Acknowledgements - Paul Smith's College



Soil moisture levels’ impact on variation in microhabitat selection and distribution between shrub species along the riparian zones of the St. Regis River in Northern New York by Gregory DavisMentor: Dr. Justin WaskiewiczPaul Smith’s College December 1, 2020A paper submitted in partial fulfillment of the requirements for the degree of Bachelor of Science in Forestry: Forest Biology at Paul Smith’s College.Abstract: Willows (Salix spp.) are commonly found along riparian zones of northern latitude water bodies and are often used in riparian restoration and bank stabilization. However, not much is known about willows on a species level, especially among thee shrubby species native to North America. This study seeks to better understand the variation in the distribution of three willows (Salix bebbiana, S. discolor, and S. petiolaris) and two other shrub species (Alnus incana and Viburnum nudum) on a species level based on soil moisture. To reach this goal this study analyzes their distribution at several points along the shore of the St. Regis River. The study analyzed vertical distance from the river as a proxy for soil moisture, saturation volume as a proxy for porosity, and bulk density. It was found that soil moisture, as approximated by vertical distance was statistically significantly linked to the distribution of different shrub species. Porosity, bulk density, and distance along the river had no statistically significant relationship. The results supported the hypothesis that willows had the highest soil moisture requirements. The sample sizes were very small and only one willow, a S. petiolaris, was found in the study sites. AcknowledgementsI would like to thank my amazing mentor Dr. Justin Waskiewicz for all his assistance and expertise throughout this project. I would also like to thank Drs. Celia Evans and Craig Milewski and Mr. Stephen Langdon for their additional expertise in plant ecology and willow habitats. I would like to thank Dr. Janet Mihuc for assistance in preparing and reviewing this report. Lastly, I would like to thank Miss Caroline Granger, Miss Dallas Olsen, Mr. Jacob Keslar, Mr. Tyler Wright, and Mr. Daniel Power, for assisting in data collection. This project could not be completed to satisfaction without their help. Table of Contents TOC \o "1-3" \h \z \u Abstract: PAGEREF _Toc55504235 \h 2Acknowledgements PAGEREF _Toc55504236 \h 3List of Tables and Figures PAGEREF _Toc55504237 \h 5List of Appendices PAGEREF _Toc55504238 \h 6Introduction PAGEREF _Toc55504239 \h 7Methods PAGEREF _Toc55504240 \h 10Site Description PAGEREF _Toc55504241 \h 10Field Sampling PAGEREF _Toc55504242 \h 11Laboratory Analysis PAGEREF _Toc55504243 \h 12Statistical Analysis PAGEREF _Toc55504244 \h 13Results PAGEREF _Toc55504245 \h 14Data Collection PAGEREF _Toc55504246 \h 14Hypothesis 1: Comparing the composition of plant communities at each distance PAGEREF _Toc55504247 \h 15Hypothesis 2: Comparing each species with their vertical distance from the river as a proxy for soil moisture PAGEREF _Toc55504248 \h 17Hypothesis 3: Comparing composition of plant community to bulk density PAGEREF _Toc55504249 \h 17Hypothesis 4: Comparing composition of plant community with soil saturation volume as a proxy for porosity PAGEREF _Toc55504250 \h 19Discussion PAGEREF _Toc55504251 \h 22Hypothesis 1: Comparing the composition of plant communities at each distance PAGEREF _Toc55504252 \h 22Hypothesis 2: Comparing each species with their vertical distance from the river as a proxy for soil moisture PAGEREF _Toc55504253 \h 22Hypothesis 3: Comparing composition of plant community to bulk density PAGEREF _Toc55504254 \h 23Hypothesis 4: Comparing composition of plant community with soil saturation volume as a proxy for porosity PAGEREF _Toc55504255 \h 24General Observation: Wild-raisin’s near significance PAGEREF _Toc55504256 \h 24Likely causes of errors PAGEREF _Toc55504257 \h 24Literature Cited PAGEREF _Toc55504258 \h 26Appendix A: Maps PAGEREF _Toc55504259 \h 30Appendix B: Diagram of Search Design PAGEREF _Toc55504260 \h 32Appendix C: Description of Twigs of Target Species PAGEREF _Toc55504261 \h 33Appendix D: Tree Data PAGEREF _Toc55504262 \h 38Appendix E: Site and Soil Data PAGEREF _Toc55504263 \h 39Appendix F: Contingency Tables PAGEREF _Toc55504264 \h 40Appendix G: R Output PAGEREF _Toc55504265 \h 41Appendix H: R Code50List of Tables and FiguresTablesTable 1. Distances along the St. Regis river with data collected………………………………...14Table 2. Number of plants in each vertical distance group………………………………………14Table 3. Distances along the St. Regis river where each species was found…………………….15Table 4. Observed vertical distances of each species……………………………………………17FiguresFigure 1. Composition of plants that are alder along the St. Regis river………………………...16Figure 2. Composition of plants that are wild-raisin along the St. Regis river…………………..16Figure 3. Composition of plants that are willows along the St. Regis river……………………..17Figure 4. Composition of plants that are alder compared to bulk density……………………….18Figure 5. Composition of plants that are wild-raisin compared to bulk density…………………19Figure 6. Composition of plants that are willow compared to bulk density……………………..19Figure 7. Composition of plants that are alder compared to soil saturation volume…………….20Figure 8. Composition of plants that are wild-raisin compared to soil saturation volume………21Figure 9. Composition of plants that are willow compared to soil saturation volume…………..21List of AppendicesAppendix A. Maps…..…………………………………………………………………………29Appendix B. Diagram of Search Design.………..……………………………………………..30Appendix C. Description of Twigs of Target Species …………………………………………31Appendix D. Tree Data…………………………………………………………………………38Appendix D. Site and Soil Table……………………………………………………………….39Appendix D. Contingency Tables ..…………………………………………………………….40Appendix D. R Output….………………………………………………………………………41Appendix D. R Code…………………………………………………………………………....50IntroductionWillows (Salix spp. L.) are difficult to properly identify to the species level (Harlow et al., 1979), so they are often studied on the genus level. Willows typically are either generalists or wetland specialists, with generalists still occupying high moisture level sites more often than low moisture level sites (Savage et al., 2009). Since willows all share a relationship for high moisture related to germination requirements (Kransy et al., 1988) it is logical to study them at the genus level taxonomic group. Willow’s relationship to soil moisture level and germination is that their seeds are low viability, meaning they are unsuccessful in sexual reproduction in sites that do not meet their soil moisture requirements for imbibing (Kransy et al., 1988). The focus on the genus level in the literature has led to a lack of knowledge of the ecology of specific Salix species and microhabitats. There is a gap in understanding the fine scale distribution of willows (Amlin & Rood, 2002). Silvertown et al. (2005) found in meadow ecosystems that related species occupy similar habitats on a coarse scale, but on a finer scale this does not hold true. Two main factors are often considered when looking at the composition of plant communities: habitat filtration that “selects” species more suited for the abiotic factors of the habitat and niche differentiation which excludes species that exist within the same niche (Maire et al., 2012). For example, habitat filtration causes willows to generally be found in riparian zones, while niche differentiation excludes species coexisting if they have all the same requirements on a fine scale. The habitat filtration relates to the beta niche (coarse scale) and excludes species from a site if they are unsuitable to the general site conditions such as moisture and nutrient levels from a coarse lens, while niche differentiation allows multiple species that are phylogenetically related to coexist by occupying different alpha niches (fine scale) and having different microhabitat requirements (Silvertown et al., 2005). One abiotic factor that determines the alpha and beta niches of a species is soil moisture content. Moisture is one of four main abiotic factors required for release of dormancy, the others being light, temperature, and nutrient levels (Finch-Savage & Leubner-Metzger, 2006). As seeds imbibe water, the stages of dormancy progress eventually to termination of dormancy and initiation of germination. The level of moisture necessary for dormancy to terminate differs on an individual basis and different species have significantly different requirements for dormancy termination (Finch-Savage & Leubner-Metzger, 2006). Water is also used in multiple functions through the life of the mature plant, at differing levels among different species of willows (Savage et al., 2009; Amlin & Rood, 2002; Martinez-Ferri et al., 2000). This differing requirement of moisture is often determined through physiological differences that correspond with differing life strategies (Savage et al., 2009; Amlin & Rood, 2002; Martinez-Ferri et al., 2000; Pockman & Sperry, 2000). For example, willows that are more generalist, such as S. discolor, have more resistance to photorespiration, energy inefficient oxidation of RuBisCo which can lead to increased toxicity within the cell, while willows that are more wetland specialist, such as S. pyrifolia, have a growth strategy of quick unrestrained growth that limits their photoprotective, protection against photorespiration and radiation, capabilities, such as photosenescence where leaves are abscised under high light drought conditions (Savage et al., 2009). This difference in physiological traits corresponding with life strategies and moisture requirements was also found to correspond with distribution of species on at least a coarse habitat level (Pockman & Sperry, 2000). The genus Salix in general requires a germination site that is consistently saturated and thus is typically found in moist areas occupying wetlands and riparian zones (Harlow et al., 1979). Compared to other species of Salicaceae, willows were found to have physiological traits correspondent to life strategies adapted to habitats with high moisture (Pockman & Sperry, 2000). One laboratory study (Savage et al., 2009) has found that willows differentiate at a species level with these physiological traits. However, there has been little field research looking at variation of distributions on an alpha niche level between different species of willow (Amlin & Rood, 2002). Within the Adirondack Park, willow populations are very sparse, and the areas that could have willows are dominated by speckled alder (Alnus incana) and wild-raisin (Viburnum nudum). One river that starts within the Adirondacks before exiting is the St. Regis River. This river passes through multiple microsites as it flows south to north. Willow populations are believed to dominate the northern, downriver portions while alder and wild-raisin dominate the southern, upriver portions (C. Milewski, personal communication, September 2020). Of all willow species explored in this study, S. bebbiana is the one most commonly found in upland sites. It is typically classified as a facultative wetland plant, but it is classified as a facultative upland plant in some areas (United States Department of Agriculture [USDA], 2019). S. petiolaris is the one most commonly found in the wettest sites. It is typically classified as a facultative wetland plant, but it is classified as an obligate wetland plant in some areas (USDA, 2019). S. discolor is only ever classified as a facultative wetland plant (USDA, 2019). These classifications were developed by the U.S. Army Corps of Engineers, the Fish and Wildlife Service (FWS), the Environmental Protection Agency, and the Natural Resource Conservation Service in 2012 based on taxonomic and distribution data from the Biota of North America Project and legacy information from the FWS (USDA, 2019). Classifications were updated in 2014 (USDA, 2019). This study will explore how soil moisture levels impact variation in microhabitat distribution among the following species of willow: S. bebbiana, S. discolor, S. petiolaris, as well as A. incana and V. nudum found in riparian habitats. It is done to better understand willows on a species level and to explore reasons why willows are not found in the Adirondack park in high numbers. Hypothesis one is that riparian shrub populations change down river, with willows found mainly in the downriver portions. Hypothesis two is that willows and other riparian shrubs will occupy different soil moisture contents with willows occupying the highest, and that among willows, S. petiolaris will occupy sites with a relatively higher soil moisture content while S. bebbiana will occupy sites with a relatively lower soil moisture content, and S. discolor will be found in between. Hypothesis three is that shrub species composition will differ based on soil bulk density with willows occupying sites with the lowest bulk density. Hypothesis four is that shrub species composition will differ based on porosity with willows occupying sites with the highest porosity. MethodsSite DescriptionThe study was conducted along the St. Regis River in northern New York. Sampling started at the dam beyond the outlet of Lower St. Regis Lake into the St. Regis river at 44.4311?, -74.2995? and ended at the outlet of the St. Regis River into the St. Lawrence at 45.0000?,-74.6370?, Google, 2020). The St. Regis River is a tributary river of the St. Lawrence River whose head water is within the Adirondack Park. The St. Regis River is approximately 125 km long. This site is a riparian ecosystem within a northern hardwood forest (Erye, 1980). The average annual temperature is 4° Celsius (Intellicast, 2018). The average monthly precipitation is 96 cm (Intellicast, 2018).The soils in the study area are predominantly spodisols, with typical soils along the riparian zones being well drained, loamy sands (Natural Resource Conservation Service [NRCS], 2019). Some parts of the study area are boggy swamps, marshes, and other wetlands. These sites have hydric soils with higher peat content and lower drainage (NRCS, 2019). Other plants found in this area make up a variety of wetland species including leatherleaf (Chamaedaphne calyculata), Labrador tea (Rhododendron groenlandicum), sweet gale (Myrica gale) a variety of sedges (Carex spp.), and a variety of trees including white pine (Pinus strobus), red maple (Acer rubrum), tamarack (Larix larcina), and white birch (Betula papyrifera) (New York Flora Association [NYFA], 2020). The riparian shrubs that are found along this river are speckled alder (Alnus incana), wild-raisin (Viburnum nudum), and willows (Salix spp.) (NYFA, 2020). All three are facultative wetland plants (United States Department of Agriculture [USDA], 2019). The native willows that have been observed in the study area include Bebb’s willow (Salix bebbiana), American pussy willow (S. discolor), heartleaf willow (S. cordata), shining willow (S. sericea), meadow willow (S. petiolaris), snow bed willow (S. herbacea), black willow (S. nigra) and balsam willow (S. pyrifolia) (NYFA, 2020; Kudish, 1975). Laurel willow (S. pentreda), crack willow (S. fragilis), basket willow (S. purpea), and weeping willow (S. babylonia) are all non-native species present, most of which are planted (Kudish, 1975). The other species are all native. With the exception of black willow which is a tree-willow, the native willows are shrub-willows. Bebb’s willow, American pussy willow, and meadow willow are the most common, widespread, native species (NYFA, 2020), therefore, the study will focus on these three species of willows. (See Appendix C for a description of twigs of target species)Field Sampling A point was selected for study approximately every 8 km along the river, +/-1 km to account for accessibility of the site. The points were determined using GPS services on Avenza Maps (Avenza, 2020). From each point I flipped a coin to either travel upstream or downstream and began a search until target plants were found or a total of 150m was traveled. A total of 15 points were selected for sampling; only seven points were sampled due to posted signs and other accessibility problems (See Appendix A for maps of sites).At each point, the first nine target plants encountered on the search were identified, and percent composition was determined among these nine individuals. Individual plants were defined by root systems, and not by stems. This was to reduce skewing the data since the target species’ form are shrubs with multiple stems. In choosing the target plants to measure I found the closest target plant, then travelled another 10m to the next target plant, thus reducing spatial-autocorrelation and the likelihood of non-independence among individuals within a site. The first three plants were within 1m vertical distance of the river. The second set of three plants were between 1m and 2m vertical distance from river; these plants are 10m apart, but the search started at the same point as the previous search. The third set of three plants were over 2m vertical distance from the river; these plants are 10m apart, but the search started at the same point as the previous searches relative to distance along the length of the river (See Appendix B for a diagram of search design). A target plant was defined as a speckled alder, wild-raisin, Bebb’s willow, American pussy willow, or meadow willow over 1.5 m in height. Twig specimens of each willow were collected to have the field identification confirmed in lab. Twig specimens were labeled 1-9 followed by their site distance along the length of the river. Percent composition at each point was calculated by dividing the number of plants of a given species by nine. The higher the composition value, the larger proportion of the plant community is comprised of that target plant.Each sampled plant had the vertical distance to the river measured. The distance was calculated using trigonometry from field measurements of horizontal distance and slope from the plant to the river. Distance to the river served as a proxy for soil moisture content (Sarvade et al., 2016). A proxy was used for soil moisture to avoid errors due to weather-based fluctuation. A soil sample was cored (2 cm diameter and variable height) and volume calculated using the cylindrical volume (V= πr2h). This core was taken next the central plant of the nine-plant grid. The single core represented the entire point along the river. Multiple cores were taken if needed to reach a total of 15 cm3. These additional cores were taken next to the original core. This sample was used to determine bulk density and porosity. Soil samples were labeled with the distance traveled downriver in kilometers from the dam where sampling began. In sites that were accessible, but had no target plants within the 150m search, the core was taken at what would be the starting point approximately 1m to 2m from the bank of the river. Laboratory AnalysisAfter field collection, each soil sample was oven dried (NRCS, 2001). Oven dried is defined by no longer losing mass while drying (NRCS, 2001). After drying, the dry mass was recorded. Bulk density was determined by dividing the dry mass by the field volume. Porosity was estimated by proxy using saturation volume. Saturation volume was determined by first separating out three sub-samples of dry soil, each having a volume of 5 cm3, from the original soil sample. Each of these sub-samples was placed into a graduated cylinder, one sub-sample at a time. Water was placed into a different graduated cylinder to determine a starting volume. Water was slowly poured from the water cylinder into the soil sample until it was fully saturated. Full saturation was defined by having standing water pool on top of the soil. The finished volume in the water cylinder of water was subtracted from the starting volume to determine the volume of water required for saturation. The saturation volume of the three sub-samples were represented by calculating an arithmetic mean. The willow twig specimens were identified with the help of a dichotomous key (Native Plant Trust, 2020). The willows were identified in the lab rather than the field so that a microscope could assist with the small details necessary in identification of willows to the species level.Statistical AnalysisNormality was tested using Shapiro-Wilk test. If the data was normally distributed a linear regression test between percent composition of the plant community composed each species and distance along the river, percent composition of the plant community composed of each species and bulk density, and percent composition of the plant community composed of each species and saturation volume will be conducted. If not, then they would have been compared through a Spearman test. Species and vertical distance were compared using both a Fisher Test and Chi-Squared Test. Statistical analyses were carried out using R (R Core Team, 2019). ResultsData CollectionSoil data were collected at 0 km, 8 km, 48 km, 64 km, 72 km, 88 km, and 120 km (Table 1). Vegetation data were collected at 8 km, 48 km, 64 km, and 88 km (Table 1). All other sites were inaccessible. Alder (Alnus incana) was present at all sites where vegetation data were collected (Table 3). Wild-raisin (Viburnum nudum) was present at 8 km and 48 km (Table 3). Only one willow was sampled, a meadow willow (Salix petiolaris), at 88 km (Table 3). At 48 km too many plants were sampled in vertical distance group (VDG) 1 (Table 2). At 48 km there were not enough target plants in either VDG 2 or VDG 3 to reach the desired sample size (Table 2). At 64 km a full sample size was reached but skewed toward VDG 1 (Table 2). At 88 km there were too few target plants to reach the desired sample size (Table 2). Table 1. Distances along the St. Regis river with data collection. “X” marks sites where data were collected, “-“ marks sites where data were not available. Sites with soil data, but not vegetation data had no target species in the search area. (n=15).Distance (km) along river from dam at St. Regis Lake outletData081624324048566472808896104112120soilxx----x-xx-x---xvegetation-x----x-x--x----Table 2. Number of target trees measured at each vertical distance group (VDG). VDG1 was under 1m vertical distance from the plant to the river. VDG2 was between 1m and 2m vertical distance from the plant to the river. VDG3 was over 2m vertical distance from the plant to the river. In field sampling I tried to have 3 plants in each group, with a total of nine plants. Distance(km)Vertical Distance Group 1Vertical Distance Group 2Vertical Distance Group 38333484116454088200Table 3. Species present at each distance. “-“ marks sites where no data were collected. No Salix discolor or S. bebbiana were found. Distance (km)Species081624324048566472808896104112120S. petiolaris ------1---Alnus incana3----5-9-1---Viburnum nudum?6----1-??-?---? Hypothesis 1: Comparing the composition of plant communities at each distanceThe p-value for the Shapiro-Wilk test looking at the distribution of the composition of alder was 0.059. This is within the range to not reject the null hypothesis of normality but is on the low end. The distribution of the composition of wild-raisin along the river was not normal (Shapiro-Wilk: p=<0.001). The distribution of the composition of willow along the river was not normal (Shapiro-Wilk: p=<0.001). Due to the low normality of the distribution of the composition of alder and the nonnormal distribution of the composition of wild-raisin and willow, they were analyzed with a Spearman correlative test (Figure 1; Figure 2; Figure 3). The relationship between distance along the river and the composition of shrubs by any species was not statistically significant (Figure 1; Figure 2; Figure 3). The relationship between distance along the river and the composition shrubs by wild-raisin is nearly statistically significant (Figure 2). Figure 1. Composition of plants that are alder along the St. Regis river. Data were taken at ~8 km intervals. (n=4, p=.75). Figure 2. Composition of plants that are wild-raisin along the St. Regis river. Data were taken at ~8 km intervals. (n=4, p=0.051, R2=0.89). Figure 3. Composition of plants that are willows along the St. Regis river. Data were taken at ~8 km intervals. (n=4, p=.225). Hypothesis 2: Comparing each species with their vertical distance from the river as a proxy for soil moistureDue to the small sample size, a Fisher test was conducted to compare observed vertical distance of each species. There was a statistically significant relationship between vertical distance and species (Table 4). Meadow willow (Salix petiolaris) had the lowest average vertical distance (0m). Wild-raisin (Viburnum nudum) had the highest average vertical distance (2.7m). Alder (Alnus incana) was in the middle (0.5m). Meadow willow was only found directly on the river, so it has an average distance of 0m.Table 4. Observed species at each vertical distance group (df=4, p=<0.001). See Appendix F for greater detail.?Vertical Distance Groupspecies1 (14)2 (8)3 (4)Willow (1)100Alder (17)1340Wild Raisin (8)044Hypothesis 3: Comparing composition of plant community to bulk densityBulk density was normally distributed (Shapiro-Wilk: p=0.357). The p-value for the Shapiro-Wilk test looking at the distribution of the composition of alder was 0.059. This is within the range to not reject the null hypothesis of normality but is on the low end. The distribution of the composition of wild-raisin along the river was not normal (Shapiro-Wilk: p=<0.001). The distribution of the composition of willow along the river was not normal (Shapiro-Wilk: p=<0.001). Due to the low normality of the distribution of the composition of alder and the nonnormal distribution of the composition of wild-raisin and willow, they were analyzed with a Spearman correlative test (Figure 4; Figure 5; Figure 6). The relationship between distance along the river and the composition of shrubs by any species was not statistically significant (Figure 4; Figure 5; Figure 6). The relationship between distance along the river and the composition of wild-raisin is nearly statistically significant (Figure 5).Figure 4. Composition of alder compared to bulk density (g/cc) (n=4, p=0.333). Data were taken at ~8km intervals. Figure 5. Composition of wild-raisin compared to bulk density (g/cc) (n=4, p=0.051, R2=.81). Data were taken at ~8km intervals.Figure 6. Composition of willow compared to bulk density (g/cc) (n=4, p=0.742). Data were taken at ~8km intervals.Hypothesis 4: Comparing composition of plant community with soil saturation volume as a proxy for porosityDistribution of soil saturation volumes were normal (Shapiro-Wilk: p=0.201). The p-value for the Shapiro-Wilk test looking at the distribution of the composition of alder was 0.059. This is within the range to not reject the null hypothesis of normality but is on the low end. The distribution of the composition of wild-raisin along the river was not normal (Shapiro-Wilk: p=<0.001). The distribution of the composition of willow along the river was not normal (Shapiro-Wilk: p=<0.001). Due to the low normality of the distribution of the composition of alder and the nonnormal distribution of the composition of wild-raisin and willow, they were analyzed with a Spearman correlative test (Figure 7; Figure 8; Figure 9). The relationship between distance along the river and the composition of shrubs by any species was not statistically significant (Figure 7; Figure 8; Figure 9). The relationship between distance along the river and the composition of wild-raisin is nearly statistically significant, but the statistics describe the composition being explained by saturation volume only 56% of the time, which is not significant (Figure 5).Figure 7. Composition of alder compared to soil saturation volume, used as a proxy for porosity (n=4, p=0.75). Data were taken at ~8km intervals.Figure 8. Composition of wild-raisin compared to soil saturation volume, used as a proxy for porosity (n=4, p=0.051, R2=.56). Data were taken at ~8km intervals.Figure 9 . Composition of willow compared to soil saturation volume, used as a proxy for porosity (n=4, p=0.225). Data were taken at ~8km intervals. DiscussionHypothesis 1: Comparing the composition of plant communities at each distanceThe hypothesis, that riparian shrub populations change traveling down river, with willows found mainly in the downriver portions, was not statistically supported by this study. The only nearly statistically significant relationship with distance traveled down river was with wild-raisin. The relationship between wild-raisin and distance had a p value of 0.051 (Figure 2). The evidence that led to this hypothesis was anecdotal (C. Milewski, personal communication, September 2020). However, the issue arises from such a small sample size. While a general observation of willows increasing down river can be seen, there is not a large enough sample size to demonstrate this with any statistical vigor. Outside of the sampling areas, there appeared to be an increase in willows, but that cannot be confirmed with the sampling. Hypothesis 2: Comparing each species with their average vertical distance as a proxy for soil moistureThe hypothesis, that willows and other riparian shrubs would occupy different soil moisture contents with willows occupying the highest, and that among willows S. petiolaris would occupy sites with a relatively higher soil moisture content while S. bebbiana would occupy sites with a relatively lower soil moisture content, and S. discolor would be found in between, was supported by this study. This relationship was statistically significant (Table 4). There was only one willow, but it was found at the water line. This supports the idea that willows require the highest comparative soil moisture among shrubs. This is supported by published laboratory experiments (Hughes et al., 1997). This is supported by published field experiments (Savage et al., 2009; Amlin & Rood, 2002; Martinez-Ferri et al., 2000; Pockman & Sperry, 2000). This is supported by the silvics of willow (USDA, 2019; Finch-Savage & Leubner-Metzger, 2006; Harlow et al., 1979). The sample size for this analysis appears to be the largest (n=26). However, this is a pseudoreplication issue. Of the four sites studied, none were proper replicates of each other. At most sites, more plants in vertical distance group (VDG) 1 were sampled than in VDG2 or VDG3. This skewed the sampling towards low vertical distance. The other issue is that only two sites (8km and 64km) had all nine plants sampled. Therefore, the sample size is not the 26 plants, but rather the 4 plots, as are the sample sizes for the other analyses. That it was a meadow willow that was found at the water table may support the hypothesis that it is the most water needy willow; this cannot be well supported with this study as it is the only willow sampled. That meadow willow is the most water needy would support the wetland classifications of the USDA (2019). The supported hypothesis that there is a difference in soil moisture levels supports the idea of niche differentiation (Maire et al., 2012). Each soil moisture level is occupied by a different species (Table 4). If there is niche differentiation, the species could survive in a greater range of environmental conditions than they do at any given site if the niches described by those environmental conditions were unoccupied (McGill et al., 2006). This situation is a possible explanation for the overlap in distances between the species. Each species has a wider fundamental niche but is being out competed in certain sites causing a smaller realized niche in those sites. Another cause for the overlap in distances is because spatial placement is not solely caused by soil moisture levels (Finch-Savage & Leubner-Metzger, 2006). Another explanation for the overlap in distances is that this study ended up looking more at beta niche than alpha niche, simply due to the scale. Alpha niche is an important, though controversial explanation as to how coexistence within niches may exist (Wilson, 2011). Alpha niche has been shown through observational, but not empirical studies (Wilson, 2011). Hypothesis 3: Comparing composition of plant community to bulk densityThe hypothesis, that shrub species composition will differ based on soil bulk density with willows occupying sites with the lowest bulk density, was not supported by the study. This is a statistically insignificant relationship (Figure 5; Figure 6; Figure 7). The bulk density did have a normal distribution between the sites, however there was not a large range in bulk densities (Figure 5; Figure 6; Figure 7). This may show that bulk density is not a major condition that impacts plant distribution. Soil conditions have an impact on the distribution of plants that reproduce by seed (Kransy et al., 1988). The lack of a relationship is not supported by Labelle and Kammermeier’s (2019) study on bulk density and balsam fir (Abies balsamea) which showed a negative relationship between high bulk density and plant growth. One possible explanation is the different conditions riparian species face compared to forest species (Latella et al., 2020; Uria-Diez et al., 2014). Another explanation is that the bulk densities were too low and too close together to show a difference (Labelle & Krammermeier, 2019). Bulk density and porosity are closely linked (Fischer et al., 2015). This may be why the relationships with bulk density and with porosity in this study are so close.Hypothesis 4: Comparing composition of plant community with soil saturation volume as a proxy for porosityThe hypothesis, that shrub species composition will differ based on porosity with willows occupying sites with the highest porosity, was not supported by the study. This is a statistically insignificant relationship (Figure 5; Figure 6; Figure 7). This may show that porosity is not a major condition that impacts plant distribution. Soil conditions have an impact on the distribution of plants that reproduce by seed (Kransy et al., 1988). Porosity should have a positive relationship with plants that require soil moisture as it increases infiltration (Fischer et al., 2015). However, if the soil moisture levels need to be constant, having high infiltration may cause a negative relationship as it increases drainage (Fischer et al., 2015). This counteractive relationship of porosity, that it can both increase and decrease soil moisture levels at different points, may be the cause of the lack of a statistically significant relationship. The lack of a relationship could also be a result of not having much data. Porosity and bulk density are closely linked (Fischer et al., 2015). This may be why the relationships with porosity and with bulk density in this study are so close. General Observation: Wild-raisin’s near significanceIn each regressive analysis the relationship with the composition of shrubs by wild-raisin was nearly statistically significant (Figure 2; Figure 5; Figure 8). Soil conditions have an impact on the distribution of plants that reproduce by seed (Kransy, et al., 1988). The reason wild raisin may have the nearest statistically significant relationship with soil conditions is that wild raisins primarily reproduce by seed while alders and willows readily reproduce asexually (Flaherty, et al., 2018; Funk, 1990; Harrington, 1990; Kransy, et al., 1988). It is important to note there were only eight wild-raisins sampled, and none of the relationships were truly statistically significant (Figure 2, Figure 5, Figure 8). Even if it were statistically significant, the relationship between composition of shrubs composed of wild-raisin compared to porosity had a R2 value of 0.56 (Figure 7). This shows that if the relationship was statistically significant, the composition of shrubs composed of wild-raisin could only be explained by porosity about half of the time. Likely causes of errorsThe main issue in data collection was the lack of data to collect. The initial issue was a lack of willows within earlier study sites. To remedy this, the study design was changed to study changes along the St. Regis River. This led to other issues in access to the study site. Most of the St. Regis River was inaccessible due to private property. From just after 8 km, until 48 km, the river was posted and therefore inaccessible. After entering the St. Lawrence valley, the private property issue did not go away; in fact only two sites, 88 km and 120 km, were accessible (See Appendix A for maps of sites). This meant that the sample sizes were extremely small. Even for hypothesis two, which is statistically significant, the sample size is so small that the statistics could show a false relationship. A way to remedy this, at least for hypothesis two, where each plant was independent, would be to sample more than nine plants per site. Overall, the small sample sizes could be showing false relationships; and variations and outliers can be amplified leading to sampling bias and skewed data. Estimated vertical distances in the field were not accurate when determining VDG compared to mathematically derived distances done in the lab. This may have led to a skewing of the data towards near river plants. This may not be significant since the order of the plants by vertical distance would not shift after removing the extra plants in VDG1. It is important to note that the statistical significance may shift due to this skewing, but the statistical significance is so high, there would not be much of a change. A way to remedy this would be to calculate vertical distances in field before choosing which trees to sample, rather than estimating which VDG they belong to. Literature CitedAdirondack Forever Wild (2020). Shrubs of the Adirondacks: northern wild raisin (Viburnum nudum L., var. cassinoides). Retrieved from ADDIN ZOTERO_BIBL {"uncited":[],"omitted":[],"custom":[]} CSL_BIBLIOGRAPHY Amlin, N. M. and Rood, S. B. (2002). Comparative tolerances of riparian willows and cottonwoods to water-table decline. Wetlands 22, 338–346.Avenza (2020) Avenza Maps. Avenza Groups Inc. , D (2020). Leaves: Salix bebbiana [Photograph]. Nature Plant Trust. , P. M. (2020). Next year's catkins developing with this year's fruit [Photograph]. Minnesota Wildflowers , F.H. (1980). Forest Cover Types of the United States and Canada. Washington, DC: Society of American ForestersFinch-Savage, W. E., & Leubner-Metzger, G. (2006). Seed dormancy and the control of germination. The New Phytologist, 171(3), 501–523.Fischer, C., Tischer, J., Roscher, C., Eisenhauer, N., Ravenek, J., Gleixner, G., … Hildebrant, A. (2015). Plant species diversity affects infiltration capacity in an experimental grassland through changes in soil properties. Plant & Soil 397, 1-16. Flaherty, K. L., Rentch, J. S., and Anderson, J. T. (2018). Wetland seed dispersal by white-tailed deer in a large freshwater wetland complex. AoB Plants 10(1), 1-8.Funk, D. T. (1990). European alder, in R.M. Burns and B.H. Honkala (Eds.)., Silvics of North America: vol. 2. Hardwoods. Washington DC: U.S. Department of Agriculture, Forest Service. Retrieved from volume_2/silvics_ v2.pdf Google (2020) Google Earth Pro (Version 7.3). Google. , A (2020a). Leaves: Salix discolor [Photograph]. Nature Plant Trust . species/salix/discolorHaines, A (2020b). Leaves: Salix petiolaris [Photograph]. Nature Plant Trust. . species/salix/petiolaris/Haines, A (2020c). Leaves: Viburnum nudum [Photograph]. Nature Plant Trust . species/viburnum/nudumHarlow, W. M., Harrar, E. S., and White, F. W. (1979). Textbook of Dendrology. New York, NY: McGraw-HillHarrington, C. A. (1990). Red alder, in R.M. Burns and B.H. Honkala (Eds.)., Silvics of North America: vol. 2. Hardwoods. Washington DC: U.S. Department of Agriculture, Forest Service. Retrieved from volume_2/silvics_ v2.pdf Hughes, F. M. R., Harris T., Richards, K., Pautou, G., Hames, A. E., Barsoum, N., … Foussadier, R. (1997). Woody riparian species responses to different soil moisture conditions: laboratory experiments on Alnus incana (L.) Moench. Global Ecology and Biogeography Letters 6, 247–256.Intellicast (2018). Historical averages of Paul Smiths, NY, USA weather. Received from , M. (1975). Paul Smith’s Flora: A preliminary vascular flora of the Paul Smiths-Saranac Lake area, the Adirondacks, New York, with notes on the climate, geology, and soils. Paul Smiths, NY: Paul Smith’s College.Krasny, M. E., Vogt, K. A., and Zasada, J. C. (1988). Establishment of four Salicaceae species on river bars in interior Alaska. Holarctic Ecology 11: 210-219.Labelle, E. R. and Kammermeier, M. (2019). Above- and belowground growth response of?Picea abies?seedlings exposed to varying levels of soil relative?bulk?density. European Journal of Forest Research 138(4):705-722Latella, M., Bertagni, M. B., Vezza, P., and Camporeale, C. (2020). An integrated methodology to study riparian vegetation dynamics: From field data to impact modeling. Journal of Advances in Modeling Earth Systems 12, 1-23. Maire, V., Gross, N., Borger, L., Proulx, R., Wirth, C., Pontes, L. S., Louault, F. (2012). Habitat filtering and niche differentiation jointly explain species relative abundance within grassland communities along fertility and disturbance gradients. The New Phytologist, 196(2), 497-509.Martínez-Ferri, E., Balaguer, L., Valladares, F., Chico, J. M., and Manrique, E. (2000). Energy dissipation in drought-avoiding and drought-tolerant tree species at midday during the Mediterranean summer. Tree Physiology 20, 131–138. McGill, B. J., Enquist, B. J., Weiher, E., and Westoby, M. (2006). Rebuilding community ecology from functional traits. Trends in Ecology and Evolution 21, 178–185.Minnesota Wildflowers (2020a) Bebb’s willow. Retrieved from . Info/shrub/bebbs-willowMinnesota Wildflowers (2020b) Meadow willow. Retrieved from . info/shrub/meadow-willow Minnesota Wildflowers (2020c). Pussy willow. Retrieved from . Info/shrub/pussy-willowMinnesota Wildflowers (2020d). Speckled alder. Retrieved from . info/shrub/speckled-alderNative Plant Trust (2020) Go Botany (Version 3.4) Native Plant Trust. Resource Conservation Service (2001). Soil Quality Test Kit Guide. Washington, DC: USDA. Natural Resource Conservation Service (2019). Web Soil Survey. Received from .sc.egov.App/HomePage.htmNew York Flora Association (2020). New York Flora Atlas. Received from , W. T. and Sperry, J. S. (2000). Vulnerability to xylem cavitation and the distribution of Sonoran desert vegetation. American Journal of Botany 87, 1287–1299. R Core Team (2019) R (Version 3.6.2). R Foundation. , S., Gupta, B. and Singh, M. (2016). Composition, diversity and distribution of tree species in response to changing soil properties with increasing distance from water source: A case study of Gobind Sagar Reservoir in India. Journal of Mountain Science 13(3), 522-533. Savage, J. A., Cavender-Bares, J., and Verhoeven, A. (2009). Willow species (genus: Salix) with contrasting habitat af?nities differ in their photoprotective responses to water stress. Functional Plant Biology 36, 300-309. Silvertown, J., McConway, K., Gowing, D., Dodd, M., Fay, M. F., Joseph, J. A., and Dolphin, K. (2005). Absence of phylogenetic signal in the niche structure of meadow plant communities. Proceedings of the Royal Society Biological Sciences 273(1582), 39-44. United States Department of Agriculture (2019). PLANTS database. Received from , J., Gazol, A., and Ibanez, R. (2014). Drivers of a riparian forest specialist (Carex remota, Cyperaceae): It is not only a matter of soil moisture. American Journal of Botany, 101(8), 1286-1292.Wilson, J. B. (2011). The twelve theories of co-existence in plant communities: the doubtful, the important and the unexplored. Journal of Vegetation Science 22, 184–195.Appendix A: MapsMap of all sites from 0 km at the dam on Lower St. Regis to 120 km down river, sites are approximately 8 km apart. Map of sites where sampling was successful. The large black points (0,72, and 120 km) had no vegetation data. Small blue points (8, 48, 64, and 88 km) had target species present with 150m of the point. Appendix B: Diagram of Search DesignDiagram of the search design. Red Stars are the starting points of each search. Green dots are target plants, at least 10m between each target plant. VDG1 is under 1m, VDG2 between 1m and 2m, VDG is over 2m. Search starts at the first plant found and does not go over 150m. Soil core is taken at plant 5. Appendix C: Description of Twigs of Target SpeciesAlnus incana (picture credit: Dziuk, 2015): leaves ovoid; prominently veined; cone-like structures; prominent lenticles; finely toothed leaves (Minnesota Wildflowers, 2020d).Salix bebbiana (picture credit: Cameron, 2020): elliptical leaves; stipules are serrate, absent in young leaves; leaf margins complete or wavy; red-tinged, white hairs on young growth, less hairy with age; stems reaching over 20 cm (Minnesota Wildflowers, 2020a)Salix discolor (picture credit: Haines, 2020a): leaves elliptical to urn shaped; stipules persistent, but are obscure on new leaves; stems green when young, reddish brown to yellowish brown by second year; stems are hairless by second year, some hair on twigs; leaves hairy when young, less hairy with time (Minnesota Wildflowers, 2020c). Salix petiolaris (picture credit: Haines, 2020b): leaves similar to Salix discolor; leaves more lanceolate in shape; stipules are not persistent; greenish to grey bark; up to 3.8 cm in diameter; leaves start hairy, lose hair with age; twigs start greenish turn reddish by second year (Minnesota Wildflowers, 2020b)Viburnum nudum (picture credit: Haines, 2020): opposite branching; lanceolate; margins are nearly complete, some fine serration; clumps of white flowers on the ends of branches; fruit are berries that ripen from red to a dark dried fleshy fruit (Adirondack Forever Wild, 2020). Appendix D: Tree DataDistance (km)tree numberspeciesvertical distance (m)notes 81ALIN0.0082ALIN0.0083ALIN0.0084VINU1.3485VINU1.4986VINU1.1687VINU4.7888VINU5.1789VINU3.51481ALIN0.85on bank hangin over bank482ALIN0.23483ALIN0.03484ALIN0.40485VINU1.12486VINU3.13641ALIN0.06642ALIN0.00643ALIN0.17644ALIN0.62645ALIN1.14646ALIN0.92647ALIN1.04648ALIN1.71649ALIN1.12881ALIN0.13882SAPE0.00?ALIN=Alnus incanaVINU=Viburnum nudumSAPE=Salix petiolarisAppendix E: Site and Soil dataDistance (km)0848647288120Comp. A. 0.000.330.561.000.000.110.00Comp. V. 0.000.670.110.000.000.000.00Comp. W.0.000.000.000.000.000.110.00Soil Height (cm)36.5839.6239.6227.4335.5339.6218.29Soil Volume (cc)104.25112.94112.9478.1995.56112.9452.12Soil Dry Mass (g)99.70104.00114.1095.50128.80135.1046.20Bulk Density (g/cc)0.960.921.011.221.351.200.89Saturation Volume 1 (mL)1.002.002.201.602.801.602.00Saturation Volume 2 (mL)0.602.602.601.602.401.802.80Saturation Volume 3 (mL)0.802.802.402.400.801.002.60Avg. Saturation Volume (mL)0.802.472.401.872.001.472.47Comp.= CompositionA.=AlderV=Wild-raisinW=Willows (combined)Appendix F: Contingency Tables?Vertical Distance Groupspecies1 (14)2 (8)3 (4)Willow (1)0.5380.3080.154Alder (17)9.6925.5382.769Wild Raisin (8)3.7692.1541.077Expected?Vertical Distance Groupspecies1 (14)2 (8)3 (4)Willow (1)100Alder (17)1340Wild Raisin (8)044ObservedChi SquaredX2=16.878 df=4 p=0.002Fisher TestP=0.00014Appendix G: R Output> library(abind, pos=16)> .Table <- matrix(c(1,0,0,13,4,0,0,4,4), 3, 3, byrow=TRUE)> dimnames(.Table) <- list("species"=c("1", "2", "3"), "distance group"=c("1",+ "2", "3"))> .Table # Counts distance groupspecies 1 2 3 1 1 0 0 2 13 4 0 3 0 4 4> .Test <- chisq.test(.Table, correct=FALSE)> .TestPearson's Chi-squared testdata: .TableX-squared = 16.878, df = 4, p-value = 0.002041> .Test$expected # Expected Counts distance groupspecies 1 2 3 1 0.5384615 0.3076923 0.1538462 2 9.1538462 5.2307692 2.6153846 3 4.3076923 2.4615385 1.2307692> remove(.Test)> fisher.test(.Table)Fisher's Exact Test for Count Datadata: .Tablep-value = 0.0001416alternative hypothesis: two.sided> remove(.Table)> capstonesoil <- + readXL("C:/Gregory/college/capstone/capstone project/data.xlsx", + rownames=FALSE, header=TRUE, na="", sheet="Soil collapsed", + stringsAsFactors=TRUE)> normalityTest(~avg.sat, test="shapiro.test", data=capstonesoil)Shapiro-Wilk normality testdata: avg.satW = 0.89872, p-value = 0.4247> normalityTest(~bulk.density, test="shapiro.test", data=capstonesoil)Shapiro-Wilk normality testdata: bulk.densityW = 0.88841, p-value = 0.3758> normalityTest(~comp..A, test="shapiro.test", data=capstonesoil)Shapiro-Wilk normality testdata: comp..AW = 0.94971, p-value = 0.7143> normalityTest(~comp.st, test="shapiro.test", data=capstonesoil)Shapiro-Wilk normality testdata: comp.stW = 0.62978, p-value = 0.001241> normalityTest(~comp.v, test="shapiro.test", data=capstonesoil)Shapiro-Wilk normality testdata: comp.vW = 0.79065, p-value = 0.08649> with(capstonesoil, cor.test(comp..A, distnace, alternative="two.sided", + method="spearman"))Spearman's rank correlation rhodata: comp..A and distnaceS = 6, p-value = 0.75alternative hypothesis: true rho is not equal to 0sample estimates:rho 0.4 > with(capstonesoil, cor.test(comp.st, distnace, alternative="two.sided", + method="spearman"))Spearman's rank correlation rhodata: comp.st and distnaceS = 2.254, p-value = 0.2254alternative hypothesis: true rho is not equal to 0sample estimates: rho 0.7745967 > with(capstonesoil, cor.test(comp.v, distnace, alternative="two.sided", + method="spearman"))Spearman's rank correlation rhodata: comp.v and distnaceS = 19.487, p-value = 0.05132alternative hypothesis: true rho is not equal to 0sample estimates: rho -0.9486833 > with(capstonesoil, cor.test(bulk.density, comp..A, alternative="two.sided", + method="spearman"))Spearman's rank correlation rhodata: bulk.density and comp..AS = 2, p-value = 0.3333alternative hypothesis: true rho is not equal to 0sample estimates:rho 0.8 > with(capstonesoil, cor.test(bulk.density, comp.sp, alternative="two.sided", + method="spearman"))Spearman's rank correlation rhodata: bulk.density and comp.spS = 7.418, p-value = 0.7418alternative hypothesis: true rho is not equal to 0sample estimates: rho 0.2581989 > with(capstonesoil, cor.test(bulk.density, comp.v, alternative="two.sided", + method="spearman"))Spearman's rank correlation rhodata: bulk.density and comp.vS = 19.487, p-value = 0.05132alternative hypothesis: true rho is not equal to 0sample estimates: rho -0.9486833 > with(capstonesoil, cor.test(avg.sat, comp..A, alternative="two.sided", + method="spearman"))Spearman's rank correlation rhodata: avg.sat and comp..AS = 14, p-value = 0.75alternative hypothesis: true rho is not equal to 0sample estimates: rho -0.4 > with(capstonesoil, cor.test(avg.sat, comp.st, alternative="two.sided", + method="spearman"))Spearman's rank correlation rhodata: avg.sat and comp.stS = 17.746, p-value = 0.2254alternative hypothesis: true rho is not equal to 0sample estimates: rho -0.7745967 > with(capstonesoil, cor.test(avg.sat, comp.v, alternative="two.sided", + method="spearman"))Spearman's rank correlation rhodata: avg.sat and comp.vS = 0.51317, p-value = 0.05132alternative hypothesis: true rho is not equal to 0sample estimates: rho 0.9486833> capstonespp <- + readXL("C:/Gregory/college/capstone/capstone project/data.xlsx", + rownames=FALSE, header=TRUE, na="", sheet="tree data", + stringsAsFactors=TRUE)> normalityTest(~vertical.distance, test="shapiro.test", data=capstonespp)Shapiro-Wilk normality testdata: vertical.distanceW = 0.76418, p-value = 0.00004592Appendix H: R Code> library (Rcmdr)Loading required package: splinesLoading required package: RcmdrMiscLoading required package: carLoading required package: carDataLoading required package: sandwichLoading required package: effectsRegistered S3 methods overwritten by 'lme4': method from cooks.distance.influence.merMod car influence.merMod car dfbeta.influence.merMod car dfbetas.influence.merMod car lattice theme set by effectsTheme()See ?effectsTheme for details.Rcmdr Version 2.6-1Attaching package: 'Rcmdr'The following object is masked from 'package:base': errorConditionError in .subset2(x, i, exact = exact) : subscript out of boundsError in .subset2(x, i, exact = exact) : subscript out of boundsError in .subset2(x, i, exact = exact) : subscript out of boundslibrary(abind, pos=16).Table <- matrix(c(1,0,0,13,4,0,0,4,4), 3, 3, byrow=TRUE)dimnames(.Table) <- list("species"=c("1", "2", "3"), "distance group"=c("1", "2", "3")).Table # Counts.Test <- chisq.test(.Table, correct=FALSE).Test.Test$expected # Expected Countsremove(.Test)fisher.test(.Table)remove(.Table)capstonesoil <- readXL("C:/Gregory/college/capstone/capstone project/data.xlsx", rownames=FALSE, header=TRUE, na="", sheet="Soil collapsed", stringsAsFactors=TRUE)normalityTest(~avg.sat, test="shapiro.test", data=capstonesoil)normalityTest(~bulk.density, test="shapiro.test", data=capstonesoil)normalityTest(~comp..A, test="shapiro.test", data=capstonesoil)normalityTest(~comp.st, test="shapiro.test", data=capstonesoil)normalityTest(~comp.v, test="shapiro.test", data=capstonesoil)with(capstonesoil, cor.test(comp..A, distnace, alternative="two.sided", method="spearman"))with(capstonesoil, cor.test(comp.st, distnace, alternative="two.sided", method="spearman"))with(capstonesoil, cor.test(comp.v, distnace, alternative="two.sided", method="spearman"))with(capstonesoil, cor.test(bulk.density, comp..A, alternative="two.sided", method="spearman"))with(capstonesoil, cor.test(bulk.density, comp.sp, alternative="two.sided", method="spearman"))with(capstonesoil, cor.test(bulk.density, comp.v, alternative="two.sided", method="spearman"))with(capstonesoil, cor.test(avg.sat, comp..A, alternative="two.sided", method="spearman"))with(capstonesoil, cor.test(avg.sat, comp.st, alternative="two.sided", method="spearman"))with(capstonesoil, cor.test(avg.sat, comp.v, alternative="two.sided", method="spearman"))capstonespp <- readXL("C:/Gregory/college/capstone/capstone project/data.xlsx", rownames=FALSE, header=TRUE, na="", sheet="tree data", stringsAsFactors=TRUE)normalityTest(~vertical.distance, test="shapiro.test", data=capstonespp) ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download