Evaluating ecosystem integrity across major forest biomes using ecological theory and MODIS time series

Abstract:

Remote sensing data of global forest extent and tree cover change are widely used for conservation assessment, planning and decision making. Yet there is an increasing recognition that these efforts must include elements of forest quality for biodiversity and ecosystem services including climate regulation through carbon storage. Such information is rarely available, especially at a global scale. To meet this need, here we present a novel framework to estimate forest ecosystem integrity based on ecological theory and time series analysis of Moderate Resolution Imaging Spectroradiometer (MODIS) data at 500 m spatial resolution. We integrated several ecologically-based metrics that provide insights into the productivity, stability, resilience and adaptive capacity of forest ecosystems based on the fraction of photosynthetically active radiation intercepted by sunlit vegetation canopy (fPAR) and the shortwave infrared water stress index (SIWSI) over the period 2003–2018. Using these metrics, we evaluated ecosystem integrity levels, ranging from intact to highly disturbed or degraded ecosystems, in two contrasting forest biomes, the southern boreal taiga (Siberia, Russia) and the tropical humid forest (Amazon basin, Brazil). The final forest integrity maps were validated against (i) MODIS land cover type product MCD12Q1, (ii) 30-m digitized annual burned and logged areas upscaled to MODIS resolution (in Siberia) or PRODES deforestation data from the Brazilian space agency (INPE) (in Amazonia), and (iii) Global Forest Change Data. Our framework is capable of providing timely and reliable information for ecosystem science, primary forest identification and conservation, and evolving national and international policies regarding land use.