Flash Drought: Review of Concept, Prediction and the Potential for Machine Learning, Deep Learning Methods

• Published in: Earth's future Vol. 10; no. 11
• Main Authors: Tyagi, Shoobhangi, Zhang, Xiang, Saraswat, Dharmendra, Sahany, Sandeep, Mishra, Saroj Kanta, Niyogi, Dev
• Format: Journal Article
• Language: English
• Published: Bognor Regis John Wiley & Sons, Inc 01.11.2022; Wiley
• Subjects: Algorithms; data‐driven drought assessment; Deep learning; Drought; drought prediction; flash droughts; Heat waves; High temperature; Hydrology; Indicators; Learning algorithms; Literature reviews; Machine learning; Parameter uncertainty; Parameterization; Prediction models
• ISSN: 2328-4277; 2328-4277
• DOI: 10.1029/2022EF002723

This paper reviews the Flash Drought concept, the uncertainties associated with FD prediction, and the potential of Machine Learning (ML) and Deep learning (DL) for future applications. For this, 121 relevant articles covering different aspects of FD ‐ definitions, key indicators, distinguishing characteristics, and the current methods for FD assessment (i.e., ‐ monitoring, prediction, and impact assessment) are examined. FD is typically a short‐term drought event ‐ characterized by the rapid progression of heat waves and precipitation deficits, causing cascading impacts on the land and surface hydrology. FD prediction is constrained by the lack of consistent FD definitions, key indicators, the limited predictability of FD at the subseasonal‐ to‐seasonal (S2S) timescale, and uncertainties associated with the current prediction methods. Some of the uncertainties in the current methods are associated with a lack of our understanding of the physical processes. They are also related to the error in the input datasets (imperfect representation of indicators), parameter uncertainty (parameterization scheme adopted by the prediction model), multicollinearity, nonlinear, and non‐stationary interactions among different indicators. Combining traditional methods and multisource fusion data with ML and DL methods shows promise to better understand FD evolution and improves prediction. Plain Language Summary Flash Drought (FD) events develop rapidly and have notable agricultural impacts. This paper reviews recent notable FD studies. Several studies highlight the need to develop an improved understanding and prediction of FD to manage its effects better. However, the lack of consistent definition, documentation and fragmented assessments have limited progress toward improved prediction models. This review considers questions related to FD definition, triggers, and how it is different from the more classical droughts. A summary of methods and their limitations to predict FD are considered. The review concludes that the lack of standard definition, and characterization of FD is a major stumbling block. It inhibits the development of representative data set that can capture the complex multiscale information that characterizes FDs. Without such data set and reference, resultant models could lead to divergent pathways and false predictions of FD. The review also suggests Machine Learning (ML) and Deep Learning (DL) approaches can extract information from multi‐scale complex datasets routinely used for traditional drought prediction. The addition of ML and DL methods with the current method can help FD prediction by extracting predictive information regarding the physical processes that trigger such events, their impacts and society's response. Key Points Lack of consistent definition and related inventories for Flash Drought (FD) is a stumbling block to developing reliable prediction models Traditional FD prediction methods need to represent the uncertainty of key indicators, their highly correlated and time‐varying interactions Combining ML and DL methods with traditional methods, multisource fusion data can provide a potential framework for improving FD prediction