Variability of whitebark pine (Pinus albicaulis Engelm.) leaf traits in the Great Basin

Trevor Carter and Alison Agneray advised by Dr.Elizabeth Leger

University of Nevada, Reno.


Citation: T. Carter and A. Agneray, “Variability of whitebark pine (Pinus albicaulis Engelm.) leaf traits in the Great Basin” Nevada State Undergraduate Research Journal. V6:I1 Spring-2020. (2020).


Whitebark pine are valued for the ecosystem services they provide in subalpine forests of the western United States and have been declining across their range. This project quantifies two leaf traits of contemporary and historical populations (sites) of whitebark pine within the Great Basin. Characteristics of historical herbarium specimens were compared against samples collected in the 2018 field season of four different populations of whitebark pine within the Great Basin and the eastern Sierra Nevada. We asked how these populations differed from each other. Little change was observed through time for any of the sites, but leaf trait values were different among populations. The Jarbidge site within the Great Basin showed the most different leaf trait values, with smaller leaf mass per area and fewer leaves than other sites, in both historical and contemporary samples. Our research suggests there are differences among populations that may reflect important differences among growing conditions, genetic variation, or a combination of these factors. Additional research is needed to determine what is driving variation among whitebark pine within the Great Basin.


Whitebark pine is valued for its ecosystem services and the wildlife habitat it provides (1). It has an expansive range in high elevation western North American forests (2), but whitebark pine has been declining across its range due to a variety of interconnected mortality agents including white pine blister rust (3), the mountain pine beetle (4, 5), and drought (6). The influence of drought is particularly crucial because it can cause direct mortality (6), lead to decreased tree vigor, and exacerbate other sources of mortality (7, 8). The effects of drought are especially felt in environments where resources are limited, such as sub-alpine ecosystems (9). However, the influence of drought on whitebark pine adaptation has not been fully examined.

The effects of drought on plants are measured by changes in their ecophysiology (10, 11). These processes can be inferred through leaf traits such as leaf quantity (12), and leaf mass per area (13). Previous studies have shown that leaf traits vary through time in both annual (e.g., Ref. (14)) and perennial species (e.g., Ref. (15)), and leaf morphology is known to respond to drought (12). Understanding the variability of a species across environmental gradients is a key first step to understanding whether populations are adapted to their local environments (16) and measuring the variation in traits over time can indicate whether morphology is changing in response to factors such as drought or harvest (17). Leaf traits are easy to measure on herbarium specimens, allowing for analysis through time, even of long-lived perennials.

Determining the distribution of trait variation across populations can better guide restoration efforts (18). Whitebark pine populations are locally adapted across geographic areas with traits related to cold adaptation, and current restoration practices for whitebark pine utilize seed transfer guidelines based on climatic factors (19). These practices include moving genetic stock across environments that do not differ by more than 1.0◦C in mean temperature (19). Whitebark pine habitats also differ in a variety of other environmental conditions such as vapor pressure deficit and precipitation (20), which haven’t been studied for this species. Work to-date on trait distribution in whitebark pine has not fully incorporated populations within the Great Basin, and additional drought-specific traits may be of interest. For example, leaf number is known to decrease as a way for trees to manage water stress (21), while leaf mass per area (LMA) is shown to be higher in more drought-resistant plants (22). The relationship between LMA and drought may be more complicated. For example, LMA was documented to decrease in Scots pine (Pinus sylvestris) during drought-induced mortality events (23).

The objectives of this research were to ask the following questions: (1) Do leaf quantity and LMA differ within contemporary whitebark pine populations in the Great Basin and the Sierra Nevada mountains? And (2) have leaf quantity, and LMA changed through time? We predicted that populations would differ and that regions with greater water limitation would have lower leaf quantity and higher LMA. We expected that leaf quantity would decrease over time in areas experiencing climate warming, and would increase in areas getting wetter over time, with LMA showing the opposite pattern. We were also interested in whether populations differed in their ability to change in these leaf traits over time, which might indicate a higher degree of plasticity within certain populations and better predict success after seed transfer (16).

Material and methods

Population selection

We studied four populations of whitebark pine: Mount Rose (MR), Eastern Sierra (ES), Ruby Mountains (RM), and Jarbidge (Jar). We chose populations based on the availability of historical specimens at the University of Nevada, Reno herbarium (Appendix Table 2). Samples were identified using the Intermountain Region Herbarium Network (24) and selected based on specimen location. Each population, defined as a different mountain range, contained a minimum of three historical samples collected at least ten years before sampling. The date of collection for specimens ranged from 1912 – 2006. While these historic specimens varied greatly in age, we included all of them due to the limited data available for historic measurements. Geographic locations were determined based on the collection information provided in the descriptions of the specimens. These locations created a starting point for identifying the contemporary population in the same locality.

Field sampling

To determine the target tree branch size for field sampling, we measured the stem diameter of herbarium specimens at the nadir of each sample, near the point where the sample was removed from the tree. We then averaged the diameters within populations to create a target diameter for contemporary sampling, which were collected at each of the four mountain ranges based on this measure. Averaging the stem diameter was done to reduce sampling biases based on branch age, and to conduct the most accurate comparison of modern samples to historical specimens. Trees (n = 30 per site) were chosen randomly within a 5-km radius. Two samples were collected per tree based on the stem diameter at varying locations of the tree. Global Positioning System (GPS) points were recorded at every replicate tree (Appendix Table 1). Samples were then stored in a plant press for at least one week to preserve leaf tissue.

Train measurements

Leaf quantity and LMA were estimated based on methods used by Abrams et al. (10). We counted every fascicle for contemporary specimens and multiplied the final count by five, assuming every fascicle had five needles. Individual needles were counted for herbarium specimens due to fascicles of herbarium specimens not being readily visible without destructive sampling. The needle count of herbarium specimens was rounded up to the next multiple of five, assuming that needle count would be a multiple of five. LMA measurements were conducted on both contemporary and historical specimens. Needles were randomly pulled from a bag of contemporary samples to ensure random selection. The herbarium director permitted us to remove a single needle from the specimen mounts; this was done for all but one delicate specimen (Appendix Table 2). If able, needles were selected at random from a small pocket at the bottom of the herbarium sheet where residual material was stored. All leaves were oven-dried for two days at 40◦C before measuring LMA, similar to work by Hultine & Marshall (25).

Climate data

Precipitation and vapor pressure deficit (VPd) data were gathered from the PRISM Climate Group (20) for each location and used as a proxy for drought stress. The trends of 30-year average precipitation were used to assess change over time for each population.

Data analysis

All of our analyses were performed using R statistical software (26). ANOVA was used to compare contemporary populations based on LMA and leaf quantity metrics. Ttests were used to compare LMA and leaf quantity through time within the same population. For the purpose of a pairwise statistical test, age was categorized as either “contemporary” or “historical.” These data were compared against the leaf trait values for the corresponding population.


Do leaf quantity and LMA differ within four contemporary populations of the Great Basin and eastern Sierra? Precipitation was variable across the four populations (Table 2). The Mount Rose population had the highest 30-year average precipitation, and Jarbidge had the lowest average rainfall. However, the VPd was less variable and showed no apparent differences among populations (Table 2). Precipitation was increasing at three sites, and decreasing at one.

Contemporary populations differed in leaf traits, with the Jarbidge population varying the most other populations (Figure 1). Average leaf quantity per replicate at Jarbidge (n = 613) was significantly lower than other populations (Jar-ES p = 0.0110; Jar-MR p < 0.0001; Jar-RM p = 0.0001). The other populations did not differ significantly from each other in average leaf quantity (Figure 2). Similarly, the LMA measurements for Jarbidge (0.0229 g/cm2) were lower than the other populations (Jar-ES p < 0.0001; Jar-MR p < 0.0001; Jar-RM p = 0.0491). Additionally, the Ruby Mountain population had slightly lower LMA than the Eastern Sierra population and was significantly smaller than the Mount Rose LMA measurements (RM-MR p = 0.0146).

Figure 1 - Leaf Quantity at Different Comtemporary Populations
Figure 1: Box plot of leaf quantity at four different populations (ES = Eastern Sierra, Jar = Jarbidge, MR = Mount Rose, RM = Ruby Mountain). Leaf quantity ranged from 263 to 2423 leaves per sample. Black dots represent individual data points, with two branches per tree for 30 trees per population. Solid back lines represent median values within a population. Whiskers represent 95% confidence intervals of measurements. TukeyHSD values are: ES-Jar = 0.011, ES-MR = 0.355, ES-RM = 0.581, Jar-MR < 0.001, Jar- RM < 0.001, RM-MR = 0.982.
Figure 2 - Leaf Mass per Area at DIfferent Contemporary Populations
Figure 2: Box plot of leaf mass per area at four different populations (ES = Eastern Sierra, Jar = Jarbidge, MR = Mount Rose, RM = Ruby Mountain). LMA ranged from 0.01729 (g/cm2) to 0.03867 (g/cm2). Black dots represent individual data points, with two leaves per tree for 30 trees per population. Solid back lines represent median values within a population. Whiskers represent 95% confidence intervals of measurements. TukeyHSD values are: ES-Jar < 0.001, ES-MR = 0.961, ES-RM = 0.055, Jar-MR < 0.001, Jar- RM = 0.049, RM-MR = 0.015.

Do leaf quantity and LMA change through time within four populations of the Great Basin and eastern Sierra? Of the four populations and two traits, only one leaf trait in one population exhibited change over time (Figure 3). The Mount Rose population showed a significant increase in leaf quantity (mean historical = 826.7, mean contemporary = 1168.6; p = 0.0312). The other populations showed no significant change of leaf quantity (Appendix Table 3). Leaf mass per area did not show a significant change over time within any population. Overall, the contemporary samples trended towards having more leaves than the historical samples, though this difference was not significant at the p = 0.05 level.

Figure 3 - Leaf Quantity at Different Populations Through Time
Figure 3: Box plot showing leaf quantify for four populations (ES = Eastern Sierra, Jar = Jarbidge, MR = Mount Rose, RM = Ruby Mountain), with contemporary (Modern) and historical (Historical) values. Filled boxes represent modern populations while outlines boxes represent historical counterparts. Leaf quantity ranged from 263 to 2423 leaves per sample. Black dots represent individual data points, which, in the field, were measured for two branches from 30 trees per population. Solid back lines represent median values within a population. Whiskers represent 95% confidence intervals of measurements. P values of contemporary vs historical t-tests are as follows: ES = 0.276, Jar = 0.161, MR = 0.031, RM = 0.648.


Understanding how leaf traits vary among populations is an important first step towards understanding how species respond to their environments and for understanding the degree of variation in traits across landscapes. In our study, we measured two leaf traits across multiple populations that differed in climatic conditions and how traits change across time. We found that the driest whitebark pine site, Jarbidge, had both lower leaf quantity and LMA compared to other populations. Lower leaf quantities decrease surface area for evapotranspiration, making plants more water-use efficient (27). However, decreased leaf mass per area has an inverse effect. Decreased LMA is associated with lower densities of mesophyll tissue within the leaves, making plants less drought-tolerant (28, 29). This is of concern for Jarbidge given its notably low precipitation of the four sites and because similar LMA decreases in Scots pine were observed during drought-induced mortality (23).

The only population to show a change between historical and contemporary samples was Mount Rose, which showed an increase in leaf quantity over 105 years. This change in leaf quantity was not driven solely by the range of collection years within the Mount Rose population, as the exclusions of the oldest and most recent specimens yield comparable results. The exclusion of the most recent specimen does not influence the interpretation of the analysis for any of the populations. Mount Rose had the highest precipitation (Table 2) which has been increasing over time (20). Climate change may be driving the increasing precipitation in this area and lengthening the growing season, allowing for trees to accumulate more biomass. Increases in rainfall caused by climate change occur through the increased water holding capacity of a warming atmosphere (30). However, the interactions between temperature, precipitation, and plant traits are not always easily predictable (31). Although some interactions may extend the period for growth due to more favorable climatic conditions (32), climate islands may be forcing subalpine plants further up mountain tops. As trees increase in altitude there is less available land surface area, decreasing suitable habitat (33). The increase in leaf quantity we observed at Mount Rose may have ecological implications as it relates to leaf area index (LAI = total crown leaf area/ground area). Increased leaf quantity increases the total leaf area, which raises LAI. Increases in LAI have been shown to interact with hydrologic functioning by intercepting and sublimating snowpack (34). Increased canopy structure may also improve habitat for a variety of bird species (35), which are often used as surrogates for ecosystem diversity (36).

Our study indicates that there is variation in the leaf traits of contemporary populations of whitebark pine. Specifically, the number of leaves and the LMA from individuals at Jarbidge are lower than other populations. Further research is needed to understand the source of this variability, which could be due to local adaptation or genetic drift, as this population is relatively isolated and may experience limited gene flow. Additionally, the Jarbidge population has experienced high beetle mortality since 2008 (37), making the future of these trees uncertain. These unique differences at Jarbidge, along with the history of drought and mountain pine beetles, make this isolated population a research priority. A reciprocal transplant or common garden experiment would be able to measure the p of local adaptation within this species. Understanding how whitebark pine is changing across its range will be critical for future restoration success.


Table 1 - Latitude, longitude, and elevation
Table 1: Latitude, longitude, and elevation for each tree sampled in the field. Site ID indicated population at which individual trees were sampled, 30 trees were sampled per site.
Table 2 - List of herbarian specimens
Table 2: List of herbarium specimens used from the University of Nevada, Reno herbarium. Four populations were available for sampling (ES = Eastern Sierra, Jar = Jarbidge, MR = Mount Rose, RM = Ruby Mountains). Year collected is based on date present on herbarium sheet. Each sheet has a distinct herbarium code listed below. Specimen 11525 was unable to have LMA data collected because of concern of damaging the specimen.
Table 3 - Average Leaf Quantity and LMA
Table 3: Average leaf quantity and LMA (g/cm2), for both each contemporary and historical population.


We would like to thank Drs.Sarah Bisbing and Peter Weisberg at the University of Nevada, Reno for their advice and comments; Dr.Michele Slaton at the US Forest Service for her instrumental assistance in the early stages of this project. This project was supported by funding from the University of Nevada, Reno, and the Bristlecone Chapter of the California Native Plant Society (Alder2018).


[1] R. E. Keane et al., Gen. Tech. Rep. RMRS-GTR279. Fort Collins, CO: US Department of Agriculture, Forest Service, Rocky Mountain Research Station. 108, 279 (2012).
[2] S. F. Arno, R. J. Hoff, Gen. Tech. Rep. INT-GTR253. Ogden, UT: US Department of Agriculture, Forest Service, Intermountain Research Station 11, 253 (1989).
[3] D. F. Tomback, P. Achuff, Forest Pathology 40, 186– 225 (2010).
[4] K. Gibson et al., US Department of Agriculture Forest Service, Northern Region, Missoula, Montana 20, 8–20 (2008).
[5] R. E. Keane, P. Morgan, J. P. Menakis, Northwest Science 68 (1994).
[6] C. I. Millar et al., Canadian Journal of Forest Research 42, 749–765 (2012).
[7] N. McDowell et al., New Phytologist 178, 719–739 (2008).
[8] C. D. Allen et al., Forest Ecology and Management 259, 660–684 (2010).
[9] R. A. Andrus, B. J. Harvey, K. C. Rodman, S. J. Hart, T. T. Veblen, Ecology 99, 567–575 (2018).
[10] M. D. Abrams, M. E. Kubiske, K. C. Steiner, Tree Physiology 6, 305–315 (1990).
[11] S. A. Anjum et al., African Journal of Agricultural Research 6, 2026–2032 (2011).
[12] H. M. Poulos, G. P. Berlyn, The Journal of the Torrey Botanical Society 134, 281–289 (2007).
[13] J. L. Funk et al., Biological Reviews 92, 1156–1173 (2017).
[14] E. A. Leger, Global Change Biology 19, 2229–2239 (2013).
[15] G. R. Guerin, H. Wen, A. J. Lowe, Biology Letters 8, 882–886 (2012).
[16] L. Benomar et al., Frontiers in Plant Science 7, 48 (2016).
[17] C. A. Jones, C. C. Daehler, PeerJ 6, e4576 (2018).
[18] J. K. McKay, C. E. Christian, S. Harrison, K. J. Rice, Restoration Ecology 13, 432–440 (2005).
[19] A. D. Bower, S. N. Aitken, American Journal of Botany 95, 66–76 (2008).
[20] PRISM Climate Group, Oregon State University, created 12 Apr 2019.
[21] P. A. Vesk, M. Westoby, The New Phytologist 160, 7–14 (2003).
[22] I. J. Wright, P. B. Reich, M. Westoby, et al., Nature 428, 821–827 (2004).
[23] R. Poyatos, D. Aguad´e, L. Galiano, M. Mencuccini, J. Martı´enez-Vilalta, New Phytologist 200, 388–401 (2013).
[24] Intermountain Regional Herbarium Network. 2019. Downloaded on 30 August 2018.
[25] K. R. Hultine, J. D. Marshall, Oecologia 123, 32–40 (2000).
[26] R Core Team. 2019. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria.
[27] N. Br´eda, R. Huc, A. Granier, E. Dreyer, Annals of Forest Science 63, 625–644 (2006).
[28] H. Poorter, U. Niinemets, L. Poorter, I. J. Wright, R. Villar, New Phytologist 182, 565–588 (2009).
[29] E. G. de la Riva, M. Olmo, H. Poorter, J. L. Ubera, R. Villar, PLoS ONE 11, e0148788 (2016).
[30] K. E. Trenberth, Climate Research 47, 123–138 (2011).
[31] M. Westoby, D. S. Falster, A. T. Moles, P. A. Vesk, I. J. Wright, Annual Review of Ecology and Systematics 33, 125–159 (2002).
[32] T. Hwang et al., Water Resources Research 54, 5359– 5375 (2018).
[33] W. Romme, M. Turner, Conservation Biology 5, 373– 386 (1991).
[34] A. N. Gelfan, J. W. Pomeroy, L. S. Kuchment, Journal of Hydrometeorology 5, 785–803 (2004).
[35] C. E. Swift, K. T. Vierling, A. T. Hudak, L. A. Vierling, Canadian Journal of Remote Sensing 43, 231–243 (2017).
[36] F. Larsen, J. Bladt, A. Balmford, C. Rahbek, Journal of Applied Ecology 49, 349–356 (2012).
[37] G. Durham et al., Forest Pest Conditions in Nevada 2010 (Nevada Division of Forestry, 2011).