Scholarlink

<p>Hydrometeorological processes associated with teleconnections can create lagged responses between extreme events over different temporal scales. Identification of causal relationships in such pairs of extremes can improve the accuracy of prediction algorithms and help with better risk mitigation. In this global analysis, we investigated whether extreme events like High Temperature and Low Precipitation (HTLP) days, High Temperature (HT) days, and Low Precipitation (LP) days can be used to predict concurrent heatwaves and short-term moderate droughts (CHWD). Using Spearman's rank correlation, mutual information, and dynamic time warping, our results indicated a strong dependence between these variables. Based on the observed relationship, we investigated the existence of causality between HTLP, HT and LP with the CHWD of the next year using Granger Causality and Transfer Entropy. Our results showed that HTLP and HT are the better predictors for equatorial regions, while HTLP and LP are the better predictors for mid-latitudes and Arctic regions. Furthermore, HTLP and HT show higher pattern similarity and temporal alignment with CHWD as compared to LP, as evidenced by lower dynamic time warping distances. Based on the causal relation obtained, CHWD prediction models are developed using a Support Vector Machine, considering HTLP/HT/LP as predictors. The model shows a high accuracy in almost <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(96\%\)</EquationSource> </InlineEquation> of grids in Europe, while Africa shows the highest accuracy in predicting CHWDs for individual years. By incorporating individual and concurrent extreme events, we established a new approach for predicting CHWDs and improving preparedness for such events.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Causality between concurrent heatwaves and droughts and high temperature and low precipitation extremes

  • 期刊论文

  • 发表日期：

    2026年7月29日

查看全文
  • Devjit Sinha,
  • Chandra Rupa Rajulapati

摘要

Hydrometeorological processes associated with teleconnections can create lagged responses between extreme events over different temporal scales. Identification of causal relationships in such pairs of extremes can improve the accuracy of prediction algorithms and help with better risk mitigation. In this global analysis, we investigated whether extreme events like High Temperature and Low Precipitation (HTLP) days, High Temperature (HT) days, and Low Precipitation (LP) days can be used to predict concurrent heatwaves and short-term moderate droughts (CHWD). Using Spearman's rank correlation, mutual information, and dynamic time warping, our results indicated a strong dependence between these variables. Based on the observed relationship, we investigated the existence of causality between HTLP, HT and LP with the CHWD of the next year using Granger Causality and Transfer Entropy. Our results showed that HTLP and HT are the better predictors for equatorial regions, while HTLP and LP are the better predictors for mid-latitudes and Arctic regions. Furthermore, HTLP and HT show higher pattern similarity and temporal alignment with CHWD as compared to LP, as evidenced by lower dynamic time warping distances. Based on the causal relation obtained, CHWD prediction models are developed using a Support Vector Machine, considering HTLP/HT/LP as predictors. The model shows a high accuracy in almost \(96\%\) of grids in Europe, while Africa shows the highest accuracy in predicting CHWDs for individual years. By incorporating individual and concurrent extreme events, we established a new approach for predicting CHWDs and improving preparedness for such events.