Remote sensing dynamic monitoring of the flood season area of Poyang Lake over the past two decades

Abstract

Drought and flood disasters occur frequently in the Poyang Lake basin during the flood season, making it significantly important to monitor the lake's flood season area using time-series remote sensing. This research employed multisource remote sensing datasets from Landsat, ALOS, and Sentinel-1 across the time span of 2000–2022. Using the MNDWI method and adaptive global threshold segmentation, the flood season area of Poyang Lake was extracted. The spatiotemporal variation characteristics of the flood season area were analysed, and correlations with precipitation and temperature were assessed. Moreover, this study explored the characterization of its response to drought and flood events. The results revealed a strong fluctuation in the lake during the flood season, with the maximum and minimum area differences exceeding 3000 ​km², and spatial changes were mainly concentrated in the southwestern lake region. There is a significant positive correlation between area changes and precipitation and a significant negative correlation with temperature. By analysing the response characteristics of the flood season area changes to drought and flood events, in years when the flood season area of Poyang Lake exceeds 4500 ​km², extreme flood disasters usually occur. Areas between 3900 ​km² and 4500 ​km² are prone to floods, areas between 2000 ​km² and 3000 ​km² are prone to drought events, and areas below 2000 ​km² typically experience extreme drought disasters.