数字农科院2.0

Machine learning-based inversion and sensitivity analysis of soil moisture in Hemerocallis cultivation

文献类型： 外文期刊

作者： Jingshu Wang;Peng He; Keming Du;Xuran Li;Lishuai Xu; Fan Yang;Rutian Bi

关键词： Sentinel-1/2; Neural network; Random forest; Soil moisture; Yield estimation

期刊名称： Agricultural Water Management

ISSN： 0378-3774

年卷期： 2025 年

页码：

收录情况： SCIE（2025版） ; ; EI（2025版）

摘要： Accurate monitoring and prediction of soil moisture (SM) are essential for identifying water-sensitive growth stages, optimizing irrigation strategies, and improving agricultural productivity in Hemerocallis cultivation. This study develops a machine learning framework integrating TSM640 sensor, multi-source remote sensing (Sentinel-1/2), and meteorological datasets to analyze SM dynamics across soil layers and their impact on yield. The Backpropagation Neural Network (BPNN) model demonstrated robust performance in SM estimation (R2 = 0.64), with evaluation metrics of mean absolute error (MAE) = 1.89, mean bias error (MBE) = 0.91, and root mean square error (RMSE) = 2.60. The model’s multi-layer architecture proved particularly effective for handling high-dimensional feature spaces. A significant 0.15 improvement in R2 was achieved through the synergistic use of Sentinel-1 and Sentinel-2 data, with corresponding reductions in error metrics (MAE: − 0.57, MBE: − 0.81, RMSE: − 0.71) compared to single-source approaches. Incorporating meteorological data showed limited predictive enhancement due to the complex timing between moisture response and meteorological conditions. For yield prediction, the Random Forest (RF) algorithm outperformed BPNN (R2 = 0.63, MAE = 298.99, MBE = 78.16, RMSE = 366.84). Feature importance analysis revealed critical growth stages: Bolting (0.42), Squaring (0.32), Filling (0.28), and Seedling (0.25). Specifically, these results identify bolting and squaring phases as particularly moisture-sensitive periods, suggesting that targeted irrigation during these developmental windows could significantly improve Hemerocallis productivity.

分类号：

  • 相关文献

[1]基于GEE云平台与Sentinel数据的高分辨率水稻种植范围提取——以湖南省为例. 桑国庆,唐志光,毛克彪,邓刚,王靖文,李佳. 2022

[2]基于Sentinel-1/2数据的中国南方单双季稻识别结果一致性分析. 杨靖雅,胡琼,魏浩东,蔡志文,张馨予,宋茜,徐保东. 2022

[3]Estimation of plant height and yield based on UAV imagery in faba bean (Vicia faba L.). Yishan Ji,Zhen Chen,Qian Cheng,Rong Liu, Mengwei Li,Xin Yan,Guan Li, Dong Wang,Li Fu,Yu Ma, Xiuliang Jin, Xuxiao Zong, Tao Yang. 2022

[4]Improving digital mapping of soil organic matter in cropland by incorporating crop rotation. Yuan Liu,Songchao Chen,Qiangyi Yu,Zejiang Cai,Qingbo Zhou,Sonoko Dorothea Bellingrath-Kimura,Wenbin Wu. 2023

[5]Mapping the Complex Crop Rotation Systems in Southern China Considering Cropping Intensity, Crop Diversity, and Their Seasonal Dynamics. Liu, Yuan,Yu, Qiangyi,Zhou, Qingbo,Wang, Cong,Bellingrath-Kimura, Sonoko Dorothea,Wu, Wenbin. 2022

[6]An automated sample generation method by integrating phenology domain optical-SAR features in rice cropping pattern mapping. Jingya Yang,Qiong Hu,Wenjuan Li,Qian Song,Zhiwen Cai,Xinyu Zhang,Haodong Wei,Wenbin Wu. 2024

[7]Deep Memory And Prediction Neural N.etwork For Video Prediction. Liu, Zhipeng,Chen, Xilin,Chai, Xiujuan,Chai, Xiujuan,Chen, Xilin,Liu, Zhipeng. 2019

[8]An Advanced Radiative Transfer And N.eural Network Scheme And E valuation For Estimating Water Vapor Content From Modis Data. Mao, KB, Shen, XY, Zuo, ZY, Ma, Y, Liu, G, Tang, HJ. 2017

[9]Evaluation Of Green Tea Sensory Q.uality Via Process Characteristics A nd Image Information. Zhu, HK, Ye, Y, He, HF, Dong, CW. 2017

[10]Sequestration potential of phytolith occluded carbon in China's paddy rice (Oryza sativa L.) systems. Li Tan,Xiaoping Fan,Guochao Yan,Miao Peng,Nan Zhang,Mujun Ye,Zixiang Gao,Alin Song,Miroslav Nikolic,Yongchao Liang. 2021

[11]Long-Term Nitrogen Application Decreases The Abundance And Copy Number Of Predatory Myxobacteria And Alters The Myxobacterial Community Structure In The Soil. Wang, Wenhui,Cui, Zhongli,Wang, Hui,Gao, Jusheng,Wang, Boren,Wang, Ning,Wang, Wenhui,Dang, Keke,Dong, Yuanhua,Dai, Wei,Guan, Ling. 2020

[12]Net Surface Shortwave Radiation Retrieval Using Random Forest Method With Modis/Aqua Data. Li, Zhao-Liang,Wu, Hua,Wu, Hua,Wu, Hua,Ying, Wangmin,Ying, Wangmin,Li, Zhao-Liang. 2019

[13]Predictive Geographical Authentication Of Green Tea With Protected Designation Of Origin Using A Random Forest Model. Deng, XF, Liu, Z, Zhan, Y, Ni, K, Zhang, YZ, Ma, WZ, Shao, SZ, Lv, XN, Yuan, YW, Rogers, KM. 2020

[14]Deciphering The Routes Of Invasion O.f Drosophila Suzukii By M eans Of Abc Random Forest. Marin, Jean-Michel,Martinez-Sanudo, Isabel,Pascual, Marta,Kimura, Masahito T.,Loeb, Gregory,Loiseau, Anne,Price, Donald K.,Singh, Nadia,Tamura, Koichiro,Zhang, Jinping,Pudlo, Pierre,Richmond, Maxi Polihronakis,Shearer, Peter,Depra, Marindia,Duyck, Pierre Francois,Estoup, Arnaud,Cattel, Julien,Chen, Xiao,Debat, Vincent,Fellous, Simon,Hufbauer, Ruth A.,Hufbauer, Ruth A.,Kenis, Marc,Foucaud, Julien,Fraimout, Antoine,Guedot, Christelle. 2017

[15]Crop Classification Based On A N.ovel Feature Filtering And E nhancement Method. Wang, LM, Dong, QH, Yang, LB, Gao, JM, Liu, J. 2019

[16]Mapping Regional Cropping Patterns By U.sing Gf-1 Wfv Sensor D ata. Song, Q, Zhou, QB, Wu, WB, Hu, Q, Lu, M, Liu, SB. 2017

[17]A Practical Method For Reducing T.errain Effect On Land S urface Temperature Using Random Forest Regression. Zhao, W, Duan, SB, Li, AN, Yin, GF. 2019

[18]The Estimation Of Crop Emergence I.n Potatoes By Uav R gb Imagery. Li, B, Xu, XM, Han, JW, Zhang, L, Bian, CS, Jin, LP, Liu, JG. 2019

[19]Net Surface Shortwave Radiation Retrieval U.sing Random Forest Method W ith Modis/Aqua Data. Ying, WM, Wu, H, Li, ZL. 2019

[20]Application Of Machine Learning Algorithms I.n Quality Assurance Of F ermentation Process Of Black Tea- Based On Electrical Properties. Zhu, HK, Liu, F, Ye, Y, Chen, L, Li, JY, Gu, AH, Zhang, JQ, Dong, CW. 2019

作者其他论文 更多>>