Citation
The role of whitebark pine (Pinus Albicaulis) as a tree island initiator in the alpine-treeline ecotone

Material Information

Title:
The role of whitebark pine (Pinus Albicaulis) as a tree island initiator in the alpine-treeline ecotone examining microclimate and microsite
Creator:
Pyatt, Jill C. ( author )
Language:
English
Physical Description:
1 electronic file (134 pages). : ;

Subjects

Subjects / Keywords:
Whitebark pine -- Conservation ( lcsh )
Whitebark pine -- Ecology ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (M.S.)--University of Colorado Denver.
Bibliography:
Includes bibliographic references.
System Details:
System requirements: Adobe Reader.
General Note:
Department of Integrative Biology
Statement of Responsibility:
by Jill C. Pyatt.

Record Information

Source Institution:
Auraria Library
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
892525382 ( OCLC )
ocn892525382

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

THE ROLE OF WHITEBARK PINE (PINUS ALBICAULIS) AS A TREE ISLAND INITIATOR IN THE ALPINE TREELINE ECOTONE: EXAMINING MICROCLIMATE AND MICROSITE by J ILL C P YATT B.A. Biology, College of Saint Benedict, 2009 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 Biology 2013

PAGE 2

ii This thesis for the Master of Science degree by Jill C Pyatt has been approved for the Department of Integrative B iology by Diana F. Tomback, Chair Michael B. Wunder Laurel M. Hartley July 25 th 2013

PAGE 3

iii Pyatt, Jill, C. ( M.S., Department of Integrative Biology Master of Science ) The Role of Whitebark P ine ( Pinus albicaulis ) as a Tree Island Initiator in the Al pine Treeline Ecotone: Examining Microclimate and M icrosite Thesis directed by Professor Diana F. Tomback ABSTRACT In the alpine treeline ecotone on the Eastern Front of the Northern Rocky Mountains whitebark pine ( Pinus albicaulis ) initiates tree isl and formation more often than any other conifer. W object for the establishment of other conifers, especially on northeast facing aspects. To explain these observations, w e hypothesize d that: 1) conifer nurse microsites diffe r in microclimate from rock and unprotected microsites more favorable for seedling survival and 2) whitebark pine microsites experience a more favorable microclimate than Engelmann spruce microsites. F rom mid July to mid September in 2010, 2011 and 2012 b locks of whitebark pine, Engelmann spruce, rock and unprotected microsites primarily on northeastern slopes, were assessed at two treeline study areas in Mon tana: Divide Mountain, Glacier National Park and Blackfeet Reservation ; and Line Creek Research N atural Area, Custer National Forest. Among microsites, we compared daily maximum and minimum values for soil and air temperatures average soil moisture, daily PAR, gust speed maximums and wind speed averages for each study area. Sky exposure for microsit es and carbon and mitrogen in microsite soils were also examined. The majority of these variables were compared using a Kruskal Wallis one way analysis of variance with a pairwise Wilcoxon rank sum test for post hoc analysis. Overall, conifer microsites l owered daily PAR and ameliorated average wind speeds in comparison to rock and unprotected microsites, suggesting that they may be important

PAGE 4

iv in seedling establishment. At Line Creek, where there was a greater variance in air temperature extremes and a high er number of freeze thaw events, conifer microsites had higher minimum soil temperatures and soil temperatures with lower variances. Engelmann spruce microsites specifically had lower maximum air temperatures and gust speeds compared to whitebark pine, roc k and unprotected microsites Results for Divide Mountain were extremely variable and inconsistent among blocks. We attribute thi s to the large variance in landscape conditions coupled with low experimental replication At both study areas, conifer microsi tes and whitebark pine microsites in particular, had reduced sky exposure. W hitebark pine microsites had a trend towards higher carbon and nitrogen content of soils than other microsites although these results were not significant Collec tively, these da ta suggest that conifer microsites have a climatic advantage over rock and unprotected microsites more tolerable for seedling establishment. Further, whitebark pine microsites may have some non climatic advantages over spruce microsites for seedling establ ishment; but we generally a freque ncy as a tree island initiator to its avian seed dispersal mode and hardiness under harsh conditions The form and content of this abstract are approved. I recommend its publication. Approved: D iana F. Tomback

PAGE 5

v ACKNOWLEDGMENTS I would like to thank my advisor, Diana Tomback for guidance and support throughout the project s entirety. I would also like to extend a huge thank you to my colleagues and field partners Sarah Blakeslee, Libby Pansing, Aaron Wagner and Soledad Diaz who worked through rain, wind and shine with me over different field seasons. The data management and analysis for this project could not have been completed without the frequent help and guidance of my committee mem ber Dr. Michael Wunder and programmer extraordinaire Jamie Costabile. Thank you to my committee member Dr. Laurel Hartley for all of her wonderful feedback and advice throughout this process Thank you to Kent Houston and the Shoshone National Forest for their c onstant support in this project I would like to extend a special thank you to our collaborators Dr. Lynn Resler and Dr. George Malanson for coming together to create such an important and valuable research endeavor F urther thanks go to Emily S mith, Lauren Franklin, Mari Majack, Chelsea Beebe, James Pembrook, Erin Pyatt, Jennifer Scott my parents and everyone from the ecology and evolution biology group for their very helpful advice and support along the way.

PAGE 6

vi TABLE OF CONTENTS CHAPT ER I. LITERATURE BACKGROUND REVIEW ................................ ............................. 1 Taxonomy and Distribution ................................ ................................ ................. 1 ................................ ................................ 1 Main Threats to Whitebark Pine ................................ ................................ .......... 2 Whitebark Pine Communities ................................ ................................ .............. 4 Treeline ................................ ................................ ................................ ............... 4 Climate Change ................................ ................................ ................................ ... 5 Tables and Figures ................................ ................................ .............................. 9 II. INTRODUCTION ................................ ................................ ................................ .. 11 Backgroun d ................................ ................................ ................................ ....... 11 Conceptual Framework ................................ ................................ ..................... 15 Hypotheses for Testing ................................ ................................ ...................... 16 Tables and Figures ................................ ................................ ............................ 17 III. METHODS ................................ ................................ ................................ ............. 19 Study Areas ................................ ................................ ................................ ....... 19 Study Design ................................ ................................ ................................ ..... 20 Field Methods ................................ ................................ ................................ ... 20 First Year (2010) ................................ ................................ ............................... 20 Second and Third Year s (2011 and 2012) ................................ .......................... 21 Microclimate Station Setup ................................ ................................ ............... 21 Data Measurements and Analysis ................................ ................................ ...... 22

PAGE 7

vii Soil Sampling ................................ ................................ ................................ .... 23 Tables and Figures ................................ ................................ ............................ 24 IV. RESULTS ................................ ................................ ................................ ............... 29 Climate and Soil Differences between Study Areas ................................ ........... 29 Comparing Microclimate among Microsites ................................ ...................... 30 Soil Temperature ................................ ................................ ................... 30 Air Temperature ................................ ................................ .................... 32 PAR ................................ ................................ ................................ ...... 32 Gust Speed ................................ ................................ ............................ 33 Wi nd Speed ................................ ................................ ........................... 33 Soil Moisture ................................ ................................ .......................... 34 Sky Exposure ................................ ................................ ................................ .... 34 C and N in Soils ................................ ................................ ................................ 35 Tables and Figures ................................ ................................ ............................ 36 V. DISCUSSION ................................ ................................ ................................ ......... 7 1 Cl imatic Stress in Study Areas ................................ ................................ ......... 7 3 Importance of Microsites ................................ ................................ ................. 7 3 Conclusion s ................................ ................................ ................................ ..... 7 8 Figures ................................ ................................ ................................ ............ 8 0 REFERENCES ................................ ................................ ................................ ....... 8 1 APPENDIX ................................ ................................ ................................ ............ 9 0 Air Temperature Results ................................ ................................ ................... 9 0 Soil Moisture Results ................................ ................................ ........................ 9 7

PAGE 8

viii AIC Scores ................................ ................................ ................................ ...... 1 0 0 Post Hoc Results ................................ ................................ ............................. 1 0 3 R Code Statistical Analysis ................................ ................................ ............. 10 8 Daily Maximum, Minim um, Variance in Temperature ......................... 10 8 D a i l y S u m P A R ................................ ................................ ................... 1 1 2 Daily Maximum Gust Speed ................................ ................................ 11 4 Daily Average Wind Speed ................................ ................................ 11 6 Percent Sky Exposure ................................ ................................ .......... 11 9 Tot al Percent N and C in Soils ................................ ............................. 1 2 0 Freeze Thaw Events ................................ ................................ ............ 1 2 1

PAGE 9

ix LIST OF TABLES Table I.1 Annual climate profiles for treeline conifers in the northern Rocky Mountains ..................... 9 II.1 Hypothesis Analysis table for microcl imate c haracteristics tested ................................ ..... 18 III.1 Block site descriptions for each study area ................................ ................................ ....... 25 III.2 ................................ ........................... 25 III.3 HOBO Onset Microclimate Sensors and installation heights about ground (cm) .............. 27 III.4 Start and end dates for Micro Station Data Loggers ................................ ......................... 27 IV.1. Daily air temperature descriptive statistics ................................ ................................ ....... 38 IV.2 Daily soil moisture descriptive statistics ................................ ................................ .......... 41 IV. 3. Daily wind speed descriptive statistics ................................ ................................ ............. 44 IV.4. Descriptive statistics for total percent nitrogen a nd carbon in soils ................................ ... 45 IV.5. Daily maximum soil temperature results (Kruskal Wallis) ................................ ............... 49 IV.6. Daily minimum soil temperature results (Kruskal Wallis) ................................ ............... 53 IV.7. Daily variance soil temperature results (Kruskal Wallis) ................................ ................. 57 IV.8. Daily PAR results (Kruskal W allis) ................................ ................................ ................. 59 IV.9. Daily maximum gust speed results (Kruskal Wallis) ................................ ....................... 62 IV.10. Daily average wind speed results (Kruskal Wallis) ................................ ........................ 65 IV.11 Sky exposure results (Chi Square Goodness of Fit) ................................ ....................... 68 IV.12 Percent nitrogen and carbon results (ANOVA) ................................ .............................. 70 A1. Dail y maximum air temperature results (Kruskal Wallis) ................................ .................. 93 A2. Daily minimum air temperature results (Kruskal Wallis) ................................ ................... 96 A3. Daily average soil moisture results (Kruskal Wallis) ................................ ....................... 100

PAGE 10

x A4. Soil temperature AIC Scores ................................ ................................ ................ 101 A5. Air temperature AIC Scores ................................ ................................ ................. 102 A6. PAR AIC scores ................................ ................................ ................................ ... 102 A7. Maximum soil temperature ( Kruskal Wallis Post Hoc Results) ............................ 103 A8. Minimum soil temperature ( Kruskal Wallis Post Hoc Results) ............................ 103 A9. Variance in soil temperature ( Kruskal Wallis Post Hoc Results) .......................... 104 A10. Maximum air temperatur e ( Kruskal Wallis Post Hoc Results) ........................... 104 A11. Minimum air temperature ( Kruskal Wallis Post Hoc Results) ............................ 105 A12. Sum PAR ( Kruskal Wallis Post Hoc Results) ................................ .................... 105 A13. Maximum Gust Speed ( Kruskal Wallis Post Hoc Results) ................................ 106 A14. Average Wind Speed ( Kruskal Wallis Post Hoc Results) ................................ ... 106 A15. Average Soil Moisture ( Kruskal Wallis Post Hoc Results) ................................ 107 A16. Soil N and C ( ANOVA Post Hoc Results) ................................ .......................... 10 7

PAGE 11

xi LIST OF FIGURES Figure I.1 Distribution of whitebark pine ( Pinus albicaulis ) ................................ ................................ 9 I.2 Projected occ urrence of whitebark pine using the Random Forests multiple repression tree for a) present distribution b) 2030 c) 2060 and d) 2090 ................................ ............................. 10 II.1. Potential microclimatic stress experienced by a seedling in an unprotected microsite ........ 17 III 1 Research study a reas ................................ ................................ ................................ ........ 24 III.2 Microstation set up ................................ ................................ ................................ ........... 26 III.3 Full microstation setup in a whitebark pine microsite ................................ ....................... 26 III.4. T he sequential steps for data detrending and maximum soil temperature analysis ............ 28 IV.1. Daily average air temperatures at Divide Mountain an d Line Creek ................................ 36 IV.2. Daily variance in air temperatures at Divide Mountain and Line Creek ............................ 37 IV.3. Number of freeze thaw events at Divide Mountain and Line Creek ................................ 39 IV.4. Daily average soil moisture at Divide Mountain and Line Creek ................................ ...... 40 IV.5. Dail y average wind speeds at Divide Mountain and Line Creek ................................ ....... 42 IV.6. Daily variance in wind speeds at Divide Mountain and Line Creek ................................ .. 43 IV.7. Total percentages of nitrogen and carbon in soils at Divide Mountain and Line Creek ..... 45 IV.8.a. Boxplots of resid ual data from 2010 daily maximum soil temperatures ......................... 46 IV.8.b. Boxplots of residual data from 2011 daily maximum soil temperatures ......................... 47 IV.8.c. Boxplots of residual data from 2012 daily maximum soil temperatures ......................... 48 IV.9.a. Boxplots of residual data from 2010 daily minimum soil temperatures ......................... 50 IV.9.b. Boxplots of residual data from 2011 daily minimum soil temperatures ......................... 51 IV.9.c. Boxplots of residual data from 2012 daily minimum soil temperatures ......................... 52 IV.10.a. Boxplots of residual data from 2010 daily variance in soil temperatures ..................... 54

PAGE 12

xii IV.10.b. Boxplots of residual data from 2011 daily variance in soil temperatures ..................... 55 IV.10.c. Boxplots of residual data from 2012 daily variance in soil temperatures ..................... 56 IV.11. Boxplots of residual data from 2010 daily PAR ................................ ............................. 5 8 IV.12.a. Boxplots of daily maximum gust speeds from 2011 ................................ .................... 60 IV.12.b. Boxplots of daily maximum gust speeds from 2012 ................................ .................... 61 IV.13.a. Boxplots of daily average wind speeds from 2011 ................................ ...................... 63 IV.13.b. Boxplots of daily average wind speeds from 2012 ................................ ...................... 64 IV.14.a. Percent sky exposure from 2010 microsites ................................ ................................ 66 IV.14.b. Percent sky exposure from 2011 microsites ................................ ................................ 67 IV.15.a. Percent total nitrogen found in microsites ................................ ................................ ... 69 IV.15.b. Percent total carbon found in microsites ................................ ................................ ..... 69 V.1.a. Summary model of results for Divide Mountain ................................ ............................. 80 V.1.b. Summary model of results for Line Creek ................................ ................................ ...... 80 A1.a. Boxplots of residual data from 2010 daily maximum air temperature .............................. 90 A1.b. Boxplots of residual data from 2011 daily maximum air temperatures ............................ 91 A1.c. Boxplots of resid ual data from 2012 daily maximum air temperatures ............................. 92 A2.a. Boxplots of residual data from 2010 daily minimum air temperatures ............................. 94 A2.b. Boxplots of residual data from 2011 and 2012 daily minimum air temperatures .............. 95 A3.a. Boxplots of daily average soil moisture from 2010 ................................ .......................... 97 A3.b.Boxplots of daily average soil moisture from 2011 ................................ .......................... 98 A3.c. Boxplots of daily average soil moisture from 2012 ................................ .......................... 99

PAGE 13

xiii LIST OF ABBREVIATIONS ATE Alpine Treeline Ecotone Divide Mtn Divide Mountain, Blackfeet Indian Reservat ion ES Engelmann spruce ( Picea engelmannii ) microsite Line Creek Line Creek Research Natural Area MPB Mountain Pine Beetle NE Northeast facing slope face RK Rock microsite UN Unprotected microsite W West facing slope face WB Whitebark pine ( Pinus albicaulis ) microsite

PAGE 14

1 CHAPTER I. LITERATURE BACKGROUND REVIEW Taxonomy and Distribution Whitebark pine ( Pinus albicaulis Engelm.) is a high elevation white pine that occurs in some of the last undisturbed land in North America. It is one of five stone pine species in the world which were traditionally classified in the section Strobus, subsection Cembrae Loudon but ha s since been placed into the new subsection Strobus Loudon (Gernandt et al., 2005 ; Tomback and Achuff 2010). The d istribution of whitebark pine is separated into eastern and western ranges. With the most southeastern stands occurring in the Wind River Range of Wyoming, the eastern range of the northern Rocky Mountains extends north through Wyoming, Idaho, and Montana to northeast of McBride, British Columbia (Arno 1986). In its eastern distribution, whitebark pine forests encompass an estimated 10 15% of the coniferous forest in the northern R ocky Mountains in North America (Arno 1986). Along the western distributi on its range extends from the Kern River in southeastern California through northeastern Oregon up through the Bulkley Mountains in northern British Columbia as far north as Smithers, B.C. ( Olgilvie, 1990; McCaughey and Schmidt 2001). Whitebark pine oc curs at high elevations forming upper subalpine and treeline communities, typically residing within elevations of 2,590 to 3,200 meters in western Wyoming and as low as distribution (Arno 1986; Arno and Hoff 1989 ; McCaughey and Schmidt 2001) ( Figure I. 1) Characteristically, stone pines have large wingless seeds and indehiscent cones Cembrae pines coevolved with their seed dispersers birds of the genus Nucifraga or the nutcr acker ( F amily Corvidae)

PAGE 15

2 birds harvest seeds in early fall, caching them throughout the landscape ( Tomback 1978; Tomback, 1982; Hutchins and Lanner 1982 ). Although the seeds are a n important food source for many birds and mammals, nutcrackers are the only bird s evolved to break into cones and disperse the heavy seeds of the stone pines ( Hutchins and Lanner 1982 ; Tomback and Linhart 1990). In late summer and early fall, the bird s harvest ripe seeds, dispersing and caching them up to 29 km away from the trees in which they were taken ( Lorenz and Sullivan 2009 ) Inserted a few centimeters into the ground, fifteen or more seeds may be placed into a single cache (Tomback 1978 198 2 ; Hutchins and Lanner, 1982 ). Many of the seeds buried are retrieved as food by nutcrackers throughout the winter and spring seasons, but the seeds that remain may germinate Nutcrackers typically bury whitebark pine seeds near large objects such as rocks logs or in niches (Tomback 1978; Tomback, 1982; Tomback 2001) T herefore many seeds are placed in ideal locations for establishment. Once established at treeline whitebark pine offers protection to other conifer species, such as subalpine fir ( Abies lasiocarpa (Lamb.) Poir.) and Engelmann spruce ( Picea engelmannii Parry e x Engelm.) (Resler, 2004 ; Tomback and Resler, 2008). The distribution and regeneration of whitebark pines mutually depend on the seed dispersal behavior of these birds. Main Threats t o Whitebark Pine There are three main threats to the future survival of whitebark pine. Perhaps the most immediate threat is Dendroctonus ponderosae or m ountain pine beetle (MPB) which is killing larger, cone bearing tree s with stems over 20 cm in diamete r (Arno 1986 ; Gibson et al. 2008 ). Secondly, f ire exclusion in areas that were once naturally regimented has drastically in creased with urbanization and development allowing succession to occur and shade tolerant species to outcompete whitebark in suba l p ine communities ( Arno 1986; Kendall and Keane 2001)

PAGE 16

3 F inally the exotic fungal disease white pine blister rust ( WPBR ) caused by the pathogen Cronartium ribicola was introduced to North America from Western Europe in the late 19 th century and early 20 th century. It has infected and killed trees in subalpine and treeline populations o f whitebark pine and other widely distributed five needled white pine across North Ameri c a (Hoff et al. 2001 ; Geils et al., 2010 ) The accelerated decline o f whitebark pin e greatly reduce s seed source s available for regeneration Also reduced is the carryi ng capacity for all the organisms that rely on pine seeds for winter survival ( Arno, 1986; Tomback and Kendall, 2001 ). In July 2011 the U.S. Fish and Wildlife Service co ncluded that whitebark pine warrants protection under the Endangered Species A ct and it has been added to the list of candidate species eligible for ESA protection (U.S. Fish and Wildlife 2011) Finding effective management strategies for whitebark pine is difficult because dynamic changes in whitebark pine populations occur as these threats interact with each other and global climate change Using a time series analysis of Landsat imagery for the Greater Yellowstone Ecosystem, Jewett et al. 2011 found that mortality of whitebark pine from MPB generally increased with warmer and drier climates. Keane et al. (2010) pointed out that although fire created a favorable place for seed infection by WPBR was too high for seed sour ces to provide an adequate genetic base for regeneration. The dispersal mechanism for whitebark pine regeneration is at great risk if populations of whitebark pine continue to decline particularly in northern Montana where infections of WPBR are highest ( McKinney et al. 2009). Although some genetic resistance from WPBR has been found in whitebark pine, and current management regeneration strategies involve planting resistant trees, these t rees are still at risk from MPB when grown (Fins et al. 2001 ; Tomb ack and Achuff 2010 ). Further, increasing temperature and drier climate

PAGE 17

4 responses from global warming could possibly heighten the threats of both WPBR and MPB (Tomback 2007). Whitebark Pine Communities Whitebark pine occur s in the subalpine and alpine t reeline ecotone (ATE) and grows slowly in climatically harsh environments. Generally whitebark pine is a hardy species that dominates in open and abiotically stressful environments within its distribution ( Arno 2001) In the ATE whitebark pine is the fi rst to establish in more open and exposed areas and has shown increased seedling survival in its second year compared to Engelmann spruce (Bansal et al. 2011) Due to its hardy seed lings and dispersal mechanism whitebark pine can persist on the coolest, windiest sites where competition is not as strong (Arno 2001; Weaver 2001). In treeline communities, where it occurs in multi tree stand patches called tree islands, and in the upper subalpine, whitebark pine growing under harsh conditions helps stabiliz e soil and extend snowpack melt into the summer ( Farnes 1990 ; Tomback and Achuff 2010 ). S uccession in the subalpine occurs when species like subalpine fir and Engelmann spruce, which are more shade tolerant, outcompete whitebark pine for resources until fire occurs (Minore, 1979; Arno, 2001). nutcrackers, pine squirrels ( Tamiasciurus spp.), grizzly bears ( Ursus arctos ), and black bears ( Ursus americanus ) ( Mattson et al ., 1991 ; Tomback and Kendall, 2001 ; McKinney and Fiedler 2010). Treeline There are numerous hypotheses as to what limits conifer growth in the alpine treeline ecotone (ATE) The sink orientated growth limitation hypothesis infers that the act of carbon processing from carbon sinks are more restricted at higher altitudes, therefore reducing conifer

PAGE 18

5 growth. Limited carbon processing may be a complex response due to drought stress (Moyes et al. 2013), and low soil temperature extremes ( Krner, 1998 ). For example, freezing temperatures can lead to drought stress caused by embolism in the xylem of conifer s, although high elevation species such as subalpine fir may be better adapted to avoid this ( Sperry and Sullivan 1992; Choat et al. 2012). Areas shelter ed by trees are known to ameliorate extreme soil temperatures (Krner 1998; Germino and Smith 2002). It has been predicted that cold soil temperatures may limit root growth which in turn would limit tree growth (Krner 1998; Smith et al. 2009). The k rummholz zone where mat forming conifer stands occur is likely caused by limited carbon gain in seedlings as well as physical damage to the apical bud due to typical patterns of high winds, snow, and ice (Arno and Hammerly 1984; Smith et al. 2003). The effects of s hort growing seasons at treeline, which are typically 90 110 days long, are another hypothesis for the upper elevational limits of whitebark pine ( Weaver and Dale 1974 ; Weaver 2001 ). In comparison to subalpine fir, whitebark pine also tends to occur on soils with higher water holding capacity and higher percentages of organic carbon and nitrogen. Subalpine fir also has a wider temperature tolerance (Weaver 1978). Climate Change Comparative data on climatic tolerances for whitebark pine, En gelmann spruce and subalpine fir are shown in Table I.1 These c limate data are representative of the general northern Rocky Mountain s and should vary locally with characteristics related to slope aspect, photoperiod, wind speeds, ultraviolet radiation, a nd precipitation (Weaver 1990; Weaver, 1994; Weaver, 2001).

PAGE 19

6 Climate on t he eastern slope of the N orthern R ocky Mountains in Montana is characterized as being drier and cooler than the western side which receives precipitation from the Pacific west coast and experiences a maritime influence ( Malanson et al. 2007; Arno and Hammerly 1984 ) Because of this, timberline on the east side exist s at lower altitudes due to longer periods of drought. Treelines throughout western Montana are known to be extensive a nd compositionally different. Along with variation in geomorphology, small patches of alpine treeline ecotone occur in isolation separated by extensive grassland which has likely given rise to the differences in ecosystem composition among treelines ( Malan son et al. 2007; Arno and Hammerly 1984). There are various hypotheses regarding the future of treeline and the effects that climate change may have on treeline ecosystem dynamics. Generally, treelines are predicted to move upward in elevation and towar d poles in arctic environments with warming climates. A computer simulation of the impact of future climate change on Pinus mugo (mountain p ine), a treeline species in Austria, predicted that treeline will encroach upwards slowly over the next 1000 years w ith its biggest limitation being species recruitment (Dullinger et al. 2004). Kharuk et al. (2008) found that increases in temperature within the past thirty years positively correlated with the upwards expansion of Pinus sibirica (Siberian stone pine) which occurs in the Western Sayan and is closely related to whitebark pine They also documented a transformation from krummh olz to upright trees Multiple climate models have predicted future distributions of whitebark pine over the next century (Hamann and Wang 2006 ; McKenney et al., 2007; Warwell et al., 2007; Schrag et al., 2008 ). The projected distributions of whitebark pine by Warwell et al. ( 2007 ) are shown in Figure I.2. In British Columbia climate models predict that whitebark pine will lose si gnificant

PAGE 20

7 amounts of current habitat by 2080 (73%), but gain in new habitat could amount to 76% ( Hamann and Wang 2006 ). In modeling the effect of future bioclimatic variables on whitebark pine in the Greater Yellowstone Area, Schrag et al. ( 2008 ) predicte d that the extent of whitebark pine distribution will decrease due to increasing temperatures decreasing geographic area above treeline, and biological factors such as WPBR and MPB outbreaks. A model of North American tree species by McKenney et al. ( 200 7 ) predicted that whitebark pine will gain a distribution range of 29.1% and move north 6.5 under a full dispersal scenario. Malanson et al. ( 2007 ) predicted that increasing soil moisture through precipitation and snow pack will likely improve treeline establishment, but this prediction does not take the loss of whitebark pine due to current biological threat s into account. As whitebark pine populations decline, treeline advancement by nutcracker seed caching could cease, particularly in environments whe re whitebark pine is the primary tree island initiator. With warming temperatures, it is possible that sites will be less harsh, and therefore whitebark pine will likely be replaced by species that currently exist lower in the subalpine ecotone (Tomback an d Resler 2007). Even if warming temperatures and greater carbon levels drive treeline to higher elevations, other limitations could arise, such as changes in predicted free thaw cycles and geomorphic formations such as rock and cliff patterns (Malanson et al. 2007). One concern is whether treeline whitebark pine will be capable of responding to global warming, specifically as populations continue to decline in numbers. The nutcracker plays an important role in whitebark pine regeneration at treel ine by caching seeds within and slightly upwards of the alpine treeline ecotone (Tomback 1998). Once whitebark pine is established, it provides protective microsites in its lee, allowing other species like Engelmann spruce and subalpine fir to move in and survive under harsh treeline conditions, in turn creating

PAGE 21

8 tree islands (Resler 2004, Resler and Tomback 2008 ). A major question is whether tree islands will develop above the current krummholz zone in respon se to global warming as numbers of whitebark pi n e at treeline are reduced by WPBR (Tom b ack and Resler 2007; Resler and Tomback 2008). One difficulty in predicting treeline response to climate change is that current models focus on broad temperature changes and the effects primarily on adult tree s tha n on seed germination and seedling survival More small scale studies focusing on tree and seedling survival facilitated by microclimate effects are needed to further understand the local responses of treelines (Smith et al. 2003, Smith et al. 2009).

PAGE 22

9 Tables and Figures Figure I.1. Distribution of whitebark pine ( Pinus albicaulis ) in North America (Tomback and Achuff 2010) Table I.1 Annual climate profiles for three treeline conifers in the northern Rocky Mountains. (Alexander and Shepperd 1990; Alexander et al., 1990; Baker 1944; Weaver and Dale 1974) Species Record Low Mean January Temperature Mean July Temperature Mean Annual Precipitation Mean Annual Snowpack Whitebark pine 50 to 40 C 9 to 5 C 10 15 C 60 180 cm 460 1270 cm (250 300 cm*) Engelmann spruce 45.6 C 12 to 17 C 4 13 C 61 114 cm 635+ cm Subalpine fir 45 C 15 to 9 C 7 13 C 61 152 cm 635+ cm East of Continental Divide

PAGE 23

10 Figure I.2. Projected occurrence of wh itebark pine using The Random Forests multiple regression tree for a) present distribution b) 2030 c) 2060 and d) 2090. Model uses an average of Hadley and CCMA GCM scenarios of 1% per year increase GGa (Warwell et al. 2007)

PAGE 24

11 CHAPTER II. INTRODUCTIO N Background Facilitative interactions are positive interactions between organisms in which one benefits while the other may or may not but is not negatively impacted ( Stachowicz 2001 ). F acilitation s are particularly important for survival in physiol ogically or physically stressful environments because they create a more favorable habitat for organisms in an otherwise challenging environment ( Stachowicz 2001 ). As climates become harsher species interactions often shift from competitive to facili tat ive (Callaway, 1998; Sthultz et al. 2007) For example, in semiarid region s of the southwestern U S pinyon pine ( Pinus edulis Engelm. ) provided shelter and facilitated the establishment and survival of a common shrub ( Fallugia paradoxa (D. Don) Endl. E s Torr. ) at sites experiencing drought stress whereas competitive interaction s between the two species existed where mo isture was more plentiful (Sthultz et al. 2007). Some of the ways that plants make surrounding environments more habitable through faci litative interactions include ameliorating soil temperature s ( Breshears 1998 ; Chambers 2001 ) increasing soil nutrients ( Callaway et al. 1991) and reducing wind speeds ( Baumeister and Callaway 2006) Here, we examine the role of whitebark pine ( Pinus albicaulis Engelm Family Pinaceae, Subgenus Strobus ) and Engelmann spruce ( Picea engelmannii Parry e x Engelm.) in facilitating the leeward development of tree islands in the alpine treeline ecotone (ATE) east of the Continental Divide in Montana. In thi s region, whitebark pine in particular functions as an important tree island initiator facilitating conifer establishment (Resler and Tomback, 2008). Whitebark pine is abundant in many ATE communities but the mechanism b ehind this functional

PAGE 25

12 role is uncle ar The physical limitations to tree establishment and growth at treeline may provide insight into these mechanisms. Facilitation occurs at high elevations because the abiotic stress of the environment is severe (Callaway et al. 2002). For alpine treelin es alone, n umerous hypotheses have been developed to understand the environmental stressors that limit tree growth Climate is an important factor in determining treeline as treelines are expected to move upwards in response to more favorable temperatures through global climate change (Dullinger et al. 2004; Kharuk et al. 2008) The winter drought hypothesis suggests that the desiccation of branch es above snowpack due to high winds and blowing ice prevent upright tree growth ( Marchand and Chabot 1978; Arno and Hammerly, 1984; Stevens and Fox 1991 ; Smith et al. 2003 ) Winter drought stress and multiple free ze thaw cycles can also lead to xylem embolisms in the stem resulting in mild to severe desiccation ( Mayr et al. 2002 ; Mayr et al. 2003 ) The s ink orientated growth limitation hypothesis suggests that carbon processing from carbon sinks may be restricted at higher elevations ( Hoch et al. 2002 ). This is a complicated process that has been tied to low soil temperatures shortened growing seasons winter injury leaf loss from extreme winds, and high radiation (Stevens and Fo x 1991; Germino and Smith 1995 ; Sveinbjrnsson, 1996; Cairns and Malanson 1997; Krner 1998 ) L ow soil temperatures have may also lead to photo inhibition ( Germino and Smit h 1999 ) and slow root growth ( Landhusser 1996 ). At the most basic level, photosynthesis rates generally increase with levels of photosynthetically active radiation (PAR), but too much heat and radiation can be detrimental for plant growth ( Germino and S mith, 2002; Moyes et al. 2013 ) Lack of soil moisture, which is the case at many treelines, is further exacerbated by high wind speeds and radiation (Holtmeier

PAGE 26

13 and Broll, 1992; Holtmeier, 2003). Figure II. 1 demonstrates the possible micro climatic stress experienced by a seedling growing without protection at treeline. The structure of the alpine treeline ecotone is limited by the processes of seed and seedling establishment, which are dependent on microhabitat microclimate including local wind patterns, and small scale processes ( Maher and Germino 2006; Malanson et al. 2007 ; Batllori et al. 2009 ). Seedling survival, required for treeline recruitment and expansion, is often dependent on objects such as rocks, conifers or topogra phic niches in the landscape that moderate harsh conditions for leeward conifer establishment ( Holtmeier and Broll, 1992; Germino et al. 2002; Resler 2004; Batllori et al ., 2009) Once a conifer becomes established as sion is further facilitated through climate mitigation Two or more conifers will establish next to and further leeward of one another, resulting in a multi tree island (Bekker, 2005) Over time, tree islands grow larger and l onger as they migrate downwind and e ventually the windward trees that first established will die off from desiccation (Marr 1977; Benedict 1984). This proce ss is particularly important fo r the establishment of shade tolerant species in the ATE of the Rocky Mountains, such as subalpi ne fir ( Abies lasiocarpa (Lamb.) Poir. ) and Engelmann spruce ( Picea engelmannii Parry e x Engelm.), which are most Httenschwiler and Smith, 1999 ). Throughout its distribution in the high elevation regions of the western U.S. and Canada whitebark pine functions as a foundation species and ecosystem engineer (Tombac k et al. 2001 ; Ellison et al. 2005) by providing many ecosystem services including soil stabilization (Tomback et al., 2001) snow pack retention (Farnes 1990) and facilitation of tree growth and tree islands (Callaway, 1998; Tomback and Resler, 2008) In the fall, whitebark pine seeds are

PAGE 27

14 harvested by nutcrackers and cached at a range of elevations (from below subalpine to al pine) near objects like rocks, logs and trees or niches in the landscape ( Tomback 1978 2001 ; Hutchins and Lanner 1982 ) Cached seeds that are not retrieved as food contribute to regeneration The hardiness of whitebark pine on harsh sites may underlie it s ecological function in facilitation. Whitebark pine survives well on dry, windswept slopes, and its seedlings survive on exposed sites (Arno and Hammerly, 1984; Maher and Germino, 2006; Tillman Sutela 2008; Bansal et al., 2011). It was found to germin ate in exposed sites more than subalpine fir and Engelmann spruce (Maher and Germino, 2006) and was found at the windward side of tree islands more than any other conifer species at certain sites within its range (Resler and Tomback, 2008). For example, wh itebark pine is a pioneer species in disturbed treeline zones near Logan Pass on the east side of the Continental Divide in Glacier National Park facilitating the establishment of less hardy conifers (Habeck, 1969 ) Whitebark pine has been found to be a solitary colonizer more than other conifer species at two treeline sites on the Eastern Front of the Rocky Mountains ( Resler and Tomback 2008; Smith McKenna unpublished data ) Populations of whitebark pine occurring in the ATE are severely declining due to spread of the exotic fungal pathogen, Cronartium ribicola, which causes white pine blister rust (WPBR), a disease that kills five needled white pines (Resler and Tomback, 2008). On the E astern Rocky Mountain Front, WPBR infected 35% of all whitebark pin es at two treeline sites east of the Continental Divide (Resler and Tomback, 2008), and was at high as 47% at treeline sites sampled in Glacier National Park, Montana (Smith et al., 2011). The consequences of WPBR whitebark pine mortality on tree island fa cilitation and treeline dynamics are not entirel y known. Complicating issues further is that the combined effects of WPBR and climate change on these ecosystems remain unclear Tomback and Resler (2007) suggest that the upwards response

PAGE 28

15 of treeline to cli mate change may be delayed if whitebark pine is no longer available to facilitate tree island establishment. Conceptual Framework At many sites throughout its Rocky Mountain distribution, whitebark pine has been found to facilitate the start of tree island s more than any other conifer ( Resler and Tomback 2008 ; Smith McKenna et al. unpublished data ). The mechanism underlying this role is not yet understood. Seed placement near objects and in microtopography (e.g,. rocks, logs or ground miches) utcracker s may provide whitebark pine more microsite protection than conifers with wind dispersed seeds, permitting its migration into harsh sites more often (Malanson et al. 2007 ) T he hardiness of whitebark pine seeds and seedlings enable the species to tolerate harsh condi tions of these high elevation site s and survive longer, out competing shade tolerant species like subalpine fir and Engelmann spruce ( McCaughey and Tomback, 2001; Maher and Germino 2006) The combination of nutcracker caching in protec tive sites and seedling hardiness may lead to whitebark pine becoming more common within these ecosystems and therefore more likely to facilitate tree islands Both Resler and Tomback ( 2008 ) and Blakeslee (2012) found that solitary whitebark pine were sig nificantly more numerous than solitary Engelmann spruce a nd solitary subalpine fir at multiple treeline sites east of the Continental Divide Finally, whitebark pine may provide a more protective microsite than other common conifers or nurse objects at tre eline leading more often to tree island establishment through microclimate moderation or some other process In this study we address the latter hypothesis that whitebark pine provides a more favorable microsite than other conifers or nurse objects at tr eeline It may well be that whitebark and the fairly common Engelmann spruce

PAGE 29

16 provide equal protection at treeline, but whitebark pine is hardier grows faster, and survives longer On the Eastern Rocky Mountain Front, the survival of Engelmann spruce and subalpine fir seedling s is higher in conifer microsites as opposed to open or rock microsites. At treeline in the Snowy Range, WY, Germino et al. ( 2002 ) describe 70 percent greater survivorship of Engelmann spruce seedlings beneath overhanging branches, co mpared to those present in bare and open microsites. Blakeslee (2012) found that germinated Engelmann spruce seeds in whitebark pine microsites had higher summer survival while planted subalpine fir and Engelmann spruce seedlings in whitebark pine microsit es had greater health and vigor than those in Engelmann spruce, rock and unprotected microsites. Here we compare whitebark pine microsites with several other common microsite types to examine differences and other factors relevant to seedling facilitation and tree island initiation. Hypotheses f or T esting In this study, w e test the following hypothes es pertaining to conditions during much of the summer growing season: 1) C onifer leeward microsites in the ATE during the growing season will be more favorable for seedling establishment and growth than rock or unprotected microsites 2) W hitebark pine leeward microsites in the ATE will be more favorable for seedling establishment and growth than Engelmann spruce leeward micro sites In this study, w e compare lee ward microclimate variables including air and soil temperature soil moisture, photosynthetically active radiation (PAR), and wind speeds and gust speeds among whitebark pine, Eng e lm a nn spruce, rock and unprotected microsites from 2010 to 2012 To clarify our hypotheses with respect to micro conditions less or more favorable for seedling establishment and growth, w e propose that compared to rock or unprotected microsites, whitebark pine and

PAGE 30

17 spruce microsites will have lower ai r and soil temperature maximum s, higher ai r and soil temperature minimums, moister so ils, greater soil carbon and nitrogen, lower amounts of daily photosynthe tically active radiation (PAR), and lower wind speeds and wind gust speeds and lower sky exposure (Table II.1) Tables and Fi gures Figure II.1. Potential microclimatic stress experienced by a seedling in an unprotected microsite at treeline in the a) daytime and b) nighttime.

PAGE 31

18 Table II.1. Hypothesis Analysis table for microclimate characteristics tested. Whit ebark pine microsites were predicted to have: Method 1). A ttenuated air and soil temperatures with lower temperature max imums higher min imums and lower variances compared to Engelmann spruce, rock and unprotected microsite s Data were plotted and com pared graphically and descriptively Kruskal Wallis comparison b y microsite using daily maximum, minimum and variance temperature values 2). Higher soil moisture than Engelmann spruce, rock and unprotected microsites. Data were plotted and compared gra phically and descriptively Kruskal Wallis comparison by microsite and study area using daily average water content values. 3 ). L ower photosy n thetically active radiation (PAR) than rock and unprotected microsites. Data were plotted and compared graphic ally and descriptively Kruskal Wallis comparison by microsite using daily PAR sums. 4 ). L ower wind speeds and lower wind gusts than Engelmann spruce, rock, and unprotected microsites. Data were plotted and compared graphically and descriptively Kruskal Wallis comparison by microsite using daily average wind speeds and maximum gust speed s. 5 ). L ower percent sky exposure than rock and unprotected microsites due to conifer height and branch overhang. Data were plotted and compare d graphically and desc riptively Chi Square Goodness of Fit comparison by microsite using exposure percentages 6 ) H igher total carbon and nitrogen in soil. Data were plotted and compared graphically and descriptively Two way ANOVA comparison by microsite and study area usi ng total percentages of C and N.

PAGE 32

19 CHAPTER III METHODS Study Areas This research was conducted at two treeline study areas in Montana over three growing seasons during 2010, 2011, and 2012, from early Ju ly through mid September, when sites could be acce ssed The Divide Mountain study area straddles the eastern boundary between Glacier National Park and Blackfeet Tri b al Land (48 39' 25.38" N, 113 23' 45.44" W; elevation approx. 2,200 m) T he Line Creek Research Natural Area Custer National Forest occ urs on the Beartooth Plateau (45 01' 47.45" N, 109 24' 09.22" W; elevation approx. 2,950 m) (Figure III.1). The two study areas are separated by about 500 km straight line distance and more than 3 o latitude. B oth study areas occur on the eastern front of the Northern Rocky Mountains and with the primary community type categorized as alpine treeline ecotone with krummholz tree islands primarily comprised of whitebark pine, subalpine fir and Engelmann spruce P revious studies show that whitebark pine has an important tree facilitative role in initiating tree islands at both study areas (Resler and Tomback 2008; Smith McKenna et al. unpublished data ). Our study site at Divide Mountain is predominately northeast facing and relatively steep. The bedrock is c omprised of white limestone of the Altyn Formation (Lesica 2002). Common vegetation included Arctostaphylos uva ursi (L.) Spreng. Dryas octopetala L. Oxytropis sericea Torr. & A. Gray Silene acaulis (L.) Jacq. Hedysarum sulpharescens Rydb. Poa alpina L., and Potentilla sp. The study site at Line Creek was also predominantly northeast facing but less steep than Divide Mountain. Soils on the Beartooth Plateau are characteristically shallow, coarse and relatively undeveloped (Nimlos et al. 1965). The lo cal geology is characterized as an uplifted Precambrian granitic mass (Bevan 1923). Common vegetation included Potentilla

PAGE 33

20 diversifolia Lehm. Lupinus argente us Pursh Aster alpigenus (Torr. & A.Gray) A.Gray Agoseris glauca (Pursh) Raf. Geum rossii Ser. Pedicularis s p p., and Carex s p p. See Table III.1 for more details on each study area. Study Design Field Methods M icroclimate was recorded immediately leeward of four solitary microsite types frequently found at treeline whitebark pine, Engelmann spru ce, rock and open or unprotected microsite. Micros ites were grouped in blocks to decrease the impacts of environmental variability (Figure III.2) Blocks were composed of a set of each solitary microsite type occurr ing in close proximity (within 20 m eter s ) to one another with microsite object heights lengths and widths ap proximately equal (Table III.2 ) Within this study, I define microsite as a space 20 c enti m eters in diameter and no taller than 15 c enti m eters directly leeward of a protective object The position of the microsite, on the leeward side of the facilitative object, exactly opposite of wind direction of origin, was estimated from the wind flagged branches of surrounding conifer tree islands. An unprotected microsite occurred in an open site without a protective niche or vegetation. First Year (2010) As a pilot study in its first year, we selected two treatment blocks at each study area, with o ne on a northeast facing slope, and one on a west facing s lope General s lope steepness and aspects are listed in Table III.1. Microclimate characteristics tested at each microsite included photosynthetically active radiation (PAR), air temperature, soil temperature, and soil moisture. All microclimate data were taken at 15 minute intervals from mid Jul y to mid Septemmber each year using Ons et HOBO

PAGE 34

21 Dataloggers (Table III.3). Sky exposure was documented for each microsite using a 180 fisheye lens on a Nikon D50 digital camera A photo was taken facing up wards towards the sky from the base of each microsi te. Percent cover was found digitally using Adobe Photoshop Elements 10 (2011). S ky exposure percentages were estimated by selecting sky and recording the sky pixel to total pixel ratio. Second and Third Year (2011 and 2012) In the second year and third y ear I assessed the same five block s each year (20 microsites total). Three blocks were placed on Divide Mountain and two at Line Creek on the northeastern slope only Tree island community sampling by collaborators, Smith McKenna et al. ( unpublished data) indicate s that whitebark pine plays an important tree island initiation role on north east facing slopes at both study areas. Focusing on the northeastern slope s only in 2011 and 2012 allow ed for replication within each study area. I added HOBO Onset win d anemometers to each microsite setup to collect wind speeds continuously ( Figure III.2 ). Wind speed measurements from the unprotected microsite acted as a co ntrol for its block, providing data for undisturbed, near ground wind conditions. This eliminated the need for sensors both in front of and behind each microsite. Due to restriction in the number of sensors that can be handled by the Micro Station Data L oggers, I eliminated PAR sensors to accommodate for the addition of wind anemometer sensors. I pred icted that PAR would show the least variation year to year. Micro climate S tation Setup I created HOBO sensor and cable protective housing using vinyl drain and PVC piping. I mounted a ir temperature and PAR sensors on a single wood block, 7.62 cm (three in c hes ) in height secured in place by a metal stake, using the mounts that came with each sensor All

PAGE 35

22 exposed wires were en cased in protective PVC piping 1 m in length enabling the Micro Station Data Logger to be placed at a distance preventing interference with the sensors in a microsite (Figure III.3). The Micro Station Data Logger was attached directly to plastic drain piping using screws and wing nuts. The sensor wires ran directly from the data logger through the drain pipe through the one meter of PVC piping to the microsite. The s ensors that I placed in the g round (soil temperature and soil moisture) were positioned at least 4 cm from other sensors or anchoring points to avoid interference in readings All microsites were geolocated with a Trimble G eo XT handheld GPS unit (GeoExplorer 2008 series) and marked with a nail spike and tag number for relocation the following year Data Measurements and Analysis Using Hoboware Pro (version 3.3.0), I set the instrumentation to record all microclimate varia bles every 15 minutes from mid July to mid September each year (Table III.4). Data recorded included average w ind speed over each 15 minute interval (based on one minute observations), and g ust speed, defined as the highest wind speed over a ny three second period within each 15 minute logging interval All analyses for this study were complete d using R statistical software version 2.14.1 (R Core Development Team, 2011) For 2011 and 2012 data, a two factor analysis was completed for each climate variable t o determine if our blocks represented true replicates The assumptions of normality and homoscedasticity were not met so I first determined the interaction effect of blocks on microsite by ranking the data, and comparing the average ranking of the climate data among micr osites and blocks in an ANOVA Then, microclimate variables were compared among microsites either separately among blocks, or combined if no interaction effect was detected, using a Kruskal Wallis one way analysis of variance. I used a Pairw ise Wilcoxon rank

PAGE 36

23 sum test for post hoc analysis. V ariables that trended seasonally (temperature and PAR) were first detrended by a quadratic or linear model of best fit, chosen by comparing A kaike information criterion (AIC) scores and using the model wit h the lowest score Residuals, representing the distance and direction from the average, were retained after detrending, and used as a method of comparing deviations among microsite classes from the mean See Figure III.4 for a stepwise example of the detr ending process. I compared the n umber of freeze thaw events among microsites for each study area using a c hi squared goodness of fit test. A detailed description of all analyses is presented in R Code Statistical Analysis (Appendix). For a general clima te description of each study area, air temperature, soil moisture and wind speeds from unprotected microsites only were averaged for each study area, each year and described graphically. Soil Sampling S oil samples were collected directly leeward of whiteb ark pine, Engelmann spruce, rock and unprotected microsites using a 2.56 cm diameter soil corer. I collected two samples per microsite from 20 microsites o f each type at each study area Different microsites were chosen for soil samples than were used for microclimate data loggers. Each core was taken to the O horizon (6 cm depth a t Divide Mountain; 15 cm depth at Line Creek ) the layer directly influencing seed and seedling growth. Due to budgeting constraints t en soil samples representing each study are a for each microsite were randomly selected, dried, ground and analyzed for total carbon and nitrogen at the EcoCore Analytical Services Lab at Colorado State University (CSU) in Fort Collins, CO. I determined the percent of total carbon and nitrogen b y ru nning each sample through a LECO CN elemental analyzer. Because s amples from Divide Mountain contained high amounts of leaf

PAGE 37

24 litter pushing the percentage of carbon beyond the limits of the elemental analyzer for soil analysis, I ran the se samples using ve getation sampled rather than soil sampled to calibrate the elemental analyzer Percentages of total carbon and nitrogen in soil were compared among microsites and study area using a one way analysis of variance, as all parametric test assumptions were met. A Tukey HSD ( honestly significant difference) test was used for post hoc analysis. Tables and Figures Figure III.1 Research study areas: Divide Mountain on the Blackfeet Rese r vation and Line Creek Research Natural Area, Custer Nation al Forest in Montana. Basemap provided by Google Maps (2013) Divide Mountain Line Creek RNA

PAGE 38

25 Table III. 1 Block site descriptions for each study area. Includes the mean and standard deviation ( M SD ) for general wind direction (), slope aspect () and slope steepness ( ). Year Study Area Block Wind Direction () Slope Aspect () Slope Steepness () 2010 Divide Mtn NE -34 4 30 7 W -273 8 26 4 Line Creek NE 254 0 85 11 10 0 W 268 0 277 21 17 1 2011 Divide Mtn 1 235 0 39 9 22 0 2 235 0 33 6 25 0 3 235 0 48 4 20 0 Line Creek 1 273 7 67 3 14 2 2 273 10 79 8 15 0 2012 Divide Mtn 1 235 0 40 0 15 0 2 235 0 48 0 18 0 3 235 0 40 0 15 0 Line Creek 1 237 8 60 0 10 0 2 2 36 13 62 5 10 0 Table III. 2 Mean height, length and width (cm) for objects within microsites used at Divide Mountain and Line Creek RNA (2010, 2011 and 2012). Includes the mean and standard deviation ( M SD ) for each variable. WB ES RK Year Study Area Height Length Width Height Length Width Height Length Width 2010 Divide Mtn 31 4 56 0 71 0 39 10 50 0 63 0 16 1 37 1 31 5 Line Creek 65 22 62 24 58 21 57 19 52 3 65 1 19 2 51 37 53 6 2011 Divide Mtn 30 3 73 17 91 30 32 8 79 26 87 21 42 18 79 53 126 63 Line Creek 57 12 67 29 59 14 50 18 83 39 83 21 21 4 56 25 57 17 2012 Divide Mtn 33 9 65 14 70 15 36 8 79 38 100 31 41 18 41 6 71 3 Li ne Creek 70 24 77 22 88 21 58 23 77 22 88 20 19 2 71 39 102 28

PAGE 39

26 Figure III.2 Microstation setup using PVC pip es for cable protection leeward wood blocks for sensor mounting HOBO Onset sensors and Micro Station Data L ogger Setups from a) unprotected microsite in 2010 with air and soil temperature, soil moisture and PAR sensors and b) whitebark pine microsite in 2011 and 2012 with air and soil temperature, soil moisture and anemometer sensors. Figure III.3. Full microstation s et up with Micro Station Data Logger in a whitebark pine microsite. a) b )

PAGE 40

27 Table III.3 HOBO Onset Microclimate Se nsors and installation heights in relation to ground. Sensor Model Height in Relation to Ground after Installation (cm) Air Temperature/ Relat ive Humidity (C/%) S THB M002 12 bit Temperature/RH Smart Sensor (2m cable) 10 Soil Temperature ( C) S TMB M002 12 Bit Temp Smart Sensor (2m cable) 3 Soil Moisture (m 3 /m 3 ) 1 S SMC M005 Soil Moisture Smart Sensor 6 Photosynthetic Light (mol/m/sec) S LIA M003 Photosynthetic Light (PAR) Smart Sensor with 3m cable 7 Wind Anemometer (m/s) 2 S WSA M003 Wind Speed Smart Sensor with 3m cable 10 1 used in 2010 only 2 used in 2011 and 2012 only Table III.4 Start and end date for Micro Station Data Loggers Year Study Area Logging Start Logging End 2010 Divide Mtn July 8 September 11 Lin e Creek July 22 September 14 2011 Divide Mtn July 9 September 12 Line Creek July 22 September 17 2012 Divide Mtn July 4 September 15 Line Creek July 17 September 17

PAGE 41

28 Figure III.4. T he sequential steps for maximum soil temperature analysis: a) raw data from Line Creek 2011, show a general decreasing trend in maximum daily soil temperatures over the growing season ; and, according to AIC scores, a quadratic model better represents the data than a linear model ( T max o 1 2 Day 2 + e where e ~ N(0,0) ); b) residual data from the quadratic model show a linear trend that remains more stable over the growing season; c) boxplots of data show residual comparisons among microsites where boxplot w hiskers represent the maximum and minimum obs ervation, outer edges of boxes represent 25 th and 75 th percentile, and the middle line represents the median. Median residuals represent the deviation (distance and direction) of each microsite type from the overall mean value.

PAGE 42

29 CHAPTER IV. RESULTS Climate and Soil Differences between Study Areas Microclimate data from 2010 through 2012 from the unprotected microsites provided descriptive information on general climatic differences between the northeastern slopes for e ach study area Divide Mountai n was generally warmer, wetter, and windier than Line Creek. Median d aily average air temperatures were higher at Divide Mountain in 2011 and 2012 (Figure IV.1); but median daily variance in temperature was greater at Line Creek in all three years ( Figure IV.2, Table IV.1), indicating that Line Creek experience d more extreme temperature swings during the growing season s than Divide Mountain. This possibility is supported by differences in the number of freeze thaw events between each study area: Line Creek had significantly more freeze thaw events than Divide Mountain both in 2011 2 = 9.55 df = 1, P = 0.0 02 2 = 39.3404, df = 1, P = 3.56e 10) (Figure IV.3). I use d soil moisture as a broad measure of precipitation throughout the growing season for each study area ; snow melt was no longer present when the sensors began recording Divide Mountain had higher median daily soil moisture than Line Creek in 2010 and 2012 but experienced lower moisture than Line Creek in 2011 ( Figure IV.4 Table IV.2 ). The driest summer for Divide Mountain and Line Creek occurred in 2011 and 20 12, respectively. Wind speed data from 2011 and 2012 indicate that Divide Mountain had higher daily wind speeds that reached greater maximums than Line Creek in 2011, but were about equal in 2012 (Figure IV.5) V ariance s in wind speeds between study areas were generally greater at Divide Mountain in 2011; although the median variance s between study areas were approximately equal (Figure IV.6, Table IV.3)

PAGE 43

30 Total percentages of carbon and nitrogen in microsites differed greatly between study areas (Figure I V.7 Table IV.4 ). Results from the two way ANOVA showed that Divide Mountain soils had significantly higher mean percentages of nitrogen ( Microsite Study area i nteraction effect F = 0.7356, df = 3, P = 0.5342; Study area F = 241.6890 df = 1 P < 2e 16 ) and carbon ( Microsite Study area interaction effect F = 0.2666, df = 3, P = 0.8492; Study area F = 253.7934 df = 1 P < 2e 16 ) than Line Creek soils Comparing Microclimate among Microsites Soil Temperature The r esults for daily soil temperature amon g microsites at both Divide Mountain and Line Creek varied significantly among and between blocks. T he interaction effect of block on microsite for soil temperature maxima, minima and variance was significant at all locations so blocks were analyzed separ ately and not treated as replicates In 2010 on the northeast slope of Divide Mountain, Engelmann spruce had significantly lower residual values for maximum soil temperature than whitebark pine, rock and unprotected microsites S oil temperature maxi ma wer e significantly lower than average (Figures IV.8 .a 8.c ) in these microsites On the west slope, both conifers had significantly lower residuals than rock and unprotected microsites and lower than average soil temperature maxim a In 2012, residuals from E ngelmann spruce microsites again were significantly below average in all three blocks compared to rock and unprotected microsites Residuals representing whitebark pine microsites were significantly lower in two out of three blocks. R esults from the Kruska l Wallis test s are found in Table IV.5. All post hoc results are found in Tables A7 A 16 Chosen soil temperature models of best fit are found in Table A 4.

PAGE 44

31 At Line Creek, conifer microsites tended to have lower residuals representing below average soil te mperature maxima (Figures IV.8 a 8 c ) However, results are inconsistent. In 2010 on the northeast slope, whitebark pine had lower than average soil temperature maxima compared to rock and unprotected microsites followed by Engelmann spruce microsites Th is occurred again, but only at one block in 2012. Only one block in both 2011 and 2012, Engelmann spruce microsites had significantly lower residuals than all other microsite types, followed by whitebark pine, while rock and unprotected microsites had high er residual data and higher than average soil temperature maxima (Table IV.5). Results for soil temperature minima in 2010, 2011 and 2012 are not consistent at Divide Mountain (Figures IV.9 a 9 c ) On the west slope in 2010, the Engelmann spruce microsite had residuals that represented below average soil temperature minima compared to all other microsite types, but this pattern was not present on the northeast slope. The effect of microsite on minimum soil temperature varies significantly among blocks in b oth 2011 and 2012 and no consistent patterns occurred At Line Creek on the northeast slope, conifer microsites generally had higher residual data for soil temperature minima than rock and unprotected microsites although results were not always significan t (Figures IV.9 .a 9.c ). This pattern also occurred at the west slope in 2010. No distinguishable differences occurred in residual data between whitebark pine and Engelmann spruce microsites All Kruskal Wallis results are presented in Table IV.6. Results f or variances in soil temperatures were also not consistent at Divide Mountain (Figures IV.10 a 10 c ). However, Engelmann spruce microsites tended to have residuals representing lower daily variances more frequently than other microsites on the northeast sl ope (Blocks 1 and 2 in both 2011 and 2012). In 2010 on the west slope, both conifers, and whitebark

PAGE 45

32 pine in particular had significantly lower residuals than rock and unprotected microsites. At Line Creek, conifer microsites had significantly lower residu als, showing that variances in soil temperature were generally below average. In 2010 and Block 2 in 2012, whitebark pine followed by Engelmann spruce had significantly lower residuals than rock and unprotected microsites, while Engelmann spruce microsite s had lower residuals at both blocks in 2011 and one in 2012. See Table IV.7 for a ll Kruskal Wallis test results Air Temperature In all three years at Divide Mountain no significant differences occurred among microsites in the comparison of residual dat a from neither daily maximum ( Figures A1 .a A1.c ) or minimum air temperatures ( Figures A2 .a A2.c ). On the northeast slope in 2011 and 2012 at Line Creek Engelmann spruce microsites had significantly lower residuals from maximum air temperature data com pared to all other microsite types except at Block 2 in 2012 where whitebark pine residuals were about even Rock microsites also experienced maximum air temperatures that were higher than average compared to all other microsites except for one block in 2012 where data from the unprotected microsite were also high No other differences among microsites occurred in daily minimum air temperatures at Line Creek All Kruskal Wallis test results are presented in Tables A1 and A2. Chosen models of best fit are found in Table A 5 PAR PAR, which was only taken in the 2010 growing season, varied slightly from the west to northeast slope aspects ( Figure IV.11 ) At Divide Mountain, Engelmann spruce microsites in general had lower daily PAR than rock and unprotected m icrosites although results were not significant On the northeastern slope aspect Kruskal Wallis test results showed that whitebark

PAGE 46

33 pine followed spruce in having significantly lower residuals for PAR than rock and unprotected microsites ( Table IV. 8 ). C hosen models of best fit are found in Table A 6 PAR results at Line Creek were consistent across slope aspects in 2010. Conifer and whitebark pine microsites in particular, had lower than average daily PAR on both west and northeast slopes Further, white bark pine PAR residuals were significant ly lower than those for Engelmann spruce. Gust Speed At Divide Mountain in 2011 and 2012, no consistent patterns were present in the differences that occurred in maximum gust speed among microsite types ( Figure s IV.1 2 a 1 2 b ). At Line Creek, d aily maximum gust speeds were attenuated in Engelmann spruce microsites in 2011 and 2012. At one block in 2011 and both blocks in 2012 Engelmann spruce microsites had significantly lower daily maximum gust speeds than whitebark pine microsites. All results from the Kruskal Wallis test are presented in Table IV. 9 Wind Speed At Divide Mountain, protective microsites moderated wind speeds better than unprotected sites ( Figure s IV.1 3 a 1 3 b ). In 2011 and 2012, whitebark pine, Enge lmann spruce, and rock microsites had significantly lower daily wind speed averages than unprotected microsites except at Block 3 in 2012 where the whitebark pine microsite had significantly lower average wind speeds than all other microsites Similar to Divide Mountain, all protective microsites at Line Creek mitigated daily average wind speeds in 2011 Results show that conifer microsites had significantly lower daily average wind speeds in comparison to rock and unprotected microsites During the 2012 growing season, Engelmann spruce microsites experienced significantly lower daily average

PAGE 47

34 wind speeds than unprotected, rock and whitebark pine microsites. All Kruskal Wallis results are shown in T able IV.10 Soil Moisture Soil moisture was extremely va riable each year at both Divide Mountain and Line Creek ( Figures A3 a A3 c ). In 2010 at Divide Mountain whitebark pine and rock microsites had significantly higher daily soil moisture averages than Engelmann spruce and unprotected microsites on the northe ast slope On the west slope, whitebark pine had a significantly lower soil moisture average than all other microsites, while the unprotected site was significantly higher These results are not reflected i n the 2011 and 2012 growing season s and inconsist ency occurred among blocks. At Line Creek, Engelmann spruce microsites were generally drier than other microsites in 2011 but not 2012 (Figures A3.a A3.c) In 2010 on the northeastern slope, daily average soil moisture did not differ significantly among microsites. All Kruskal Wallis test are shown in Table A3. Sky Exposure At Divide Mountain on the northeast slope i n 2010 and 2011, whitebark pine microsites had lower percentages of sky exposure than all other microsite types, but results wer e only significantly different in 2011( Table IV.1 1 ). Sky exposures on the northeast slope for the Engelmann spruce, rock, and unprotected microsite were approximately equal in 2010, but differed in 2011, with whitebark pine microsites having lower sky expo sure than rock and unprotected microsites at two blocks ( Figures IV.1 4 .a 1 4 .b ). Sky exposure data were missing for whitebark pine on the west slope of Divide Mountain in 2010 and Engelmann spruce from B lock 2 on the northeast slope of Divide Mountain in 2011.

PAGE 48

35 On the northeast slope of Line Creek in 2010 and 2011, whitebark pine and Engelmann spruce microsites had lower percentages of sky exposure tha n rock and unprotected microsites ( Figures IV.1 4 .a 1 4 .b). Unprotected microsites contained the highest p ercentage of sky exposure both years On the west slope, conifers had lower sky exposure than rock and unprotected microsites but sky exposure in the Engelmann spruce microsite was slightly lower than the whitebark pine microsite All chi square test resul ts are shown in Table IV. 1 1 Because the same microsites were used in 2012 as 2011, w e expect ed that sky exposure for 2012 would not differ. C and N in Soils At both Divide Mountain and Line Creek, soils collected from whitebark pine microsites had slig htly higher mean percentages of total nitrogen than soils from Engelmann spruce, rock and unprotected microsites ( Figure IV. 1 5 .a) However results from the two way ANOVA show that differences among microsites were not significan t ( Table IV. 1 2 ). Similarly, in comparison to all other microsites, whitebark pine microsites had soils with greater percentages of total carbon but results were not significant ( Figure IV. 1 5 .b ).

PAGE 49

36 Tables and Figures Figure IV.1. Daily average air tempera tures (C) for the northeast slope of Divide Mountain and Line Creek for the growing season s of a) 2010, b) 2011 and c) 2012. Data provided were averaged from unprot ected microsites only.

PAGE 50

37 Figure IV.2. Daily variance in air tem peratures (C 2 ) for the northeast slope of Divide Mountain and Line Creek for the growing season of a) 2010, b) 2011 and c) 2012. Data provided were averaged from unprotected microsites only.

PAGE 51

38 Table IV.1 Daily air temperature ( C) descriptive statistics for Divide Mountain and Line Creek. Data provided were averaged from unprotected microsites only. Study Area Year Median Variance Max Min Divide Mtn 2010 11.407 22.701 19.730 0.771 Line Creek 10.926 16.649 16.548 0.505 Divid e Mtn 2011 14.012 13.166 21.094 3.658 Line Creek 11.354 7.977 17.140 2.400 Divide Mtn 2012 14.235 16.709 19.971 2.266 Line Creek 12.618 12.897 18.047 2.333

PAGE 52

39 Figure IV.3. Number of freeze thaw events on the northeastern slope at Divide Mountain and Line Creek for 2011 and 2012 between mid July and mid September (Table III 4). One freeze thaw event occurs every time air temperature dropped below 0C and then rose above 0C. Data provided were averaged from unprotected microsite s only.

PAGE 53

40 Figure IV.4. Daily average soil moisture (m 3 m 3 ) for the northeastern slopes of Divide Mountain and Line Creek for the growing seasons of a) 2010, b) 2011 and c) 2012. Data provided were averaged from unprotected microsites only.

PAGE 54

41 Table IV.2 Daily soil moisture (m 3 m 3 ) descriptive statistics for Divide Mountain and Line Creek. Data provided were averaged from unprotected microsites only. Study Area Year Median Max Min Divide Mtn 2010 0.104 0.155 0.116 Line Creek 0.065 0.139 0.023 Divide Mtn 2011 0.009 0.152 0062 Line Creek 0.036 0.113 0.002 Divide Mtn 2012 0.092 0.208 0.160 Line Creek 0.021 0.092 0.080

PAGE 55

42 Figure IV.5 Daily average wind speeds (m/s) on the northeastern slope for Divide Mountain and Line Creek for the growing season of a) 2011 and b) 2012. Data provided were averaged from unprotected microsites only.

PAGE 56

43 Figure IV.6 Daily variance in wind speeds (m/s) 2 for the n ortheastern slope on Divide Mountain and Line Creek for the growing season of a) 2011 and b) 2012. Data provided were averaged from unprotected microsites only.

PAGE 57

44 Tab le IV.3 Daily wind speeds (m/s) descriptive statistics for Divi de M oun t ai n and Line Creek. Data provided were averaged from unprotected microsites only. Study Area Year Median Variance Max Min Divide Mtn 2011 0.823 0.426 2.862 0.016 Line Creek 0.416 0.122 1.703 0.000 Divide Mtn 2012 0.420 0.477 3.268 0.000 Line Creek 0.422 0.179 2.160 0.000

PAGE 58

45 Figure IV.7. Total percentages of a) nitrogen and b) carbon in soils collected from northeast slope unprotected microsites at Divide Mountain and Line Creek. Data provided were averaged from all micr osites. Table IV.4 Descriptive statistics for total percent nitrogen and carbon in soils at each study area in 2011. Variable Study Area N Mean SD 95% Confidence Interval Min Max Lower Bound Upper Bound Nitrogen Divide Mtn 40 0.455 0.098 0.42 5 0.48 6 0.2665 0.6991 Line Creek 40 0.19 2 0.03 8 0. 180 0. 203 0.09 1 0.282 Carbon Divide Mtn 40 9.074 2.504 8.298 9.850 4.705 14.690 Line Creek 40 2.34 5 0.571 2.168 2.522 0.89 6 3.721

PAGE 59

46 Figure IV.8.a Boxplots of residual data from 2010 daily maximum soil temperatures (C) are shown fo r Divide Mountain a) northeast slope, b) west slope and Line Creek c) northeast slope, d) west slope

PAGE 60

47 Figure IV.8.b. Boxplots of residual data from 2011 daily maximum soil temperatures are shown for Divide Mountain a) Block 1, b) Block 2 c) Block 3 and Line Creek d ) Block 1 and e) Block 2

PAGE 61

48 Figure IV.8.c. Boxplots of residual data from 2012 daily maximum soil temperatures are shown for Divide Mountain a) Block 1, b) Block 2 c) Block 3 and Line Creek d) Block 1 and e) Block 2

PAGE 62

49 Table IV.5 Daily m aximum soil temperature results from a Kruskal Wall is test, comparing the residual data from a model of best fit (either linear or quadratic) for daily soil temperatur e maximums among microsites, ( = 0.05). Year Study Area Block N Df P value 2 2010 Divide Mtn NE 1 3 1.73 e 05 24.76 4 W 1 3 5.93 e 13 59.98 2 Line Creek NE 1 3 < 2.2 e 16 135.0 10 W 1 3 < 2.2 e 16 83.43 3 2011 Divide Mtn 1 (NE) 1 3 < 2.2e 16 86. 786 2 (NE) 1 3 0.0004 4* 18.00 4 3 (NE) 1 3 7.975e 11 50.00 4 Line Creek 1 (NE) 1 3 < 2.2e 16* 139.8 30 2 (NE) 1 3 < 2.2e 16* 84.608 2012 Divide Mtn 1 (NE ) 1 3 1.756e 15 71.80 2 2 (NE) 1 3 < 2.2e 16* 144.846 3 (NE) 1 3 < 2.2e 16* 124.669 L ine Creek 1 (NE) 1 3 < 2.2e 16* 110.23 4 2 (NE ) 1 3 < 2.2e 16* 123.893

PAGE 63

50 Figure IV.9.a. Boxplots of residual data from 2010 daily m in imum soil temperatures (C) are shown fo r Divide Mountain a) northeast slope, b) west slope and Line Creek c) northeast slope, d) west slope.

PAGE 64

51 Figure IV.9.b Box plots of residual data from 2011 daily min imum soil temperatures are shown for Divide Mountain a) Block 1, b) Block 2 c) Block 3 and Line Creek d) Block 1 and e) Block 2

PAGE 65

52 Figure IV.9.c. Boxplots of residual data from 2012 daily min imum soil temperatures are shown for Divide Mountain a) Block 1, b) Block 2 c) Block 3 and Line Creek d) Block 1 and e) Block 2

PAGE 66

53 Table IV. 6 Minimum soil temperature re sults from a Kruskal Wallis test, comparing the residuals from a model of best fit (either linear or quadratic) for daily soil temperature minimums among microsites, ( = 0.05). Year Study Area Block N d f P value 2 2010 Divide Mtn NE 1 3 0.5159 2.28 3 W 1 3 0.001649 15.20 6 Line Creek NE 1 3 1.45e 10 48.791 W 1 3 3.90e 13 60.83 6 2011 Divide Mtn 1 (NE) 1 3 1.331e 13 63.01 9 2 (NE) 1 3 0.04248 8.17 8 3 (NE) 1 3 7.862e 06 26.40 1 Line Creek 1 (NE) 1 3 7.348e 10 45.47 1 2 (NE) 1 3 0.000 2268 19.392 2012 Divide Mtn 1 (NE ) 1 3 0.000456 17.92 4 2 (NE) 1 3 0.001578 15.299 3 (NE) 1 3 0.05513 7.59 7 Line Creek 1 (NE) 1 3 2.729e 08 38.07 2 2 (NE ) 1 3 1.314e 06 30.10 2

PAGE 67

54 Figure IV.10 .a. Boxplots of residual data from 2010 daily variance in soil temperatures (C) are shown for Divide Mountain, a) northeast slope, b) west slope, and Line Creek c) northeast slope and d) west slope

PAGE 68

55 Figure IV.10.b. Boxplots of residual data from 2011 daily variance i n soil temperatures are shown for Divide Mountain a) Block 1, b) Block 2 c) Block 3 and Line Creek d) Block 1 and e) Block 2

PAGE 69

56 Figure IV.10.c. Box plots of residual data from 2012 daily variance in soil temperatures are shown for Divide Mou ntain a) Block 1, b) Block 2 c) Block 3 and Line Creek d) Block 1 and e) Block 2

PAGE 70

57 Table IV.7. Daily variance soil temperature results from a Kruskal Wall is test, comparing the residual data from a model of best fit (either linear or quad ratic) for daily variance in soil temperature s among microsites, ( = 0.05). Year Study Area Block N df P value 2 2010 Divide Mtn NE 1 3 2.493e 12* 57.062 W 1 3 3.828e 14 65.549 Line Creek NE 1 3 <2.2e 16* 168.365 W 1 3 <2.2e 16* 111.105 2011 Divide Mtn 1 (NE) 1 3 <2.2e 16* 165.688 2 (NE) 1 3 1.81e 08* 38.914 3 (NE) 1 3 <2.2e 16* 95.068 Line Cree k 1 (NE) 1 3 <2.2e 16* 148.071 2 (NE) 1 3 <2.2e 16* 80.217 2012 Divide Mtn 1 (NE ) 1 3 <2.2e 16* 122.240 2 (NE) 1 3 <2.2e 16* 186.575 3 (NE) 1 3 <2.2e 16* 123.124 Line Creek 1 (NE) 1 3 <2.2e 16* 116.411 2 (NE ) 1 3 <2.2e 16* 115.934

PAGE 71

58 Figure IV. 11 Boxplots of residual data from 2010 daily PAR are shown for Divide Mountain a) western slope, b) northeastern slope and c) coordinated raw data comprised of daily PAR ( umol/m 2 /sec) for the northeastern slope Line Creek data are sho wn for d) western slope e) northeastern slope, and f) coordinated data comprised of daily PAR for the northeastern slope.

PAGE 72

59 Table IV.8 PAR results from a Kruskal Wallis test, comparing the residuals from a model of best fit (either linear or quadratic) for daily amount of PAR among microsites, ( = 0.05). Year Study Area Block N df P value 2 2010 Divide Mtn NE 1 3 1.40 e 15 72.268 Divide Mtn W 1 3 0.006817 12.1717 Line Creek NE 1 3 < 2.2e 16 86.0993 Line Creek W 1 3 < 2.2e 16 148.338

PAGE 73

60 Figure IV.1 2 a. Boxplots of daily maximum gust speeds (m/s) in 201 1 are shown for Divide Mountain a) Block 1, b) Block 2, and c) Block 3 and Line Creek d) Block 1 and e) Block 2.

PAGE 74

61 Figure IV. 12 .b. Boxplots of daily maximum gust speeds (m/s) in 2012 are shown for Divide Mountain a) Block 1, b) Block 2, and c) Block 3 and Line Creek d) Block 1 and e) Block 2.

PAGE 75

62 Table IV. 9 Daily maximum wind gust speed (m/s) results from a Kruskal Wallis test, ( = 0.05). Year Study Area Block N df P value 2 2011 Divide Mtn 1 (NE) 1 3 0.0004 18.121 0 2 (NE) 1 3 0.1388 5.4966 3 (NE) 1 3 0.002 4 14.4396 Line Creek 1 (NE) 1 3 0.015 7 10.3678 2 (NE) 1 3 3.818e 09 42.1013 2012 Divide Mtn 1 (NE ) 1 3 1.455e 07 34.6355 2 (NE) 1 3 0.0073 12.0231 3 (N E) 1 3 0.000761 16.8437 Line Creek 1 (NE) 1 3 0.0001521 20.2297 2 (NE ) 1 3 5.64e 11 50.7101

PAGE 76

63 Figure IV.1 3 a. Boxplots of daily average wind speeds (m/s) in 2011 are shown for Divide Mountain a) Block 1, b) Block 2, and c) Block 3 and Line Creek d) Block 1 and e) Block 2.

PAGE 77

64 Figure IV.1 3 b. Boxplots of daily average wind speeds (m/s) in 2012 are shown for Divide Mountain a) Block 1, b) Block 2, and c) Block 3 and Line Creek d) Block 1 and e) Block 2.

PAGE 78

65 Table IV.10 Daily average wind speed (m/s) results from a Kruskal Wallis test, ( = 0.05). Year Study Area Block N df P value 2 2011 Divide Mtn 1 (NE) 1 3 <2.2e 16* 99.8519 2 (NE) 1 3 2.84e 10* 47.4122 3 (NE) 1 3 1.037e 13* 63.5254 L ine Creek 1 (NE) 1 3 1.391e 12 58.2489 2 (NE) 1 3 <2.2e 16 119.3188 2012 Divide Mtn 1 (NE ) 1 3 < 2.2e 16 156.0889 2 (NE) 1 3 < 2.2e 16 92.5072 3 (NE) 1 3 2.872e 11 52.0858 Line Creek 1 (NE) 1 3 8.774e 09 40.3979 2 (NE ) 1 3 < 2.2e 16 7 9.5411

PAGE 79

66 Figure IV. 1 4 .a. Percent sky exposure for 2010 microsites at a) Divide Mountain, NE slope, b) Line Creek, NE slope, c) Divide Mountain, W slope, and d) Line Creek, W slope (n = 1). Whitebark pine (WB) data missing for Divide Mountai n, W slope

PAGE 80

67 Figure IV.1 4 .b. Average percent sky exposure for 2011 m icrosites at Divide Mountain a) Block 1, b) Block 2, c) Block 3 and Line Creek d) Block 1 and e) Block 2. Engelmann spruce (ES) data missing for Divide Mountain, Block 2.

PAGE 81

68 Table IV.11 Chi square goodness of fit test results comparing sky exposure (%) among microsites ( = 0.05). Year Study Area Block N df P value 2 2 010 Divide Mtn NE 1 3 0.60 1.8 6 W 1 2 0.40 0.1 3 Line Creek NE 1 3 2.49e 04* 19. 20 W 1 3 7.77 e 06 26.42 2011 Divide Mtn 1 1 3 0.02 9.63 2 1 2 6.62 e 07 28.4 6 3 1 3 1.46 e 08 39.36 Line Creek 1 1 3 8.43e 04* 16.6 3 2 1 3 1.20 e 05 2 5.52

PAGE 82

69 Figure IV.15 .a. Percent total nitrogen found in soils at a) Divide Mountain and b ) Line Creek. Soils collected from microsites in 2011. Figure IV.15 .b. Percent total carbon found in soils at a) Divide Mountain and b ) Line Creek. Soils col lected from microsites in 2011.

PAGE 83

70 Table IV.12 Percent nitrogen and carbon two way ANOVA results, comparing per cent nutrients among microsites and study areas ( = 0.05). Nutrient Study Area n df P value F % N Study Area on Microsite ( interaction effect ) 3 0.534 0.736 Study Area 40 1 < 2 e 16 241.689 Microsite 10 3 0.475 0.843 % C Study Area on Microsite ( interaction effect ) 3 0.849 0.267 Study Area 4 0 1 < 2 e 16 253.793 Microsite 10 3 0.845 0.2 72

PAGE 84

71 CHAPTER V. DISCUSSION While many studies have focused on factors that limit treeline growth and allow for a response to global warming at a broad scale (Krner, 1998; Krner, 2003), studies that clarify local processes are lacking. My studies within two treeline ecotones on the East ern Front of the Rocky Mountains Montana examine the microclimate of four common microsites, including which directly impact seed germination and seedling growth. Currently, m ost temperature data used in describing treeline conditi on s are estimated using a moist adiabatic lapse rate from the closest meteorological station. These stations are typically found at a distance from treeline and at lower elevations; and therefore their application to treeline is not entirely accurate (Hol tmeier 2003). For our study area s we provide a continuous and detailed local climate description for the majority of the 2010, 2011 and 2012 growing seasons. At several treeline sites throughout the E astern Rocky Mountains, whitebark pine was found to i nitiate tree islands through facilitation more frequently than any other species of conifers but the explanation for this relationship is unknown (Resler and Tomback, 2008; Smith McKenna et al., unpublished data; Tomback et al. unpublished data). There a re several possible causations, not necessarily mutually exclusive, that could account for these observations: First w hitebark pine may facilitate tree islands more frequently than other conifers, simply because it occurs as solitary trees more frequently than Engelmann spruce and subalpine fir at these sites (Blakeslee, 2012). Its abundance may result from its general hardiness, and/or from its unique ( Tomback, 1982; Arno and Hoff, 1990; Tom back, 2001; Maher and Germino, 2006 ). However, there are treelines where whitebark pine is abundant and yet not a frequent tree island initiator (Habeck 1969; Tomback et

PAGE 85

72 al., unpublished MS. ). At both Divide Mountain and the Line Creek study areas sampl ing from randomly distributed plots indicated that there were significantly greater numbers of solitary whitebark pine than solitary occurrences of other conifers ( Divide Mtn, 64%; Line Creek, 83%) (Blakeslee 2012) These observations alone suggest that t he probability is higher that whitebark pine might serve as a tree island initiator, especially if the likelihood of its survival is greater. Alternatively, whitebark pine through facilitation may provide a higher quality microsite than other conifers or other microsites for example, by creating a micro climate that is more favorable for seed germination and seedling establishment. The objective of this study was to test th e latter explanation by examining the abiotic conditions represented by several typ es of treeline microsites. First I predicted that conifers in general would have leeward microsites with reduced climate extremes in comparison to rock and unprotected microsites, and secondly, I predicted that whitebark pine microsites would p rovide the most favorable micro climates. Our results suggest that at both study areas, conifer microsites displayed different and generally more favorable microclimates in their lee compared to rock and unprotected microsites. Overall, conifer microsites were charac terized by reduced maximum air and soil temperatures, higher minimum soil temperatures, lower PAR, and reduced wind and gust speeds compared to rock and unprotected microsites. Microsites with no protection, and those leeward of rocks generally displayed harsher microclimates. Results also varied somewhat between study areas, with greater differences among microsites under the more extreme climatic regime of the Line Creek study area than the Divide Mountain study area. Despite these differences, whitebark pine microsites in general do not appear to produce a mor e favorable leeward microsite with respect to microclimate for conifer recruitment than Engelmann spruce microsites.

PAGE 86

73 From analysis of fisheye lens exposure photos, it appears that the whitebark pine krummholz canopies reduce sky exposure more than all other microsites, and there is some indication that soils leeward of a krummholz whitebark pine may have higher carbon and nitrogen content, but this requires further testing. Climatic Stress in Study A reas Which study area has the most extreme conditions is difficult determine: both appear to have different but challenging conditions for conifer growth an d survival. Based on our micro climate dat a from unprotected microsites, the ATE at Divide Mountain is moister warmer, but windier than the ATE at Line Creek. Wind patterns at Divide Mountain also appear to be more complex Extreme wind may explain why conifers at Divide Mountain show more prominent krummholz growth and severe windward needle kill than those at Line Creek. However, during the growing season Line Creek showed higher temperature variation, higher PAR and less mo isture, presumable resulting in more climatic stress. High wind speeds, and winter winds in particular, can result in conifer ne edle dehydration and mortality (Hadley and Smith 1986). Winter winds and snow pack distribution patterns also influence pattern s of conifer growth at treeline ( Httenschwiler and Smith, 1999; Batllori et al., 2009). Investigation into these effects may pro vide additional insight into differences in the effectiveness of facilitation among microsites between study areas. Importance of Microsites Seedling survival in the ATE is challenging, and first year seedling survival is extremely low (Cui and Smith, 19 91). Our results support previous findings that objects are important for the fa cilitation of seedling survival because they provide protection from some components of the extreme climate often present in these harsh ecosystems

PAGE 87

74 ( C allaway, 1998; Httenschwiler and Smith, 199 9 ; Germino et al., 2002). However, our study goes farther in characterizing in detail the quality of several common kinds of microsites (Figures V.1.a and V.1.b) At Divide Mountain whitebark pine and Engelmann spruce were average wind speeds. At some locations, Engelmann spruce microsites also experienced a reduction in soil temperature variance. Conifer microsites at Line Creek had lower daily PAR, but also experienced more moderate temperature and wind extremes. Rocks at Line Creek actually had increased air temperature maximums, likely caused by heat radiating off their surface. Cold soil temperatures are one of the major factors that can limit seedling growth at treeline ( Landhusser 1996; Germino and Smith, 1999) At Divide Mountain, Engelmann spruce microsites ameliorated soil temperature variance and conifer microsites at Line Creek both reduced variance and inc reased m in imum soil temperatures. In other words, microsites leeward of rock and unprotected microsites experience more extreme swings each day in soil temperatures. Compared to Divide Mountain, Line Creek had greater variance in air temperature each day, and more freeze thaw events during the growing season. Temperature amelioration may be more important at Line Creek because greater temperat ure extremes occurred commonly throughout the gr owing season. While conifers can withstand extreme temperature lows through the winter season due to cold acclimation ( Sakai and Okada, 1971 ), cold hardiness is greatly reduced during the growing season, when summer conditions leave conifers vulnerable to cold spells ( Kalberer et al., 2006 ). Frequent summer freezes will le ad to visible cold injury, damaging the apical meristem and needles of conifer seedlings ( Christersson and Fircks 1988 ). Therefore,

PAGE 88

75 shelter by facilitation of leeward conifer microsites may be important for tree island recruitment in sites that experience high variations in daily temperatures. Too much solar radiation, along with high soil temperatures can be detrimental for seedling growth, and can easily cause desiccation and mortality of small Engelmann spruce and subalpine fir seedlings (Cui and Smit h, 1991; Germino and Smith, 2002). My radiation data show that both whitebark pine and Engelmann spruce microsites reduce photosynthetically active radiation (PAR) better than rock and unprotected microsites at both study areas likely due to their physica l volume. Interestingly, Engelmann spruce microsites had lower PAR than whitebark pine on the northeast slopes at both Divide Mountain and Line Creek. This is reversed on the western slopes, where whitebark pine microsites had significantly lower PAR than all other microsite types, including Engelmann spruce. The 2010 results for percentages of sky exposure reflected the amount of PAR observed in each microsite at Line Creek but not Divide Mountain, where whitebark pine had lower and Engelmann spruce, rock and unprotected sites had approximately equal amounts of PAR. Reduced s ky exposure is important for seedling warming at night because long wave irradiance is increased and will cool seedlings (Germino and Smith, 1999 ). Our data suggest that conifer microsi tes may be better at insul a ting seedlings during nighttime temperature lows. Error in sky exposure data could have easily occurred from uneven ground, or a slight angle of the camera. Sky exposure estimates also did not take into account the density of nee dles, which would be reflected in PAR values. W hi tebark pine shoots and needles, measure at our study areas, are generally longer than those of Engelmann spruce (Blakeslee, 2012) so the lower levels of percent sky exposure in whitebark pine micros ites may be explained by their longer branches, longer needles and more irregular canopies (Blakeslee, 2012 )

PAGE 89

76 Maximum gust speeds represent the absolute highest wind speeds experienced in microsites throughout the growing season. Wind speeds experienced by seed lings close to the ground surface are not nearly as strong as wind speeds even two meters above the surface (Holtmeier, 2003). Still, extremely high wind speeds can impact seedling survival in discrete events, causing physical injury to conifer growth by b reaking branches, uprooting seedlings and killing and blowing off needles, which reduces a conifers ability to photosynthesize (Holtmeier, 2003) In 2011 and 2012 at Line C reek, the only microsites type that reduced maximum gust speeds w ere leeward of Enge lmann spruce. It is unclear why whitebark pine and rock microsites did not block wind differently from unprotected microsites. Alt hough solitary Engelmann spruce are less abundant than whitebark pine at Line Creek, one explanation is that its growth form d oes appear denser. Further more t he barrier to wind represented by a protective object is likely not linked to wind speed in a linear fashion. The ability of a low growing plant to modify wind speeds decreases as wind speeds become more extreme (Geiger, 19 50). As opposed to maximum gust speeds, daily average wind speeds represent the wind conditions experienced mo st frequently within a microsite on a daily basis. Although not as strong as wind gusts, these more constant wind speeds can influence microsite conditions by altering air and soil temperature and soil moisture through convection ( Holtmeier, 2003 ). both Divide Mountain and Line Creek in 2012 and 2012 However, whitebark pine microsites di d not reduce average wind speeds in 2012 as they did in 2011 at Line Creek. Unprotected microsites had higher d aily average wind speeds, which over time can cause needle desiccation, higher transpiration rates and increased soil evaporation rates, resulti ng in higher drought stress (Holtmeier, 2003) The

PAGE 90

77 wind speeds on Divide Mountain were lower in 2012 than in 2011, which may suggest a non linear influence on some micro sites. From 2010 to 2012, soil moisture among microsites w as extremely variable and no distinct patterns emerged at either study area although significant differences did occur among microsites. These results suggest that the presence of microtographic features can significantly alter the amount of soil moisture present in a microsite, alt topography across our study areas was too great for us to capture with such few replicates. Microsite soil moisture shoul d be directly correlated with our findings for wind speeds, PAR, and soil and air temperature, but they are not in this study. Problems associated with sensor placement could also account for some discrepancies ; soil packed around each sensor was extremely rocky, and air pockets along the sensor could have created anomalies in our results. Soil type, along with soil nutrients also influences the water holding capacity of soils. Although results were not significant, whitebark pine microsites contained highe r overall percentages of nitrogen and carbon. Higher percentages of nitrogen may lead to increased seedling growth, while higher carbon content in soils increases water holding capacity (Holtmeier, 2003). To further understand the role of the abiotic micr osite in seedling establishment at treeline, threshold clima tes for species survival are still widely unknown, although some information has been forthcoming in recent years. For example, Krner (1998) predicts that the global treeline temperature threshol d exists between 5.5 and 7.5 C. However, this is a broad generalization that could vary locally, and does not take individual species requirements into account. Quantifying stress at these elevations is extremely difficult because climate var ies greatly year to year and stress can occur both discretely and continuously.

PAGE 91

78 Without wide repetition in our study, it is hard to know how much the microsites we tested reflect the majority microsites of their types on the northeastern aspec t at each study area. D ifferences occurred in microclimate variables among blocks, illustrating how even slight differences in topography, slope steepness, or wind exposure can greatly affect microsite s within very short distances. At Divide Mountain, more ridges and gullies wer e present at our study area. More importantly, the mountain itself funneled prevailing westerly winds around its base from the north and from the south resulting in complex wind patterns Choosing microsites in blocks was one of the ways we tried to redu ce some of this natural variation. We also tried to use and slope aspect but this precluded choosing objects at random. Conclusions In the ATE at two study a reas on eastern front of the northern Rocky Mountains, initiation. Our data show that leeward conifer microsites moderate microclimate better than rock or exposed mi crosites. At Divide Mountain, where temperatures were less variable and less extreme but wind speeds were greater, protective leeward microsites reduced wind speeds; conifers in general decreased PAR better than rocks. At Line Creek, where temperatures rea ched greater extremes but wind speeds were lower, microsites leeward of conifers moderated temperature extremes, reduced wind speeds, and reduced PAR better than solitary rocks or unprotected microsites. Further, our data show the ecological importance of a species like whitebark pine which can withstand harsh climate, act as a pioneer, and establish well at treeline. The micro climate data presented here repres ent a continuous description of near ground level treeline climate at

PAGE 92

79 sites on the northeastern slopes at t wo Rocky Mountain Front study areas du ring much of the growing season. They also represent the climat ic conditions under which a seed must germinate and a seedling must withstand in order to establish, survive and in turn, act as a facilitator. Results from this study also provide comparative data on local climate for two very different treeline environments on the eastern front of the Rocky Mountains that have generally similar community composition but somewhat different structural composition and ecology (Smith McKenna et al. unpublished data) The data presented here demonstrate the differences that can occur in climate thresholds for seedlings and the different challenges faced by the same species at different locations. This project repr esents a n important step in understanding the mechanism behind tree island establishment in whitebark pine dominated treelines, but several factors warrant further investigation. Whitebark pine microsites showed both higher total nitrogen and carbon, but r esults were not significant and sample sizes were small. This, along with mycorrhizae relationships, is something that needs to be revisited as more studies show the significance of mycorrhizae on species survival Furthermore, Blakeslee (2012) found that whitebark pine had longer annual shoot growth than Engelmann spruce and subalpine fir. One explanation is that whitebark pine is more efficient in carbon allocation. Differences among whitebark pine and spruce and other treeline species in physiological ef ficiencies in water and nutrient uptake, and how this potentially affects their surviva l, should be further addressed.

PAGE 93

80 Figures Figure V.1.a. Summa ry of results for Divide Mountain MT. Model shows differences in climate n comparison to unprotected microsites. Figure V.1.b. Summary of results for Line Creek RNA, MT. Model shows differences in climate

PAGE 94

81 REFERENCES Alexander, R.R. and Shepperd, W.D. (199 0) Picea engelmannii Parry ex Engelm. In: Burns, R.M., and Honkala, B.H. (tech. coords). Silvics of North America: 1. Conifers; 2. Hardwoods. USDA Forest Service, Agricultural Handbook 654. US Government Printing Office, Washington, D.C., 187 203. Alexand er, R.R., Shearer, R.C. and Shepperd, W.D. (1990). Abies lasiocarpa (Hook.) Nutt. Silvics of North America: 1. Conifers; 2. Hardwoods. USDA Forest Service, Agricultural Handbook 654. US Governm ent Printing Office, Washington, D.C., 877. Arno, S.F. (1986). Whitebark pine stone crops a diminishing source of wildlife food? Western Journal of Applied Forestry 1 92 94. Arno, S.F. (2001). Community types and natural disturbance processes. In: Tomb ack, D.; Arno, S. F.; Keane, R. E. (editors). Whitebark pine communities: ecology and restoration Washington, DC: Island Press. 74 88. Arno, S.F. and Hammerly, R.P. (1984). Timberline: mountain and arctic forest frontiers Seattle, Washington: The Mounta ineers. 57 79, 200 207. Arno, S.F. and Hoff, R.J. (1989). Silvics of Whitebark Pine ( Pinus albicaulis ). General Technical Report INT 253. Ogden, UT: U.S. Department of Agriculture, Forest Service, Intermountain Research Station. 11. Arno, S.F. and Hoff, R.J. (1990). Pinus albicaulis Engelm. Whitebark Pine. In Burns, R.M. and Honkala, B.H. (tech. coords). Silvics of North America: 1. Conifers; 2. Hardwoods. USDA Forest Service, Agricultural Handbook 654. US Government Printing Office, Washington, D.C., 26 8 279. Baker, F.S. (1944). Mountain climates of the western United States. Ecological Monographs 14 225 254. Bansal, S., Reinhardt, K. and Germino, M.J. (2011). Linking carbon balance to establishment patterns: comparison of whitebark pine and Engelman n spruce seedlings along an herb cover exposure gradient at treeline. Plant Ecology 212 219 228. Batllori, E., Camarero, J.J., Ninot, J.M., and Gutirrez, E. (2009). Seedling recruitment, survival and facilitation in alpine Pinus uncinata tree line eco tones. Implications and potential responses to climate warming. Global Ecology and Biogeography 18 (4), 460 472. Baumeister, D. and Callaway, R.M. (2006). Facilitation by Pinus flexilis during succession: a hierarchy of mechanisms benefits other plant sp ecies. Ecology 87 (7), 1816 1830.

PAGE 95

82 Bekker, M.F. (2005). Positive feedback between tree establishment and patterns of subalpine forest advancement, Glacier National Park, Montana, USA. Arctic, Antarctic, and Alpine Research 37 (1), 97 107. Benedict, J.B. (1984). Rates of tree island migration, Colorado Rocky Mountains, USA. Ecology 65 (3). 820 823. Bevan, A. (1923). Summary of the geology of the Beartooth Mountains, Montana. Journal of Geology 31 441 465. Blakeslee, S.C. (2012). Assessing whitebark p ine vigor and facilitation roles in the alpine Breshears, D.D., Nyhan, J.W., Heil, C.E., and Wilcox, B.P. (1998). Effects of woody plants on microclimate in a semiarid woodlan d: soil temperature and evaporation in canopy and intercanopy patches. International Journal of Plant Sciences 159 (6), 1010 1017. Cairns, D.M., and Malanson, G. P. (1997). Environmental variables influencing the carbon balance at the alpine treeline: a modeling approach. Journal of Vegetation Science 9 (5), 679 692. Callaway, R.M. (1998). Competition and facilitation on elevation gradients in subalpine forest of the northern Rocky Mountains, USA. Oikos 82 561 573. Callaway, R.M., Brooker, R.W., Chol er, P., Kikvidze, Z., Lortie, C.J., Michalet, R., ... and Cook, B. J. (2002). Positive interactions among alpine plants increase with stress. Nature 417 (6891), 844 848. Callaway, R.M., Nadkarni, N.M., and Mahall, B.E. (1991). Facilitation and interferen ce of Quercus douglasii on understory productivity in central California. Ecology 72 (4), 1484 1499. Callaway, R.M., Sala, A. and Keane, R. E. (2000). Succession May Maintain High Leaf Area: Sapwood Ratios and Productivity in Old Subalpine Forests. Ecosys tems 3 (3), 254 268. Campbell, E.M. and Antos, J.A. (2000). Distribution and severity of white pine blister rust and mountain pine beetle on whitebark pine in British Columbia. Canadian Journal of Forest Research. 30 1051 1059. Chambers, J.C. (2001). P inus monophylla establishment in an expanding Pinus Juniperus woodland: Environmental conditions, facilitation and interacting factors. Journal of Vegetation Science 12 (1), 27 40. Choat, B., Jansen, S., Brodribb, T.J., Cochard, H., Delzon, S., Bhaskar, R., ... and Zanne, A. E. (2012). Global convergence in the vulnerability of forests to drought. Nature 491 752 755.

PAGE 96

83 Christersson, L., and Fircks, H.A.V. (1988). Injuries to conifer seedlings caused by simulated summer frost and winter desiccation. Silva Fennica 22 (3), 195 201. Cui, M., and Smith, W.K (1991). Photosynthesis, water relations and mortality in Abies lasiocarpa seedlings during natural establishment. Tree Physiology 8 (1), 37 46. Dullinger, S., Dirnbock, T. and Grabherr, G. (2004). Modeling climate change driven treeline shifts: relativ e effects of temperature increase, dispersal and invisibility. Journal of Ecology. 92 241 252. Ellison, A.M., Bank, M.S., Clinton, B.D., Colburn, E.A., Elliott, K., Ford, C.R., ... and Webster, J. R. (2005). Loss of foundation species: consequences for the structure and dynamics of forested ecosystems. Frontiers in Ecology and the Environment 3 (9), 479 486. Farnes, P.E. (1990). SNOTEL and snow course data: describing the hydrology of whitebark pine ecosystems. In: Proceedings Symposium on Whitebark P ine Ecosystems: Ecology and Management of a High mountain Resource. Compiled by Schmidt, W.C.; McDonald, K.J. General Technical Report INT 270. Ogden, Utah, USA: USDA Forest Service, Intermountain Research Station, pp. 302 304. Fins, L., Byler, J., Fergu son, D., Harvey, A., Mahalovich, M.F., McDonald, Geral, I. Miller, D., Schwandt, J. and Zack, A. (2001). Return of the giants: Restoring white pine ecosystems by breeding and aggressive planting of blister rust resistant white pines. Station Bulletin 72. M oscow, ID: University of Idaho, Wildlife and Range Experiment Station. 21 Geiger, R. (1950). The Climate Near The Ground Harvard University Press, Cambridge, MA. Geils, B.W., Hummer, K.E., and Hunt, R.S. (2010). White pines, Ribes and blister rust: a review and synthesis. Forest Pathology. 40 147 185. Germino, M.J. and Smith W.K. (1999). Sky exposure, crown architecture, and low temperature photoinhibition in conifer seedlings at alpine treeline. Plant, Cell, & Environment. 22 407 415. Germino, M.J., Smith, M.K. and Resor, A.C. (2002). Conifer seedling distribution and survival in an alpine treeline ecotone. Plant Ecology 162 157 168. Gibson, K. E., Skov, K., Kegley, S., Jorgensen, C., Smith, S. and Witcosky, J. (2008). Mountain pine beetle i mpacts in high elevation five needle pines: Current trends and challenges. Rep. R1 08 020.Missoula, MT: U.S. Department of Agriculture, Forest Service, Northern Region, Forest Health Protection. 32 Hadley, J.L., and Smith, W.K. (1986). Wind effects on ne edles of timberline conifers: seasonal influence on mortality. Ecology 67 (1), 12 19.

PAGE 97

84 Hamann, A. and Wang, T. (2006). Potential effects of climate change on ecosystem and tree species distribution in British Columbia. Ecology 87 2773 2786. Httenschwi ler, S., and Smith, W.K. (1999). Seedling occurrence in alpine treeline conifers: a case study from the Central Rocky Mountains, USA. Acta Oecologica 20 (3), 219 224. Hoch, G., Popp, M., and Krner, C. (2002). Altitudinal increase of mobile carbon pools i n Pinus cembra suggests sink limitation of growth at the Swiss treeline. Oikos 98 (3), 361 374. Hoff, R.J., Ferguson, D.E., McDonald, G.I. and Keane, R.E. (2001). In: Tomback, D., Arno, S. F., Keane, R. E. (editors). Whitebark pine communities: ecology an d restoration. Washington, DC: Island Press. 346 366. Holtmeier, F.K. (2003). Mountain timberlines: ecology, patchiness, and dynamics. Mountain timberlines: ecology, patchiness, and dynamics 57 58 Holtmeier, F.K., and Broll, G. (1992). The influence of tree islands and microtopography on pedoecological conditions in the forest alpine tundra ecotone on Niwot Ridge, Colorado Front Range, USA. Arctic and Alpine Research 24 (3), 216 228. Holtmeier, F.K., and Broll, G. (2010). Wind as an ecological agent at treelines in North America, the Alps, and the European Subarctic. Physical Geography 31 (3), 203 233. and establishment of whitebark pine. Oecologia 55 192 201. Jewett, J.T., Lawrence, L.A., Marshall, L.A., Gessler, P.E., Powell, S.L. and Savage, S.L. (2011). Spatiotemporal relationships between climate and whitebark pine mortality in the greater Yellowstone ecosystem. Forest Science. 75 (4), 320 334. Ka lberer, S.R., Wisniewski, M., and Arora, R. (2006). Deacclimation and reacclimation of cold hardy plants: current understanding and emerging concepts. Plant Science 171 (1), 3 16. Kendall, K. and Keane, R.E. (2001). The decline of whitebark pine. In: Tom back, D., Arno, S. F., Keane, R. E. (editors). Whitebark pine communities: ecology and restoration. Washington, DC: Island Press. 123 145. Krner, C. (1998). A re assessment of high elevation treeline positions and their explanation. Oecologia 115 445 4 59. Krner, C. (2003). Carbon limitation in trees. Journal of Ecology 91 (1), 4 17. Kharuk, V.I., Dvinskaya, M.L., Im, S.T., and Ranson, K.J. (2008). Tree vegetation of the forest tundra ecotone in the western Sayan mountains and climatic trends. Russ ian Journal of Ecology 39 (1), 8 13.

PAGE 98

85 Landhusser, S. M., Wein, R.W., and Lange, P. (1996). Gas exchange and growth of three arctic tree line tree species under different soil temperature and drought preconditioning regimes. Canadian Journal of Botany 74 (5), 686 693. Lanner, R.M. (1990). Biology, taxonomy, evolution, and geography of stone pines of the world. Pages 14 23 in W. C. Schmidt and K. J. McDonald, compliers. Whitebark pine ecosystems Ecology and management of a high mountain resource. USDA Fore st Service, Intermountain Research Station, General Technical Report INT 270, Ogden, Utah. Larson, E.R. (2011). Influences of the biophysical environment on blister rust and mountain pine beetle, and their interactions, in whitebark pine forests. Journal of Biogeography 38 (3), 453 470. Lesica, P. (2002). A flora of Glacier National Park, Montana Oregon State University Press. Lorenz, T.J., Aubry, C. and Shoal, R. (2008). A review of the literature on seed fate in whitebark pine and the life history t General Technical Report PNW GTR 742. in the Cascade Range. The Condor. 111 (2), 326 340. M aher, E.L., and Germino, M.J. (2006). Microsite differentiation among conifer species during seedling establishment at alpine treeline. Ecoscience 13 (3), 334 341. Malanson, G.P., Butler, D.R., Fagre, D.l B., Walsh, S. J., Tomback, D.F., Daniels, L.D., R esler, L.M., Smith, W.K., Weiss, D.J., Peterson, D.L., Bunn, A.G., Hiemstra, C.A., Liptzin, D., Bourgeron, P.S., Shen, Z. and Millar, C.I. (2007). Alpine treeline of western north America: linking organism to landscape dynamics. Physical Geography 28 (5), 378 396. Marchand, P.J. and Chabot, B.F. (1978). Winter water relations of tree line plant species on Mt. Washington, New Hampshire. Arctic and Alpine Research 10 (1), 105 116. Marr, J.W. (1977). The Development and Movement of Tree Islands Near the Upp er Limit of Tree Growth in the Southern Rocky Mountains. Ecology 1159 1164. Mattson, D.J., Blanchard, B.M. and Knight, R.R. (1991). Food habits of Yellowstone grizzly bears, 1977 1987. Canadian Journal of Zoology 69 1619 1629. Mayr, S., Gruber, A., an d Bauer, H. (2003). Repeated freeze thaw cycles induce embolism in drought stressed conifers (Norway spruce, stone pine). Planta 217 (3), 436 441. Mayr, S., Wolfschwenger, M., and Bauer, H. (2002). Winter drought induced embolism in Norway spruce ( Picea abies ) at the Alpine timberline. Physiologia Plantarum 115 (1), 74 80.

PAGE 99

86 McCaughey, W.W. and Schmidt, W.C. (2001). Taxonomy, Distribution, and History In: Tomback, D.; Arno, S. F.; Keane, R. E. (editors). Wh itebark pine communities: ecology and restoration. Washington, DC: Island Press. 30 40. McDonald, G.I. and Hoff, R.J. (2001). Blister rust: an introduced plague. In: Tomback, D., Arno, S. F., Keane, R. E. (editors). Whitebark pine communities: ecology an d restoration. Washington, DC: Island Press. 193 220. McKenney, D.W., Pedlar, J.H., Lawrence, K., Campbell, K. and Hutchinson, M.F. (2007). Potential impacts of climate change on the distribution of North American trees. BioScience 57 (11), 939 948. Mc Kinney, S.T. and Fiedler, C.E. (2010). Tree Squirrel habitat selection and predispersal seed predation in a declining subalpine conifer. Oecologia 2 (3), 697 707. McKinney, S.T., Fiedler, C.E. and Tomback, D.F. (2009). Invasive pathogen threatens bird pi ne mutualism: implications for sustaining a high elevation ecosystem. Ecological Applications 19 (3), 597 607. Miegroet, H.V., Hysell, M.T. and Johnson, A.D. (2000). Soil microclimate and chemistry of spruce fir tree islands in northern Utah. Soil Scienc e Society of America Journal 64 1515 1525. Minore, D. (1979). Comparative autecological characteristics of northwestern tree species A literature review. USDA Forest Service, Pacific Northwest Forest and Range Experiment Station General Technical Repo rt PNW 87, Portland, Oregon. Montana. (2013). Google Maps. Google. Retrieved from https://maps.google.com/maps?q=Montana&hl=en&sll=38.997934, 105.550567&sspn=4.54166,9.876709&hnear=Montana&t=m&z=6 Moyes, A.B., Castanha, C., Germino, M.J. and Kueppers, L .M. (2013). Warming and the dependence of limber pine ( Pinus flexilis ) establishment on summer soil moisture within and above its current elevation range. Oecologia 171 271 282. Nimlos, T.J., McConnel, R.C. and Pattie, D.L. (1965). Soil temperature and moisture regimes in Montana alpine soils. Northwest Science 39 (4), 129 137. Olgivie, R.T. (1990). Distribution and ecology of whitebark pine in western Canada. Pp. 60 in Proceedings Symposium on whitebark pine ecosystems: ecology and management of a h igh mountain resource. Schmidt, W.C., and K. J. McDonald, Compilers. General Technical Report, INT 270, USDA Forest Service, Intermountain Research Station. Ogden, UT.

PAGE 100

87 R Development Core Team (2011). R: A language and environment for statistical computi ng. R Foundation for Statistical Computing,Vienna, Austria. ISBN 3 900051 07 0, URL http://www.R project.org/ Resler, L.M. (2004). Conifer establishment sites on a periglacial landscape, Glacier National Park, Montana. ETD Collection for Texas State Unive rsity Paper AAI3143407. Resler, L.M ., Butler, D.R. and Malanson, G.M. (2005). Topographic shelter and conifer establishment and mortality in an alpine environment, Glacier National Park, Montana. Physical Geography 26 (2), 112 125. Resler, L.M. and Tom back, D.F. (2008). Blister rust prevalence in krummholtz whitebark pine: implications for treeline dynamics, northern Rocky Mountains, Montana, U.S.A. Arctic, Antarctic, and Alpine Research 40 (1), 161 170. Sakai, A. and Okada, S. (1971). Freezing resist ance of conifers. Silvae Genetics 20 91 97. Schoettle, A.W., Sniezko, R.A. and Burns, K.S. (2008). Sustaining Pinus flexilis ecosystems of the southern Rocky Mountains (USA) in the presence of Cronartium ribicola and Dendroctonus ponderosae in a changin g climate. Schrag, A.M., Bunn, A.G. and Graumlich, L.J. (2008). Influence of bioclimatic variables on tree line conifer distribution in the Greater Yellowstone Ecosystem: implications for species of conservation concern. Journal of Biogeography 35 (4), 698 710. Smith, W.K., Germino, M.J., Hancock, T.E. and Johnson, D.M. (2003). Another perspective on altitudinal limits of alpine timberlines. Tree Physiology 23 1101 1112. Smith, W.K., Germino, M.J., Johnson, D.M. and Reinhardt, K. (2009). The altitud e of alpine treeline: a bellwether of climate change effects. Botany Review. 75 163 190. Smith, E.K., Resler, L.M., Vance, E.A., Carstensen, L.W. and Kolivras, K.N. (2011). Blister rust incidence in treeline whitebark pine, Glacier National Park, USA: en vironmental and topographic influences. Arctic, Antarctic, and Alpine Research 43 (1), 107 117. Smith, C.M., Wilson, B., Rasheed, S., Walker, R.C., Carolin, T. and Shepherd, B. (2008). Whitebark pine and white pine blister rust in the Rocky Mountains of Canada and northern Montana. Canadian Journal of Forest Research 38 (5), 982 995. Sperry, J.S., and Sullivan, J.E. (1992). Xylem embolism in response to freeze thaw cycles and water stress in ring porous, diffuse porous, and conifer species. Plant Physio logy 100 (2), 605 613. Stachowicz, J.J. (2001). Mutualism, Facilitation, and the Structure of Ecological Communities: Positive interactions play a critical, but underappreciated, role in ecological communities by

PAGE 101

88 reducing physical or biotic stresses in e xisting habitats and by creating new habitats on which many species depend. Bioscience 51 (3), 235 246. Sthultz, C.M., Gehring, C.A. and Whitham, T.G. (2007). Shifts from competition to facilitation between a foundation tree and a pioneer shrub across sp atial and temporal scales in a semiarid woodland. New Phytologist 173 (1), 135 145. Sveinbjrnsson, B., Kauhanen, H. and Nordell, O. (1996). Treeline ecology of mountain birch in the Tornetrsk area. Ecological bulletins 65 70. Tillman Sutela, E., Kaup pi, A., Karppinen, K. and Tomback, D. F. (2008). Variant maturity in seed structures of Pinus albicaulis (Engelm.) and Pinus sibirica (Du Tour): key to a soil seed bank, unusual among conifers? Trees 22 (2). 225 236. Tomback, D.F. (1978). Foraging strate Living Bird. 16 123 161. hypothesis. Journal of Animal Ecology. 51 451 467. Nucifraga colum biana ). In: The Birds of North America, Inc., No. 331. Ed. By Poole, A.; Gill, F. Philadephia, PA: The Birds if North America, Inc. 24 104. In: Tomback, D.F., Arno, S.F., Keane, R. E. editors. Whitebark Pine Communities: Ecology and Restoration. Island Press, Washington, D.C., U.S.A. Tomback, D.F. (2007). Whitebark pine: Ecological importance and future outlook. Proceedings Whitebark Pine: A Pacific Coast Perspective. USDA Forest S ervice, Region 6. In press 6 19. Tomback, D.F., Arno, S.F. and Keane, R.E. (2001). The Compelling Case for Management Intervention. Pages 3 11. In: Tomback, D.F., Arno, S.F., Keane, R.E. (editors). Whitebark Pine Communities: Ecology and Restoration. Is land Press, Washington, D.C., U.S.A. Tomback, D.F. and Achuff, P. (2010). Blister rust and western forest biodiversity: ecology, values, and outlook for white pines. Forest Pathology. 40 186 225. Tomback, D.F. and Kendall, K.C. (2001). Biodiversity loss es: the downward spiral. Pages 243 262. In: Tomback, D.F., Arno, S.F., Keane, R.E. editors. Whitebark Pine Communities: Ecology and Restoration. Island Press, Washington, D.C., U.S.A. Tomback, D.F. and Linhart, Y.B. (1990). The evolution of bird disperse d pines. Evolutionary Ecology 4 185 219.

PAGE 102

89 Tomback, D.F. and Resler, L.M. (2007). Invasive pathogens and alpine treeline: consequences for treeline dynamics. Physical Geography 28 (5), 397 418. U.S. Fish and Wildlife Service. (2011). Whitebark Pine to be Designated a Candidate for Endangered Species Protection. Visited April 12, 2013. Retrieved from: http://www.fws.gov/mountain prairie/species/plants/whitebarkpine/PressRelease07182011.pdf Vander Wall, S.B. (1982). An experimental analysis of cache rec overy in Clark's nutcracker. Animal Behavior 30 (1), 84 94. Warwell, M.V., Rehfeldt, G.E. and Crooskston, N. (2007). Modeling contemporary climate profiles of whitebark pine ( Pinus albicaulis ) and predicting responses to global warming. Proceedings White bark Pine: A Pacific Coast Perspective, USDA Forest Service, Region 6 in press. 139 142. Weaver, T. (1978). Change in soils along a vegetation altitudinal gradient of the northern Rocky Mountains. Pages 14 29 in C. T. Youngberg, editor. Forest soils and land use. Soil Science Society America and Forest and Wood Science Department, Colorado State University, Fort Collins. Weaver, T. (1990). Climates of subalpine pine woodlands. Pages 72 79 in W. C. Schmidt and K. J. McDonald, compilers. Whitebark pine e cosystems Ecology and management of a high mountain resource. USDA Forest Service, Intermountain Research Station, General Technical Report INT 270, Ogden, Utah. Weaver, T. (1994). Climates where stone pines grow, a comparison. Pages 85 89 in W.C. Schmidt and F.K. Holtmeier, compilers. Proceedings International workshop on subalpine stone pines and their environment: The status of our knowledge. USDA Forest Service, Intermountain Research Station, General Technical Report INT GTR 309, Ogden, Utah. Weaver, T. (2001). Whitebark pine and its environment. In: Tomback, D., Arno, S. F., Keane, R. E. (editors). Whitebark pine communities: ecology and restoration. Washington, DC: Island Press: 41 74. Weaver, T. and Dale, D. (1974). Pinus albicaulis in central Mon tana, environment, vegetation and production. American Midland Naturalist 92 222 230 Weih, M. and Karlsson, P.S. (1999). The nitrogen economy of mountain birch seedlings: implications for winter survival. Journal of Ecology 87 (2), 211 219.

PAGE 103

90 APPENDI X Air Temperature Results Figure A1.a Boxplots of residual data from 2010 daily maximum air temperatures (C) are shown fo r Divide Mountain, a) northeast slope b) west slope, and Line Creek c) northeast slope and d) west slope.

PAGE 104

91 Figure A1.b. Boxplots of residual data from 2011 daily maximum air temperatures are shown for a) Divide Mountain and b) Line Creek Interaction effect of block on microsite were not significant for each study area so respective blocks were treated as repl icates ( Divide F = 7.99, df = 6, p = 0.1962 and Line Creek, F = 1.88, df = 3, p = 0.1318).

PAGE 105

92 Figure A1.c. Boxplots of residual data from 2012 daily maximum air temperatures are shown for Divide Mountain a) Block 2, b) Block 2, c) Block 3 and Line Creek d) Block 1 and e) Block 2

PAGE 106

93 Table A1 M ax imum air temperature results from a Kruskal Wallis test, comparing the residuals from a model of best fit (either linear or quadratic) for daily air temperature m ax imums amon g microsites, ( = 0.05). Year Study Area Block N df P value 2 2010 Divide Mtn NE 1 3 0.2302 4.306 W 1 3 0.206 4.571 Line Creek NE 1 3 1.10 e 07 35.207 W 1 3 0.2803 3.8314 2011 Divide Mtn 1 3 3 3 0.5632 2.044 Line Creek 1, 2 2 3 < 2.2 e 16 113.560 2012 Divide Mtn 1 1 3 0.4746 2.504 2 1 3 0.1922 4.73 5 3 1 3 0.3151 3.544 Line Creek 1 1 3 <2.2e 16* 110.234 2 1 3 2.253e 05* 24.214

PAGE 107

94 Figure A2.a Boxplots of residual data from 2010 daily m in imum air temperatures (C) are shown fo r Divide Mountain, a) northeast slope, b) west sl ope, and Line Creek c) northeast slope and d) west slope.

PAGE 108

95 Figure A2. b Boxplots of residual data from daily minimum air temperatures are shown for 2011 a) Divide Mountain and b) Line Creek and 2012 c) Divide Mountain and e) Line Creek Inte raction effect of block on microsite were not significant for each study area so respective blocks were treated as replicates ( 2011: Divide F = 1.04, df = 6, p = 0.3976 and Line Creek, F = 0.371, df = 3, p = 0.7743 ; 2012: Divide F = 0.355, df = 6, p = 0. 9071 and Line Creek, F = 0.167, df = 3, p = 0.9187 ).

PAGE 109

96 Table A2. M in imum air temperature results from a Kruskal Wallis test, comparing the residuals from a model of best fit (either linear or quadratic) for daily air temperature m in imums amo ng microsites, ( = 0.05). Year Study Area Block N df P value 2 2010 Divide Mtn NE 1 3 0.9406 0.398 W 1 3 0.4331 2.7419 Line Creek NE 1 3 9.44e 01 0.3829 W 1 3 0.001373 15.5938 2011 Divide Mtn 1 3 3 3 0.9151 0.517 Line Creek 1, 2 2 3 0.08126 6.7234 2012 Divide Mtn 1 3 3 3 0.1211 5.8121 Line Creek 1, 2 2 3 0.1786 4.9082

PAGE 110

97 Soil Moisture Results Figure A3.a. Boxplots of daily average soil moisture (m 3 m 3 ) for 2010 are shown for Divide Mountain, a) northeast slope, b) west slope, and Line Creek c) northeast slope and d) west slope.

PAGE 111

98 Figure A3.b Boxplots of daily average soil moisture (m 3 m 3 ) in 2011 are shown for Divide Mountain a) Block 1, b) Block 2 and c) Block 3 and Line Creek d) Block 1 and e) Bloc k 2.

PAGE 112

99 Figure A3.c Boxplots of daily average soil moisture (m 3 m 3 ) in 2012 are shown for Divide Mountain a) Block 1, b) Block 2 and c) Block 3 and Line Creek d) Block 1 and e) Block 2.

PAGE 113

100 Table A3 Daily average soil moistu re (m 3 m 3 ) results from a Kruskal Wallis test, ( = 0.05). Year Study Area Block N df P value 2 2010 Divide Mtn NE 1 3 < 2.2e 16 172.0353 W 1 3 < 2.2e 16 147.7006 Line Creek NE 1 3 0.1308 5.6339 W 1 3 4.245e 10 46.5917 2011 Divide Mtn 1 1 3 0.1178 5.8768 2 1 3 3.289e 10 47.1128 3 1 3 < 2.2e 16* 82.0322 Line Creek 1 1 3 < 2. 2e 16* 115.4292 2 1 3 0.000214 19.5138 2012 Divide Mtn 1 1 3 < 2.2e 16* 165.508 2 1 3 3.289e 10* 47.1128 3 1 3 < 2.2e 16* 134.1374 Line Creek 1 1 3 < 2.2e 16* 119.836 2 1 3 < 2.2e 16* 171.3403

PAGE 114

101 Table A4 AIC scores fitting models to soil temperature data ( *Delineates lowest score and fit model chosen. ) Variable Year Study Area Block # AIC (Quadratic) AIC (Linear) Maximum 2010 Divide Mtn NE 1240.1* 1291.5 W 1479.3* 1521.8 Line Creek NE 1384.5 1382.7* W 1415.6 1413.2* 2011 Divide Mtn 1 1298.7 1332.5 2 1107.3 1161.6 3 1170.6 1241.3 Line Creek 1 1251.9 1256.8 2 1208.3 1220.9 2012 Divide Mtn 1 1469.4 1527.2 2 1518.0 1557.9 3 1436.2 1457.0 Line Creek 1 1457.7 1464.7 2 1482.7 1489.2 Minimum 2010 Divide Mtn NE 1046.8 1137.2 W 1143. 4* 124 5.0 Line Creek NE 948. 4* 951.1 W 1023.2 1024.5 2011 Divide Mtn 1 1118.9* 1140.2 2 991.1* 1027.1 3 970.6* 1010.6 Line Creek 1 963.0* 964.9 2 951.8* 956.3 2012 Divide Mtn 1 1257.9 1369.5 2 1281.9 1373.9 3 1112.7 1172.2 Line Creek 1 1133.7 1155.7 2 1131.7 1154.8 Variance 2010 Divide Mtn NE 1070.5* 1078.0 W 1348.2* 1366.6 Line Creek NE 1428.8 1426.0* W 1483.4 1482.6* 2011 Divide Mtn 1 1292.6* 1296.3 2 856.3* 874.2 3 1103.1* 1129.7 Line Creek 1 1353.6 1352.1* 2 1299.3* 1302.4 2012 Divide Mtn 1 1515.6* 1522.7 2 1775.5* 1776.5 3 1731.3 1730.2* Line Creek 1 2222.3 2220.0* 2 2078.2 2076.4*

PAGE 115

102 Table A5 AIC score s fitting models to air temperature data (*Delineates lowest score and fit model chosen.) Variable Year Study Area Block # AIC (Quadratic) AIC (Linear) Maximum 2010 Divide Mtn NE 1557.55* 1588.94 W 1613.75* 1655.13 Line Creek NE 1316.20* 1327.38 W 1310.33 1307.15* 2011 Divide Mtn 1 1500.62* 1501.21 2 1515.09 1513.71* 3 1398.12* 1410.18 Line Creek 1 1165.04* 1187.33 2 1228.74* 1248.45 2012 Divide Mtn 1 1744.38 1759.19 2 1749.79* 1762.13 3 1284.55* 1319.33 Line Creek 1 1457.70* 1464.71 2 1363.46* 1371.63 Minimum 2010 Divide Mtn NE 1313.59* 1340.41 W 1306.45* 1353.93 Line Creek NE 1076.55* 1096.53 W 1083.92* 1089.43 2011 Divide Mtn 1 1326.51* 1328.85 2 1339.86* 1340.57 3 1240.30* 1241.23 Li ne Creek 1 1124.50 1122.59* 2 1131.63 1129.82* 2012 Divide Mtn 1 1517.26* 1530.15 2 1526.22* 1542.03 3 1103.03 1100.23* Line Creek 1 1273.49* 1280.68 2 1275.02* 1282.58 Table A6 AIC scores fitting models to sum PAR data (*Del ineates lowest score and fit model chosen.) Variable Year Study Area Block # AIC (Quadratic) AIC (Linear) Sum 2010 Divide Mtn NE 5709.0 5706.9* W 5822.4* 5834.0 Line Creek NE 4900.7 4897.9* W 4842.9 4841.4*

PAGE 116

103 Post Hoc Results Table A 7 Maximum s oil t emperature (C) post hoc results from a pairwise Wilcoxon Rank Sum Test Year Study Area Block N ES RK ES UN ES WB RK UN RK WB UN WB 2010 Divide Mtn NE 1 0.0001 0.00024 0.00597 1 1 1 W 1 2.70 e 09 0.00089 1 0.00079 4.10e 09 0.00087 Line Creek NE 1 4.40 e 11 2.80 e 11 1.40 e 09 1 < 2 e 16 < 2 e 16 W 1 1 4.30 e 13 0.041 1.00 e 12 2.12 e 01 1.10 e 10 2011 Divide Mtn 1 1 1.8e 12 0.22 3.0e 10* 8.5e 10* 1 2.7e 07* 2 1 1 0.0012* 0.6849 0.0057* 1 0.0470* 3 1 0.54613 0.00093* 6.9e 09* 0.08791 1.3e 06* 0.00158* Line Creek 1 1 9.1e 16* <2e 16* 1.6e 10* 3.6e 07* 3.8e 07* 1.9e 13* 2 1 2.0e 13* 4.4e 13* 2.2e 12* 1 1 1 2012 Divide Mtn 1 1 1.4e 07 1.2e 05* 7.7e 13* 0.04* 0.27 2.9e 06* 2 1 5.3e 12* <2e 16* 0.71 7.1e 10* 5.7e 10* <2e 16* 3 1 8.7e 08* <2e 16* 7.2e 05* 1.4e 13* 0.098 2.0e 15* Line Creek 1 1 5.2e 14* 3.7e 16* 1.7e 06* 0.36 1.9e 06* 5.3e 09* 2 1 6.1e 06* 1 2.3e 14* 1.8e 07* <2e 16* 5.8e 14* Table A 8 Minimum soil temperature (C) post hoc results from a pairwise Wilcoxon Rank Sum Test Year Study Area Block N ES RK ES UN ES WB RK UN RK WB UN WB 2010 Divide Mtn NE 1 1 0.96 1 1 1 1 W 1 0.0073 0.0198 0.01 1 1 1 Line Creek NE 1 0.00022 0.00518 0.44694 1 1.00 e 08 6.60 e 07 W 1 1.70 e 05 3.40 e 09 1 0.11 8.10e 05 9.70e 09 2011 Divide Mtn 1 1 1.1e 10* 0.00042* 1 3.5e 05* 1.9e 08* 0.0058* 2 1 0.24 1 0.22 0.32 1 0.29 3 1 0.0003* 0.07282 1 0.54922 0.00011* 0.02431* Line Creek 1 1 0.49 1.7e 08* 1 4.3e 05* 1 2.8e 07* 2 1 0.0039* 0.0167* 1 1 0.0085* 0.0412* 2012 Divide Mtn 1 1 0.00028* 0.20658 1 0.23237 0.02022* 1 2 1 1 0.0085* 0.0025* 0.2146 0.4146 1 3 1 0.063 0.264 0.905 1 1 1 Line Creek 1 1 0.0484* 1.7e 07* 0.8757 0.0067* 0.8280 9.0e 06* 2 1 0.03014* 0.95108 0.0306* 1 1.9e 06* 0.00095*

PAGE 117

104 Table A 9 Variance in soil temperature (C) post hoc results from a pairwise Wilcoxon Rank Sum Test Year Study Area Block N ES RK ES UN ES WB RK UN RK WB UN WB 2010 Divide Mtn NE 1 1.1e 09* 8.9e 10* 7.5e 07* 1 1 0.86 W 1 3.7e 08* 0.059 0.03 8* 0.001* 7.7e 11* 4.8e 06* Line Creek NE 1 5.7e 15* 8.3e 15* 1.4e 15* 1 <2e 16* <2e 16* W 1 8.1e 05* <2e 16* 0.0058* 9.4e 14* 1 7.9e 15* 2011 Divide Mtn 1 1 <2e 16* 4.0e 11* <2e 16* <2e 16* 4.5e 12* 4.4e 05* 2 1 0.00026* 4.9e 07* 5.1e 06* 0.15532 1 0.98435 3 1 0.043* 0.065 2.6e 12* 8.2e 07* 1.2e 14* 1.0e 08* Line Creek 1 1 3.5e 15* <2e 16* 1.8e 10* 5.3e 10* 8.1e 06* 6.9e 15* 2 1 2 .1e 12* 3.3e 12* 1.9e 10* 1 0.05* 0.13 2012 Divide Mtn 1 1 2.7e 16* 9.3e 12* <2e 16* 0.00099* 0.57291 1.7e 07 2 1 <2e 16* <2e 16* 1.8e 06* 1.7e 13* 1.2e 10* <2e 16* 3 1 0.00087* <2e 16* 0.12862 4.3e 16* 0.64297 <2e 16* Line Creek 1 1 1.5e 11* <2e 16* 1.1e 05* 1e 04* 2.2e 05* 8.7e 12* 2 1 1.5e 06* 0.221 1.5e 12* 0.039* <2e 16* 1.1e 15* Table A 1 0 M aximum air temperature (C) post hoc results from a pairwise Wilcoxon Rank Sum Test Year Study Area Block N ES RK ES UN ES WB RK UN RK WB UN WB 2010 Divide Mtn NE 1 1 0.28 1 1 1 0.95 W 1 0.53 1 1 0.35 1 1 Line Creek NE 1 1.20 e 05 1 0.056 3.60 e 06 0.041 0.021 W 1 1 1 0.36 1 1 1 2011 Divide Mtn 1 3 3 1 1 1 1 1 1 Line Creek 1,2 2 < 2 e 16 2.10 e 11 0.0001 9.20 e 05 1.00 e 10 0.0071 2012 Divide Mtn 1 1 1 0.95 1 1 1 1 2 1 1 1 1 1 0.43 0.53 3 1 1 1 0.80 1 1 0.45 Line Creek 1 1 5.2e 14 3.7e 16* 1.7e 06* 0.36 1.9e 06* 5.3e 09* 2 1 0.0223* 1 0.3489 0.2286 5.2e 05* 0.0056*

PAGE 118

105 Table A 11 Minimum air temperature (C) post hoc results from a pairwise Wilcoxon Rank Sum Test Year Study Area Block N ES RK ES UN ES WB RK UN RK WB UN WB 2 010 Divide Mtn NE 1 1 1 1 1 1 1 W 1 1 0.87 1 1 1 1 Line Creek NE 1 1 1 1 1 1 1 W 1 0.1669 0.1408 1 1 0.0094 0.0096 2011 Divide Mtn 1 3 3 1 1 1 1 1 1 Line Creek 1, 2 2 0.099 0.37 1 1 1 1 2012 Divide Mtn 1 3 3 1 0.14 1 0.48 1 0.51 Line Creek 1, 2 2 0.23 1 1 1 1 1 Table A 12 Sum PAR ( mol/m 2 /sec ) post hoc results from a pairwise Wilcoxon Rank Sum Test Year Study Area Block N ES RK ES UN ES WB RK UN RK WB UN WB 2010 Divide Mtn NE 1 2.10 e 10 4.30 e 12 0.0031 0.1489 0.005 2.20 e 05 W 1 0.0462 0.0096 0.1138 1 1 1 Line Creek NE 1 0.024 4.60 e 05 2.30 e 07 0.068 2.80 e 11 3.40 e 14 W 1 5.00 e 04 1.30 e 12 1.80 e 14 2.00 e 07 < 2 e 16 < 2 e 16

PAGE 119

106 Table A 13 Maximum gust speed (m/s) post hoc results from a pairwi se Wilcoxon Rank Sum Test Year Study Area Block N ES RK ES UN ES WB RK UN RK WB UN WB 2011 Divide Mtn 1 1 0.9095 0.2110 0.1872 0.0026* 0.0023* 1 2 1 1 0.55 1 0.17 1 0.51 3 1 1 0.0238* 1 0.2805 0.7046 0.0021* Line Creek 1 1 0.035* 0.029* 0.287 1 1 1 2 1 0.00012 1.7e 07* 0.21592 1 0.01149* 2.6e 05* 2012 Divide Mtn 1 1 0.0017 0.1493 1 1.7e 07* 0.0038* 0.0239* 2 1 1 0.0520 1 0.0649 1 0.0081* 3 1 0.0696 0.9017 1 1 0.0017* 0.0186* Line Creek 1 1 1 0.67874 0.00026* 1 0.00479* 0.02010* 2 1 8.7e 08* 7.5e 07* 1.6e 08* 0.70 1 0.31 Table A 1 4 Average wind speed (m/s) post hoc results from a pairwise Wilcoxon Rank Sum Test Year Study Area Block N ES RK ES UN ES WB RK UN RK WB UN WB 2011 Divide Mtn 1 1 3.7e 05* 2.6e 09* 0.00505* <2e 16* 1 .3e 10* 0.00052* 2 1 1 1.7e 05* 0.094 2.4e 06* 0.464 5.1e 09* 3 1 1 3.5e 08* 0.44 1.2e 07* 0.13 2.9e 12* Line Creek 1 1 2.6e 05* 8.1e 11* 0.817 0.014* 0.018* 4.3e 07* 2 1 8.1e 09* 1.5e 14* 1 6.0e 05* 1.3e 11* <2e 16* 2012 Divide Mtn 1 1 2.3e 08 5.5e 15 0.069 <2e 16 9.9e 15 9.5e 12 2 1 1.000 1.0e 09 0.053 1.4e 12 0.042 <2e 16 3 1 0.66 0.61 3.9e 05* 1.00 9.6e 09* 4.3e 10* Line Creek 1 1 0.00022 2.2e 06 2.3e 07 0.58563 0.19194 1.00000 2 1 1.6e 09 2.0e 13 1.7e 08 0.00045 1 .00000 0.00229

PAGE 120

107 Table A15. Results from post hoc comparisons of daily average soil moisture (m 3 m 3 ) among microsites using a pairwise wilcoxon rank sum test. Year Study Area Block N ES RK ES UN ES WB RK UN RK WB UN WB 2010 Divide Mtn NE 1 < 2e 16 1.3e 15 < 2e 16 1.3e 08 0.019 1.2e 12 W 1 1 3.4e 06 < 2e 16 1.4e 06 < 2e 16 < 2e 16 Line Creek NE 1 0.53 1 1 0.28 1 0.58 W 1 1 2.2e 06 0.0015* 1.5e 06 4.9e 05 0.0143 2011 Divide Mtn 1 1 1 0.20 1 0.66 1 0.56 2 1 1.4e 07* 2.0e 05* 1 1 5.6e 06* 0.00051* 3 1 1.4e 11* 3.6e 13* 3.5e 05* 0.0034* 0.2482 4.8e 06* Line Creek 1 1 1.2e 13* <2e 16* <2e 16* 0.81 0.35 1 2 1 0.002* 0.26996 0.00031* 1 1 0.3311 2012 Divide Mtn 1 1 <2e 16* 0.5295 0.0042* <2e 1 6* <2e 16* 1 2 1 1.4e 07* 2.0e 05* 1 1 5.6e 06* 0.00051* 3 1 <2e 16* 0.073 1 5.8e 16* <2e 16* 0.078 Line Creek 1 1 <2e 16* <2e 16* 1 1 1e 09* 1e 09* 2 1 6.6e 09* 1.2e 15* 1.3e 13* <2e 16* 4.5e 07* <2e 16* Table A 16 Post hoc results from of nitrogen and carbon among microsites. Nutrient Study Area N ES RK ES UN ES WB RK UN RK WB UN WB % N Divide Mtn 10 0.678 0.963 0.616 0.917 1 0.879 Line Creek 10 0.607 0. 968 0.814 0.855 0.174 0.547 % C Divide Mtn 10 0.986 0.877 0.930 0.977 0.993 0.999 Line Creek 10 0.492 0.948 0.550 0.805 0.047* 0.258

PAGE 121

108 R Code Statistical Analysis Daily Maximum, Minimum and Variance Soil Temperature (Repeated for Air Temperature Maximum and Minimum) ### 2010, 2011, and 2012 Maximum/ Minimum Soil Temperature### library(gdata) library(gtools) ##Import data (Maximum > (Sheet = 4); Minimum > (Sheet = 3)) d.1 < read.xls(file.choose(),sheet=3 ) d.2 < read.xls(file.ch oose(),sheet=3 ) d.3 < read.xls(file.choose(),sheet=3 ) lc.1 < read.xls(file.choose(),sheet=3 ) lc.2 < read.xls(file.choose(),sheet=3 ) ##Check models with AIC scores ##Block 1 (Divide) AIC(lm(d.1$Minimum~d.1$Day*I(d.1$Day^2))) AIC(lm(d.1$Minimum~d.1$Day)) ##Bl ock 2 (Divide) AIC(lm(d.2$Minimum~d.2$Day*I(d.2$Day^2))) AIC(lm(d.2$Minimum~d.2$Day)) ##Block 3 (Divide) AIC(lm(d.3$Minimum~d.3$Day*I(d.3$Day^2))) AIC(lm(d.3$Minimum~d.3$Day)) ##Block 4 (LineCreek) AIC(lm(lc.1$Minimum~lc.1$Day*I(lc.1$Day^2))) AIC(lm(lc.1 $Minimum~lc.1$Day)) ##Block 5 (LineCreek) AIC(lm(lc.2$Minimum~lc.2$Day*I(lc.2$Day^2))) AIC(lm(lc.2$Minimum~lc.2$Day)) ##Select model ##Quadratic (T max o 1 2 Day 2 + e) d.1.q< (lm(d.1$Minimum~d.1$Day*I(d.1$Day^2))) d.2.q< (lm(d.2$Minimum~d.2$Day *I(d.2$Day^2))) d.3.q< (lm(d.3$Minimum~d.3$Day*I(d.3$Day^2))) lc.1.q< (lm(lc.1$Minimum~lc.1$Day*I(lc.1$Day^2))) lc.2.q< (lm(lc.2$Minimum~lc.2$Day*I(lc.2$Day^2)))

PAGE 122

109 ##OR ##Linear (T max o 1 Day + e) d.1.q< (lm(d.1$Minimum~d.1$Day)) d.2.q< (lm(d.2$Minimu m~d.2$Day)) d.3.q< (lm(d.3$Minimum~d.3$Day)) lc.1.q< (lm(lc.1$Minimum~lc.1$Day)) lc.2.q< (lm(lc.2$Minimum~lc.2$Day)) ##Residuals around model res.1< d.1.q$residuals res.2< d.2.q$residuals res.3< d.3.q$residuals res.4< lc.1.q$residuals res.5< lc.2.q$residu als ##Rank Residuals for Two factor ANOVA rank.1 < rank(res.1) rank.2 < rank(res.2) rank.3 < rank(res.3) rank.4 < rank(res.4) rank.5 < rank(res.5) ##Create Divide Dataframe (2011 D.Micros ite< rep(c("UN","RK","ES","WB"), 187) D.Site< rep("Divide",256+256+236) B1< rep("1",256) B2< rep("2",256) B3< rep("3",236) D.Block< c(B1,B2,B3) D.rank< c(rank.1,rank.2,rank.3) D.Total< data.frame(D.Block,D.Microsite,D.rank) ##Check Interaction of Block on Microsite ##Divide Only anova(lm(D.Total$D.rank~D.Total$D.Microsite*D.Total$D.Block)) ##Create LineCreek Dataframe (2011 L.Microsite< rep(c("UN","RK","ES","WB"), 112) L.Site< rep("LineCreek",2 24+224) B4< rep("4",224) B5< rep("5",224) L.Block< c(B4,B5) L.rank< c(rank.4,rank.5) L.Total< data.frame(L.Block,L.Microsite,L.rank) ##Check Interaction of Block on Microsite

PAGE 123

110 ##Line Creek Only anova(lm(L.Total$L.rank~L.Total$L.Microsite*L.Total$L.Block)) ##Dataframe for both study areas (2011 Microsite< c(D.Microsite,L.Microsite) Site< c(D.Site,L.Site) Block< c(B1,B2,B3,B4,B5) rank.Res< c(rank.1,rank.2,rank.3,rank.4,rank.5) Total< data.frame(Si te,Block,Microsite,rank.Res) ##Check Interaction of Block on Microsite ##Both study areas anova(lm(Total$rank.Res~Total$Microsite*Total$Block)) ##When interaction is significant > Treat Blocks at non replicates ##DIVIDE, BLOCK 1 ##NONPARAMETRIC KRUSKAL WALLIS TEST kruskal.test(res.1~d.1$Treatment) pairwise.wilcox.test(res.1, d.1$Treatment, p.adj="bonferroni", exact=F) ##DIVIDE, BLOCK 2 ##NONPARAMETRIC KRUSKAL WALLIS TEST kruskal.test(res.2~d.2$Treatment) pairwise.wilcox.test(res.2, d.2$Treatment, p.adj ="bonferroni", exact=F) ##DIVIDE, BLOCK 3 ##NONPARAMETRIC KRUSKAL WALLIS TEST kruskal.test(res.3~d.3$Treatment) pairwise.wilcox.test(res.3, d.3$Treatment, p.adj="bonferroni", exact=F) ##BOXPLOTS ##Divide Mtn par(mar=c(5,4,2,2)) par(mfrow=c(1,3)) par(pin= c(2,2),mai=c(.9,.7,.3,.3)) d.1$Treatment< factor(d.1$Treatment, levels = c("WB", "ES","RK", "UN")) d.2$Treatment< factor(d.2$Treatment, levels = c("WB", "ES","RK", "UN")) d.3$Treatment< factor(d.3$Treatment, levels = c("WB", "ES","RK", "UN")) boxp lot( res.1~d.1$Treatment,ylim=c( 6,6),font.lab=2,font.sub=2,cex=1.25,ylab="Residuals",col=c("forestgreen","firebr ick3","darkorchid3","gold"),cex.axis=1.5,cex.sub=1.75,cex.lab=1.75,sub="Block 1")

PAGE 124

111 boxplot(res.2~d.2$Treatment,ylim=c( 6,6),font.lab=2,font.sub=2,cex =1.25,ylab="Residuals",col=c("forestgreen","firebr ick3","darkorchid3","gold"),cex.axis=1.5,cex.sub=1.75,cex.lab=1.75,sub="Block 2") boxplot(res.3~d.3$Treatment,ylim=c( 6,6),font.lab=2,font.sub=2,cex=1.25,ylab="Residuals",col=c("forestgreen","firebr ick3","d arkorchid3","gold"),cex.axis=1.5,cex.sub=1.75,cex.lab=1.75,sub="Block 3") ##LINE CREEK, BLOCK 1 ##NONPARAMETRIC KRUSKAL WALLIS TEST kruskal.test(res.4~lc.1$Treatment) pairwise.wilcox.test(res.4, lc.1$Treatment, p.adj="bonferroni", exact=F) ##LINE CREEK, BLOCK 2 ##NONPARAMETRIC KRUSKAL WALLIS TEST kruskal.test(res.5~lc.2$Treatment) pairwise.wilcox.test(res.5, lc.2$Treatment, p.adj="bonferroni", exact=F) ##BOXPLOTS ##Line Creek par(mar=c(6,5,2,2)) par(mfrow=c(1,2)) par(pin=c(2,2),mai=c(1.5,1.3,.3,.3)) lc. 1$Treatment< factor(lc.1$Treatment, levels = c("WB", "ES","RK", "UN")) lc.2$Treatment< factor(lc.2$Treatment, levels = c("WB", "ES","RK", "UN")) boxplot(res.4~lc.1$Treatment,ylim=c( 8,6),font.lab=2,font.sub=2,cex=1.25,ylab="Residuals",col=c("forestgreen ","firebrick3"," darkorchid3","gold"),cex.axis=1.5,cex.sub=1.75,cex.lab=1.75,sub="Block 1") boxplot(res.5~lc.2$Treatment,ylim=c( 8,6),font.lab=2,font.sub=2,cex=1.25,ylab="Residuals",col=c("forestgreen","firebrick3"," darkorchid3","gold"),cex.axis=1.5,cex.sub =1.75,cex.lab=1.75,sub="Block 2")

PAGE 125

112 Daily Sum PAR library(gdata) library(gtools) ## D.NE d.ne< read.xls(file.choose(),sheet=5) head(d.ne) AIC(lm(d.ne$Sum~d.ne$Day*I(d.ne$Day^2))) AIC(lm(d.ne$Sum~d.ne$Day)) lin.d.ne< lm( d.ne$ Sum~ d.ne$ Da y) ## BOXPLOT plot( d.ne$ Treatment,lin.d.ne$residuals,ylab="Residuals",main="PAR sum") par(mar=c(5,4.5,2,2)) par(mfrow=c(2,2)) d.ne$Treatment < factor(d.ne$Treatment, levels = c("WB", "ES","RK", "UN")) boxplot(lin.d.ne$r esiduals~d.ne$Treatment,ylim=c( 4 5000,40000), ylab="Residuals", xlab="Divide NE",col=c("lightcyan","mistyrose","lavender","cornsilk")) ##Kruskal Wallis Test kruskal.test(lin.d.ne$residuals~ d.ne$ Treatment) pairwise.wilcox.test(lin.d.ne$residuals, d.ne$ Treatment, p.adj="bonferroni", exact= F) ## D. W d.w< read.xls(file.choose(),sheet=5) head(d.w) par(mfrow=c(1,3)) plot( d.w $ Day, d.w $ Sum,pch=as.numeric(Treatment),col=c("green","blue","red","gray")) AIC(lm( d.w $ Sum~ d.w $ Day*I( d.w $ Day^2))) AIC(lm( d.w $ Sum~ d.w $ Day)) lin.d.w< lm(d.w $Sum~d.w$Day) ## BOXPLOT plot( d.w $ Treatment, lin.d.w$residuals) d.w$Treatment < factor(d.w$Treatment, levels = c("WB", "ES","RK", "UN"))

PAGE 126

113 boxplot(lin.d.w$residuals~d.w$Treatment,ylim=c( 45000,40000),xlab="Divide W",col=c("lightcyan","mistyrose","lavende r","cornsilk")) kruskal.test(lin.d.w$residuals~ d.w $ Treatment) pairwise.wilcox.test(lin.d.w$residuals, d.w $ Treatment, p.adj="bonferroni", exact=F) ##LN. NE lc.ne< read.xls(file.choose(),sheet=5) head(lc.ne) AIC(lm( lc.ne $ Sum~ lc.ne $ Day*I( lc.ne $ Day^ 2))) AIC(lm( lc.ne $ Sum~ lc.ne $ Day)) lin.lc.ne< lm(lc.ne$Sum~lc.ne$Day) ## BOX PLOT plot( lc.ne $ Treatment,lin.lc.ne$residuals) lc.ne$Treatment < factor(lc.ne$Treatment, levels = c("WB", "ES","RK", "UN")) boxplot(lin.lc.ne$residuals~lc.ne$Treatment, yl im = c( 45000, 40000), ylab = "Residuals",xlab="Line Creek NE",col=c("lightcyan","mistyrose","lavender","cornsilk")) kruskal.test(lin.lc.ne$residuals~ lc.ne $ Treatment) pairwise.wilcox.test(lin.lc.ne$residuals, lc.ne $ Treatment, p.adj="bonferroni", exact= F) ## LC. W lc.w< read.xls(file.choose(),sheet=5) head(lc.w) AIC(lm( lc.w $ Sum~ lc.w $ Day*I( lc.w $ Day^2))) AIC(lm( lc.w $ Sum~ lc.w $ Day)) lin.lc.w< lm(lc.w$Sum~lc.w$Day) plot(Day,lin.lc.w$residuals,pch=as.numeric(Treatment),col=c("green","blue","red" ,"gray ")) ## BOXPLOT plot(Treatment,lin.lc.w$residuals) lc.w$Treatment < factor(lc.w$Treatment, levels = c("WB", "ES","RK", "UN")) boxplot(lin.lc.w$residuals~lc.w$Treatment,ylim=c( 45000,40000),xlab="Line Creek W",col=c("lightcyan","mistyrose","lavende r","cornsilk"))

PAGE 127

114 kruskal.test(lin.lc.w$residuals~ lc.w $ Treatment) pairwise.wilcox.test(lin.lc.w$residuals, lc.w $ Treatment, p.adj="bonferroni", exact=F) Daily Maximum Gust Speeds ### 2011 and 2012 Maximum Gust Speeds ### library(gdata) library(gtool s) #Import data d.1< read.xls(file.choose(),sheet=4) d.2< read.xls(file.choose(),sheet=4) d.3< read.xls(file.choose(),sheet=4) lc.1< read.xls(file.choose(),sheet=4) lc.2< read.xls(file.choose(),sheet=4) ##Rank Averages rank.1 < rank(d.1$Maximum) rank.2 < rank(d.2$Maximum) rank.3 < rank(d.3$Maximum) rank.4 < rank(lc.1$Maximum) rank.5 < rank(lc.2$Maximum) ##Create Divide Dataframe (2011 D.Microsite< rep(c("UN","RK","ES","WB"), 187) D.Site< rep("Divide",256+256+236) B1< rep("1",256) B2< rep("2",256) B3< rep("3",236) D.Block< c(B1,B2,B3) D.Rank< c(rank.1,rank.2,rank.3) D.Total< data.frame(D.Block,D.Microsite,D.Rank) ##Create LineCreek Dataframe (2011 >(" L.Microsite< rep(c("UN","RK","ES","WB"), 112) L.Site< rep("LineCreek",224+224) B4< rep("4",224) B5< rep("5",224) L.Block< c(B4,B5) L.Rank< c(rank.4,rank.5) L.Total< data.frame(L.Block,L.Microsite,L.Rank) ##Dataframe of all data (2011

PAGE 128

115 Microsite< rep(c("UN","RK","ES","WB"),299) Site< c(D.Site,L.Site) Block< c(B1,B2,B3,B4,B5) T.Rank< c(rank.1,rank.2,rank.3,rank.4,rank.5) Total< data.frame(Site,Block,Microsite,T.Rank) ##When in teraction is significant > Treat Blocks at non replicates ##DIVIDE, BLOCK 1 ##NONPARAMETRIC KRUSKAL WALLIS TEST kruskal.test(d.1$Maximum~d.1$Treatment) pairwise.wilcox.test(d.1$Maximum, d.1$Treatment, p.adj="bonferroni", exact=F) ##DIVIDE, BLOCK 2 ##NONPA RAMETRIC KRUSKAL WALLIS TEST kruskal.test(d.2$Maximum~d.2$Treatment) pairwise.wilcox.test(d.2$Maximum, d.2$Treatment, p.adj="bonferroni", exact=F) ##DIVIDE, BLOCK 3 ##NONPARAMETRIC KRUSKAL WALLIS TEST kruskal.test(d.3$Maximum~d.3$Treatment) pairwise.wilco x.test(d.3$Maximum, d.3$Treatment, p.adj="bonferroni", exact=F) ##Box Plots ##Divide Mtn par(mar=c(5,4,2,2)) par(mfrow=c(1,3)) par(pin=c(2,2),mai=c(.9,.7,.3,.3)) d.1$Treatment< factor(d.1$Treatment, levels = c("WB", "ES","RK", "UN")) d.2$Treatment< fac tor(d.2$Treatment, levels = c("WB", "ES","RK", "UN")) d.3$Treatment< factor(d.3$Treatment, levels = c("WB", "ES","RK", "UN")) boxplot(d.1$Maximum~d.1$Treatment,ylim=c(0,18),font.lab=2,font.sub=2,cex=1.25,yla b="Residuals",col=c("forestgreen","firebrick3", "darkorchid3","gold"),cex.axis=1. 5,cex.sub=1.75,cex.lab=1.75,sub="Block 1") boxplot(d.2$Maximum~d.2$Treatment,ylim=c(0,18),font.lab=2,font.sub=2,cex=1.25,yla b="Residuals",col=c("forestgreen","firebrick3","darkorchid3","gold"),cex.axis=1. 5,cex.sub=1.75,cex. lab=1.75,sub="Block 2") boxplot(d.3$Maximum~d.3$Treatment,ylim=c(0,18),font.lab=2,font.sub=2,cex=1.25,yla b="Residuals",col=c("forestgreen","firebrick3","darkorchid3","gold"),cex.axis=1. 5,cex.sub=1.75,cex.lab=1.75,sub="Block 3") ##LINE CREEK, BLOCK 1 ##NON PARAMETRIC KRUSKAL WALLIS TEST kruskal.test(lc.1$Maximum~lc.1$Treatment) pairwise.wilcox.test(lc.1$Maximum, lc.1$Treatment, p.adj="bonferroni", exact=F)

PAGE 129

116 ##LINE CREEK, BLOCK 2 ##NONPARAMETRIC KRUSKAL WALLIS TEST kruskal.test(lc.2$Maximum~lc.2$Treatment) pa irwise.wilcox.test(lc.2$Maximum, lc.2$Treatment, p.adj="bonferroni", exact=F) ##Box Plot ##Line Creek par(mar=c(6,5,2,2)) par(mfrow=c(1,2)) par(pin=c(2,2),mai=c(1.5,1.3,.3,.3)) lc.1$Treatment< factor(lc.1$Treatment, levels = c("WB", "ES","RK", "UN")) lc .2$Treatment< factor(lc.2$Treatment, levels = c("WB", "ES","RK", "UN")) boxplot(lc.1$Maximum~lc.1$Treatment,ylim=c(0,18),font.lab=2,font.sub=2,cex=1.25,yl ab="Residuals",col=c("forestgreen","firebrick3","darkorchid3","gold"),cex.axis=1 .5,cex.sub=1.75,cex. lab=1.75,sub="Block 1") boxplot(lc.2$Maximum~lc.2$Treatment,ylim=c(0,18),font.lab=2,font.sub=2,cex=1.25,yl ab="Residuals",col=c("forestgreen","firebrick3","darkorchid3","gold"),cex.axis=1 .5,cex.sub=1.75,cex.lab=1.75,sub="Block 2") Daily Average Wind Speed s ### 2011 and 2012 Average Wind Speeds ### library(gdata) library(gtools) ##Import data d.1< read.xls(file.choose(),sheet=2) d.2< read.xls(file.choose(),sheet=2) d.3< read.xls(file.choose(),sheet=2) lc.1< read.xls(file.choose(),sheet=2) lc.2< read.xls(f ile.choose(),sheet=2) ##Rank Averages rank.1 < rank(d.1$Average) rank.2 < rank(d.2$Average) rank.3 < rank(d.3$Average) rank.4 < rank(lc.1$Average) rank.5 < rank(lc.2$Average) ##Create Divide Dataframe (2011 >("W D.Microsite< rep(c("UN","RK","ES","WB"), 187)

PAGE 130

117 D.Site< rep("Divide",256+256+236) B1< rep("1",256) B2< rep("2",256) B3< rep("3",236) D.Block< c(B1,B2,B3) D.Rank< c(rank.1,rank.2,rank.3) D.Total< data.frame(D.Block,D.Microsite,D.Rank) ##C reate LineCreek Dataframe (2011 L.Microsite< rep(c("UN","RK","ES","WB"), 112) L.Site< rep("LineCreek",224+224) B4< rep("4",224) B5< rep("5",224) L.Block< c(B4,B5) L.Rank< c(rank.4,rank.5) L.Tota l< data.frame(L.Block,L.Microsite,L.Rank) ##Dataframe of all data (2011 Microsite< rep(c("UN","RK","ES","WB"),299) Site< c(D.Site,L.Site) Block< c(B1,B2,B3,B4,B5) T.Rank< c(rank.1,rank.2,rank.3 ,rank.4,rank.5) Total< data.frame(Site,Block,Microsite,T.Rank) ##Check Interaction of Block on Microsite ##All anova(lm(Total$T.Rank~Total$Microsite*Total$Block)) ##Divide Only anova(lm(D.Total$D.Rank~D.Total$D.Microsite*D.Total$D.Block)) ##Line Creek Only anova(lm(L.Total$L.Rank~L.Total$L.Microsite*L.Total$L.Block)) ##When interaction is significant > Treat Blocks at non replicates ##DIVIDE, BLOCK 1 ##NONPARAMETRIC KRUSKAL WALLIS TEST kruskal.test(d.1$Average~d.1$Treatment) pairwise.wilc ox.test(d.1$Average, d.1$Treatment, p.adj="bonferroni", exact=F) ##DIVIDE, BLOCK 2 ##NONPARAMETRIC KRUSKAL WALLIS TEST kruskal.test(d.2$Average~d.2$Treatment) pairwise.wilcox.test(d.2$Average, d.2$Treatment, p.adj="bonferroni", exact=F) ##DIVIDE, BLOCK 3 ##NONPARAMETRIC KRUSKAL WALLIS TEST kruskal.test(d.3$Average~d.3$Treatment)

PAGE 131

118 pairwise.wilcox.test(d.3$Average, d.3$Treatment, p.adj="bonferroni", exact=F) ##Boxplot ##Divide Mtn par(mar=c(5,4,2,2)) par(mfrow=c(1,3)) par(pin=c(2,2),mai=c(.9,.7,.3,.3)) d.1 $Treatment< factor(d.1$Treatment, levels = c("WB", "ES","RK", "UN")) d.2$Treatment< factor(d.2$Treatment, levels = c("WB", "ES","RK", "UN")) d.3$Treatment< factor(d.3$Treatment, levels = c("WB", "ES","RK", "UN")) boxplot(d.1$Average~d.1$Treatment,ylim= c(0,3.5),font.lab=2,font.sub=2,cex=1.25,ylab ="Avg Wind Speed (m/s)",col=c("forestgreen","firebrick3","darkorchid3","gold"),cex.axis=1.5,cex.su b=1.75,cex.lab=1.75,sub="Block 1") boxplot(d.2$Average~d.2$Treatment,ylim=c(0,3.5),font.lab=2,font.sub=2,cex=1.25, ylab ="Avg Wind Speed (m/s)",col=c("forestgreen","firebrick3","darkorchid3","gold"),cex.axis=1.5,cex.su b=1.75,cex.lab=1.75,sub="Block 2") boxplot(d.3$Average~d.3$Treatment,ylim=c(0,3.5),font.lab=2,font.sub=2,cex=1.25,ylab ="Avg Wind Speed (m/s)",col=c("fores tgreen","firebrick3","darkorchid3","gold"),cex.axis=1.5,cex.su b=1.75,cex.lab=1.75,sub="Block 3") ##LINE CREEK, BLOCK 1 ##NONPARAMETRIC KRUSKAL WALLIS TEST kruskal.test(lc.1$Average~lc.1$Treatment) pairwise.wilcox.test(lc.1$Average, lc.1$Treatment, p.adj=" bonferroni", exact=F) ##LINE CREEK, BLOCK 2 ##NONPARAMETRIC KRUSKAL WALLIS TEST kruskal.test(lc.2$Average~lc.2$Treatment) pairwise.wilcox.test(lc.2$Average, lc.2$Treatment, p.adj="bonferroni", exact=F) ##Boxplot ##Line Creek par(mar=c(6,5,2,2)) par(mfrow =c(1,2)) par(pin=c(2,2),mai=c(1.5,1.3,.3,.3)) lc.1$Treatment< factor(lc.1$Treatment, levels = c("WB", "ES","RK", "UN")) lc.2$Treatment< factor(lc.2$Treatment, levels = c("WB", "ES","RK", "UN")) boxplot(lc.1$Average~lc.1$Treatment,ylim=c(0,3.5),font.lab =2,font.sub=2,cex=1.25,yla b="Avg Wind Speed

PAGE 132

119 (m/s)",col=c("forestgreen","firebrick3","darkorchid3","gold"),cex.axis=1.5,cex.su b=1.75,cex.lab=1.75,sub="Block 1") boxplot(lc.2$Average~lc.2$Treatment,ylim=c(0,3.5),font.lab=2,font.sub=2,cex=1.25,yla b="Avg Wind Speed (m/s)",col=c("forestgreen","firebrick3","darkorchid3","gold"),cex.axis=1.5,cex.su b=1.75,cex.lab=1.75,sub="Block 2") Percent Sky Exposure ##CHI SQUARE GOODNESS OF FIT ##SKY EXPOSURE 2010 (Repeated for each block, 2011) ## (O E) 2 /E ##Divide NE WB .d.ne< ((78.23 89.39)^2)/(89.39) ES.d.ne< ((93.07 89.39)^2)/(89.39) RK.d.ne< ((93.27 89.39)^2)/(89.39) UN.d.ne< ((92.98 89.39)^2)/(89.39) chi.sq.d.ne< (WB.d.ne+ES.d.ne+RK.d.ne+UN.d.ne) chi.sq.d.ne pchisq(chi.sq.d.ne,3,lower.tail=FALSE) ##Divide W ES.d. w< ((90.37 91.26)^2)/(91.26) RK.d.w< ((93.97 91.26)^2)/(91.26) UN.d.w< ((89.45 91.26)^2)/(91.26) chi.sq.d.w< (ES.d.w+RK.d.w+UN.d.w) chi.sq.d.w pchisq(chi.sq.d.ne,2,lower.tail=FALSE) ##Line Creek NE WB.l.ne< ((55.38 76.82)^2)/(76.82) ES.l.ne< ((60.34 76 .82)^2)/(76.82) RK.l.ne< ((92.37 76.82)^2)/(76.82) UN.l.ne< ((99.22 76.82)^2)/(76.82) chi.sq.l.ne< (WB.l.ne+ES.l.ne+RK.l.ne+UN.l.ne) chi.sq.l.ne pchisq(chi.sq.l.ne,3,lower.tail=FALSE)

PAGE 133

120 ##Line Creek W WB.l.w< ((43.28 58.51)^2)/(58.51) ES.l.w< ((36.9 58.5 1)^2)/(58.51) RK.l.w< ((67.75 58.51)^2)/(58.51) UN.l.w< ((86.11 58.51)^2)/(58.51) chi.sq.l.w< (WB.l.w+ES.l.w+RK.l.w+UN.l.w) chi.sq.l.w pchisq (chi.sq.l.w,3,lower.tail=FALSE) Total Percent Nitrogen and Carbon in Soils Two Way ANOVA ##Carbon and Nitrogen in Soils ##Import Soil Data library(gtools) library(gdata) ##All Data soils=read.xls(file.choose(),sheet=3) ##Divide only soils.d=read.xls(file.choose(),sheet=1) ##Beartooth only soils.b=read.xls(file.choose(),sheet=2) ##Variance Tests ##Nitrogen bart lett.test(soils$X.N,soils$microsite) ##Carbon bartlett.test(soils$X.C,soils$microsite) ##Normality Tests (repeat with Line Creek) ##Nitrogen (repeat with carbon) library(nortest) ##Whitebark inds< soils.d[,1]=="WB" w_avg< soils.d[inds,2] ad.test(w_avg) ##Spruce inds< soils.d[,1]=="ES" s_avg< soils.d[inds,2] ad.test(s_avg) ##Rock inds< soils.d[,1]=="RK" r_avg< soils.d[inds,2] ad.test(r_avg) ##Unprotected

PAGE 134

121 inds< soils.d[,1]=="UN" u_avg< soils.d[inds,2] ad.test(u_avg) ##Two Way ANOVA ##Nitrogen anova(lm(so ils$X.N~soils$microsite*soils$site)) summary(lm(soils$X.N~soils$microsite*soils$site)) ##Carbon anova(lm(soils$X.C~soils$microsite*soils$site)) summary(lm(soils$X.C~soils$microsite*soils$site)) ##Post hoc not performed due to insignificant results ##95% C.I. for Divide Mtn, % Nitrogen (Repeat for Line Creek and then again with Carbon) ##Margin of Error, z*(sd/sqrt(n)) me< qnorm(0.975)*(sd(soils.d$X.N)/sqrt(40)) ##Lower and Upper Bounds mean(soils.d$X.N) me mean(soils.d$X.N)+ me Number of Freeze Thaw Events ##CHI SQUARE GOODNESS OF FIT ## FREEZE THAW EVENTS (comparing study area) ## (O E) 2 /E ## STUDY AREA 2011 LineCreek .11 < (( 67 49 )^2)/( 49 ) DivideMtn .11 < (( 37 49 )^2)/( 49 ) chi.sq .11 < ( LineCreek.11+DivideMtn.11 ) chi.sq .11 pchisq(chi.sq. 11 1 ,lower.tai l=FALSE) ## (O E) 2 /E ## STUDY AREA 2012 LineCreek.12 < (( 137 94)^2)/(94 ) DivideMtn.12 < (( 51 94 )^2)/( 94) chi.sq .12 < ( LineCreek.12+DivideMtn.12 ) chi.sq .12 pchisq(chi.sq. 12 1 ,lower.tail=FALSE)