Scholarlink

<p>Climate change is recognized as an influential phenomenon on many natural processes, including the hydrological cycle. One of the key components of this hydrological cycle is evapotranspiration, which will be affected by climatic changes. Given the importance of evapotranspiration in water resource management and agricultural planning, the present study aims to investigate the effect of climate change on this process in the Kermanshah Plain, Located in the west of Iran. In this regard, using the GEESEBAL model, the amount of wheat evapotranspiration was calculated over a period of 10 years in the Kermanshah plain. Furthermore, to examine and forecast the effects of climate change, the MIROC6 and MRI-ESM2-0 models under SSP126, SSP245, and SSP585 scenarios were analyzed. Ultimately, utilizing Machine Learning methods, actual evapotranspiration for the period from 2025 to 2065 was estimated. Based on the results of the simulated models MIROC6 and MRI-ESM2-0, precipitation in future periods will have significant fluctuations. This study integrates remote sensing-based ET estimation with CMIP6 climate projections and machine learning techniques (Random Forest, XGBoost, and Gradient Boosting) to forecast future wheat evapotranspiration under different SSP scenarios. The results of investigation showed that in both models, temperature will increase and relative humidity will decrease. On the other hand, the MIROC6 model tends to predict a decrease in wind speed, whereas the MRI-ESM2-0 model predicts a slight increase across of all scenarios. The examination of the percentage changes in evapotranspiration using Random Forest, XGBOOST, and GBOOST methods indicated that the greatest decrease and increase will occur in March and November respectively, emphasizing the significance of the impact of climatic changes on evapotranspiration. Given the high values of R<sup>2</sup> and NSE, low MAE, and RMSE close to zero, it can be concluded that the performance of the Random Forest, XGBOOST, and GBOOST models in predicting actual evapotranspiration in the future has been satisfactory.</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

Assessing future wheat evapotranspiration dynamics in the Kermanshah plain using CMIP6 and GEESEBAL: integration of SSP scenarios and machine learning-based projections

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月3日

查看全文
  • Neda Baboli,
  • Maryam Hafezparast Mavaddat,
  • Houshang Ghamarnia

摘要

Climate change is recognized as an influential phenomenon on many natural processes, including the hydrological cycle. One of the key components of this hydrological cycle is evapotranspiration, which will be affected by climatic changes. Given the importance of evapotranspiration in water resource management and agricultural planning, the present study aims to investigate the effect of climate change on this process in the Kermanshah Plain, Located in the west of Iran. In this regard, using the GEESEBAL model, the amount of wheat evapotranspiration was calculated over a period of 10 years in the Kermanshah plain. Furthermore, to examine and forecast the effects of climate change, the MIROC6 and MRI-ESM2-0 models under SSP126, SSP245, and SSP585 scenarios were analyzed. Ultimately, utilizing Machine Learning methods, actual evapotranspiration for the period from 2025 to 2065 was estimated. Based on the results of the simulated models MIROC6 and MRI-ESM2-0, precipitation in future periods will have significant fluctuations. This study integrates remote sensing-based ET estimation with CMIP6 climate projections and machine learning techniques (Random Forest, XGBoost, and Gradient Boosting) to forecast future wheat evapotranspiration under different SSP scenarios. The results of investigation showed that in both models, temperature will increase and relative humidity will decrease. On the other hand, the MIROC6 model tends to predict a decrease in wind speed, whereas the MRI-ESM2-0 model predicts a slight increase across of all scenarios. The examination of the percentage changes in evapotranspiration using Random Forest, XGBOOST, and GBOOST methods indicated that the greatest decrease and increase will occur in March and November respectively, emphasizing the significance of the impact of climatic changes on evapotranspiration. Given the high values of R2 and NSE, low MAE, and RMSE close to zero, it can be concluded that the performance of the Random Forest, XGBOOST, and GBOOST models in predicting actual evapotranspiration in the future has been satisfactory.