Remote Monitoring of Vegetation Managed for Dust Control on the Dry Owens Lakebed, California

Abstract

A monitoring program was developed to assess the cover of saltgrass managed for dust control on the saline dry Owens Lake. Although the original intent was to manage the vegetation as total cover that included green and senesced leaf and stem material, aged leaves that make up a large proportion of total cover were not differentiable spectrally from the background salt and lakebed. Hence, greenness-based indices were explored for detection of plant recruitment. Since all plant cover begins as green and growing, greenness indices provide a measure of all future cover whether living or senesced. The criteria for judging compliance were changed so that spatially variable vegetation cover measured as a milestone will need to be met in the future. A derivative of NDVI, NDVIx, calculated using scene statistics, proved highly accurate, to about 0.001 of this index and with an average signal to noise ratio of 64. This high level of accuracy allowed detection of small changes in vegetation growth and vigor. Performance according to the benchmark-as-par standard was determined through combined use of cumulative distribution functions and derivative maps.