Scholarlink

<p>Climate variability impacts agricultural productivity, requiring advanced models to assess these effects. Traditional econometric models often fail to capture the nonlinear climate-yield relationships. This study addresses this gap by combining wavelet decomposition with the Nonlinear Autoregressive Distributed Lag (NARDL) model to improve the accuracy and interpretation of climate-yield dynamics. A new model, the dual wavelet enhanced NARDL (DWaveNARDL), is proposed, which uses wavelet transformation to decompose climate and yield data, followed by NARDL to capture both short- and long-term asymmetric effects. Using crop yield data (1966–2022) and climatic variables from six districts in West Bengal, India, the study shows that wavelet-based models outperform traditional approaches. The DWaveNARDL model reduced AIC and BIC by approximately 12–18%. Among the four wavelet filters tested, LA8 and BL14 produced the best model-fitting performance, as reflected by lower AIC and BIC values. Their superior performance is attributed to smoother signal approximation and better noise localization, which allow the model to preserve important low-frequency climate–yield information while more effectively capturing asymmetric effects. The DWaveNARDL model provides a robust framework for understanding climate impacts on crop yields, aiding agricultural planning and policy decisions for climate resilience.</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

Exploring the impact of climate change on crop yields: A deep dive with dual wavelet enhanced nonlinear ARDL modeling

  • 期刊论文

  • 发表日期：

    2026年7月11日

查看全文
  • Md Yeasin,
  • Anita Sarkar,
  • Ranjit Kumar Paul,
  • Ranjit Kumar Upadhyay,
  • Bitan Mondal,
  • Arti Thakur,
  • Himadri Sekhar Roy,
  • Prakash Kumar

摘要

Climate variability impacts agricultural productivity, requiring advanced models to assess these effects. Traditional econometric models often fail to capture the nonlinear climate-yield relationships. This study addresses this gap by combining wavelet decomposition with the Nonlinear Autoregressive Distributed Lag (NARDL) model to improve the accuracy and interpretation of climate-yield dynamics. A new model, the dual wavelet enhanced NARDL (DWaveNARDL), is proposed, which uses wavelet transformation to decompose climate and yield data, followed by NARDL to capture both short- and long-term asymmetric effects. Using crop yield data (1966–2022) and climatic variables from six districts in West Bengal, India, the study shows that wavelet-based models outperform traditional approaches. The DWaveNARDL model reduced AIC and BIC by approximately 12–18%. Among the four wavelet filters tested, LA8 and BL14 produced the best model-fitting performance, as reflected by lower AIC and BIC values. Their superior performance is attributed to smoother signal approximation and better noise localization, which allow the model to preserve important low-frequency climate–yield information while more effectively capturing asymmetric effects. The DWaveNARDL model provides a robust framework for understanding climate impacts on crop yields, aiding agricultural planning and policy decisions for climate resilience.