Remote sensing diagnosis of ecosystem resilience dynamics in agricultural heritage landscapes

This dataset supports the publication <i>“Remote sensing diagnosis of ecosystem resilience dynamics in agricultural heritage landscapes”</i>. It provides processed remote sensing products, derived indicators, and analysis codes used to quantify and map ecosystem resilience across polder-based agricultural heritage landscapes in the Taihu Basin, China, from 2005 to 2024.The dataset includes:<b>EVI Time Series (2005–2024):</b><br>A long-term Enhanced Vegetation Index (EVI) dataset derived from MODIS imagery.<b>LSWI Time Series (2005–2024):</b><br>A Land Surface Water Index (LSWI) time series produced from MODIS surface reflectance data to capture spatiotemporal variations in surface moisture and hydrological dynamics relevant to agricultural resilience.<b>Long-term Resilience Indicator:</b><br>Pixel-wise resilience metrics estimated from EVI time series using autoregressive modeling (AR(1)) and recovery trajectory analysis to capture long-term stability and recovery capacity.<b>Short-term Resilience Index:</b><br>Disturbance–recovery responses computed over short temporal windows to detect transient resilience fluctuations following perturbations.<b>Environmental Drivers Dataset:</b><br>Spatial layers representing key environmental and anthropogenic factors influencing resilience.<b>R Code for Indicator Calculation:</b><br>Scripts used for preprocessing, resilience computation, and visualization. All analyses were performed in R (v4.3.2) using packages <code>raster</code>, <code>dplyr</code>, <code>mgcv</code>, and <code>ggplot2</code>.<br>