Ian M. Wright 1 December, 2011 GRG 386G GIS Final Project: Compositional Changes in Vegetation and Categorization of Areas for Monitoring Invasive Grass Spread at Chaparral Wildlife Management Area, TX Introduction Species invasions constitute one of the most damaging threats to the conservation of native communities and ecosystems worldwide. Grasses are some of the most pervasive global invaders and several species have colonized and spread in the United States. Grass invasions can have many negative consequences for native biotas in the areas they invade. Grasses often choke out native vegetation and homogenize the landscape. This loss of native plant species often results in a reduction of herbivore species diversity as well as the diversity of predators that feed on these herbivores. This homogenization can put once resilient communities at further risk of collapse due to other anthropogenic disturbances such as habitat destruction. Grass invasions often alter fire regimes, typically in ways they further reinforce grass invasions. Simplified habitats may also be easier to invade, resulting in a positive feedback loop that can quickly and irreversibly deprive an area of its native fauna. One poorly understood grass invader in south Texas is buffel grass (Pennisetum ciliare). Once planted as livestock fodder in the 1960’s, it has spread along roadsides and now effects most of south Texas. The impacts this invasion is having on native systems remains unknown. More broadly, the impacts these types of invasion have to entire communities remains little studied. In part, using ArcGIS I hope to enhance our knowledge of what these impacts are and how they come about through studying buffel grass invasions across south Texas. As part of my dissertation work I will investigate the effects that buffel grass invasions have on the native arthropod communities of Texas and the food webs they sustain. I will conduct this study on the Chaparral Wildlife Management Area (CWMA) – a 15,200 acre habitat preserve straddling La Salle and Dimmit counties near the town of Cotulla in south Texas. The biological sampling is beyond the scope of this assignment, however, ArcGIS is an important step in understanding the current situation and in targeting areas for future study. Specifically through this project I hope to address the following questions: 1) What is the vegetational change at CWMA from 2004 to 2010? 2) How much area is at risk for buffel grass invasion? Data Collection In order to conduct spatial analyses of vegetation change at CWMA I needed to obtain data relevant to its 15,200 acre area. Unfortunately, this site is underutilized as a study area and is fairly remote. As such, very little data exists for CWMA and I had to digitize much of myown. The first and most important data I needed was aerial photos of the field site that I could use to quantify vegetation cover. I was able to download this data from the Texas Natural Resources Information System (TNRIS – available at http://data.tnris.org/datadownload/). TNRIS maintains these photos as 4-band RGB raster tiles in the NAD83 UTM Zone 14N Projected Coordinate System. These orthoimage rasters are available variously in natural color or color infrared, 0.5 m or 1 m resolution, and for 1996, 2004-6, and 2008-10. To cover the area of the WMA, I needed to download and extract to their own folder 6 rasters; 4 from the western edge of Dimmit county and 2 from the eastern edge of La Salle county. I downloaded all of those rasters that were at 1 m resolution, which included the years 1996, 2004, 2008, and 2010. However upon visually inspecting these rasters in ArcMap it appeared that many of the rasters for particular years were quite different from one another despite sharing boundaries. Thus, I assume that these photos were taken at different times, under different atmospheric conditions, or with different equipment, etc. Because of this I restricted my analysis to the years 2004 and 2010. Unfortunately, this was the only relevant data available to me. A number of access roads criss-cross the WMA and are important avenues for grass invasions. However, to my knowledge, no one has digitized these roads and made them available as a shapefile. The most precise road map I was able to download (TXDOT_Roadways.shp – available from the Texas Department of Transportation at www.txdot.gov) had main roads around the area digitized but not the smaller access roads throughout the site. Thus, in order to determine how much roadside area is available for grass invasion I had to digitize these roads myself. The same is true for hydrology of the area; no water features occurred in CWMA on even the finest scale hydrologic map of Texas (the National Hydrography Dataset available from nhd.usgs.gov). Upon inspecting the orthophotos, some areas of the WMA appeared to have temporary rivers or water drainages but as I could not be certain of these features and because they are not entirely relevant to this project I neglected to digitize any water features. The final data source I obtained was a soil map of Dimmit and La Salle counties (available as two shape files with multiple polygons from soils.usda.gov). These soil polygon features, however, were labeled and symbolized differently on different side of the county line. In order to correct this in order to effectively use soil data in any analyses I would have to merge and re-attribute many of the polygons. While this will likely be an important component of my future dissertation research, this tedious task was out of the purview of this assignment and thus no soil data was used. All other data were created by me by digitizing lines of polygons that I could discern from off of the aerial photographs. Data Prerocessing Virtually no data preprocessing was required for this project. The orthophoto rasters were available as MrSID files and simply needed to be loaded into ArcMap from ArcCatalog. Some data post processing will be required prior to conducting field research at the site. The data in question are random points placed by ArcMap into polygons of putative study areas. These points will be used to center vegetation and insect survey plots and will need to be loaded in to a GPS unit prior to departing for field work. To do this I simply need to open the .dbf table of this points in Excel, save them as an Excel file, and manually or automatically (through software) input these