MA-SARNet is a one-shot SAR image forecasting framework built with a modified MA-Net (Figure 1), to predict pixel-level backscatter values of SAR images in response to meteorological and geomorphic driving forces. The model performance was evaluated against the benchmark performance created with the temporal persistence assumption and was compared against the label image, label water extent, and two independently validated flood maps. Compared to the benchmark performance, model predictions show an increase of 31.9% and 17.8% for the mean and median AAI (Assemble Accuracy Index) and an increase of 37.9% and 15.1% for the mean and median NSE (Nash-Sutcliffe Efficiency) on the test set. Results showed that the flood extent derived from backscatter predictions is more robust against misclassifications caused by pixel-level noise compared to the flood map derived using the real backscatters and those from two additional flood map repositories (Figure 2). Results from spatial and temporal robustness tests demonstrate that the model has sufficient generalization potential for real-time, physically informed, deep learning-based earth surface prediction tasks to facilitate fast response to and mitigation for future floods.
Related Articles
- Li, Z., Xiang, Z., Demiray, B.Z., Sit, M. and Demir, I., 2023. MA-SARNet: A One-Shot Forecasting Framework for SAR Image Prediction with Physical Driving Forces. Water Resources Research.
DOI: https://doi.org/10.31223/X5765J (in review)
Figure 1
Figure 2
