Analysis of ecological resilience to evaluate the inherent maintenance capacity of a forest ecosystem using a dense Landsat time series

Abstract

The inherent maintenance capacity of forest ecosystems is an important indicator to evaluate environmental quality. The development of remote sensing time series analysis provides a new perspective and methodology to evaluate such maintenance capacities, that take account of the gradual and subtle changes contained in forest growth over long time periods. This study analyzed forest ecosystem resilience, to evaluate the inherent maintenance capacity of a forest ecosystem in Hengdong County, Hunan Province, China, using a dense Landsat time series from 1988 to 2018. The key resilience metrics of disturbance magnitude (M), degree of resilience, and recovery rate (RR) were developed, to describe forest ecosystem resilience. These metrics were derived using a dynamic time warping algorithm, based on calculations of the distance among the inter-annual time series, which detected forest disturbance-regrowth dynamics. Through this methodology, we obtained an overall accuracy of 87.14% (in the degree of resilience) for the spatio-temporal resilience characteristics of the forest ecosystem. Disturbance was ever-present and 59.72% of the change processes varied within a relatively small range. Furthermore, we found RR to increase concomitantly with M and large M values caused a decreased degree of resilience, indicating a very small likelihood of forest greenness restoration under large M values in the red soil erosion region. We also found that the degree of resilience and RR showed an overall increasing trend, indicating an enhancement in the inherent maintenance capacity of the ecosystem. Our methodology provides a significant potential for the evaluation of the regional environmental quality, while providing insights into the comprehensive management of soil erosion in forest ecosystems.