Machine Learning Modeling of Climate Variability Impact on River Runoff

Mateusz Norel,Zbigniew W Kundzewicz,Krzysztof Krawiec

doi:10.3390/w13091177

Mateusz Norel, Zbigniew W Kundzewicz + Show 1 more

Open Access

https://doi.org/10.3390/w13091177

Copy DOI

Abstract

The hypothesis of this study was one of existence of spatially organized links between the time series of river runoff and climate variability indices, describing the oscillations in the atmosphere–ocean system: ENSO (El Niño–Southern Oscillation), PDO (Pacific Decadal Oscillation), AMO (Atlantic Multidecadal Oscillation), and NAO (North Atlantic Oscillation). The global river flow reconstructions (ERA-20-CM-R) for 18 study areas on six continents and climate variability indices for the period 1901–2010 were used. The split-sample approach was applied, with the period 1901–2000 used for training and 2001–2010 used for testing. The quality measures used in this paper were mean absolute error, dynamic time warping, and top extreme events error. We demonstrated that a machine learning approach (convolution neural network, CNN) trained on climate variability indices can model the river runoff better than the long-term monthly mean baseline, both in univariate (per-cell) and multivariate (multi-cell, regionalized) settings. We compared the models to the baseline in the form of heatmaps and presented results of ablation experiments (test time ablation, i.e., jackknifing, and training time ablation), which suggested that ENSO is the primary determinant among the considered indices.

Highlights

Destructive floods kill, on average, thousands of people worldwide per year, and cause material loss of the order of tens of billions to hundreds of billions of USD [1]
In the “Interpretation of Change in Flood-related Indices based on Climate Variability” (FloVar) project, we examined the variability related to high river runoff and floods and sought relationships between it and the indices of climate variability
The work conducted within the “Interpretation of Change in Flood-related Indices based on Climate Variability” (FloVar) project demonstrated that the natural climate variability alone carries important and useful information relevant to the spatio-temporal field of river runoff, being of considerable and broad relevance

Summary

Introduction

Destructive floods kill, on average, thousands of people worldwide per year, and cause material loss of the order of tens of billions to hundreds of billions of USD [1]. There have been many reports on dramatic deluges, in less developed countries. In terms of both percentage of national population and national GDP affected, has been found to exist in Southeast Asia (Vietnam, Cambodia, Bangladesh) [2]. Large floods have been reported in recent years in large countries of Asia— China, India, and Pakistan, as well as in Africa, the Americas, and Australia. In Europe, many destructive floods have been recorded in recent decades, with the costliest one, in August 2002, affecting Germany, Austria, and the Czech Republic in particular. Despite all the progress made in monitoring, forecasting, and safeguarding, there is no place on earth where this challenge has been definitely solved [3]

Objectives

Results

Conclusion