A Contemporary Review on Deep Learning Models for Drought Prediction

Abstract

Deep learning models have been widely used in various applications, such as image and speech recognition, natural language processing, and recently, in the field of drought forecasting/prediction. These models have proven to be effective in handling large and complex datasets, and in automatically extracting relevant features for forecasting. The use of deep learning models in drought forecasting can provide more accurate and timely predictions, which are crucial for the mitigation of drought-related impacts such as crop failure, water shortages, and economic losses. This review provides information on the type of droughts and their information systems. A comparative analysis of deep learning models, related technology, and research tabulation is provided. The review has identified algorithms that are more pertinent than others in the current scenario, such as the Deep Neural Network, Multi-Layer Perceptron, Convolutional Neural Networks, and combination of hybrid models. The paper also discusses the common issues for deep learning models for drought forecasting and the current open challenges. In conclusion, deep learning models offer a powerful tool for drought forecasting, which can significantly improve our understanding of drought dynamics and our ability to predict and mitigate its impacts. However, it is important to note that the success of these models is highly dependent on the availability and quality of data, as well as the specific characteristics of the drought event.

The Drought Regime in Southern Africa: A Systematic Review

“…Some other methods to estimate drought characteristics, especially hydrological drought, are time series analysis/modeling, regionalization procedures to estimate/extrapolate lowflow/drought characteristics spatial distribution (simple estimation methods, multivariate analysis, regional regression models, hydrological mapping procedures), and frequency analysis using probability distribution analysis, extreme value analysis, regional frequency analysis, severity-area-frequency curves [85]. Following the recent and global trend of using machine learning methods, several researchers have proposed different algorithms for drought modeling, hazard monitoring, forecasting, and impacts, e.g., [87][88][89][90][91]. These methods have the advantage of not relying on prior knowledge of the phenomena/processes and being data-driven, i.e., the models are calibrated based on previous experiences [92].…”

Analysis, characterization, prediction, and attribution of extreme atmospheric events with machine learning and deep learning techniques: a review

“…Very recently, DL algorithms have been also applied to different problems in drought prediction. In Gyaneshwar et al (2023), a review of the most important DL algorithms with application in drought prediction is presented. The work also includes a number of ML approaches for drought prediction.…”