Changes with land cover and land use are closely integrated with water and other ecological processes at the land surface. Nowhere is that more apparent than in the Edwards Aquifer region of southcentral Texas. The Edwards Aquifer contributing and recharge zones cover approximately 18,000 square kilometers in parts of 15 counties in Texas and includes San Antonio and Austin, the nation's eighth and nineteenth largest cities, respectively. Population growth within the counties that intersect the Edwards Aquifer contributing and recharge zones has taken place over the last two decades, with the logical translation being an expanded infrastructure. This implies that a greater amount of impervious surface coverage and other land cover changes have occurred. This work quantified the changes in land cover within the Edwards Aquifer contributing and recharge zones between the years 1986 and 2000. Increasing trends in impervious surface area and woodland growth were identified. Additionally, a new ArcView software tool was developed to process SSURGO soil data for use within the ArcView SWAT model. Hydrologic modeling for the Upper Sabinal River watershed, located within the Edwards Aquifer region, revealed that the high resolution SSURGO data produces different results when used in place of the existing STATSGO soils data. Finally, an index of urbanization was developed and evaluated to assist investigators in identifying potential areas of urbanization.

Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco var. menziesii), a keystone species in western Washington, faces threats on plantations across this region from the invasive species Scotch broom (Cytisus scoparius (L.) Link), whose invasions on recently established stands can lead to mortality of Douglas-fir through overtopping. The susceptibility of sites to Scotch broom achieving dominance over Douglas-fir has been demonstrated as highly site dependent, however the site conditions that cause this have not yet been identified. Scotch broom has a demonstrated average maximum height of 3m, thus, after Douglas-fir exceeds this height, its risk of being overtopped is significantly reduced. This thesis strives to identify sites that were at the greatest risk Douglas-fir being overtopped by Scotch broom by first, identifying what factors improved growth of Douglas-fir during the period when they are at the greatest risk, and second, identifying factors that led to Douglas-fir outcompeting Scotch broom on sites they cohabitated. In Chapter 1, we utilized LiDAR scans, Soil Survey Geographic Database characteristics, and management histories to identify conditions that improved growth for Douglas-fir in ages 3-8. Individual tree detection was used to measure Douglas-fir heights, and a correction algorithm for LiDAR measured young Douglas-fir heights was established from field validation data. We identified that young Douglas-fir had improved growth on sites with lower elevation, flatter slopes, and finer textured soils. The factors identified were then transformed into four potential site index models based on mean stand elevation class, Mean stand elevation class and clay class, textural class and slope class, and textural class and Mean stand elevation class. In Chapter 2, we used paired field plots to examine Douglas-fir and Scotch broom competition on 19 sites across western Washington. Each site had 2 plots with only Douglas-fir and 2 plots with Douglas-fir and Scotch broom. Elevation, soil texture, and soil nutrient composition for carbon, nitrogen and available phosphorous were examined for influence on height and growth rate of both species. We identified that Scotch broom presence was negatively related to Douglas-fir height growth and that sites with either higher percentages of silt, lower concentrations of phosphorous, or higher percentages of Carbon were more likely to have growth patterns close to or exceeding Scotch broom. / Master of Science / Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco var. menziesii) is a valuable timber species across western Washington that is commonly grown in plantations. In western Washington when Douglas-fir are planted on a site they often face competition from the invasive shrub Scotch broom (Cytisus scoparius (L.) Link). When Scotch broom invasions occur on a recently planted Douglas-fir stands, they can cause death of the trees if Scotch broom plants grow taller than the young Douglas-fir and obstruct their access to light, a process called overtopping. The risk of Douglas-fir being overtopped has been shown to be dependent on location, however what causes a location to be at risk of overtopping is yet unknown. Scotch broom has a demonstrated average maximum height of 3m, thus, after Douglas-fir exceeds this height, its risk of being overtopped is significantly reduced. This study aims to identify sites that were at the greatest risk Douglas-fir being overtopped by Scotch broom by first, identifying what sites generate the best Douglas-fir growth when they are young and at risk of being overtopped, and second, identifying site characteristics led to Scotch broom growing faster than Douglas-fir on sites they both occur on. To identify sites that produced greater young Douglas-fir height growth we used publicly available soil data from the Soil Survey Geographic Database and company management histories to predict tree heights measured through aerial laser scanning (LiDAR). We found that sites with soil textures that had higher percentages of smaller particles, were on lower elevations, and had gentler hillslopes could all produce greater Douglas-fir height growth. When attempting to identify what causes Douglas-fir to be at risk of being overtopped by Scotch broom we used plots with and without Scotch broom on a variety of field sites. This allowed us to not only identify which characteristics of sites where Douglas-fir was being outgrown by Scotch broom, but also identify if Scotch broom was changing how Douglas-fir grew. We found that reductions in Douglas-fir growth were related to Scotch broom being present and that increases in soil silt percentages, decreases in soil phosphorus concentrations, and increases in soil carbon percentages were related to Douglas-fir having height growth closer to or exceeding that of scotch broom.

No description available.

The National Wetlands Inventory (NWI) is the definitive source for wetland resources in the United States. The NWI production unit in Hadley, MA has begun to upgrade their digital map database, integrating descriptors for assessment of wetland functions. Updating is conducted manually and some automation is needed to increase production and efficiency. This study assigned landscape position descriptor codes to NWI wetland polygons and correlated polygon environmental properties with public domain terrain, soils, hydrology, and vegetation data within the Coastal Plain of Virginia. Environmental properties were applied to a non-metric multidimensional scaling technique to identify similarities within individual landscape positions based on wetland plant indicators, primary and secondary hydrology indicators, and field indicators of hydric soils. Individual NWI landscape position classes were linked to field-validated environmental properties. Measures provided by this analysis indicated that wetland plant occurrence and wetland plant status obtained a stress value of 0.136 (Kruskal's stress measure = poor), which is a poor indicator when determining correlation among wetland environmental properties. This is due principally to the highly-variable plant distribution and wetland plant status found among the field-validated sites. Primary and secondary hydrology indicators obtained a stress rating of 0.097 (Kruskal's stress measure = good) for correlation. The hydrology indicators measured in this analysis had a high level of correlation with all NWI landscape position classes due the common occurrence of at least one primary hydrology indicator in all field validated wetlands. The secondary indicators had an increased accuracy in landscape position discrimination over the primary indicators because they were less ubiquitous. Hydric soil characteristics listed in the 1987 Manual and NTCHS field indicators of hydric soils proved to be a relatively poor indicator, based on Kruskal's stress measure of 0.117, for contrasting landscape position classes because the same values occurred across all classes. The six NWI field–validated landscape position classes used in this study were then further applied in a public domain digital data analysis. Mean pixel attribute values extracted from the 180 field-validated wetlands were analyzed using cluster analysis. The percent hydric soil component displayed the greatest variance when compared to elevation and slope curvature, streamflow and waterbody, Cowardin classification, and wetland vegetation type. Limitations of the soil survey data included: variable date of acquisition, small scale compared to wetland size, and variable quality. Flow had limitations related to its linear attributes, therefore is often found insignificant when evaluating pixel values that are mean of selected pixels across of wetland landscape position polygons. NLCD data limitations included poor quality resolution (large pixel size) and variable classification of cover types. The three sources of information that would improve wetland mapping and modeling the subtle changes in elevation and slope curvature that characterize wetland landscapes are: recent high resolution leaf-off aerial photography, high-quality soil survey data, and high-resolution elevation data. Due to the data limitations and the choice of variables used in this study, development of models and rules that clearly separate the six different landscape positions was not possible, and thus automation of coding could not be attempted. / Master of Science

This research work concentrate on developing digital soil maps to support field based plant phenotyping research. We have developed soil organic matter content (OM), cation exchange capacity (CEC), natural soil drainage class, and tile drainage line maps using topographic indices and aerial imagery. Various prediction models (universal kriging, cubist, random forest, C5.0, artificial neural network, and multinomial logistic regression) were used to estimate the soil properties of interest.