Browsing by Author "Savage, Shannon Lea"

Now showing 1 - 2 of 2

Mapping changes in Yellowstone's geothermal areas
(Montana State University - Bozeman, College of Agriculture, 2009) Savage, Shannon Lea; Chairperson, Graduate Committee: Rick L. Lawrence; Stephan G. Custer (co-chair)
Yellowstone National Park (YNP) contains the world's largest concentration of geothermal features, and is legally mandated to protect and monitor these natural features. Remote sensing is a component of the current geothermal monitoring plan. Landsat satellite data have a substantial historical archive and will be collected into the future, making it the only available thermal imagery for historical analysis and long-term monitoring of geothermal areas in the entirety of YNP. Landsat imagery from Thematic Mapper (TM) and Enhanced Thematic Mapper Plus (ETM+) sensors was explored as a tool for mapping geothermal heat flux and geothermally active areas within YNP and to develop a change analysis technique for scientists to utilize with additional Landsat data available from 1978 through the foreseeable future. Terrestrial emittance and estimates of geothermal heat flux were calculated for the entirety of YNP with two Landsat images from 2007 (TM) and 2002 (ETM+). Terrestrial emittance for fourteen summer dates from 1986 to 2007 was calculated for defined geothermal areas and utilized in a change analysis. Spatial and temporal change trajectories of terrestrial emittance were examined. Trajectories of locations with known change events were also examined. Relationships between the temporal clusters and spatial groupings and several change vectors (distance to geologic faults, distance to large water bodies, and distance to earthquake swarms) were explored. Finally, TM data from 2007 were used to classify geothermally active areas inside the defined geothermal areas as well as throughout YNP and a 30-km buffer around YNP. Estimations of geothermal heat flux were inaccurate due to inherent limitations of Landsat data combined with complexities arising from the effects of solar radiation and spatial and temporal variation of vegetation, microbes, steam outflows, and other features at each geothermal area. Terrestrial emittance, however, was estimated with acceptable results. The change analysis showed a relationship between absolute difference in terrestrial emittance and earthquake swarms, with 34% of the variation explained. Accuracies for the classifications of geothermally active areas were poor, but the method used for classification, random forest, could be a suitable method given higher resolution thermal imagery and better reference data.
Vegetation dynamics in Yellowstone's Northern Range : 1985 - 1999
(Montana State University - Bozeman, College of Agriculture, 2005) Savage, Shannon Lea; Chairperson, Graduate Committee: Rick L. Lawrence.
The Northern Range (NR) of Yellowstone National Park (YNP) is currently a critical area of analysis in the Greater Yellowstone Ecosystem (GYE). The complex dynamic ecosystems in the NR provide outstanding field laboratories for long-term scientific investigations to evaluate management techniques. Grassland and shrubland ecosystems serve as important habitat for a wide variety of animal species in the NR and empirical knowledge of these systems enables managers to better understand the ecological complexities and make informed management decisions. Accurate vegetation maps are useful tools for these land managers, as is the ability to detect changes in vegetation over time. An inexpensive and easily reproducible method for mapping rangelands over the NR was developed utilizing Landsat Enhanced Thematic Mapper Plus (ETM+) satellite imagery. A classification hierarchy of non-forest vegetation was produced with 5 levels, ranging from very broad vegetation types such as woodland, shrubland, or herbaceous vegetation (Level 1) to specific vegetation types such as aspen (Populus tremuloides), tufted hairgrass/sedge (Deschampsia cespitosa/Carex spp.), or big sagebrush/Idaho fescue (Artemisia tridentata/Festuca idahoensis) (Level 5). A 1999 base map of nonforest vegetation in the NR was created using classification tree analysis (CTA) combined with boosting. Overall accuracies of the final maps ranged from 72.30% for Level 5 to 83.65% for Level 1, providing evidence that this method can be successful for mapping non-forest vegetation in the NR. A 1985 Landsat Thematic Mapper (TM) satellite image was chosen for performing a change detection analysis from 1985 to 1999. Tasseled Cap space was utilized to choose a threshold of change in a change vector analysis (CVA). Areas of no change in the 1999 image were used to train areas of potential change on the 1985 image to produce a final map of the NR for 1985. Overall accuracies of the final maps ranged from 72.60% for Level 5 to 88.73% for Level 1. Managers are able to analyze this change information and modify their management techniques as needed. With the 1999 base map the CVA method enables managers to detect vegetation change in the NR on a regular basis.