Inductive Predictions of Extreme Hydrologic Events in The Wabash River Watershed

TL;DR

This paper introduces a bidirectional LSTM model with timestep reduction for efficient, spatially-inductive prediction of extreme hydrologic events like droughts, using long-term data from the Wabash River Watershed.

Contribution

The study presents a simple, fast-trained LSTM approach that generalizes to new locations, outperforming complex models in predicting extreme hydrologic events.

Findings

Model predicts soil water and stream flow accurately.

Extreme event prediction is effective across different geographical locations.

The approach reduces training time compared to attention-based models.

Abstract

We present a machine learning method to predict extreme hydrologic events from spatially and temporally varying hydrological and meteorological data. We used a timestep reduction technique to reduce the computational and memory requirements and trained a bidirection LSTM network to predict soil water and stream flow from time series data observed and simulated over eighty years in the Wabash River Watershed. We show that our simple model can be trained much faster than complex attention networks such as GeoMAN without sacrificing accuracy. Based on the predicted values of soil water and stream flow, we predict the occurrence and severity of extreme hydrologic events such as droughts. We also demonstrate that extreme events can be predicted in geographical locations separate from locations observed during the training process. This spatially-inductive setting enables us to predict extreme events in other areas in the US and other parts of the world using our model trained with the Wabash Basin data.