Over the past few years I have had the opportunity to work on a number of research topics, including vegetation modeling, landscape monitoring, and invasive species mapping and prediction to name a few. These opportunites have given me a solid understanding and appreciation for the use of geospatial data in many fields. The use of this type of data will not subside and the industry trend shows a strong increase in the use of and the uses for geospatial information.
Selected topics include: Invasive Species / Monitoring / State and Transition / WebApps / Fire
Using MODIS Data for Identification and Mapping of Bromus tectorum across the Great Basin Region of North America.
Bromus tectorum, also known as cheatgrass, is a wide ranging invasive weed in the Intermountain West. This invasive grass encroaches on native shrub and shrub-steppe environments, providing fuel for fire, out competing native plants post fire, and increasing fire return intervals. The conversion of native shrublands to Cheatgrass is a significant problem in the Great Basin region of the United States. MODIS 16-day composite data from 2000 to 2004 is being integrated into a database to model current distributions of Cheatgrass over the approximately 518,000 square kilometer region known as the Great Basin. Field observations consisting of approximately 15,000 individual sites recording Cheatgrass percent ground cover (from absence to dominance) are undergoing comparison against MODIS temporal NDVI data. Using a Computer Aided Regression Tree and general linear models estimations of Cheatgrass cover are being generated for this region.
Remote Sensing of Watersheds and Lake Ice Coverage as a Template for Monitoring and Assessment of Aquatic Resources in the Arctic Network of Parks.
Rocky Mountains Cooperative Ecosystem Studies Unit) The National Park Service has accumulated a variety of spatial/synoptic data for Gates of the Arctic National Park and Preserve (GAAR). We propose to use previously collected remotely sensed images and existing GIS data layers for GAAR to develop and assess techniques to catalogue aquatic ecosystems and classify watersheds. We will classify water bodies and surrounding watersheds using and unsupervised cluster analysis, then manually analyze results to determine if existing NPS GIS coverages are adequate for such an analysis at the appropriate spatial scale. Additionally, we will use remotely sensed images and existing GIS coverages to inventory lakes in GAAR.
Identification of Temporal Trends in Vegetation Communities Following Distrubances Using Remotely Sensed Imagery.
It has been hypothesized that remotely sensed imagery may be analyzed, using the state and transition theory, to identify variations in vegetation communities due to natural and anthropogenic disturbances. Additionally, the state and transition spectral characteristics of vegetation communities can be applied to land management decisions.
United States Department of Agriculture – Agriculture Research Service Forage and Range Research Laboratory online GIS data extraction and modeling tools
The USDA-ARS FRRL utilizes many physical characteristics (slope, aspect, elevation, soil type, climate data, and other physical attributes) as research variables from the environment in which plants are sampled. Much of this data must be painstakingly extracted through on-the-ground field processes, or extracted from coarse, out-of-date paper maps. Significant amounts of accurate data are freely available through multiple sources such as universities and research organizations that would enhance and expedite variable collection and save time and resources. These data are to be culled, organized, and placed on the internet for rapid, public access utilizing custom internet mapping and extraction tools
Identification of Fire Scars Using Remotely Sensed Data and Geographic Information Systems.
This project has been implemented with the goal of accurately identifying and extracting fire scars from Landsat TM and ETM imagery in the shrub steppe and semi-desert region of Northern and Central Utah. By organizing, radiometrically correcting, and processing a temporally dense image database, fire scars hare being identified using a fire extraction algorithm. GIS is being employed to apply accuracy assessments and map algebra functions on the extracted fire scar areas. Rapid identification of multiple burn areas will help land managers red-flag locations requiring primary attention to mitigate further environmental degradation.