水土保持学报2025，Vol.39Issue(2)：111-124,14.DOI:10.13870/j.cnki.stbcxb.2025.02.014
淮北平原砂姜黑土区典型作物根区土壤水分的来源及其量化
The Sources and Quantification of Soil Water in the Root Zone of Typical Crops in the Shajiang Black Soil Region of Huaibei Plain
摘要
Abstract
[Objective]The sources and quantitative characteristics of soil moisture in the root zone of crops in the Huaibei Plain from 1990 to 2022 were explored to provide scientific guidance for crop irrigation management and water resource utilization.[Methods]The contribution of precipitation and groundwater to soil water in the root zone was calculated by Pearson correlation analysis and multiple regression,and soil water quantitative simulation was performed using multiple linear regression and long short-term memory network(LSTM)model.[Results]1)For summer corn,in the 0-10 cm soil layer,except for the 55%contribution of precipitation during the seedling-jointing period,the contribution of groundwater in other stages could reach 61%—68%;in the soil layer below 10 cm,the soil water in the root zone mainly came from soil water in the adjacent upper layer,accounting for more than 74%.In the later stage,the root system gradually growed to 100 cm,especially in the 40-80 cm soil layer,the contribution of groundwater to soil water increased significantly,accounting for 9%—22%.2)For winter wheat,in the 0-10 cm soil layer,except for the sowing-seedling period and the greening-jointing period,the contribution of groundwater in other stages could reach 54%—73%;in the soil layer below 10 cm,the soil water in the root zone mainly came from soil water in the adjacent upper layer,accounting for more than 63%.However,in the 40-80 cm soil layer,the contribution of groundwater and precipitation increased.3)The LSTM model was more stable and accurate than multiple linear regression in quantitatively simulating soil water.The model evaluation index could reach R2>0.6 and MSE<10,which could better reflect the dynamic changes of soil moisture.The application of the above simulation results at other stations in the Huaibei Plain further verified the feasibility of the model.[Conclusion]The main source of soil water in the 0-10 cm layer for summer corn is groundwater,and the main source of soil water in the 0-10 cm layer for winter wheat is precipitation;the main source of soil water below 10 cm is soil water in the adjacent upper layer(including precipitation,irrigation water,etc.).Performing quantitative simulation based on the sources of soil water in the root zone at different growth stages,and dynamically adjusting the irrigation plan through the changing trend of soil moisture in the root zone of the crop is helpful to optimize irrigation management and improve water resource utilization efficiency.关键词
淮北平原/作物/土壤水/相关性/多元回归/深度学习Key words
Huaibei Plain/crops/soil water/correlation/multiple regression/deep learning分类
农业科技引用本文复制引用
朱婉婷,方逸敏,朱永华,王启猛,吕海深,刘名文,王怡宁,陆玟,彭宇杰,许海婷..淮北平原砂姜黑土区典型作物根区土壤水分的来源及其量化[J].水土保持学报,2025,39(2):111-124,14.基金项目
国家自然科学基金项目(42071033,41830752) （42071033,41830752）
