Scholarlink

<p>Climate variability has significant impacts on agricultural productivity in the Northeastern region (NER), especially during the monsoon season. This study investigates the trends and implications of key climate parameters — maximum temperature (T<sub>max</sub>), minimum temperature (T<sub>min</sub>), rainfall anomaly index (RAI), net shortwave radiation flux (NSWRF), soil moisture (SM), and relative humidity (RH) — on rice yield (Kharif) throughout the period from 1991–2020. Results indicate a significant increasing trend in T<sub>min</sub> (0.02&#xa0;°C&#xa0;yr⁻<sup>1</sup> in Phase-II), which exhibits the strongest positive correlation (r = 0.70, <i>p</i> &lt; 0.0001) with rice yield. A shift from a drying trend in Phase-I (-0.034&#xa0;yr⁻<sup>1</sup>) to a wetting trend in Phase-II (0.18&#xa0;yr⁻<sup>1</sup>) contributes to improved soil moisture and humidity, enhancing rice productivity. Despite an overall increasing yield trend (30.7&#xa0;kg&#xa0;ha⁻<sup>1</sup>&#xa0;yr⁻<sup>1</sup>), extreme dry conditions in 2006–07 results in yield anomalies of -6.5% and -4.8% respectively. Machine learning models, Random Forest (RF) and Extreme Gradient Boosting (XGBoost) are used to predict rice yield, with RF demonstrating superior accuracy (R<sup>2</sup> = 0.78, RMSE = 117.03, MAPE = 5.70%) compared to XGBoost (R<sup>2</sup> = 0.71, RMSE = 133.25, MAPE = 6.42%). Both models identify T<sub>min</sub> and T<sub>max</sub> as the most influential variables, followed by NSWRF and RAI. Overall, this study highlights the critical role of T<sub>min</sub> in improving rice yields and offers valuable insights for developing sustainable agricultural strategies and decision-making processes to ensure long-term food security amidst evolving climatic conditions.</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

Rice yield responses to climate variability in Northeast India using machine learning approach

  • 期刊论文

  • 发表日期：

    2025年4月2日

查看全文
  • Niki Gogoi,
  • Binita Pathak,
  • Rizwan Rehman,
  • Sristisri Upadhyaya,
  • Pranami Mahanta,
  • Anindita Borah,
  • Krishnanka Jyoti Baishya,
  • Kalyan Bhuyan

摘要

Climate variability has significant impacts on agricultural productivity in the Northeastern region (NER), especially during the monsoon season. This study investigates the trends and implications of key climate parameters — maximum temperature (Tmax), minimum temperature (Tmin), rainfall anomaly index (RAI), net shortwave radiation flux (NSWRF), soil moisture (SM), and relative humidity (RH) — on rice yield (Kharif) throughout the period from 1991–2020. Results indicate a significant increasing trend in Tmin (0.02 °C yr⁻1 in Phase-II), which exhibits the strongest positive correlation (r = 0.70, p < 0.0001) with rice yield. A shift from a drying trend in Phase-I (-0.034 yr⁻1) to a wetting trend in Phase-II (0.18 yr⁻1) contributes to improved soil moisture and humidity, enhancing rice productivity. Despite an overall increasing yield trend (30.7 kg ha⁻1 yr⁻1), extreme dry conditions in 2006–07 results in yield anomalies of -6.5% and -4.8% respectively. Machine learning models, Random Forest (RF) and Extreme Gradient Boosting (XGBoost) are used to predict rice yield, with RF demonstrating superior accuracy (R2 = 0.78, RMSE = 117.03, MAPE = 5.70%) compared to XGBoost (R2 = 0.71, RMSE = 133.25, MAPE = 6.42%). Both models identify Tmin and Tmax as the most influential variables, followed by NSWRF and RAI. Overall, this study highlights the critical role of Tmin in improving rice yields and offers valuable insights for developing sustainable agricultural strategies and decision-making processes to ensure long-term food security amidst evolving climatic conditions.