forest area of the fragmented forest
classes are smaller than the percent of the total forest of core and edge classes, the transitions
have significance for biodiversity conservation. For example, for both temperate and tropical
forests during both time periods, aside from transitioning to no forest or remaining the same
fragmentation class, the majority of bridge forests turned into branch forests. The bridge
fragmentation class indicates forests that connect two different core areas, and the branch
fragmentation class represents forests that connect at one end to an edge, perforation, bridge,
or loop. This means that relatively high percentages of forests that originally acted as a
connector from one large, intact region of forest to another large, intact region of forest
(bridge) now connect to a fragmented patch of forest rather than a core forest (branch). This
transition indicates that a core forest at one end of the original bridge connector has been
fragmented, and corridors between two interior forests are being lost.
53


Additionally, a relatively high percentage of both bridge in edge forests and loop in
edge forests turned into edge forests. A bridge in edge forest is a connector within the edge
fragmentation class that connects two core areas, and a loop in edge forest is a connector
within an interior edge (meaning an edge within a perforation in the core of a forest) that
connects the same core area. When these two classes transition to the edge fragmentation
class, they are no longer connecting to a core forest area. This, again, contributes to the loss
of corridors connecting core forests.
Lastly, a relatively high percentage of loop in perforation forests turned into
perforations. A loop in perforation fragmentation class connects the same core forest. They
are located within perforations within the interior of the forest. When loop in perforation
forests transition to perforations, that forest area is completely lost and turns into an area of
no forest within the interior of a core forest. This indicates the loss of the forest that connects
the two ends of areas of no forest within core forests (perforations).
All of the transitions above indicate the loss of corridors between two areas of large,
intact core forests. As mentioned before, although these types of fragmentation classes make
up less total forest area, they are important to preserve due to their effect as biological
corridors. These corridors aid in the conservation of biodiversity, particularly in a
fragmented landscape (Fahrig & Merriam 1994). Therefore, the results of this study can aid
in identifying these important corridors to aid in conservation efforts in a fragmented
landscape.
54


Transitions Between Fragmentation Classes During 2002-2008 and 2008-2013 Periods
Differentiating Between Internal and External Fragmentation Classes
As mentioned previously, differentiating between internal and external fragmentation
classes generated results very similar to the results without differentiating the two. Each
fragmentation class had a relatively high percent change to no forest, tropical forests
experienced more change in fragmentation classes than temperate forests, and the 2002-2008
time period experienced more change in fragmentation classes than the 2008-2013 time
period. Much of the changes detected in the internal fragmentation classes are change to
external core forests. This is likely because all of the internal fragmentation classes are
within perforations within the core of a forest, and as mentioned previously, due to the
original scale of the data, the aggregation to broad forest types of temperate and tropical
forests, and the 250x250 meter cell size used in this analysis, fine details within the
perforations of a forest become difficult to detect. Future studies conducted over smaller
geographical areas and at higher resolution are more likely to detect differences in trends
between internal and external fragmentation classes.
Limitations and Sources of Error
When evaluating the results and discussion presented above, the following sources of
error should be kept in mind. First, there may be errors in the land use/land cover
classifications between the Series III, IV and V data sets. The original land use/cover data
sets are at a scale of 1:250,000. At this scale, small forest areas are difficult to detect and the
shape of the forest area borders are not reported extremely accurately. Second, the
generalization effects introduced by using a 250x250 meter cell size may decrease the
amount of detail defining the edges of forests. A sensitivity analysis using a smaller cell size
55


performed on a smaller region of Mexico may help in quantifying the effects of cell size on
the results of this analysis. Third, there is a generalization effect when grouping more
detailed forest cover classes into the broader tropical and temperate forests classes used in
this study. This could indirectly contribute to smooth or loose detail on the shape of the
forest area borders. Fourth, there is a difference in length between the 2002-2008 period (six
years) and 2008-2013 period (five years) as well as the fact that the INEGIs Series III, IV
and V were created using inputs (e.g. satellite images) collected over a period of time that
may extend to more than one year (Nino & Victoria 2013). Even with each of these four
possible sources of error, none by itself or in combination is likely large enough to change
the clear and large trends identified at the national level in the results of this study.
56


CHAPTER VI
CONCLUSIONS AND RECOMMENDATIONS
This study used the MSPA method to define the fragmentation classes of temperate
and tropical forests in Mexico for 2002, 2008, and 2013. This study additionally explored
the transitions between each fragmentation class during the two time periods of 2002-2008
and 2008-2013 to gain an understanding of the evolution of forest fragmentation in Mexico
during these times. There is extensive research regarding the effects of forest fragmentation
on specific species and ecosystems in Mexico, but there is a lacking of literature on the state
of fragmentation of Mexican forests on a national scale, along with cartographic
representations of forest fragmentation. Additionally, there is no literature detailing the
change of specific forest fragmentation classes over time. This study helps to fill in these
gaps in the current literature using the most up to date land cover data and methodology to
define levels of forest fragmentation.
MSPA Method
Using the MSPA method to identify forest fragmentation levels has several
advantages. It is conceptually simple to understand by diverse stakeholders, and
computational easy to implement using freely available software. These characteristics
facilitate the periodical replication of fragmentation assessments at the national level, as well
as at smaller geographical extents. The MSPA method produces cartographic products that
allow the exploration of the spatial relationships of forest fragmentation over time, as well as
the analysis of the relationships between fragmentation patterns to other environmental and
anthropogenic factors. For example, the study of the potential correlation between
fragmentation classes and environmental factors such as topography, and anthropogenic
57


factors such as distance to urban areas, agricultural land uses, and socioeconomic factors
such as dominant economic activity or poverty levels. The MSPA results provide scientists,
managers and the general public a visual representation of the forest fragmentation that they
can visually explore and correlate to their empirical knowledge and on-the-ground
experiences. Finally, the MSPA output assigns unique identifiers to each fragmentation class
which facilitates further analysis, such as the cross referencing of fragmentation classes
between dates as it was done in this study.
Summary of Findings
This study provides insights into how forest fragmentation has evolved between the
2002-2008 and 2008-2013 periods at the national level in Mexico in both tropical and
temperate forests. All fragmentation classes for both periods experienced a higher percent
change to no-forest than to other fragmentation classes. The smaller, isolated fragmentation
classes (islets) and elongated, narrow fragmentation classes (branch, bridge, and loop)
experienced the highest percent change to no forest from one period to the next. Tropical
forests experienced a higher percent change to no forest than temperate forests across both
periods. The percent change of all fragmentation classes to no forest was less during the
2008-2013 than during the 2002-2008 period. There was loss of important corridors that
connect core forest areas in both forest types and time periods. When differentiating between
internal and external fragmentation classes, the results generated were very similar to the
results when the two were not differentiated. The transitions of the internal fragmentation
classes to other classes or to no forest were very small, frequently less than one percent
change.
58


Differentiating between internal and external fragmentation classes did not provide
significant additional benefits at the national scale used in this study. The scale of the source
data (1:250,000) does not portrait in detail the small gaps in the forest core areas nor the
shape of the edges of the forest areas, and hence the internal fragmentation classes are
underrepresented. However, studies at finer resolutions and larger scales over smaller
geographical areas would be capable of detecting small perforations and fine details of the
internal forest areas, and hence benefits from differentiating between internal and external
fragmentation classes could be realized.
Implications for Science and Policy
The results of this study provide tabular data and cartographic results quantifying the
degree of forest fragmentation in both temperate and tropical forests at a national scale in
Mexico. Additionally, it provides tabular data and cartographic results quantifying the
percent of each forest fragmentation class that transitioned to all other fragmentation classes
across two time periods of 2002-2008 and 2008-2013. This information adds landscape level
forest fragmentation data to a robust amount of literature that quantifies forest fragmentation
and its effects on specific species and ecosystems at local or regional scales. These results
can be incorporated into national level forest management efforts, such as the REDD+ and
PES programs. The cartographic results can be understood by a variety of stakeholders more
effectively than only tabular data. Additionally, the cartographic results can be used to
identify areas with a high degree of fragmentation, areas with high amounts of forest
transitioning to no forest, or areas with a high loss of corridor forests. Targeted policy
measures can be incorporated into a variety of conservation initiatives such as PES programs,
Forest Management Plans, REDD+ efforts, and education and awareness in these identified
59


regions or communities. The effectiveness of conservation efforts can be tracked using the
methodology used in this study to provide the detailed percentages of the composition of
temperate and tropical forests as well as the percent change of forest fragmentation classes
over time. Overall, the results of this study can be used to see how conservation efforts are
working across the landscape in Mexico, and identify areas where resources need to be
allocated the most to conserve the forests in Mexico.
Future Recommendations
This study does not differentiate between natural and anthropogenic causes of forest
fragmentation which is important for conservation purposes. Future studies can use GIS
systems to overlay the cartographic results of this study with maps of anthropogenic
activities such as urban developments, agricultural uses, and cattle ranching to better
understand the driving factors of forest fragmentation in different geographic areas.
Additionally, the fragmentation classes can be cross-referenced with socioeconomic data at
the county or state level in order to enhance the understanding of the relationships between
socioeconomic factors and activities with land use change and forest fragmentation
processes, and their impacts on forest ecosystems over time. Lastly, since the beginning of
this study an updated version of the Guidos Toolbox containing the MSPA has been released
(Guidos Toolbox 2.3, Revision 1). The newer version incorporates more image analysis
tools that can provide further information such as differentiating between forest areas natural
edge interface and an anthropogenic/artificial edge interface, this is valuable information that
would enhance forest conservation efforts.
Due to the influence of forest fragmentation on ecological processes and its impact on
the sustainability of the remaining forest areas in Mexico, assessments and monitoring of the
60


level of fragmentation of the forests should be incorporated into national forest inventories
and forest ecosystems health reports. This information can enhance the prioritization and
targeting of conservation efforts of the remaining forests in the country.
61


REFERENCES
Bogaert, J., Ceulemans, R., & Salvador-Van Eysenrode, D. (2004). Decision tree algorithm
for detection of spatial processes in landscape transformation. Environmental
management, 33( 1), 62-73.
Cayuela, L., Benayas, J. M. R., & Echeverria, C. (2006). Clearance and fragmentation of
tropical montane forests in the Highlands of Chiapas, Mexico (1975-2000). Forest
Ecology and Management, 226(1), 208-218.
de Mexico, S. A. T. (1999). Tropical rain forest fragmentation, howler monkeys (Alouatta
palliata), and dung beetles at Los Tuxtlas, Mexico. American Journal of
Primatology, 48, 253-262.
Donovan TM, Flather CH 2002. Relationship among North American songbird trends,
habitat fragmentation, and landscape occupancy. Ecological Applications, 12, 364-
374.
Estrada, A., & Coates-Estrada, R. (1996). Tropical rain forest fragmentation and wild
populations of primates at Los Tuxtlas, Mexico. International journal of
primatology, 17(5), 759-783.
Estrada, A., Coates-Estrada, R., & Meritt, D. (1993). Bat species richness and abundance in
tropical rain forest fragments and in agricultural habitats at Los Tuxtlas,
Mexico. Ecography, 16(A), 309-318.
Estreguil, C., Caudullo, G., de Rigo, D., San Miguel, J. (2012). Forest Landscape in Europe:
Pattern, Fragmentation and Connectivity. European Commission Joint Research
Centre Institute for Environment and Sustainability.
Fahrig, L. (2003). Effects of habitat fragmentation on biodiversity. Annual review of
ecology, evolution, and systematics, 487-515.
Fahrig, L., & Merriam, G. (1994). Conservation of fragmented populations. Conservation
biology, 5(1), 50-59.
FAO (Food and Agriculture Organization of the United Nations). 2010. Global Forest
Resources Assessment 2010. FAO, Rome.
FAO. (2013). Case studies on renumeration of positive externalities (RPE)/payments for
environmental services (PES). Available:
http://www.fao.org/fileadmin/user_upload/pes-project/docs/FAO_RPE-PES_PSAH-
Mexico.pdf
Fazey, T, Fischer, J., & Lindenmayer, D. B. (2005). What do conservation biologists
publish? Biological conservation, 124(1), 63-73.
62


Fischer, J., & Lindenmayer, D. B. (2007). Landscape modification and habitat fragmentation:
a synthesis. Global Ecology and Biogeography, 16(3), 265-280.
Garcia-Gigorro, S., & Saura, S. (2005). Forest fragmentation estimated from remotely sensed
data: is comparison across scales possible? Forest Science, 57(1), 51-63.
Groom, M. J., G. K. Meffe, and C. Ronald (2005). Principles of Conservation Biology. 3rd
Edition. Sinauer Associates, 779 pp. 213-252.
Groombridge, B., Jenkins, M.D. (2000). Global biodiversity. In: United Nations Environment
Programme (Ed.), Earths Living Resources in the 21st Century.
Haila, Y. (1999). Islands and fragments. In: Hunter MLJ (editor). Maintaining biodiversity
in forest ecosystems. Cambridge University Press, Cambridge, UK. pp. 234-264.
Hall, F. G., Botkin, D. B., Strebel, D. E., Woods, K. D., & Goetz, S. J. (1991). Large-scale
patterns of forest succession as determined by remote sensing .Ecology, 628-640.
Harper, K. A., S.E. MacDonald, P.J. Burton, J. Chen, K.D. Brosofske, S.C. Saunders, E.S.
Euskirchen, D. Roberts, M.S. Jaith, and E. Per-Anders. 2005. Edge influence on
forest structure and composition in fragmented landscapes. Conservation Biology, 19\
768-782.
Harrison, S., and E. Bruna (1999). Habitat fragmentation and large-scale conservation: what
do we know for sure? Ecography, 22\ 225-232.
INEGI (Instituto Nacional de Geografia e Informatica) 2009. Guia para la interpretacion de la
cartografia uso del suelo y vegetacion, escala 1:250,000 Serie III. INEGI, Mexico. 77
pp. Available online at:
http://www.inegi.org. mx/prod_serv/contenidos/espanol/bvinegi/productos/geografia/p
ublicaciones/guias-carto/sueloyveg/l_250_IIESuelo_Vegeta.pdf Accessed August
2015
INEGI (Instituto Nacional de Geografia e Informatica) 2012. Guia para la interpretacion de la
cartografia uso del suelo y vegetacion, escala 1:250,000 Serie IV. INEGI, Mexico.
132 pp. Available online at:
http://www.inegi.org. mx/prod_serv/contenidos/espanol/bvinegi/productos/geografia/p
ublicaciones/guias-carto/sueloyveg/l_250_IV/l_250_IV.pdf. Accessed August 2015
63


INEGI (Instituto Nacional de Geografia e Informatica) 2014. Guia para la interpretation de la
cartografia uso del suelo y vegetacion, escala 1:250,000 Serie V. INEGI, Mexico.
200 pp. Available online at:
http://www.inegi.org.mx/geo/contenidos/recnat/usosuelo/doc/guia_interusosuelov.pdf
Accessed August 2015 (Site to download digital GIS
layer.http://www.inegi.org.mx/geo/contenidos/recnat/usosuelo/ Accessed March
2015.
Kupfer, J. A. 2012. Landscape ecology and biogeography: Rethinking landscape metrics in a
post-FRAGSTATS landscape. Progress in Physical Geography, 36(3): 400-420.
Kupfer, J.A. 2006. National assessment of forest fragmentation in the US. Global
Environmental Change, 16: 72-82.
Laurance, W. F., Camargo, J. L., Luizao, R. C., Laurance, S. G., Pimm, S. L., Bruna, E. M.,
Phillip, P.C., Williamson, B., Benitez-Malvido, J., Vasconcelos, H.I., Van Houtan,
K.S., Zartman, C.E., Boyle S.A., Didham, R.K., Andrade, A., Lovejoy, T. E. (2011).
The fate of Amazonian forest fragments: a 32-year investigation. Biological
Conservation, 144(1), 56-67.
Laurance, W. F., Delamonica, P., Laurance, S. G., Vasconcelos, H. L., & Lovejoy, T. E.
(2000). Conservation: rainforest fragmentation kills big trees. Nature, 404(6780),
836-836.
Lindenmayer, D. B., & Fischer, J. (2007). Tackling the habitat fragmentation
panchreston. Trends in Ecology & Evolution, 22(3), 127-132.
Mas, J. F., Velazquez, A., Diaz-Gallegos, J. R., Mayorga-Saucedo, R., Alcantara, C., Bocco,
G., Castro, R., Fernandez, T., & Perez-Vega, A. (2004). Assessing land use/cover
changes: a nationwide multidate spatial database for Mexico. International Journal of
Applied Earth Observation and Geoinformation, 5(4), 249-261.
McAlpine, C. A., & Eyre, T. J. (2002). Testing landscape metrics as indicators of habitat loss
and fragmentation in continuous eucalypt forests (Queensland, Australia). Landscape
Ecology, 77(8), 711-728.
Melo, F. P., Martinez-Sal as, E., Benitez-Malvido, J., & Ceballos, G. (2010). Forest
fragmentation reduces recruitment of large-seeded tree species in a semi-deciduous
tropical forest of southern Mexico. Journal of Tropical Ecology, 2 6(01), 35-43.
Metzger, J. P., & Decamps, H. (1997). The structural connectivity threshold: an hypothesis in
conservation biology at the landscape scale. Acta Oecologica,18(\), 1-12.
Metzger, J. P., & Muller, E. (1996). Characterizing the complexity of landscape boundaries
by remote sensing. Landscape Ecology, 11(2), 65-77.
64


Moreno-Sanchez, R., F. Moreno-Sanchez, and J. M. Torres-Rojo (2011). National
assessment of the evolution of forest fragmentation in Mexico. Journal of Forestry
Research 22: 167-174.
Moreno-Sanchez, R., J.M. Torres-Rojo, F. Moreno-Sanchez, S. Hawkins, J. Little, S.
MacPartland. 2012. National assessment of the fragmentation, accessibility and
anthropogenic pressure on the forests in Mexico. Journal of Forestry Research
23(4): 529-541.
Moreno-Sanchez, R., T. Buxton-Torres, K. Silbernagel, F. Moreno-Sanchez. 2014.
Fragmentation of the forests in Mexico: National level assessments for 1993, 2002
and 2008. International Journal of Statistics and Geography 5(2): 4-17.
Murcia, C, (1995). Edge effects in fragmented forests: implications for conservation. Trends
in Ecology & Evolution 10: 58-62.
Myers, N. (1988). Threatened biotas:" hot spots" in tropical forestsFnvironmentalist, 5(3),
187-208.
Neel, M. C., McGarigal, K., & Cushman, S. A. (2004). Behavior of class-level landscape
metrics across gradients of class aggregation and area. Landscape ecology, 19(4),
435-455.
Nino Alcocer, M. and Victoria Hernandez, A. (2013). Informacion de uso del suelo y
vegetacion escala 1:250,000 Serie V (Conjunto Nacional). Memorias del Congreso
SELPER (Sociedad Latinoamericana de Percepcion Remota y Sistemas de
Informacion Espacial). Available online at:
http://langif.uaslp.mx/selper/documentos/CD_SELPER_2013/MEMORIAS_SELPER
_PDF/Estudios_Tematicos/ID_008.pdf
Ochoa-Gaona, S., & Gonzalez-Espinosa, M. (2000). Land use and deforestation in the
highlands of Chiapas, Mexico. Applied Geography, 20(1), 17-42.
Ochoa-Gaona, S., Gonzalez-Espinosa, M., Meave, J. A., & Sorani, V. (2004). Effect of forest
fragmentation on the woody flora of the highlands of Chiapas, Mexico. Biodiversity
& Conservation, 13(5), 867-884.
O'neill, R. V., Milne, B. T., Turner, M. G., & Gardner, R. H. (1988). Resource utilization
scales and landscape pattern. Landscape Ecology, 2(1), 63-69.
Ostapowicz, K., Vogt, P., Riitters, K. H., Kozak, J., & Estreguil, C. (2008). Impact of scale
on morphological spatial pattern of forest. Landscape ecology, 23(9), 1107-1117.
65


Ribeiro, M. C., Metzger, J. P., Martensen, A. C., Ponzoni, F. J., & Hirota, M. M. (2009). The
Brazilian Atlantic Forest: How much is left, and how is the remaining forest
distributed? Implications for conservation. Biological conservation, 142(6), 1141-
1153.
Riitters, K. H., Wickham, J. D., O'neill, R. V., Jones, K. B., Smith, E. R., Coulston, J. W., ...
& Smith, J. H. (2002). Fragmentation of continental United States
forests. Ecosystems, 5(8), 0815-0822.
Riitters, K., Wickham, J., O'Neill, R., Jones, B., & Smith, E. (2000). Global-scale patterns of
forest fragmentation. Conservation Ecology, 4(2), 3.
Schleuning, M., Farwig, N., Peters, M. K., Bergsdorf, T., Bleher, B., Brandi, R., Dalitz, H.,
Fischer, G., Freund, W., Gikungu, M. W., Hagen, M., Garcia, F. H., Kagezi, G. H.,
Kaib, M., Kraemer, M., Lung, T., Naumann, C., M., Schaab, G., Templin, M., Uster,
D., Wagele, J. W., & Bohning-Gaese, K. (2011). Forest fragmentation and selective
logging have inconsistent effects on multiple animal-mediated ecosystem processes in
a tropical forest. PLoS One, 6(11), e27785-e27785.
Schmiegelow FKA, Monkkonen M. (2002). Habitat loss and fragmentation in dynamic
landscapes: Avian perspectives from the boreal forest. Ecological Applications, 12:
375-389
Soille, P, Vogt, P. (2008). Morphological segmentation of binary patterns. Pattern
Recognition Letters 30, 4:456-459, doi: 10.1016/j.patrec.2008.10.015
Tabarelli, M., Da Silva, J. M. C., & Gascon, C. (2004). Forest fragmentation, synergisms and
the impoverishment of neotropical forests. Biodiversity & Conservation, 13(7), 1419-
1425.
Thompson, J., Brokaw, N., Zimmerman, J. K., Waide, R. B., Everham III, E. M., Lodge, D.
J., Taylor, C.M., Garcia-Montiel, D., & Fluet, M. (2002). Land use history,
environment, and tree composition in a tropical forest. Ecological applications, 12(5),
1344-1363.
Trejo, I., & Dirzo, R. (2000). Deforestation of seasonally dry tropical forest: a national and
local analysis in Mexico. Biological Conservation, 94(2), 133-142.
Tscharntke T, Steffan-Dewenter I, Kruess A, Thies C. (2002). Contribution of small habital
fragmentes to conservation of insect communities of grassland-cropland landscapes.
Ecological Applications, 12: 354-363.
Turner EM (1996). Species loss in fragments of tropical rain forest: A review of the
evidence. Journal of Applied Ecology, 33: 200-209.
66


UN-REDD. (2013). The REDD desk Mexico. Available:
http://theredddesk.Org/countries/mexico#forest
Vellend, M., Verheyen, K., Jacquemyn, H., Kolb, A., Van Calster, H., Peterken, G., &
Hermy, M. (2006). Extinction debt of forest plants persists for more than a century
following habitat fragmentation. Ecology, 57(3), 542-548.
Vogt, P. (n.d.). MSPA Guide.
Vogt, P., Riitters, K. H., Estreguil, C., Kozak, J., Wade, T. G., & Wickham, J. D. (2007).
Mapping spatial patterns with morphological image processing. Landscape
Ecology, 22(2), 171-177.
Wilcove, D. S., McLellan, C. H., & Dobson, A. P. (1986). Habitat fragmentation in the
temperate zone. Conservation biology, 6, 237-256.
Zipperer, W. C. (1993). Deforestation patterns and their effects on forest patches. Landscape
Ecology, 5(3), 177-184.
67


Full Text

PAGE 1

EVOLUTION OF FOREST FRAGMENTATION IN TEMPERATE AND TROPICAL FORESTS IN MEXICO FOR 2002, 2008, AND 2013 by ELIZABETH CLAY B.S., University of North Carolina Wilmington, 2012 A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Master of Science Environmental Science 2016

PAGE 2

ii This thesis for the Master of Science degree by Elizabeth Clay h as been approved for the Environmental Science Program by Rafael Moreno Chair Peter Anthamatten Juan Manuel Torres Rojo November 18 2015

PAGE 3

iii Clay, Elizabeth (MS, Environmental Science) Evolution of Forest Fragmentation in Temperate and Tropical Forests in Mexico for 2002, 2008, and 2013 Thesis directed by Associate Professor Rafael Moreno ABSTRACT studying forest fragmentation in Mexico is fundamen rich biodiversity as well as the forest ecosystem functions and services This thesis develop s a national level assessment of the fragmentation of the temperate and tropical forests in Mexico for three dates 2002, 2 008 and 2013 (corresponding to the land use/cover layers known as Series III, IV and V created by the Instituto Nacional de Estadstica y Geografa INEGI in Mexico). Then, the forest fragmentation classes identified for a date (e.g. 2002) are cross referen ced with t he forest fragmentation classes identified for the next date (e.g. 2008 ) with the purpose of assessing how the forest area that falls in each of the fragmentation classes identified in this study are evolving in time. The same cross reference p rocess is carried out for 2008 2013 to analyze the changes during this period This analytical process explores the trends in the relationship between fragmentation levels and changes in fragmentation of the forests that occur in different forest fragment ation classes over time. The Morphological Spatial Pattern Analysis (MSPA) method and the GUIDOS Toolbox will be used to identify the forest fragmentation classes. The INEGI Series III, IV and V land cover layers are used to extract the areas covered by te mperate and tropical forest s. A raster overlay with unique identifiers technique and the Tabulate Areas Tool in ArcGIS were used to carry out the cross reference analysis. Understanding forest fragmentation is integral for

PAGE 4

iv biodiversity conservation. The results from this study can enhance the prioritization of conservation efforts of the r emaining forests in Mexico. The form and content of this abstract are approved. I recommend its publication. Approved: Rafael Moreno

PAGE 5

v ACKNOWLEDGMENTS I would like to express my gratitude to my advisor and thesis chair, Dr. Rafael Moreno who provided me endless support throughout my research through h is immense kn owledge, patience, and passion for the subject. His guidance helped me develop my research from the start to the finished product. I would also like to thank the rest of my thesis committee; Dr. Peter Anthamatten and Dr. Juan Manuel Torres Rojo for taking their time to contribute their individual expertise to better my thesis

PAGE 6

vi TABLE OF CONTENTS CHAPTER I. INTRODUCTION ................................ ................................ ............................ 1 II. REVIEW OF LITERATURE ................................ ................................ ........... 3 Forest Fragmentation ................................ ................................ ........................ 3 Characte rizing Spatial Patterns of Forests ................................ ........................ 7 Effects of Geographic Scale on Forest Fragmentation ................................ ... 11 Conclusions ................................ ................................ ................................ ..... 12 III. METHODOLOGY ................................ ................................ ......................... 13 Introduction ................................ ................................ ................................ ..... 13 Data S ets ................................ ................................ ................................ ......... 13 Morphological Spatial Pattern Analysis ................................ ......................... 15 Analyzing the Transitions Between Fragmentation Classes D uring the 2002 2008 and 2008 2013 Periods ................................ ................................ .......... 20 IV. RESULTS ................................ ................................ ................................ ....... 23 Fragmentation Levels in 2002, 2008, and 2013 with Intext Parameter Off ... 23 Fragmentation Levels in 2002, 2008, and 2013 with Intext Parameter On .... 26 Transitions Between Fragmentation Classes During 2002 2008 and 2008 2013 Periods Without Differentiating Between Internal and External Fragmentation Classes ................................ ................................ .................... 30 Transitions Between Fragmentation Classes During 2002 2008 and 2008 2013 Periods Differentiating Between Internal and External Fragmentation Classes ................................ ................................ ................................ ............. 39 V. DISCUSSION ................................ ................................ ................................ 47 Defining Fragmentation Classes of Temperate and Tropical Forests ............. 47

PAGE 7

vii Transitions Between Fragmentation Classes During 2002 2008 and 2008 2013 Periods Without Differentiating Between Internal and External Fragmentation Classes ................................ ................................ .................... 49 Transitions Between Fragmentation Classes During 2002 2008 and 2008 2013 Periods Differentiating Between Internal and External Fragmentation Classes ................................ ................................ ................................ ............. 55 Limitations and Sources of Error ................................ ................................ .... 55 VI. CONCLUSIONS AND RECOMMENDATIONS ................................ ......... 57 REFERENCES ................................ ................................ ............................... 62

PAGE 8

viii LIST OF TABLES TABLE 1. Vegetation Types from INEGI's Series III, IV, and V included in the definition of temperate forests and tropical forests ................................ ................................ ............... 14 2. Summary table of input parameters for MSPA. ................................ ................................ 19 3. MSPA outputs with unique values for each fragmentation class ................................ .... 20 4. Areas of each fragmentation class for temperate and tropical forests in 2002. ................ 24 5. Areas of each fragmentation class for temperate and tropical forests in 2008. ................ 24 6. Areas of each fragmentation class for temperate and tropical forests in 2013. ................ 25 7. Areas of each external and internal fragmentation class for temperate and tropical forests in 2002. ................................ ................................ ................................ ............................. 27 8. Areas of each external and internal fragmentation class for temperate and tropical forests in 2008. ................................ ................................ ................................ ............................. 28 9. Areas of each external and internal fragmentation class for temperate and tropical for ests in 2013. ................................ ................................ ................................ ............................. 29 10. Fragmentation classes transition matrix for tropical forests in the 2002 2008 period without differentiating between internal and external class es. ................................ ......... 31 11. Fragmentation classes transition matrix for temperate and forests in the 2002 2008 period without differentiating between internal and external classes. ................................ ......... 32 12. Fragmentation classes transition matrix for tropical forests in the 2008 2013 period without differentiating between internal and external classes. ................................ ......... 32 13. Fragmentation classes transition matrix for temperate forests in the 2008 2013 period without differentiating between internal and external classes. ................................ ......... 33 14. Fragmentation classes transition matrix for tropical forests in the 2002 2008 period differentiating between internal and external classes. ................................ ...................... 40 15. Fragmentation classes transition matrix for temperate forests in the 2002 2008 period differentiating between internal and external classes. ................................ ...................... 42

PAGE 9

ix 16. Fragmentation classes transition matrix for tropical forests in the 2008 2013 period differentiating between internal and external classes. ................................ ...................... 43 17. Fragmentation classes transition matrix for temperate forests in the 2008 2013 period differentiating between internal and external classes. ................................ ...................... 45

PAGE 10

x LIST OF FIGURES FIGURES 1. Seven basic output classes from the MS PA analysis ................................ ........................ 15 2. Foreground connectivity of eight versus four ................................ ................................ .. 16 3. Effects of increased edgew idth on MSPA results ................................ ............................ 17 4. Effects of transition pixels on (left) or off (right) on MS PA results ................................ 18 5. Effects of intext on (left) and intext off (ri ght) on MSPA results ................................ .... 19 6. MSPA output for tropical forests in 2002. Similar cartographic outputs were generated for 2008 and 2013 temperate and tropical forests. ................................ ............................ 26 7. MSPA cartographic output for fragmentation classes of the tropical forests in 2002 differentiating between external and internal fragmentation classes. Similar outputs were created for the 2008 and 2013 temperat e and tropical forests. ................................ ......... 30 8. Example of fragmentation classes in 2008 that were edge forest in 2002 for tropical forests. Similar results were generated for tempe rate and tropical forests for all dates and all fragmentation classes. ................................ ................................ ................................ .. 34 9. Example of fragmentation classes in 2008 that were external edge in 2002 for tropical fo rests. Similar results were generated for temperate and tropical forests for all dates and all fragmentation classes. ................................ ................................ ................................ .. 46

PAGE 11

1 CHAPTER I INTRODUCTION The effects of forest fragmentation in Mexico have been an extensively researched topic, particularly for specific plants or animals at a local or regional scale. However, forest fragmentation research in Mexico at the national scale focuses only on specific types of forests (Trejo & Dirzo 2000). Additionally, local and national studies do not use a common methodology to compare results over time. Lastly, there is a need for spatial reference when analyzing forest fragmentation to better understand influences on fragmentation and changes over time. This analysis will address these shortcomings by providing a spatially referenced national level assessment analyzing the changes in forest fragmentation over time for both temperate and tropical forests in Mexico. The study will use the most up to date land cover data for Mexico, and identify forest fragmentation classes using the latest methodology and tool that has been sanctioned and agreed by a large number of nati onal and international forestry agencies. The purpose of this study is to provide a national level assessment of the state of forest fragmentation for temperate and tropical forests in Mexico for 2002, 2008, and 2013, as well as analyze changes in fragmen tation classes of temperate and tropical forests in Mexico from 2002 2008 and again from 2008 2013 in relation to the fragmentation class of the previous date, and to determine the area of each of these new forest fragmentation classes. This study will an swer the following research questions: 1. What are the levels of forest fragmentation for both temperate and tropical forests in 2002, 2008 and 2013 in Mexico?

PAGE 12

2 2. How has each fragmenta tion class changed from 2002 2008 and 2008 2013 in temperate and tropical for ests in relation to the fragmentation class of the previous date? The results of this analysis will include maps of forest fragmentation for each date (2002, 2008 and 2013), as well as maps of the changes in fragmentation classes from one date to the next Also, tabular data will be provided to accompany each of these cartographic products in order to visualize and quantify changes in forest fragmentation in Mexico over time. The rest of this thesis is organized as follows: Review of Literature Methodolo gy Results Discussion Conclusions and Recommendations

PAGE 13

3 CHAPTER II REVIEW OF LITERATURE Forest Fragmentation Introduction to forest fragmentation Forest fragmentation, or habitat fragmentation, is a very broad term given to many different processes and effects of landscape change (Lindenmayer & Fischer 2007). It is used as an umbrella term for ecological processes, patterns of vegetation cover, and biotic responses that result fr om changes in landscape patterns (Lindenmaye r & Fischer 2007). There are many ways to define forest fragmentation, but it can be generally defined as a patches of sm aller total area, isolated from each other by a matrix of habitats unlike the However, there are numerous publications on the efforts to define fragmentation and on the theoretical approaches that h ave been taken to unders tand and assess it (Harrison & Bruna 1999, Haila 1999, Kupfer 2006, Lindenmayer & Fischer 2007, and Kupfer 2012). Forest fragmentation is one of the most researched topics by conser vation biologists (Fazey et al. 2005). There is a substant ial body of lit erature studying the effects of fragmentation on the functioning of ecosystems and conservation of diverse flora, fauna and ecological processes at different temporal and spatial scales (Laurance et al. 2011, Harrison & Bruna 1999, Turner 1996, Donovan & Flather 2002, Schmiegelow & Monkkonen 2002, Thompson et al. 2002, Tscharntke et al. 2002, Fahrig 2003, Groom et al. 2005, and Schleuning et al. 2011). Forests can become fragmented due to a variety of reasons such as expanded agricultural land, urban sett lements, transportation infrastructure, and fi re occurrence (Estreguil et al. 2012 ). Landscape modification and habitat (including

PAGE 14

4 forests) fragmentation are key drivers of global species and biodiversity loss, and are believed to negatively affect virtually all taxonomic groups of animals and plants, as well as key ecosystem components and functions for long periods of time (Tabarell i et al. 2004, Vellend et al. 2006, and Fischer & Lindenmayer 2007). I ncrease in fragmentation has also been identified as a major threat to the conservation of forest ecosystems as a whole (Murcia 1995, Laurance et al. 2000, Ribeiro et al. 2009, Harper e t al. 2005, and Laurance et al. 2011) Additionally, deforestation and forest fragmentation contribute to processes such as the greenhouse effect and climate change and result in negative changes in regional hydrology and bi ogeochemical cycles (Mas et al 2004). Extent of forest fragmentation in Mexico There have been a variety of studies examining the extent of forest fragmentation in Mexico particularly at local and regional scales. Many of these studies are focused on agriculture and urban developments have contributed to increased forest fragmentation (Ochoa Gaona & Gonzalez Espinosa 2000). Landsat satellite imagery has been used to determine annual deforestation rates and spatial patterns in the Chiapas h ighlands in Chiapas, Mexico. Using Landsat imagery to determine deforestation rates in the Chiapas highlands, it was determined that this area of Mexico is experiencing deforestation at a faster rate than the national average ( Ochoa Gaona & Gonzalez Espinosa 2000 ). Converting forests to agricultural uses has resulted in the amount of primary forest to decrease while the area of disturbed f orest, secondary vegetation, and developed areas have increased ( Ocho a Gaona et al. 2006 ). Landsat imagery analyzed during the period of 1975 2000 in the Chiapas highlands determined that the amount of continuous forest cover has decreased, while the

PAGE 15

5 amou nt of fragmented forest patches increased ( Cayuela et al. 2006 ). Additionally, the deforestation rates continuously increased over time in the Chiapas Highlands ( Cayuela et al. 2006 ). These studies using satellite imagery can estimate the deforestation r ates and general spatial patterns of forests (such as patch sizes and degree of isolation), but the specific degree of fragmentation of the forests cannot be analyzed using these methods. Trejo & Dirzo ( 2000 ) conducted a national level assessment of changes in seasonally tropical dry forests in Mexico in the early 1990s using land cover maps classified as : i ntact altered seasonally dr y tropical forests (forest fragments), degraded seasonally dry tropical forests (smaller and more isolated fragments), or converted seasonally dry tropical forests. Although these results provide a national level assessment, they focus only on one forest type (seasonally dry tropical forests) and do not specify the specific types of spatial patterns of the fragmented forest, only whether or not the forest was fragmented or not. National level assessments on patterns of forest fragmentation in Mexico have b een performed in the past (Moreno Sanchez et al. 2014, 2012, and 2011). Moreno Sanchez ( 2011 ) provides tabular data on the change in forest fragmentation in Mexico over time, but no spatial reference. Moreno Sanchez ( 2012 ) provided a national level asses sment on changes forest fragmentation in Mexico in relation to anthropogenic pressure. Updated land cover data has been released since the previous papers on forest fragmentation in Mexico have been published. Additionally, a new methodology for defining the spatial pattern of the forests will be used in this study in which more specific fragmentation classes can be determined. This methodology has been sanctioned and agreed upon by a large number of national and international forestry agencies.

PAGE 16

6 Overall t here is much more extensive research regarding forest fragmentation and deforestation in the southern tropical forests of Mexico than the rest of the nation. Additionally, there is much more research quantifying the deforestation rates rather than prov iding quantifiable metrics regarding spatial patterns of forests in Mexico. Effects of f orest fragmentation in Mexico the five biologically richest cou ntries (Groombridge and Jenkins 2000), t herefore, studying forest fragmentation in Mexico is immensely import rich biodiversity and ecosystem services. There have been many studies examining the effects of forest fragmentation on specific spec ies or ecosystems in Mexico. For example, the tropical rainforests in Los Tuxtlas, Veracruz are heavily fragmented due to anthropogenic activities, particularly agriculture and the conversion of the forest to pasture for cattle grazing (Myers 1988) This fragmentation and disappearance of rain forest in Los Tuxtlas has resulted in 80 90 percent reduction of the area of natural habitat for howler monkeys (Estrada & Coates Estrada 1996). Dung beetles in Los Tuxtlas are dependent on howler monkey dung which contribute to nutrient cycling in the soil. With a decrease in howler monkeys due to reduced habitat, dung beetle populations could experience a decrease and alter the biogeochemical cycles in the rainforest (de Mexico 1999). Additionally, specie s richness of bat species in the Los Tuxtlas region is more likely to decline as forest fragments become more isolated (Estrada et al. 1993). Since bat species in the order of Chiroptera constitute 40 50 percent of the mammal species in the region, they a re vital for species richness and biodiversity of mammals in these ecosystems (Estrada et al. 1993). Fragmented forest landscapes in

PAGE 17

7 that most larger mammals wer e eliminated from forest fragments less than 30 hectares in size therefore influencing seed dispersal of larger seedlings and effecting the regeneration of the forests (Melo et al. 2010). Again, there is an abundance of research describing the patterns an d effects of forest fragmentation in the specific regions and forest types in Mexico, and a lack of literature on forest fragmentation on the national scale. Characterizing Spatial Patterns of Forests Introduction It is important to not only consider the s ize of forest patches when researching forest fragmentation, but to quantify the degree of fragmentation by measuring the spatial pattern of the forest on the landscape (Fahrig 2003). The spatial pattern of the forest can be defined as the spatial distrib ution of the forest across the landscape (Estreguil et al. 2012 ). As mentioned previously, many studies on forest fragmentation examine patch size and degree of isolation as determinates of forest fragmentation. Examining only forest patch size as a mea sure of fragmentation does not always accurately portray the effects of forest fragmentation on the ecosystem (McAlpine & Eyre 2002 ). When analyzing the effects of forest fragmentation on woody vegetation in the Chiapas highlands, it was found that fragm entation size, shape of the forest fragments, and possibly the forest matrix all contributed to the species richness within forest fragments ( Ochoa Gaona et al. 2006 ). Ochoa Gaona et al. (2006) suggested that there may be too much emphasis on the size and isolation of forest patches in forest conservation plans, and the need of more emphasis on the effects of the shape of forests fragments on biodiversity.

PAGE 18

8 Comparison of m ethods Two major concepts of characterizing spatial patterns of forest have been implemented through different methodologies in different areas of the world. These applications do not only determine fragmentation based on patch size and isolation of forest patches, but they deter mine the different spatial patterns of patches of forests as well as differentiate between external and internal fragmentation. Differentiating between external and internal forest fragmentation is an important aspect of describing fragmentation because i nternal forest fragmentation increases the amount of edge within the forest, rather than on the boundary of the external forest patch (Vogt et al. 2007). These interior edge effects may impact species differently than edge effects on exterior forests ( Zip perer 1993 ). The first method is a landscape level estimation of sp atial patterns of fragmentation. Metzger and Muller ( 1996 ) used Landsat Thematic Mapper (TM) images to create indices that reflect the typology and complexity of forest boundaries in southeastern Brazil. This methodology was able to determine the complexity of elongated and fragmented forest patterns in Brazil (Metzger & Muller 1996). Metzger and Decamps (1997) use d similar methodology to derive landscape level estimations of the spa tial patterns of fragmentation to determine connectivity of the forests. Their model was based on three components of the landscape; percolation measurement, the complexity and quality of corridor networks, and the per meability of the habitat matrix (Metz eger & Decamps 1997). Incorporating the complexity of the forest fragments allows for a better understanding of how forest fragmentation effects species survival and biodiversity. Bogaert et al. (2004) created an algorithm to define fragmentation classes differentiating between internal and external fragmentation at the landscape level through observed changes in patch area, number of

PAGE 19

9 patches, and patch perimeter in different landscapes However, using patches to define fragmentation across large geograp hic areas (typically landscape level) may produce inaccurate estimates of patch size and shape due to the fact that there will be a large amount of patches in the spatial extent, which may require altering of the data and losing information about the small forest patches (Vogt et al. 2007). Additionally these methods of defining forest fragmentation can assign landscapes with different arrangements of forest the same value of fragmentation ( Neel et al. 2004 ) An alternative to landscape level estimation of spatial patterns of forest fragmentation is pixel level mapping of internal and external forest fragments. Pixel level mapping of forest fragmentation offers higher sensitivity to forest shapes than landscape level analysis and can more accurately portray changes over time (Vogt et al. 2007). One application of pixel level mapping is t hrough a method developed by Rii t ters et al. (2000). This method is based on image convolution and uses windows surrounding each forest pixel to measure the amount of forest and its occurrence as adjacent fore st pixels (Rii t ters et al. 2000 ). The model classifies each forest location according to the type of fragmentation that exists in the surrounding landscape that can be defined at differen t scales (different window sizes). This defines forest fragmentation as a property of the landscape that contains the forest, rather as a property of the forest itself (Rii t ters et al. 2002). This methodology results in specific forest fragmentation clas ses such as interior, perforated, edge, transitional, and patch (Rii t ters et al. 2000). The moving windows methodology has been used in several different landscape scale studies First, this method was used to analyze global forest fragmentation (Rii t ter s et al. 2000). It was also used to determine forest fragmentation in the continental United States

PAGE 20

10 (Rii t ters et al. 2002) as well as to analyze forest fragmentation changes over time in Mexico (Moreno Sanchez et al. 2014). Another primary method used fo r determining th e spatial patterns of forests using pixel level mapping is through morphological image processing. Morphological image processing analyzes the shape and form of objects in an image at the pixel level of a binary map offering the same adva ntages of mapping at the pixel level as image convolution in a moving window (Soille & Vogt 2008). However, m orphological image processing is not dependent on the size of window used and therefore offers more accuracy across all geographic scales at the p ixel level than the image convolution technique (Vogt et al. 2007). When comparing image convolution versus morphological image processing when evaluating forest fragmentation in a national park in Italy, the morphological image processing remained more a ccurate with changes in scale of the input map than the image convolution. For example, small patches of forest remained small patches, regardless of the scale allowing for accurate landscape level analysis at the pixel scale (Vogt et al. 2007). Morpholo gical image processing has been used to evaluate forest fragmentation and connectivity in Europe ( Estreguil et al. 2012 ) and by the United States Department of Agriculture (USDA) to evaluate forest fragmentation across the United States as well as in select cities ( available online at http://www.forestthreats.org/research/tools/landcover maps/mspa ). It is important to note that with pixel level mapping, the resolution (pixel size) used for analysis is important in influencing the results, and input ma ps at different resolutions should not be used to directly compare results ( Ostapowicz et al. 2008 ). For example, an input map with a pixel size of 250 meters by 250 meters is going to define the spatial patterns of the forest differently than an input ma p of the same geographic area with a

PAGE 21

11 pixel si ze of 100 meters by 100 meters. The 250 meter resolution will capture more forest area per pixel than the 100 meter resolution, so therefore these two results should not be used to make comparisons of the spati al patterns of the same geographic forest area Based on previous studies and available literature, pixel level mapping of forest fragmentation using morphological image processing is the most accurate at the landscape scale. Therefore, the Morphological Spatial Pattern Analysis (MSPA) contained in the GUIDOS Toolbox (Soille & Vogt 2008) will be used in this study to define forest fragmentation in Mexico. Effects of Geographic S ca le on Forest Fragmentation The geographic scale at which forest fragmenta tion studies are performed vary greatly which impacts the conclusions on the degree of fragmentation and its e ffects on the ecosystem (Fahrig 2003). Typically, studies performed at a smaller geographic scale indicate more interior forest than studies performed at a larger sc ale (Riitters et al. 2000). For example, when Riitters et al. (2000) used 81 square kilometer windows for a global assessment on forest fragmentation, about one third of the total g lobal forest is characterized as interior, whereas forests characterized as an edge or patch dominate when using 59,049 square kilometer windows. It is important to analyze forest fragmentation at large scales in order to define the level of fragmentation at the l andscape level (Riitters et al. 2000). With the increased availability of satellite data, it is possible to examine forest patterns at larger spatial and temporal scales ( Hall et al. 1991 ). Landscape scale analysis of forest patterns can relate fragmentation to the reduction of the area of the remaining forest patches, increased isolation of the fragments loss of overall connectivity, and increased edge effect and di sturbances from surrounding pressures ( Garcia Gigorro & Saura 2005 ). The resu lting

PAGE 22

12 spatial patterns of the forest on a landscape scale can be used to determine how changes in the spatial pattern impacts ecosystem processes such as habitat provision, gene flow, pollination, wildlife dispersal, or pest propagation ( Estreguil et al. 2012 ). Conclusions Overall, there have been numerous studies describing the effects of forest fragmentation in Mexico on specific species or ecosystems. These studies provide insight on the importance of habitat conservation, but do not provide a national level analysis on the state of forest fragmentation in Mexico. Many studies in Mexico do not quantify forest fragmentation in relation to the spatial patterns of the forest. Rather, forest fragmentation is commonly described as the size, amount of core area, and degree of isolation of forest patches. Previous studies examining forest fr agmentation in Mexico on a national scale do not provide spatial reference, only describe one forest type, or do not use the m ost updated MSPA methodology Additionally there is no literature relating the changes in area in each fragmentation class to all other fragmentation classes over time to get a more specific picture of how the f ragmentation of the forests are changing in Mexico. This study will fill these gaps in the literature by providing a national level assessment of forest fragmentation in Mexico using the most updated MSPA methodology in order to describe forest fragmentation in terms of the spatial patterns of the forest. Additionally, this study will relate each fragmentation class in one date to the fragmentation classes of the next date during the two time periods of 2002 2008 and 2008 2013 in order to determine specifically how forest fragmentation is evolving over ti me in Mexico. Lastly, this study will provide cartographic results for the classes of forest fragmentation as well as for the transitions of each forest fragmentation class from one date to the next.

PAGE 23

13 CHAPTER III METHODOLOGY Introduction The research que stions in this study are answered using land cover/land use data for Mexico. Several tools are used to carry out the two fold research questions. First, the geographic information system ArcGIS 10.2 (ESRI, Redlands CA) was used to prepare the land cover/ land use data to be able to define the levels of fragmentation. Next, the levels of forest fragmentation are defined using the freely available Morphological Spatial Pattern Analysis (MSPA) contained in the GUIDOS Toolbox (Soille & Vogt 2008). ArcGIS was then used again to perform the cross referencing of the fragmentation class from one date to the fragmentatio n classes of the following date The specific processes are detailed below. Data S ets L and use and land cover data from the National Institute of Geography and Informatics (INEGI) in Mexico (INEGI 2014, 2012, 2009). The most recent version s of land use and land cover data from INEGI will be used for the analysis. The original data are in vector format and consists of all vegetation types and l and uses in Mexico for the years 2002, 2008, and 2013, known as Series III, Series IV, and Series V respectively. INEGI has homogenized the land use/ land cover classes in these layers, hence temporal changes can be evaluated. The INEGI land cover vector data are at a scale of 1: 250,000 and provided in Lambert Conformal Conic GRS 1980 projection with linear units of meters. For the purpose of a national level assessment, the specific land use/land cover classes were reclassified to the two forest types of temperate and tropical forests for each date. Table 1 specifies the original vegetation types included in the temperate and tropical forests classes defined in this

PAGE 24

14 study. The temperate and tropical forests vector data sets were converted to raster fo rmat using a cell size of 250 x 250 meters for each date. The cell size was chosen based on the scale of the original data layers (1:250,000), the level of locational certainty of the features in the original layers, and the consideration that, although r esponses by plants, animals and ecosystem functions to edge effects vary widely, 250 meters represents a conservative estimate of the level of penetration of edge effect for many species and processes (Harper et al. 2005). Table 1 Vegetation Types from INEGI's Series III, IV, and V included in the definition of temperate forests and tropical forests Temperate Forests INEGI's CVE_UNION code Description BA Bosque de Oyamel BB Bosque de Cedro BC Bosque Cultivado BG Bosque de Galeria BI Bosque inducido BJ Bosque de Tascate BM Bosque Mesofilo de Montana BP Bosque de Pino BPQ Bosque de Pino Encino BQ Bosque de Encino BQP Bosque de Encino Pino BS Bosque de Ayarin Tropical Forests SAP Selva Alta Perennifolia SAQ Selva Alta Subperennifolia SBC Selva Baja Caducifolia SBK Selva Baja Espinosa SBP Selva Baja Perennifolia SBQ Selva Baja Subperennifolia SBS Selva Baja Subcaducifolia SG Selva de Galeria SMC Selva Mediana Caducifolia INEGI's CVE_UNION code Description SMP Selva Mediana Perennifolia SMQ Selva Mediana Subperennifolia SMS Selva Mediana Subcaducifolia VSA/PT Vegetacion Secundaria de Selvas Arborea/Vegetacion de Peten VSA/SAP Vegetacion Secundaria de Selvas Arborea/ Selva Alta Perennifolia VSA/SAQ Vegetacion Secundaria de Selvas Arborea/ Selva Alta Subperennifolia VSA/SBK Vegetacion Secundaria de Selvas Arborea/ Selva Baja Espinosa VSA/SBQ Vegetacion Secundaria de Selvas Arborea/ Selva Baja Subperennifolia VSA/SG Vegetacion Secundaria de Selvas Arbo rea/ Selva de Galeria VSA/SMQ Vegetacion Secundaria de Selvas Arborea/ Selva Mediana Subperennifolia VSA/SMS Vegetacion Secundaria de Selvas Arborea/ Selva Mediana Subcaducifolia VSA/BS Vegetacion Secundaria de Selvas Arborea/Bosque de Ayarin

PAGE 25

15 Morphological Spatial Pattern Analysis Interface for the Description of image Objects and their Shapes (GUIDOS) Toolbox was of forest fragmentation for Spatial Pattern Analysis (MSPA) (Soille & Vogt 2008) within the GUIDOS Toolbox was used to define the levels of forest fragmentation for each date in Mexico. The MSPA determines forests fragmentation classes based on the geometry and connectivity of a binary input image as well as user specified input parameters (Soille and Vogt, 2008). The MSPA results in seven basic classes based on pix el value. The seven classes include: 1. Core: Interior foreground area excluding foreground perimeter 2. Islet: Disjoint foreground object and too small to contain core 3. Loop: Connected at more than one end to the same Core area 4. Bridge: Connected at more than o ne end to different Core areas 5. Perforation: Internal foreground object perimeter 6. Edge: External foreground object perimeter 7. Branch: Connected at one end to Edge, Perforation, Bridge, or Loop Figure 1 shows an example of the seven basic classes that result from the MSPA analysis. Figure 1 Seven basic output classes from the MSPA analysis (Soille & Vogt 2008)

PAGE 26

16 The MSPA requires a specific input format in order to run the analysis. Using ArcGIS, the temperate and tropical forest layers for each date were formatted to input into the MSPA. First, each raster was reclassified to a binary image contain ing a backgro und (no forest value = 1 ) and a foreground (forest value = 2 ). Each raster was then be exported as an 8 bit GeoTiff with no compression to be imported into the GUIDOS Toolbox. In the GUIDOS Toolbox, an MSPA batch process was performed for temperate and tropical forests for each date. User defined input parameters must be selected when running the MSPA. The first parameter option is the foreground connectivity, options of eight or four. This option determines how the center pixel in sets of three by three pixels is connected to its adjacent neighboring pixels (Vogt, n.d.). A foreground connectivity of four results in neighboring pixels sharing a common pixel border, and a foreground connectivity of eight results in neighboring pixels sharing a comm on pixel border and corner (Vogt, n.d.). Figure 2 shows the difference between a foreground connectivity of type eight and four cells. For this analysis, a foreground connectivity of eight was used. The second parameter option is the edgewidth. The edg ewidth defines the thickness of the non fragmentation class. Figure 2 Foreground connectivity of eight versus four (Vogt n.d.)

PAGE 27

17 The actual distance of the edgewidth is the number of defined edge pixels multiplied by the resolution of the input data. Increased edgewidth results in a larger edge area and smaller core area. An edgewidth of one was used for this analysis, meaning the edge will have a thickness of 250 meters since the input raster resolution is 250 x 250 meters. Figure 3 displays the effects of edgewidth on the MSPA results. Figure 3 Effects of increased edgewidth on MSPA results (Vogt n.d. ). The third parameter option is transition. Transition pixels are the pixels of an edge or perforation where the core area intersects with a loop or a bridge. Turning the transition pixels off hides the transition pixels and the perforation and edge wil l be closed core boundaries. Turning the transition pixels on illustrates all detected connections, but may result in a more confusing image (Vogt, n.d.). Transition pixels were left on for this analysis in order to display all connections. Figure 4 dis plays the effec ts of leaving transition pixels on and turning the transition pixels off. The fourth parameter option is intext. The intext option distinguishes internal features from external fragmentation classes. Internal classes are features enclosed by a perforation. When intext is enabled, it adds a value of 100 to all outputs to separate internal

PAGE 28

18 Figure 4 Effects of transition pixels on (left) or off (right) on MSPA results (Vogt n.d.). This analysis was performed with both intext on and intext off to compare the results of distinguish ing and not distinguishing between internal and external forest classes. It is important to note that even with the intext parameter off, the MSPA still distinguishes perforations within the interior of a forest and forest fragments within a perforation, just not as detailed as wit h the intext parameter set to on. This still allows perforations within the interior of a forest to be differentiated than forest fragments separate from interior forests, which introduce different edge effects and ecological implications than forest patc hes outside of interior forests ( Zipperer 1993 ). Figure 5 displays the effects of the intext parameter. The fifth and final parameter option is statistics. The statistics option produces a text file of summary statistics from the MSPA. Statistics was tu rned on for this analysis. Table 2 provides a summary of all input parameters that will be used for the MSPA in this analysis. The MSPA results in GeoTiff raster images for temperate and tropical forests for all dates. Once complete, the output GeoTiffs were imported back into ArcGIS for analysis. Each value in the MSPA output represents a unique identifier for the different forest fragmentation classes, and these unique values were used to perform analyses.

PAGE 29

19 Figure 5 Effects of intext on (left) and intext off (right) on MSPA results (Vogt n.d.). Table 2 Summary table of input parameters for MSPA. Parameter Input for Analysis Foreground Connectivity 8 Edge Width 1 Transition On Intext On & Off Statistics On Table 3 displays the different fragmentation classes, their output color, and their unique values. Note the added value of 100 for internal versus external fragmentation classes. The colors may be manipulated once imported into ArcGIS, bu t the values will remain the same for each fragmentation class.

PAGE 30

20 Table 3 MSPA outputs with unique values for each fragmentation class (Vogt n.d.). Class Color RGB Value [byte] internal/external 1) Core 000/200/000 117/17 2) Islet 160/060/000 109/9 3) Perforation 000/000/255 105/5 4) Edge 000/000/000 103/3 5a) Loop 255/255/000 165/65 5b) Loop in Edge 255/255/000 167/67 5c) Loop in Perforation 255/255/000 169/69 6a) Bridge 255/000/000 133/33 6b) Bridge in Edge 255/000/000 135/35 6c) Bridge in Perforation 255/000/000 137/37 7) Branch 255/140/000 101/1 Background 220/220/220 100/0 Missing 255/255/255 129/129 Analyzing the Transitions Between Fragmentation Classes D uring the 2002 2008 and 2008 2013 Periods The MSPA results and ArcGIS were used to answer the second research question, from 2008 to 2013 in temperate and tropical forests in relation to the fragmentation class of the previous date? First, each forest fragmentation class from the ArcGIS rasters created from the MSPA results with the intext parameter set to off for both temperate and tropical forests in 2002 were isolated by reclassifying the fragmentation raster to gi ve the desired isolated fragmentation was performed for each fragmentation class for temperate and tropical forests in 2002. For each fragmentation class, this process will result in a raster containing only the desired fragmentation class from 2002. using the 2002 rasters, the value of three was reclassified to one and all other values were

PAGE 31

21 reclassi fied to NoData for tropical and temperate forests Therefore, the output raster only contained the edge fragmentation class from 2002. Next, the 2002 isolated fragmentation classes were clipped to the MSPA results for temperate and tropical forests in 2 008. This was done by extracting the values from the 2008 MSPA results using the 2002 isolated fragmentation classes as a mask. This results in the 2002 fragmentation areas with the 2008 fragmentation values in order to visualize what level of fragmentat ion the 2002 isolated classes turned into in 2008. From these results, the area of each 2008 fragmentation class in the 2002 masks were calculated using the unique values for each class using the Zonal Geometry tool in ArcGIS The resulting areas were in units of square meters, and were divided by 10,000 in order to get the results in units of hectares. This procedure was repeated by isolating all fragmentation classes for temperate and tropical forests in 2008 and clipping them to the fragmentation clas ses in 2013, and then calculating the areas in 2013. Additionally, these same procedures were repeated with the ArcGIS rasters generated from the MSPA results with the intext parameter set to on in order to differentiate between internal and external fores ts. In order to evaluate how the temperate and tropical forests have evolved from one date to the next, transition matrices were created for temperate forests from 2002 2008, tropical forests from 2002 2008, temperate forests from 2008 2013, and tropical f orests from 2008 2013. This wa s accomplished by first using the Zonal Geometry tool in ArcGIS to determine the area (in hectares) of each fragmentation class for the first date in each time period (i.e. 2002 for the 2002 2008 period and 2008 for the 2008 2013 period). Next, the percent change of each fragmentation class from the first date to each fragmentation class of

PAGE 32

22 the next date was calculated. This results in a transition matrix describing what percent of each fragmentation class in the first date changed into each fragmentation class from the next date for temperate and tropical forests for both time periods ( for example, eight percent of the branch fragmentation class in 2002 turned into edge in 2008) These processes result in a comprehensive n ational level assessment of what level of fragmentation the temperate and tropical forests of Mexico have changed into from 2002 2008 and 2008 2013

PAGE 33

23 CHAPTER IV RESULTS The results first discuss results generated from the MSPA for both temperate and tropical forests for 2002, 2008, and 2013 with the intext parameter set to off. Then the results from the MSPA for both forest types and all dates are discussed with the inte xt parameter set to on. Next the results from the transitions between fragmentation classes for both time periods with the intext parameter set to off are discussed, organized by major trends in the data. Lastly, the results from the transitions between fragmentation classes for both time periods with the intext parameter set to on are discussed. Fragmentation Levels in 2002, 2008, and 2013 with Intext Parameter Off Tables 4, 5 and 6 present the forest fragmentation classes resulting from the MSPA method for temperate and tropical forests in 2002, 2008 and 2013 respectively. These results were generated with the MSPA intext parameter set to off (meaning that fragmentation classes were not differentiated between internal an d external). This results in 11 fragmentation classes with unique identifiers Each tab le shows the area in hectares of each fragmentation class and the percent of the total forest area that falls within each fragmentation class for the temperate and tropical forests. There were a total 18,070,262 hectares of tropical forests and 22,126,451 hectares of tem perate forests in 2002 (Table 4) Approximately 73 percent of the total tropical forest area was classified as core (interior) forests, 17.2 percent were classified as edge, and a bout 10 percent fell in the fragmented classes (all other fragmentation classes except core and edge). Approximately 70 percent of the temperate forest areas were classified as core, 18.4 percent were classified as edge, and about 11 percent fell in the f ragmented categories.

PAGE 34

24 Table 4 Areas of each fragmentation class for temperate and tropical forests in 2002 2002 Tropical Forests 2002 Temperate Forests Fragmentation Class Area (ha) Percent of Total Forest Area (ha) Percent of Total Forest Branch 781,794 4.33 1,234,488 5.58 Edge 3,108,631 17.20 4,074,182 18.41 Perforation 531,331 2.94 602,025 2.72 Islet 67,088 0.37 134,444 0.61 Core 13,240,831 73.27 15,549,201 70.27 Bridge 137,556 0.76 223,613 1.01 Bridge in Edge 113,981 0.63 167,731 0.76 Bridge in Perforation 1,425 0.01 3,269 0.01 Loop 46,131 0.26 76,556 0.35 Loop in Edge 32,150 0.18 42,394 0.19 Loop in Perforation 9,344 0.05 18,550 0.08 Total Forest Area 18,070,262 22,126,451 There was a total of 17,957,907 hectares of tropical forests and 21,222,037 hectares of te mperate forest in 2008 (Table 5) Approximately 73 percent of the tropical forests were classified as core forests, 17.2 percent were edge, and about 10 percent fell in the fragmented classes. App roximately 69 percent of the temperate forests were classified as core, 19.4 percent as edge, and approximately 11 percent of fell in the fragmented classes. Table 5 Areas of each fragmentation class for temperate and tropical fore sts in 2008. 2008 Tropical Forests 2008 Temperate Forests Fragmentation Class Area (ha) Percent of Total Forest Area (ha) Percent of Total Forest Branch 827,375 4.61 1,217,513 5.74 Edge 3,087,438 17.19 4,117,194 19.40 Perforation 529,050 2.95 511,944 2.41 Islet 80,475 0.45 138,550 0.65 Core 13,030,538 72.56 14,686,731 69.21 Bridge 163,400 0.91 225,575 1.06 Bridge in Edge 130,438 0.73 170,938 0.81 Bridge in Perforation 2,744 0.02 3,025 0.01 Loop 56,825 0.32 83,356 0.39 Loop in Edge 37,013 0.21 49,019 0.23 Loop in Perforation 12,613 0.07 18,194 0.09 Total Forest Area 17,957,907 21,222,037

PAGE 35

25 There was a total of 18,425,013 hectares of tropical forests and 21,271,030 hectares of tem perate forests in 2013 (Table 6) Approximately 73 percent of the tropical forests were classified as core forest, 16.7 percent as edge, and about 10 percent fell in the fragmented classes. Approximately 69 percent of the temperate forests were classified as core, 19.5 percent as edge, and approximately 11 percent f ell in the fragmented classes. Table 6 Areas of each fragmentation class for temperate and tropical forests in 2013. 2013 Tropical Forests 2013 Temperate Forests Fragmentation Class Area (ha) Percent of Total Forest Area (ha) Percent of Total Forest Branch 847,975 4.60 1,222,613 5.75 Edge 3,078,619 16.71 4,147,906 19.50 Perforation 545,919 2.96 499,644 2.35 Islet 79,594 0.43 138,288 0.65 Core 13,447,188 72.98 14,716,174 69.18 Bridge 170,400 0.92 223,944 1.05 Bridge in Edge 135,581 0.74 170,694 0.80 Bridge in Perforation 2,813 0.02 2,825 0.01 Loop 61,756 0.34 82,881 0.39 Loop in Edge 39,650 0.22 48,931 0.23 Loop in Perforation 15,519 0.08 17,131 0.08 Total Forest Area 18,425,013 21,271,030 Overall, there was a loss of 112,355 hectares of tropical forests between 2002 and 2008, and a reported gain of 467,106 hectares between 2008 and 2013. There was a loss of 904,414 hectares of temperate forests between 2002 and 2008, and a reported gain of 48,993 hectares betw e en 2008 and 2013. In addition to tabular data, the MSPA results in spatially referenced fragmentation classes for the entirety of Mexico. Figure 6 shows a zoomed in sample area of the MSPA output for tropical forests in 2002 Similar cartographic outputs were generated for the temperate and tropical fo rests for 2002, 2008, and 2013.

PAGE 36

26 Figure 6 MSPA output for tropical forests in 2002. Similar cartographic outputs were generated for 2008 and 2013 temperate and tropical f orests. Fragmentation Levels in 2002, 2008, and 2013 with Intext Parameter On Tables 7, 8, and 9 present forest fragmentation classes resulting from the MSPA for temperate and tropical forests in 2002, 2008, and 2013 respectively differentiating between external and internal fragmentation classes (MSPA parameter intext set to on). This results in a t otal of 19 fragmentation classes with unique identifiers. Each tab le shows the area in hectares of each fragmentation class and the percent of the total forest area that falls within each fragmentation class for the temperate or tropical forests. In 2002 approximately 73 percent of the tropical forests were core forests (Table 7), of which almost all were external core forests (73.15 percent external versus 0.13 percent internal core). Approximately 17 percent of the tropical forests were external e dge and 0.13 percent were internal edge (meaning it was an edge forest within a perforation). Approximately 10 percent of the tropical forests were classified in the fragmented classes (96 percent in the external classes and 4 percent in the internal).

PAGE 37

27 Table 7 Areas of each external and internal fragmentation class for temperate and tropical forests in 2002. 2002 Tropical Forests 2002 Temperate Forests Fragmentation Class Area (ha) Percent of Total Forest Area (ha) Percent of Total Forest External Branch 742,894 4.11 1,171,469 5.29 External Edge 3,085,494 17.07 4,046,650 18.29 External Islet 66,313 0.37 132,863 0.60 External Core 13,217,810 73.15 15,526,460 70.17 External Bridge 133,350 0.74 213,194 0.96 External Bridge in Edge 111,344 0.62 161,906 0.73 External Loop 39,013 0.22 59,050 0.27 External Loop in Edge 32,038 0.18 42,269 0.19 Internal Branch 38,900 0.22 63,019 0.28 Internal Edge 23,138 0.13 27,531 0.12 Internal Perforation 531,331 2.94 602,025 2.72 Internal Islet 775 0.00 1,581 0.01 Internal Core 23,019 0.13 22,738 0.10 Internal Bridge 4,206 0.02 10,419 0.05 Internal Bridge in Edge 2,638 0.01 5,825 0.03 Internal Bridge in Perforation 1,425 0.01 3,269 0.01 Internal Loop 7,119 0.04 17,506 0.08 Internal Loop in Edge 113 0.00 125 0.00 Internal Loop in Perforation 9,344 0.05 18,550 0.08 Approximately 70 percent of temperate forests in 2002 were external core and only 0.1 percent internal core, and 18.2 percent were external and 0.1 percent were internal edge. Approximately 11 percent of the temperate forests were classified in the fragmented classes (97 percent in the external classes and 3 percent in the internal). In 2008 (Table 8 ), approximately 73 percent of the tropical forests were core forests of which almost all were external core forests ( 72.46 percent external versus 0.11 percent internal core). Approximately 17 percent of the tropical forests were external edge and 0.13 percent were internal edge (meaning it was an edge forest wit hin a perforation).

PAGE 38

28 Table 8 Areas of each external and internal fragmentation class for temperate and tropical forests in 2008. 2008 Tropical Forests 2008 Temperate Forests Fragmentation Class Area (ha) Percent of Total Forest Area (ha) Percent of Total Forest External Branch 779,788 4.34 1,165,412 5.49 External Edge 3,064,719 17.07 4,097,744 19.31 External Islet 79,688 0.44 137,381 0.65 External Core 13,011,530 72.46 14,671,090 69.13 External Bridge 156,419 0.87 218,488 1.03 External Bridge in Edge 126,856 0.71 166,675 0.79 External Loop 45,450 0.25 67,663 0.32 External Loop in Edge 36,869 0.21 48,981 0.23 Internal Branch 47,588 0.26 52,100 0.25 Internal Edge 22,719 0.13 19,450 0.09 Internal Perforation 529,050 2.95 511,944 2.41 Internal Islet 788 0.00 1,169 0.01 Internal Core 19,006 0.11 15,644 0.07 Internal Bridge 6,981 0.04 7,088 0.03 Internal Bridge in Edge 3,581 0.02 4,263 0.02 Internal Bridge in Perforation 2,744 0.02 3,025 0.01 Internal Loop 11,375 0.06 15,694 0.07 Internal Loop in Edge 144 0.00 38 0.00 Internal Loop in Perforation 12,613 0.07 18,194 0.09 Approximately 10 percent of the tropical forests were classified in the fragmented classes (96 percent in the external classes and 4 percent in the internal). Approximately 69 percent of temperate forests in 2008 were external core and only 0.07 percent internal core, and 19.31 percent were external edge and 0.09 percent were internal edge. Approximately 11 percent of the temperate forests were clas sified in the fragmented classes (97 percent in the external classes and 3 percent in the internal). In 2013 (Table 9 ), approximately 73 percent of the tropical forests were core forests of which almost all were external core forests ( 72.86 percent external versus 0.12 percent internal core). Approximately 17 percent of the tropical forests were external edge and 0.14

PAGE 39

29 percent were internal edge (meaning it was an edge forest within a perforation). Approximately 10 percent of the tropical fo rests were classified in the fragmented classes (96 percent in the external classes and 4 percent in the internal). Approximately 69 percent of temperate forests in 2013 were external core and only 0.07 percent internal core, and 19.42 percent were extern al edge and 0.08 percent were internal edge. Approximately 11 percent of the temperate forests were classified in the fragmented classes (97 percent in the external classes and 3 percent in the internal). Table 9 Areas of each ext ernal and internal fragmentation class for temperate and tropical forests in 2013. 2013 Tropical Forests 2013 Temperate Forests Fragmentation Class Area (ha) Percent of Total Forest Area (ha) Percent of Total Forest External Branch 796,106 4.32 1,172,244 5.51 External Edge 3,053,500 16.57 4,130,775 19.42 External Islet 79,069 0.43 137,244 0.65 External Core 13,424,288 72.86 14,702,307 69.12 External Bridge 162,800 0.88 217,200 1.02 External Bridge in Edge 131,506 0.71 166,844 0.78 External Loop 48,775 0.26 67,506 0.32 External Loop in Edge 39,438 0.21 48,894 0.23 Internal Branch 51,869 0.28 50,369 0.24 Internal Edge 25,119 0.14 17,131 0.08 Internal Perforation 545,919 2.96 499,644 2.35 Internal Islet 525 0.00 1,044 0.00 Internal Core 22,900 0.12 13,869 0.07 Internal Bridge 7,600 0.04 6,744 0.03 Internal Bridge in Edge 4,075 0.02 3,850 0.02 Internal Bridge in Perforation 2,813 0.02 2,825 0.01 Internal Loop 12,981 0.07 15,375 0.07 Internal Loop in Edge 213 0.00 38 0.00 Internal Loop in Perforation 15,519 0.08 17,131 0.08

PAGE 40

30 Overall, across all dates, a very small percentage of the tropical and temperate forests fall in the interior fragmenta tion classes. Figure 2 shows a zoomed in example of the MSPA cartographic output for the t ropical forests in 2002 differentiating between external and internal fragmentation classes. Similar cartographic outputs were generated for temperate and tropical forests for 2002, 2008, and 2013. Figure 7 MSPA cartographic output for fragmentation classes of the tropical forests in 2002 differentiating between external and internal fragmentation classes. Similar outputs were created for the 2008 and 2013 temperate and tropical forests. Transitions Between Fragmentation Classes During 2002 2008 and 2008 2013 Periods Without Differentiating Between Internal and External Fragmentation Classes Results from analyzing the transitions between fragmentation classes without differentiating between internal and external classes a re organized into three major trends in the data; the percent change from any of the fragmentation classes to no forest for both dates and time periods, differences in transitions between temperate and tropical forests, and

PAGE 41

31 differences in transitions betwe en the two time periods. Additionally, results from transitions between fragmented classes (all classes aside from core and edge) will be described. First, the transition matrices are displayed for first tropical and temperate forests in the 2002 2008 ti me period and next for tropical and temperate forests in the 2008 to 2013 time period in Tables 10, 11, 12, and 13. These transition matrices display the percent of the fragmentation class in the first date than turned into each fragmentation class for th e next date (i.e. what percent of each fragmentation class in 2002 turned int o each fragmentation class in 2008 and again for the 2008 to 2013 period ). Table 10 Fragmentation classes transition matrix for tropical for ests in the 2002 2008 period without differentiating between internal and external classes % Changed to each 2008 fragmentation class Fragmentation Class in 2002 Area in 2002 (ha) No Forest Branch Edge Perforation Islet Core Bridge Bridge in Edge Bridge in Perforation Loop Loop in Edge Loop in Perforation Branch 781,794 25 59 8 1 1 4 2 1 0 1 0 0 Edge 3,108,631 17 3 68 2 0 9 1 1 0 0 0 0 Perforation 531,331 13 2 12 53 0 18 0 0 0 0 0 0 Islet 67,088 26 9 4 0 57 1 1 0 0 1 0 0 Core 13,240,831 7 0 3 1 0 89 0 0 0 0 0 0 Bridge 137,556 21 9 7 0 0 6 51 3 0 2 0 0 Bridge in Edge 113,981 17 4 15 0 0 11 5 45 0 0 1 0 Bridge in Perforation 1,425 8 4 5 12 0 16 6 10 30 2 1 6 Loop 46,131 22 11 7 1 1 5 9 1 0 42 2 0 Loop in Edge 32,150 16 3 18 2 0 10 2 8 0 2 37 2 Loop in Perforation 9,344 13 2 5 13 0 11 2 3 1 4 10 36

PAGE 42

32 Table 11 Fragmentation classes transition matrix for temperate and forests in the 2002 2008 period without differentiating between internal and external classes. % Changed to each 2008 fragmentation class Fragmentation Class 2002 Area in 2002 (ha) No Forest Branch Edge Perforation Islet Core Bridge Bridge in Edge Bridge in Perforation Loop Loop in Edge Loop in Perforation Branch 1,234,488 16 76 4 0 1 2 1 0 0 0 0 0 Edge 4,074,182 12 2 80 0 0 5 0 0 0 0 0 0 Perforation 602,025 13 2 13 63 0 8 0 0 0 0 0 0 Islet 134,444 12 3 2 0 81 1 0 0 0 0 0 0 Core 15,549,201 5 0 3 1 0 91 0 0 0 0 0 0 Bridge 223,613 12 5 4 0 0 3 73 2 0 1 0 0 Bridge in Edge 167,731 10 2 8 0 0 6 3 69 0 0 1 0 Bridge in Perforation 3,269 16 2 7 6 0 7 2 9 47 1 0 3 Loop 76,556 14 5 4 1 0 2 5 1 0 67 1 0 Loop in Edge 42,394 11 2 10 1 0 5 1 3 0 1 65 0 Loop in Perforation 18,550 12 2 4 10 0 7 1 2 1 3 7 50 Table 12 Fragmentation classes transition matrix for tropical forests in the 2008 2013 period without differentiating between internal and external classes. % Changed to each 2013 fragmentation class Fragmentation Class 2008 Area in 2008 (ha) No Forest Branch Edge Perforation Islet Core Bridge Bridge in Edge Bridge in Perforation Loop Loop in Edge Loop in Perforation Branch 827,375 6 89 2 0 0 2 1 0 0 0 0 0 Edge 3,087,438 4 1 89 1 0 4 0 0 0 0 0 0 Perforation 529,050 8 1 10 62 0 17 0 0 0 0 0 0 Islet 80,475 10 4 2 0 82 1 0 0 0 0 0 0 Core 13,030,53 8 3 0 1 1 0 95 0 0 0 0 0 0 Bridge 163,400 5 3 3 1 0 2 84 1 0 2 0 0 Bridge in Edge 130,438 4 1 5 1 0 3 2 83 0 0 1 0 Bridge in Perforation 2,744 8 3 2 8 0 8 4 18 41 1 1 6 Loop 56,825 8 3 3 2 0 3 4 0 0 75 1 0 Loop in Edge 37,013 4 1 5 1 0 4 1 5 0 1 76 1 Loop in Perforation 12,613 6 1 3 7 0 10 0 2 1 2 9 58

PAGE 43

33 Table 13 Fragmentation classes transition matrix for temperate forests in the 2008 2013 period without differentiating between internal and external classes. Fragmentation classes transition matrix temperate forests 2008 2013 % Changed to each 2013 fragmentation class Fragmentation Class 2008 Area in 2008 (ha) No Forest Branch Edge Perforation Islet Core Bridge Bridge in Edge Bridge in Perforation Loop Loop in Edge Loop in Perforation Branch 1,217,513 1 98 1 0 0 0 0 0 0 0 0 0 Edge 4,117,194 1 0 98 0 0 1 0 0 0 0 0 0 Perforation 511,944 1 0 3 94 0 2 0 0 0 0 0 0 Islet 138,550 3 1 0 0 95 0 0 0 0 0 0 0 Core 14,686,731 0 0 0 0 0 99 0 0 0 0 0 0 Bridge 225,575 1 1 1 0 0 1 96 0 0 0 0 0 Bridge in Edge 170,938 1 0 2 0 0 1 0 96 0 0 0 0 Bridge in Perforation 3,025 0 0 0 1 0 3 0 6 89 0 0 1 Loop 83,356 1 1 2 0 0 1 1 0 0 94 0 0 Loop in Edge 49,019 1 0 3 0 0 2 0 1 0 0 93 0 Loop in Perforation 18,194 1 0 0 2 0 4 0 0 0 1 3 89 In addition to the transition matrices created for temperate and tropical forests, cartographic outputs were generated for each cross referenced fragmentation class for both time periods. The resulting maps each contain the area of the isolated fragmentation class from the first date with the fragmentation classes from the next date. Figure 8 is an example of the cartographic results showing the 2008 fragmentation classes that were edge forest in 2002. The lines on the map were edge forest in 2002 and the different colors represent the fragmentation classes in 2008, therefore displaying what the 2002 edge forest turned into in 2008. Similar results were generated for each fragmentation class for both temperate and tropical fore sts for both time periods.

PAGE 44

34 Figure 8 Example of fragmentation classes in 2008 that were edge forest in 2002 for tropical forests. Similar results were generated for temperate and tropical forests for all dates and all fragmentation classes. Transition to no forest From 2002 to 2008 for both temperate and tropical forests, the percent of each fragmentation class changing to no forest is much higher than the percent change to another fragmentation class. In tropical for ests, islets experienced the most percent change to no forest (26 percent) and core experienced the least percent change to no forest (seven percent). In temperate forests, branch and bridge in perforation forests experienced the most change to no forest (both at 16 percent) and core forests experienced the least change to no forest (five percent). From 2008 to 2013, temperate and tropical forests experienced similar transitions in that a relatively high percent of each fragmentation class transitioned t o no forest in 2013, but the changes are not as drastic as from 2002 to 2008 due to much less overall change in

PAGE 45

35 fragmentation classes. In tropical forests, islets again experienced to most change to no forest (10 percent) and core forest experiencing the least percent change to no forest (three percent). In temperate forests islets also experienced the most transition to no forest, although much lower overall percent change (three percent). Temperate core forests experienced less than one percent change to no forest during this time period. Isolated fragmentation classes (islets) and narrow, elongated fragmentation classes (branch, bridge, loop) were the most vulnerable to transition to no forest, while core forests were consistently the least likely to change to no forest. Differences in forest type Although both tropical and temperate forests experienced a higher percent change to no forest than other fragmentation classes, tropical forests experienced much more transition to no forest than temperate fo rests during both time periods. From 2002 2008, tropical forests fragmentation classes experienced a minimum of seven percent to a maximum of 26 percent transition to no forest, whereas temperate forests fragmentation classes experienced minimum of five p ercent to a maximum of 16 percent transition to no forest. From 2008 2013, tropical forests fragmentation classes experienced a minimum of three percent to a maximum of 10 percent transition to no forest, whereas temperate forests experienced zero to thre e percent transition to no forest. The results from the MSPA indicated that temperate forests had more total forest area but higher percentages of that forest were edge and fragmented than in tropical forests. When analyzing the transitions of between fra gmentation classes from one date to the next, it is evident that a higher percent of remaining tropical forests are being fragmented and turning into no forest than the remaining temperate forests.

PAGE 46

36 Additionally, temperate forests experienced less overall change in fragmentation classes across both time periods than tropical forests. For example, 59 percent of tropical branch forest in 2002 remained branch in 2008, whereas 76 percent of temperate branch forest in 2002 remained branch in 2008. Likewise, 8 9 percent of tropical branch forest in 2008 remained branch in 2013, and 98 percent of temperate branch forest in 2008 remained branch in 2013. These findings are consistent across all fragmentation classes. This indicates less change in temperate forest s than in tropical forests across both time periods. Differences between time periods When analyzing the transitions between fragmentation classes, it is evident that there is signi ficant more change from 2002 2008 than there is from 2008 2013. Across al l fragmentation classes, higher percentages remained the same fragmentation class from 2008 to 2013 than from 2002 to 2008. For example, using the example from above, from 59 percent of tropical branch forest remained branch from 2002 2008, whereas 89 per cent of tropical branch forest remained branch from 2008 2013. Similarly, 76 percent of temperate branch forest remained branch from 2002 2008, whereas 98 percent of temperate branch forest remained branch from 2008 2013. These results were consistent ac ross all fragmentation classes. Overall, c ore forests consistently changed the least in both tropical and temperate forests during both periods. This result is to be expected as these forests are buffered from edge effects and represent large compacted fo rest areas that are less subject to fragmentation and deforestation F ragment ation class to fragment ation class transitions Analyzing the changes from fragmentation class to fragmentation class is looking at how each fragmentation class has evolved in relation to the other fragmentation classes, not

PAGE 47

37 determining how they changed in relation to turning into no forest. These resu lts describe what the majority of each fragmentation class transitioned to aside from remaining the same class or turning into no forest. Therefore, aside from remaining the same class or turning into no forest, in tropical forests from 2002 2008, the maj ority of: Branch turned into edge Edge turned into core Perforation turned into core Islet turned into branch Core turned into edge Bridge turned into branch Bridge in edge turned into edge Bridge in perforation turned into core (much of it also turning into perforation) Loop turned into branch Loop in edge turned into edge Loop in perforation turned into perforation Aside from remaining the same class or turning into no forest, in temperate forests from 2002 2008, the majority of: Branch turned into edge Edge turned into core Perforation turned into edge Islet turned into branch Core turned into edge Bridge turned into branch

PAGE 48

38 Bridge in edge turned into edge Bridge in perforation turned into bridge in edge Loop turned into bridge or branch Loop in edge tur ned into edge Loop in perforation turned into perforation From 2008 2013, aside from remaining the same class or turning into no forest, in tropical forests the majority of: Branch turned into edge or core Edge turned into core Perforation turned into core Islet turned into branch Core turned into edge or perforation Bridge turned into branch or edge Bridge in edge turned into edge Bridge in perforation turned into perforation or core Loop turned into bridge Loop in edge turned into edge or bridge in edge Loop in perforation turned into core From 2008 2013, aside from remaining the same class or turning into no forest, in temperate forests the majority of: Branch turned into edge Edge turned into core Perforation turned into edge

PAGE 49

39 Islet turned into branch Le ss than one percent of core changed into a different fragmentation class Bridge turned into branch, edge, or core Bridge in edge turned into edge Bridge in perforation turned into bridge in edge Loop turned into edge Loop in edge turned into edge Loop in p erforation turned into core Transitions Between Fragmentation Classes During 2002 2008 and 2008 2013 Periods Differentiating Between Internal and External Fragmentation Classes Similar results were generated with internal and external fragm entation classes differentiated as were generated when internal and external classes were not differentiated. The same major trends of transitioning to no forest, differences between forest types, and differences in time periods were identified. Since the results of the transitions between fragmentation classes differentiating between internal and external classes are similar to the results without differentiating between external and internal classes, they are summarized briefly by time period. 2002 2008 time p eriod Table 14 presents the results from cross referencing each fragmentation class in 2002 to the fragmentation classes in 2008 for tropical forests with the intext parameter set to on, differentiating between internal and external fragmentation classes. These results are presented in the form of a transition matrix, indicating the percent of the fragmentation class in 2002 that turned into each fragmentation class in 2008.

PAGE 50

40 Table 14 Fragmentation classes transition matrix for tropical forests in the 2002 2008 period differentiating between internal and external classes Similar results were generated with internal and external fragm entation classes differentiated as were generated when internal and external classes were not differenti ated. From 2002 to 2008 in tropical forests, a high percent of fragmentation classes changed into no forest in 2008. From the external classes, external islet forests experienced the most change to no forest (26 percent) and extern al core experienced the least change to no forest

PAGE 51

41 (seven percent). From the internal classes, internal loop in edge forests experienced the most change to no forest (33 percent), and both internal bridge in edge and internal bridge in perf oration experienced the least (both eigh t percent). Of the external fragmentation classes, external core changed the least, with 88 percent remaining external core. External loop in edge forests changed the most, with 37 percent remaining external loop in edge and 63 percent changing to a diff erent fragmentation class Of the internal fragmentation classes, internal perforation forests changed the least, with 53 percent remaining internal perforation in 2008. Internal core forests changed the most, with only 27 percent remaining internal core Table 15 presents the results from cross referencing each fragmentation class in 2002 to the fragmentation classes in 2008 for temperate forests with the intext parameter set to on, differentiating between internal and external fragmentation classes. Th ese results are presented in the form of a transition matrix, indicating the percent of the fragmentation class in 2002 that turned into each fragmentation class in 2008. For temperate forests from 2002 to 2008, i n the external fra gmentation classes exter nal branch forests experienced the most change to no forests (15 percent ) and external core the least (five percent). From the internal fragmentation classes, internal islet forests experienced the most change to no forest (39 percent), and internal loop in perforation the least (12 percent). Of the external fragmentation classes, external core changed the least with 89 percent remaining external core, and external loop in edge forests changing the most with 65 percent remaining external loop in edge. Of the internal fragmentation classes, internal perforations changed the least (63 percent remaining internal perforation) and intern al loop in edge forests changed the most (only 20 percent remaining internal loop in edge)

PAGE 52

42 Table 15 Fragmentation classes transition matrix for temperate forests in the 2002 2008 period differentiating between internal and external classes 2008 2013 time p eriod Table 16 presents the results from c ross referencing each fragmentation class in 2008 to the fragmentation classes in 2013 for tropical forests with the intext parameter set to on, differentiating between internal and external fragmentation classes. These results are

PAGE 53

43 presented in the form o f a transition matrix, indicating the percent of each fragmentation class in 2008 that turned into each fragmentation class in 2013. Table 16 Fragmentation classes transition matrix for tropical forests in the 2008 2013 period diff erentiating between internal and external classes From 2008 to 2013, of the tropical external fragmentation classes, external islets experienced the most change to no forest (10 percent) and external core experienced the least (three percent). Of the internal fragmentation classes, internal islets experi enced the most

PAGE 54

44 change to no forest (44 percent) and internal loop in perforations the least (six percent). Within each external fragmentation class, external core forests changed the least (95 percent remaining external core) and external loop in edge fore sts changed the most (76 percent remaining the same class). For the internal fragmentation classes, internal loop in edge forests changed the least (70 percent remaining internal loop in edge) and internal islets changed the most (only 34 percent remainin g internal islets). Table 17 presents the results from cross referencing each fragmentation class in 2008 to the fragmentation classes in 2013 for temperate forests with the intext parameter set to on, differentiating between internal and external fragmentation classes. These results are presented in the form of a transition matrix, indicating the percent of the fragmentation class in 2008 that turned into each fragmentation class in 2013. Of the temperate forest fragmentation classes from 2008 to 2013 similar to 2002 to 2008, there was much less overall change of fragmentation classes. Of the external forests, external islets experienced the most change to no forest (three percent) and external core the least (approximately zero percent). The pe rcent change to no forest for internal fragmentation classes were overall very low, with most classes experiencing zero to one percent change. However, internal islets experienced a 10 percent change to no forest. Of the external forests, external core e xperienced the least change (99 percent remaining external core) and external loop in edge forests experiencing the most (93 percent remained the same class). Of the internal forests, internal loop in edge forests experienced the least change (approximat ely 100 percent remained the same class) and internal core experienced the most change (84 percent remained internal core).

PAGE 55

45 Table 17 Fragmentation classes transition matrix for temperate forests in the 2008 2013 period dif ferentiating between internal and external classes Cartographic r esults In addition to the transition matrices created for temperate and tropical forests, cartographic outputs were generated for each cross referenced fragmentation class for both time pe riods differentiating between internal and external forests The resulting maps each contain the area of the isolated fragmentation class from the first date with the fragmentation classes from the next date. Figure 9 is an example of the cartographic results showing the

PAGE 56

46 2008 fragmentation classes that were external edge forest in 2002. The lines on the map were external edge forest in 2002 and the different colors represent the fragmentation classes in 2008, therefor e displaying what the 2002 edge forest turned into in 2008. Similar results were generated for each fragmentation class for both temperate and tropical forests for both time periods. Figure 9 Example of fragmentation classes in 2008 that were external edge in 2002 for tropical forests. Similar results were generated for temperate and tropical forests for all dates and all fragmentation classes.

PAGE 57

47 CHAPTER V DISCUSSION Defining Fragmentation Classes of Temperate and Tropical Forests Across all dates, there is overall less area of tropical forests than there are temperate forests. When using the MSPA to define the fragmentation of these forests, although there are more overall temperate forests, they are more fragmented than tropical forests. There are less temperate core forests across all dates than tropical core forests. In 2002, 73.27 percent of tropical forests were core and 70.27 percent of temperate forests were core. In 2008, 72.56 percent of tropical forests were core and 6 9.21 percent of temperate forests were core. In 2013, 72.98 percent of tropical forests were core and 69.18 percent of temperate forests were core. Additionally, across all dates temperate forests contain a higher percent of the forest as edge than tropi cal forests. In 2002, 17.2 percent of tropical forests were edge and 18.41 percent of temperate forests were edge. In 2008, 17.19 percent of tropical forests were edge and 19.4 percent of temperate forests were edge. In 2013, 16.71 percent of tropical f orests were edge and 19.5 percent of temperate forests were edge. With the exception of the perforation and bridge in perforation fragmentation classes, temperate forests also contain a higher percent of each fragmented forest class (all classes except co re and edge) than tropical forests. Therefore, although there is more total forest area of temperate forests, less of the temperate forests are interior core and more are edge and fragmented forest classes. The MSPA results generated with the intext pa rameter set to on differentiating between internal and external fragmentation classes, are very similar to the results with intext set to off. Across all dates temperate forests contained less percent of the total forest area as core forest, more percent of forest as edge, and higher percentages of fragmented forest

PAGE 58

48 classes. The percent of total forest that were internal fragmentation classes were very small, many less than 0.1 percent. This result is in part explained by the scale of the original data (1:250,000), the generalization effects created when grouping several forest types into the broader tropical and temperate forest classes used in this study, and the coarse resolution used in the analysis (cell size of 250x250 meters). Under these conditio ns, small perforations and fine details in the internal edges of the forest areas are difficult to detect. Quantifying the degree of fragmentation in remaining forests is important for biodiversity conservation. Although there is more forest area of temp erate forests in Mexico, they are more fragmented than tropical forests. Species are less able to adapt to changing spatial patterns of the forest and a more fragmented landscape may reduce their ability to survive over time ( O'neill et al. 1988 ). This is an important factor to consider when managing forests for conservation. The overall amount of forest is important for species survival, but even if there are larger amounts of forest area, the fragmented spatial patterns may affect the ability for differ ent species to survive. For example, there may be more area of small, isolated forest patches that have a large proportion of edge forests compared to interior forests. This may increase predator activity and introduce invasive species, which can influen ce species survival and decrease biodiversity. Therefore, understanding that there is more total forest area of temperate forests than tropical forests in Mexico, but temperate forests have a higher percent of that forest area as fragmented can aid in th e conservation practices of remaining forests as well as aid in understanding how the composition of the forest may be effecting biodiversity. Conservation measures can be targeted to improve the degree of forest fragmentation based on the combination of the tabular data describing the percent of the total forest of each fragmentation class with the spatial reference generated

PAGE 59

49 from the cartographic results generated from this study to identify explicit areas in Mexico with a high degree of fragmentation. T ransitions Between Fragmentation Classes During 2002 2008 and 2008 2013 Periods Without Differentiating Between Internal and External Fragmentation Classes When analyzing the transitions between fragmentation classes without differentiating between interna l and external classes, clear trends can be identified. Notably, f or both time periods and forest types analyzed, the percent change from any of the fragmentation classes to no forest is high compared to the change from one fragmentation class to other fr agmentation class. Additionally, there were distinct differences between transitions in temperate versus tropical forests as well as between the two time periods. Lastly, there were important trends when analyzing the transitions between fragmentation cl asses, aside from the changes to no forest or remaining the same fragmentation class from one date to the next. Transition to no forest Fragmentation classes in both temperate and tropical forests during both time periods experienced a higher percent chang e to no forest than to other fragmentation classes. Overall, isolated fragmentation classes (islets) and narrow, elongated fragmentation classes (branch, bridge, loop) were the most vulnerable to transition to no forest, while c ore forests were consistentl y the least li kely to change to no forest. This is to be expected, as smaller isolated forest patches, as well as elongated narrow forest patches, are more likely to change to other land uses than forests that are farther from edges and span large compacte d areas (i.e. core class). Generally, the edge class shows less tendency to change to no forest than the elongated forest patches.

PAGE 60

50 This information relates the degree of fragmentation to the likelihood of transitioning to no forest. It is clear through th is analysis that the smaller, isolated and narrow, elongated fragmentation classes are more likely to transition to no forest over time. Targeted conservation efforts can be implemented to preserve these forest types. With the cartographic results, regio ns with increased amounts of islet, bridge, branch, and loop fragmentation classes can be identified in order to focus these kinds of conservation efforts. Although these fragmentation classes are typically smaller in area than core or edge forests, they provide important corridors and connections from larger, more intact forest patches. These connections and corridors add to the overall landscape structure which is an important Fahrig & Merriam 1994 ). Therefore, although they are not contributing as much to the overall area of forest cover as larger core forests, smaller fragmentation classes such as islets, bridge, branch, and loop are important to conserve since they provide important connections and corridors that are l ost once these areas are deforested. Additionally, it is imperative to preserve intact interior forest areas so they do not become smaller, isolated fragmentation classes that are more susceptible to becoming no forest. Incorporating land use information to the cartographic results provide additional information to determine what pressures are causing the transition to no forest in different areas across Mexico. Differences in forest type Tropical forests experienced a higher percent change to no forest as well as to different fragmentation classes during both time periods. The results from the MSPA indicated that temperate forests had more total forest area but higher percentages of that forest were edge and fragmented than in tropical forests. When analyzing the transitions of

PAGE 61

51 between fragmentation classes from one date to the next, it is evident that a higher percent of remaining tropical forests are being fragmented and turning into no forest than the remaining temperate forests. The differences in transitions of fragmentation classes in temperate and tropical forests reinforces empirical knowledge and results of studies stating that tropical forests, more than temperate forests, are su bjected to higher anthropogenic pressures and risks of deforestat ion (Moreno Sanchez et al. 2012) Moreno Sanchez et al. (2012) found that although temperate forests have high amounts of fragmentation, they typically contain a larger proportion of forest farther away from anthropogenic influences and therefore experience lower levels of anthropogenic pressure that is associated with a high risk of deforestation. These results reinforce the fact that tropical forests are at a higher risk to becoming def orested as well as more fragmented than temperate forests. However, the results from the MSPA also reinforce that temperate forests are currently more fragmented than tropical forests. Therefore, policy and conservation efforts need to target both types of forests, but may need to be implemented differently. For example, since temperate forests are currently more fragmented but are experiencing less change to no forest and other fragmentation classes than tropical forests, conservation efforts should be concentrated to rehabilitate the already fragmented forest and to mitigate further fragmentation. For tropical forests, it is essential to realize what anthropogenic pressures are causing increased fragmentation. Land use and socioeconomic data can be in corporated with the results from this study to obtain a thorough understanding of the pressures driving fragmentation in tropical forests.

PAGE 62

52 Differences between time periods Overall, there was much more change in forest fragmentation during the 2002 2008 time period than the 2008 2013 time period. This could reflect successful policy measures to improve forest conservation in Mexico. Two major payment for ecosystem services (PES) initiatives were undertaken by the National Forestry Commission (CONAFOR) i n 2003 and 2004. One was program for hydrological services and the other a program for carbon sequestration and biodiversity (FAO 2013). These programs were merged into one PES National Program in 2006. The PES compensates landowners for implementing la nd management and conservation practices. CONAFOR pays landowners a fixed compensation per hectare of forest for five years. Sustainable forest practices include preserving land cover and avoiding land use change, developing a Best Management Practices P rogram, using surveillance in order to reduce illegal logging and hunting in forests, and providing signage to inform neighbors about what is permitted on the land. The amount of compensation is determined based on the economic value of converting forests to corn production, as well as on the level of risk of the type of forest to deforestation. For example, conserving cloud forests provides the highest compensation rate per hectare. Since these programs began in 2003, more than 3.2 million hectares of f orests have been conserved (FAO 2013). The PES programs may have contributed to the decreased amount of change in forest fragmentation during the 2008 2013 period. Additionally, the Inter Ministerial Climate Change Commission (CICC) was formed in 2005 to change policies, programs, and strategies ( UN REDD 2013). In 2009, a United Nations Reducing Emissions from Deforestation and Forest Degradation (UN REDD+) working group was established within the CICC ( UN R EDD 2013). The UN REDD+ program aims

PAGE 63

53 to create financial value for preserving carbon stocks in forests through reduction of deforestation and forest degradation (fragmentation), as well as integrate forest conservation and sustainable management strategie s for countries. Efforts to preserve forests in Mexico for carbon storage starting with the creation of the CICC in 2005 and the integration of the UN REDD+ program in 2009 may have also contributed to the decrease in overall change in fragmentation level s during the 2008 2013 time period in this study. It is also known that the overall rate of deforestation has decreased by 55 percent from 2000 2010 (FAO 2010). All of this information reflects successful policy measures in forest conservation, which ref lect in the findings of this study that the time period of 2008 2013 experienced much less overall change in fragmentation than from 2002 2008 in both temperate and tropical forests. F ragment ation class to fragment ation class transitions Although the ov erall percentages of the total forest area of the fragmented forest classes are smaller than the percent of the total forest of core and edge classes, the transitions have significance for biodiversity conservation. For example, for both temperate and tro pical forests during both time periods, aside from transitioning to no forest or remaining the same fragmentation class, the majority of bridge forests turned into branch forests. The bridge fragmentation class indicates forests that connect two different core areas, and the branch fragmentation class represents forests that connect at one end to an edge, perforation, bridge, or l oop This means that relatively high percentages of forests that originally acted as a connector from one large, intact region of forest to another large, intact region of forest (bridge) now connect to a fragmented patch of forest rather than a core forest (branch). This transition indicates that a core forest at one end of the original bridge connector has been fragmented, and corridors between two interior forests are being lost.

PAGE 64

54 Additionally, a relatively high percentage of both bridge in edge forests and loop in edge forests turned into edge forests. A bridge in edge forest is a connector within the edge fragmentation clas s that connects two core areas, and a loop in edge forest is a connector within an interior edge (meaning an edge within a perforation in the core of a forest) that connects the same core area. When these two classes transition to the edge fragmentation c lass, they are no longer connecting to a core forest area. This, again, contributes to the loss of corridors connecting core forests. Lastly, a relatively high percentage of loop in perforation forests turned into perforations. A loop in perforation fra gmentation class connects the same core forest. They are located within perforations within the interior of the forest. When loop in perforation forests transition to perforations, that forest area is completely lost and turns into an area of no forest w ithin the interior of a core forest. This indicates the loss of the forest that connects the two ends of areas of no forest within core forests (perforations). All of the transitions above indicate the loss of corridors between two areas of large, intact core forests. As mentioned before, although these types of fragmentation classes make up less total forest area, they are important to preserve due to their effect as biological corridors. These corridors aid in the conservation of biodiversity, particu larly in a fragmented landscape ( Fahrig & Merriam 1994 ). Therefore, the results of this study can aid in identifying these important corridors to aid in conservation efforts in a fragmented landscape.

PAGE 65

55 Transitions Between Fragmentation Classes During 2002 2008 and 2008 2013 Periods Differentiating Between Internal and External Fragmentation Classes As mentioned previously, differentiating between internal and external fragmentation classes generated results very similar to the results without differentiating the two. Each fragmentation class had a relatively high percent change to no forest, tropical forests experienced more change in fragmentation classes than temperate forests, and the 2002 2008 time period experienced more change in fragmen tation classes than the 2008 2013 time period. Much of the changes detected in the internal fragmentation classes are change to external core forests. This is likely because all of the internal fragmentation classes are within perforations within the cor e of a forest, and as mentioned previously, due to the original scale of the data, the aggregation to broad forest types of temperate and tropical forests, and the 250x250 meter cell size used in this analysis, fine details within the perforations of a for est become difficult to detect. F uture studies conducted over smaller geographical areas and at higher resolution are more likely to detect differences in trends between internal and external fragmentation classes. Limitations and Sources of Error When ev aluating the results and discussion presented above, the following sources of error should be kept in mind. First, there may be errors in the land use/ land cover classifications between the Series III, IV and V data sets The original land use/cover data sets are at a scale of 1:250,000. At this scale, small forest areas are difficult to detect and the shape of the forest area borders are not reported extremely accurately. Second, the generalization effects introduced by using a 250x250 meter cell size may decrease the amount of detail defining the edges of forests. A sensitivity analysis using a smaller cell size

PAGE 66

56 performed on a smaller region of Mexico may help in quantifying the effects of cell size on the results of this analysis. Third, there is a generalization effect when grouping more detailed forest cover classes into the broader tropical and temperate forests classes used in this study. This could indirectly contribute to smooth or loose detail on the sha pe of the forest area borders. Fourth, there is a difference in length between the 2002 2008 period ( six years) and 2008 2013 period (five and V were created using inputs (e.g. satellite images) collected over a period of time that may extend to more than o ne year (Nino & Victoria 2013) Even with each of these four possible sources of error, none by itself or in combination is likely large enough to change the clear and large trends identified at the national level in the results of th is study.

PAGE 67

57 CHAPTER VI CONCLUSIONS AND RECOMMENDATIONS This study used the MSPA method to define the fragmentation classes of temperate and tropical forests in Mexico for 2002, 2008, and 2013. This study additionally explored the transitions between each fragmentation class during the two time periods of 2002 2008 and 2008 2013 to gain an understanding of the evolution of forest fragmentation in Mexico during these times. There is extensive research regarding the effects of forest fragmentation on specif ic species and ecosystems in Mexico, but there is a lacking of literature on the state of fragmentation of Mexican forests on a national scale, along with cartographic representations of forest fragmentation. Additionally, there is no literature detailing the change of specific forest fragmentation classes over time. This study helps to fill in these gaps in the current literature using the most up to date land cover data and methodology to define levels of forest fragmentation. MSPA Method Using the MSPA method to identify forest fragmentation levels has several advantages. It is conceptually simple to understand by diverse stakeholders, and computational easy to implement u sing freely available software. T hese characteristics facilitate th e periodical replication of fragmentation assessments at the national level, as well as at smaller geographical extents. The MSPA method produces cartographic products that allow the exploration of the spatial relationships of forest fragmentation over tim e, as well as the analysis of the relationships between fragmentation patterns to other environmental and anthropogenic factors. For example, the study of the potential correlation between fragmentation classes and environmental factors such as topography and anthropogenic

PAGE 68

58 factors such as distance to urban areas, agricultural land uses, and socioeconomic factors such as dominant economic activity or poverty levels. The MSPA results provide scientists, managers and the general public a visual representati on of the forest fragmentation that they can visually explore and correlate to their empirical knowledge and on the ground experiences. Finally, the MSPA output assigns unique identifiers to each fragmentation class which facilitates further analysis, suc h as the cross referencing of fragmentation classes between dates as it was done in this study. Summary of Findings This study provides insights into how forest fragmentation has evolved between the 2002 2008 and 2008 2013 periods at the national level in Mexico in both tropical and temperate forests. All fragmentation classes for both periods experienced a higher percent change to no forest than to other fragmentation classes. The smaller, isolated fragmentation classes (islets) and elongated, narrow frag mentation classes (branch, bridge, and loop) experienced the highest percent change to no forest from one period to the next. Tropical forests experience d a higher percent change to no forest than temperate forests across both periods. The percent change o f all fragmentation classes to no forest was less during the 2008 2013 than during the 2002 2008 period. There was loss of important corridors that connect core forest areas in both forest types and time periods. When differentiating between internal and external fragmentation classes, the results generated were very similar to the results when the two were not differentiated. The transitions of the internal fragmentation classes to other classes or to no forest were very small, frequently less than one percent change.

PAGE 69

59 Differentiating between internal and external fragmentation classes did not provide significant additional benefits at the national scale used in this study. The scale of the source data (1:250,000) does not portrait in detail the small gaps in the forest core areas nor the shape of the edges of the forest areas, and hence the internal fragmentation classes are underrepresented. However, studies at finer resolutions and larger scales over smaller geographical areas would be capable of det ecting small perforations and fine details of the internal forest areas, and hence benefits from differentiating between internal and external fragmentation classes could be realized. Implications for Science and Policy The results of this study provid e tabular data and cartographic results quantifying the degree of forest fragmentation in both temperate and tropical forests at a national scale in Mexico. Additionally, it provides tabular data and cartographic results quantifying the percent of each fo rest fragmentation class that transitioned to all other fragmentation classes across two time periods of 2002 2008 and 2008 2013. This information adds landscape level forest fragmentation data to a robust amount of literature that quantifies forest fragm entation and its effects on specific species and ecosystems at local or regional scales. These results can be incorporated into national level forest management efforts, such as the REDD+ and PES programs. The cartographic results can be understood by a variety of stakeholders more effectively than only tabular data. Additionally, the cartographic results can be used to identify areas with a high degree of fragmentation, areas with high amounts of forest transitioning to no forest, or areas with a high l oss of corridor forests. Targeted policy measures can be incorporated into a variety of conservation initiatives such as PES programs, Forest Management Plans, REDD+ efforts, and education and awareness in these identified

PAGE 70

60 regions or communities. The eff ectiveness of conservation efforts can be tracked using the methodology used in this study to provide the detailed percentages of the composition of temperate and tropical forests as well as the percent change of forest fragmentation classes over time. Ov erall, the results of this study can be used to see how conservation efforts are working across the landscape in Mexico, and identify areas where resources need to be allocated the most to conserve the forests in Mexico. Future Recommendations This study does not differentiate between natural and anthropogenic causes of forest fragmentation which is important for conservation purposes. Future studies can use GIS systems to overlay the cartographic results of this study with maps of anthropogenic activitie s such as urban developments, agricultural uses, and cattle ranching to better understand the driving factors of forest fragmentation in different geographic areas. Additionally, the fragmentation classes can be cross referenced with socioeconomic data at the county or state level in order to enhance the understanding of the relationships between socioeconomic factors and activities with land use change and forest fragmentation processes, and their impacts on forest ecosystems over time. Lastly, since the beginning of this study an updated version of the Guidos Toolbox containing the MSPA has been released (Guidos Toolbox 2.3, Revision 1). The newer version incorporates more image analysis tools that can provide further information such as differentiating between forest areas natural edge interface and an anthropogenic/artificial edge interface, this is valuable information that would enhance forest conservation efforts. Due to the influence of forest fragmentation on ecological processes and its impact on the sustainability of the remaining forest areas in Mexico, assessments and monitoring of the

PAGE 71

61 level of fragmentation of the forests should be incorporated into national forest inventories and forest ecosystems health reports. This information can enhance the prioritization and targeting of conservation efforts of the remaining forests in the country.

PAGE 72

62 REFERENCES Bogaert, J., Ceulemans, R., & Salvador Van Eysenrode, D. (2004). Decision tree algorithm for detection of spatial processes in landscape transformation. Environmental management, 33 (1), 62 73. Cayuela, L., Benayas, J. M. R., & Echeverra, C. (2006). Clearance and fragmentation of tropical montane forests in the Highlands of Chiapas, Mexico (1975 2000). Forest Ecology and Management, 226 (1) 208 218. de Mxico, S. A. T. (1999). Tropical rain forest fragmentation, howler monkeys (Alouatta palliata), and dung beetles at Los Tuxtlas, Mexico. American Journal of Primatology, 48 253 262. Donovan TM, Flather CH 2002. Relationship among North Am erican songbird trends, habitat fragmentation, and landscape occupancy. Ecological Applications, 12, 364 374. Estrada, A., & Coates Estrada, R. (1996). Tropical rain forest fragmentation and wild populations of primates at Los Tuxtlas, Mexico. Internation al journal of primatology, 17 (5), 759 783. Estrada, A., Coates Estrada, R., & Meritt, D. (1993). Bat species richness and abundance in tropical rain forest fragments and in agricultural habitats at Los Tuxtlas, Mexico. Ecography, 16 (4), 309 318. Estregui l, C., Caudullo, G., de Rigo, D., San Miguel, J. (2012). Forest Landscape in Europe: Pattern, Fragmentation and Connectivity. European Commission Joint Research Centre Institute for Environment and Sustainability. Fahrig, L. (2003). Effects of habitat fr agmentation on biodiversity. Annual review of ecology, evolution, and systematics, 487 515. Fahrig, L., & Merriam, G. (1994). Conservation of fragmented populations. Conservation biology, 8 (1), 50 59. FAO (Food and Agriculture Organization of the United Nations). 2010. Global Forest Resources Assessment 2010. FAO, Rome. FAO. (2013). Case studies on renumeration of positive externalities (RPE)/payments for environmental services (PES). Available: http://www.fao.org/fileadmin/user_upload/pes project/docs/ FAO_RPE PES_PSAH Mexico.pdf Fazey, I., Fischer, J., & Lindenmayer, D. B. (2005). What do conservation biologists publish? Biological conservation, 124 (1), 63 73.

PAGE 73

63 Fischer, J., & Lindenmayer, D. B. (2007). Landscape modification and habitat fragmentation: a synthesis. Global Ecology and Biogeography, 16 (3), 265 280. Garca Gigorro, S., & Saura, S. (2005). Forest fragmentation estimated from remotely sensed data: is comparison across scales possible? Forest Science, 51 (1), 51 63. Groom, M. J., G. K. Meffe, and C. Ronald (2005). Principles of Conservation Biology. 3rd Edition. Sinauer Associates, 779 pp. 213 252. Groombridge, B., Jenkins, M.D. (2000). Global biodiversity. In: United Nations Environment entury. in forest ecosystems. Cambridge University Press, Cambridge, UK. pp. 234 264. Hall, F. G., Botkin, D. B., Strebel, D. E., Woods, K. D., & Goetz, S. J. (1 991). Large scale patterns of forest succession as determined by remote sensing. Ecology, 628 640. Harper, K. A., S.E. MacDonald, P.J. Burton, J. Chen, K.D. Brosofske, S.C. Saunders, E.S. Euskirchen, D. Roberts, M.S. Jaith, and E. Per Anders. 2005. Edge in fluence on forest structure and composition in fragmented landscapes. Conservation Biology, 19 : 768 782. scale conservation: what Ecography, 22 : 225 232. INEGI (I nstituto Nacional de Geografa e Informtica) 2009. Gua para la interpretacin de la cartografa uso del suelo y vegetacin, escala 1:250,000 Serie III. INEGI, Mxico. 77 pp. Available online at: http://www.inegi.org.mx/prod_serv/contenidos/espanol/bvine gi/productos/geografia/p ublicaciones/guias carto/sueloyveg/1_250_III/Suelo_Vegeta.pdf Accessed August 2015 INEGI (Instituto Nacional de Geografa e Informtica) 2012. Gua para la interpretacin de la cartografa uso del suelo y vegetacin, escala 1:250, 000 Serie IV. INEGI, Mxico. 132 pp. Available online at: http://www.inegi.org.mx/prod_serv/contenidos/espanol/bvinegi/productos/geografia/p ublicaciones/guias carto/sueloyveg/1_250_IV/1_250_IV.pdf. Accessed August 2015

PAGE 74

64 INEGI (Instituto Nacional de Geo grafa e Informtica) 2014. Gua para la interpretacin de la cartografa uso del suelo y vegetacin, escala 1:250,000 Serie V. INEGI, Mxico. 200 pp. Available online at: http://www.inegi.org.mx/geo/contenidos/recnat/usosuelo/doc/guia_interusosuelov.pdf Accessed August 2015 (Site to download digital GIS layer.http://www.inegi.org.mx/geo/contenidos/recnat/usosuelo/ Accessed March 2015. Kupfer, J. A. 2012. Landscape ecology and biogeography: Rethinking landscape metrics in a post FRAGSTATS landscape P rogress in Physical Geography, 36 (3): 400 420. Kupfer, J.A. 2006. National assessment of forest fragmentation in the US. Global Environmental Change, 16 : 72 82. Laurance, W. F., Camargo, J. L., Luizo, R. C., Laurance, S. G., Pimm, S. L., Bruna, E. M., Phillip, P.C., Williamson, B., Benitez Malvido, J., Vasconcelos, H.I., Van Houtan, K.S., Zartman, C.E., Boyle S.A., Didham, R.K., Andrade, A., Lovejoy, T. E. (2011 ). The fate of Amazonian forest fragments: a 32 year investigation. Biological Conservation, 144 (1), 56 67. Laurance, W. F., Delamnica, P., Laurance, S. G., Vasconcelos, H. L., & Lovejoy, T. E. (2000). Conservation: rainforest fragmentation kills big tre es. Nature, 404 (6780), 836 836. Lindenmayer, D. B., & Fischer, J. (2007). Tackling the habitat fragmentation panchreston. Trends in Ecology & Evolution, 22 (3), 127 132. Mas, J. F., Velzquez, A., Daz Gallegos, J. R., Mayorga Saucedo, R., Alcntara, C., Bocco, G., Castro, R., Fernandez, T., & Prez Vega, A. (2004). Assessing land use/cover changes: a nationwide multidate spatial database for Mexico. International Journal of Applied Earth Observation and Geoinformation, 5 (4), 249 261. McAlpine, C. A., & E yre, T. J. (2002). Testing landscape metrics as indicators of habitat loss and fragmentation in continuous eucalypt forests (Queensland, Australia). Landscape Ecology, 17 (8), 711 728. Melo, F. P., Martnez Salas, E., Bentez Malvido, J., & Ceballos, G. (2 010). Forest fragmentation reduces recruitment of large seeded tree species in a semi deciduous tropical forest of southern Mexico. Journal of Tropical Ecology,26 (01), 35 43. Metzger, J. P., & Dcamps, H. (1997). The structural connectivity threshold: an hypothesis in conservation biology at the landscape scale. Acta Oecologica,18 (1), 1 12. Metzger, J. P., & Muller, E. (1996). Characterizing the complexity of landscape boundaries by remote sensing. Landscape Ecology, 11 (2), 65 77.

PAGE 75

65 Moreno Sanchez, R., F. Moreno Sanchez, and J. M. Torres Rojo (2011). National assessment of the evolution of forest fragmentation in Mexico. Journal of Forestry Research 22 : 167 174. Moreno Sanchez, R., J.M. Torres Rojo, F. Moreno Sanchez, S. Hawkins, J. Little, S. MacPartland. 2012. National assessment of the fragmentation, accessibility and anthropogenic pressure on the forests in Mexico. Journal of Forestry Research 23 (4): 529 541. Moreno Sanchez, R., T. Buxton Torres, K. Silbernagel, F. Moreno Sanchez. 2014. Fragmentation of the forests in Mexico: National level assessments for 1993, 2002 and 2008. International Journal of Statistics and Geography 5 (2): 4 17. Murcia, C, (1995). Edge effects in fragmented forests: implications for conservation. Trends in Ecology & Evolutio n 10 : 58 62. Myers, N. (1988). Threatened biotas:" hot spots" in tropical forests. Environmentalist, 8 (3), 187 208. Neel, M. C., McGarigal, K., & Cushman, S. A. (2004). Behavior of class level landscape metrics across gradients of class aggregation and area. Landscape ecology, 19 (4), 435 455. Nio Alcocer, M. and Victoria Hernandez, A. (2013). Informacin de uso del suelo y vegetacin escala 1:250,000 Serie V (Conjunto Nacional). Memorias del Congreso SELPER (Sociedad Latinoamericana de Percepcin Remot a y Sistemas de Informacin Espacial). Available online at: http://langif.uaslp.mx/selper/documentos/CD_SELPER_2013/MEMORIAS_SELPER _PDF/Estudios_Tematicos/ID_008.pdf Ochoa Gaona, S., & Gonzlez Espinosa, M. (2000). Land use and deforestation in the highl ands of Chiapas, Mexico. Applied Geography, 20 (1), 17 42. Ochoa Gaona, S., Gonzlez Espinosa, M., Meave, J. A., & Sorani, V. (2004). Effect of forest fragmentation on the woody flora of the highlands of Chiapas, Mexico. Biodiversity & Conservation, 13 (5), 867 884. O'neill, R. V., Milne, B. T., Turner, M. G., & Gardner, R. H. (1988). Resource utilization scales and landscape pattern. Landscape Ecology, 2 (1), 63 69. Ostapowicz, K., Vogt, P., Riitters, K. H., Kozak, J., & Estreguil, C. (2008). Impact of sca le on morphological spatial pattern of forest. Landscape ecology, 23 (9), 1107 1117.

PAGE 76

66 Ribeiro, M. C., Metzger, J. P., Martensen, A. C., Ponzoni, F. J., & Hirota, M. M. (2009). The Brazilian Atlantic Forest: How much is left, and how is the remaining fores t distributed? Implications for conservation. Biological conservation, 142 (6), 1141 1153. Riitters, K. H., Wickham, J. D., O'neill, R. V., Jones, K. B., Smith, E. R., Coulston, J. W., ... & Smith, J. H. (2002). Fragmentation of continental United States f orests. Ecosystems, 5 (8), 0815 0822. Riitters, K., Wickham, J., O'Neill, R., Jones, B., & Smith, E. (2000). Global scale patterns of forest fragmentation. Conservation Ecology, 4 (2), 3. Schleuning, M., Farwig, N., Peters, M. K., Bergsdorf, T., Bleher, B. Brandl, R., Dalitz, H., Fischer, G., Freund, W., Gikungu, M. W., Hagen, M., Garcia, F. H., Kagezi, G. H., Kaib, M., Kraemer, M., Lung, T., Naumann, C., M., Schaab, G., Templin, M., Uster, D., Wagele, J. W., & Bohning Gaese, K. (2011). Forest fragmentatio n and selective logging have inconsistent effects on multiple animal mediated ecosystem processes in a tropical forest. PLoS One, 6 (11), e27785 e27785. Schmiegelow FKA, Mnkknen M. (2002). Habitat loss and fragmentation in dynamic landscapes: Avian persp ectives from the boreal forest. Ecological Applications, 12 : 375 389 Soille, P, Vogt, P. (2008). Morphological segmentation of binary patterns. Pattern Recognition Letters 30 4:456 459, doi: 10.1016/j.patrec.2008.10.015 Tabarelli, M., Da Silva, J. M. C., & Gascon, C. (2004). Forest fragmentation, synergisms and the impoverishment of neotropical forests. Biodiversity & Conservation, 13 (7), 1419 1425. Thompson, J., Brokaw, N., Zimmerman, J. K., Waide, R. B., Everham III, E. M., Lodge, D. J., Taylor, C.M ., Garcia Montiel, D., & Fluet, M. (2002). Land use history, environment, and tree composition in a tropical forest Ecological applications, 12 (5), 1344 1363. Trejo, I., & Dirzo, R. (2000). Deforestation of seasonally dry tropical forest: a national and local analysis in Mexico. Biological Conservation, 94 (2), 133 142. Tscharntke T, Steffan Dewenter I, Kruess A, Thies C. (2002). Contribution of small habital fragmentes to conservation of insect communities of grassland cropland landscapes. Ecological App lications, 12 : 354 363. Turner IM. (1996). Species loss in fragments of tropical rain forest: A review of the evidence. Journal of Applied Ecology, 33 : 200 209.

PAGE 77

67 UN REDD. (2013). The REDD desk Mexico. Available: http://theredddesk.org/countries/mexico#forest Vellend, M., Verheyen, K., Jacquemyn, H., Kolb, A., Van Calster, H., Peterken, G., & Hermy, M. (2006). Extinction debt of forest plants persists for more than a century following habitat fragmentation. Ecolog y, 87 (3), 542 548. Vogt, P. (n.d.). MSPA Guide. Vogt, P., Riitters, K. H., Estreguil, C., Kozak, J., Wade, T. G., & Wickham, J. D. (2007). Mapping spatial patterns with morphological image processing. Landscape Ecology, 22 (2), 171 177. Wilcove, D. S., M cLellan, C. H., & Dobson, A. P. (1986). Habitat fragmentation in the temperate zone. Conservation biology, 6 237 256. Zipperer, W. C. (1993). Deforestation patterns and their effects on forest patches. Landscape Ecology, 8 (3), 177 184.