Development and Evaluation of Synthetic High-Resolution Satellite Imagery for Effectiveness Monitoring
Edward M Olexa
To evaluate habitat conditions and trends, land management agencies in the WLCI region require objective, detailed information describing the characteristics of vegetation dynamics, such as changes in biomass, species composition, or the timing of green-up. Assessing the efficacy of management activities and the duration of their effectiveness has been problematic due to the lack of high-resolution spatial and temporal satellite imagery capable of revealing patterns in vegetation responses and changes in forage production. Vegetation indices, such as the normalized difference vegetation index (NDVI) derived from satellite imagery, can be used to monitor seasonal and interannual changes in plant phenology and biomass associated with habitat altering activities and climatic conditions; however, investigation across large spatial extents at fine temporal and spatial scales was impractical until recently.
Data-fusion methods that blend temporally high-frequency satellite data with spatially high-resolution satellite data can provide the fine-resolution, spatiotemporal data required to evaluate habitat responses to management activities at the landscape level. Specifically, we blended Moderate Resolution Imaging Spectroradiometer (MODIS sensor aboard the Terra satellite; 250-m to 1-km pixels; acquired every day) data with high-resolution Thematic Mapper (TM sensor aboard the Landsat 5 satellite; 30-m pixels; acquired every 16 days) data to develop fine-resolution spatiotemporal data for evaluating habitat responses to management activities at the landscape level. This method has been successfully applied to heterogeneous landscapes, including forested areas and croplands, but had not been tested in areas dominated by semiarid shrub-steppe.
Initiated in FY2011 (see pages 84-88 in Bowen and others, 2013), this project culminated in FY2014 with a journal article detailing our methods and outcomes. We collaborated with the National Park Service and the BLM to test our methods for southwestern Wyoming, including a portion of the WLCI region. Predicted NDVI estimates were accurate and highly correlated with actual values across several land cover types. The synthetic imagery produced by blending the data provides the high spatial resolution managers need to identify treatment areas, the high temporal resolution often needed to track dynamic vegetation conditions, and the large spatial extent needed for monitoring at landscape scales. Our work suggests that managers of semiarid rangelands have alternatives to traditional methods when it comes to assessing and monitoring the living plant biomass of rangelands.
Bowen, Z.H., Aldridge, C.L., Anderson, P.J., Assal, T.J., Biewick, L.R.H., Blecker, S.W., Boughton, G.K., Bristol, S., Carr, N.B., Chalfoun, A.D., Chong, G.W., Clark, M.L., Diffendorfer, J.E., Fedy, B.C., Foster, K., Garman, S.L., Germaine, S., Hethcoat, M.G., Holloway, J., Homer, C., Kauffman, M.J., Keinath, D., Latysh, N., Manier, D., McDougal, R.R., Melcher, C.P., Miller, K.A., Montag, J., Olexa, E.M., Potter, C.J., Schell, S., Shafer, S.L., Smith, D.B., Stillings, L.L., Sweat, M.J., Tuttle, M., and Wilson, A.B., 2013, U.S. Geological Survey science for the Wyoming Landscape Conservation 60 Initiative -- 2011 annual report: U.S. Geological Survey Open-File Report 2013-1033, 145 p., at http://pubs.usgs.gov/of/2013/1033/.
U.S. Geological Survey, 2005, USGS MODIS direct broadcas -- Reception and processing at the USGS Center for Earth Resources Observation and Science: U.S. Geological Survey Fact Sheet 2005-3119, 1 p., at http://pubs.usgs.gov/fs/2005/3119/report.pdf.
U.S. Geological Survey, 2012, Landsat -- A global land-imaging mission: U.S. Geological Survey Fact Sheet 2012-3072, 4 p., at http://pubs.usgs.gov/fs/2012/3072/.
Product Completed in FY2014
- Olexa, E.M., and Lawrence, R.L., 2014, Performance and effects of land cover type on synthetic surface reflectance data and NDVI estimates for assessment and monitoring of semi-arid rangeland: International Journal of Applied Earth Observation and Geoinformation, v. 30, p. 30-41, at http://www.sciencedirect.com/science/article/pii/S0303243414000105, doi:10.1016/j.jag.2014.01.008.
Products Completed in FY2011
- Olexa, E.M., Lawrence, R.L., and Watts, J.D., 2011, Filling the gaps -- Using STARFM to create synthetic Landsat images for dates when Landsat imagery is unavailable [abstract submitted for the AmericaView 2011 Fall Technical Meeting, Cleveland, Oh., to be presented October 10-11, 2011].
- A database of preliminary NDVI accuracy results, stratified by land cover/land use, was provided to the BLM Kemmerer Field Office.