Big data and remote sensing for multi-decadal drought impact assessment on Shorea robusta

Intense droughts in recent years are a global concern. The duration and timing of drought for forests have not been researched as much as crops. This study aims to utilize big earth data to analyse the consequences of drought on forest based on drought and vegetation indices derived from remote sensing. The study area is Sal Forest ( Shorea robusta ) in Bangladesh. Vegetation indices such as NDVI and EVI are correlated with drought indices for the years 2000–2020. The data is collected via Google Earth Engine and the R statistical software environment is used for data analysis. Seasonal adjustment was conducted to remove the biases due to seasonality. Results showed that the timing for the loss of forest vitality is dependent on drought severity. Lagged correlation was used to identify the lags between drought severity and vitality via drought and vegetation indices. From 2000 to 2020, the intensity of drought was moderate (0.95). There is a lag between drought and a decrease in forest vitality. This lag decreases as the severity and duration of drought increase. Severe droughts in Sal Forest create an almost immediate impact with worsening conditions but show a lag of 3–4 months and 2–3 months for low and moderate drought severity, respectively. The NDVI anomaly shows that drought has adverse effects during growth periods. The methodology of this study can be replicated across various forest types, and further improvements would require in situ data coupled with phenological observations.