多源驱动与季节性集成学习的SMAP表层土壤水分降尺度方法

Downscaling SMAP surface soil moisture using multi-source drivers and seasonal ensemble learning

摘要

针对土壤水分主动被动卫星(soil moisture active passive,SMAP)土壤水分产品空间分辨率较粗、难以满足区域精细化应用需求的问题,本文以河南省为研究区,提出一种多源驱动与季节性集成学习相结合的1km日尺度表层土壤水分(surface soil moisture,SSM)降尺度方法.以SMAP L4 SSM为目标变量,融合遥感植被与能量因子、气象降水因子、土壤属性及地形因子,采用“9km训练-1km推理”策略,构建基于Top-k优选与Stacking的季节性集成模型SOBLE-Top5.结果表明,分季节建模整体优于全年统一建模;SOBLE-Top5在春、夏、秋、冬测试集上的R²分别为0.9389、0.9078、0.8977和0.9632,均方根误差(RMSE)分别为0.0208、0.0291、0.0222和0.0133m³/m³.站点验证表明,降尺度结果能够较好表征不同地类下SSM时序变化.SHAP分析显示,SSM驱动机制具有明显季节差异,整体呈现“冷季土壤属性主导、暖季多因子协同驱动”的特征.研究可为区域农业干旱监测和水资源管理提供高分辨率土壤水分数据支持.

Abstract

To address the problem that soil moisture active passive (SMAP) soil moisture products have coarse spatial resolution and are insufficient for regional fine-scale applications, this study proposes a 1 km daily-scale surface soil moisture (SSM) downscaling method that integrates multi-source driving factors with seasonal ensemble learning, sampling Henan Province as the study area. SMAP L4 SSM was used as the target variable, and remote sensing vegetation and energy factors, meteorological precipitation factors, soil properties, and topographic factors were integrated. A "9 km training-1 km inference" strategy was adopted, and a seasonal ensemble model named SOBLE-Top5 was constructed based on Top-k selection and Stacking. The results show that seasonal modeling generally outperformed annual unified modeling. The R ² values of SOBLE-Top5 on the spring, summer, autumn, and winter test sets were 0.9389, 0.9078, 0.8977 and 0.9632, respectively, while the root mean square error (RMSE) values were 0.0208, 0.0291, 0.0222, and 0.0133 m ³/m ³, respectively. Site-based validation indicated that the downscaled results effectively captured the temporal variations of SSM under different land-cover types. SHAP analysis revealed pronounced seasonal differences in the driving mechanisms of SSM, generally characterized by "soil-property dominance in cold seasons and multi-factor synergistic control in warm seasons." This study can provide high-resolution soil moisture data support for regional agricultural drought monitoring and water resource management.

关键词

SMAP / 表层土壤水分 / 降尺度 / 季节性建模 / 集成学习

Key words

SMAP / surface soil moisture / downscaling / seasonal modeling / ensemble learning

参考文献

[1] 王晓燕, 李净, 邢立亭. 基于3种机器学习方法的农业干旱监测比较[J]. 干旱区研究, 2022, 39(1): 322-332.

[2] Wang XY, Li J, Xing L T. Comparative agricultural drought monitoring based on three machine learning methods[J]. Arid Zone Research, 2022, 39(1): 322-332.

[3] 王晓东, 陈金华, 陈曦, 等 . 淮河流域农田旱涝逐日监测指标优化及适用性分析[J]. 农业工程学报, 2021, 37(23): 117-126.

[4] Wang XD, Chen JH, Chen X, et al. Optimization and applicability analysis of daily farmland drought and flood monitoring index in Huaihe River Basin[J]. Transactions of the Chinese Society of Agricultural Engineering, 2021, 37(23): 117-126.

[5] Entekhabi D, Njoku EG, O'Neill P, et al. The soil moisture active passive (SMAP) mission[J]. Proceedings of the IEEE, 2010, 98(5): 704-716.

[6] Reichle RH, De Lannoy GJM, Liu Q, et al. Global assessment of the SMAP level-4 surface and root-zone soil moisture product using assimilation diagnostics[J]. Journal of Hydrometeorology, 2017, 18(12): 3217-3237.

[7] Ur Rahman K, Shang SH, Balkhair K, et al. Catchment-scale drought propagation assessment in the Indus Basin of Pakistan using a combined approach of principal components and wavelet analyses[J]. Journal of Hydrometeorology, 2023, 24(4): 601-624.

[8] Colliander A, Fisher JB, Halverson G, et al. Spatial downscaling of SMAP soil moisture using MODIS land surface temperature and NDVI during SMAPVEX15[J]. IEEE Geoscience and Remote Sensing Letters, 2017, 14(11): 2107-2111.

[9] Colliander A, Reichle RH, Crow WT, et al. Validation of soil moisture data products from the NASA SMAP mission[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2022, 15: 364-392.

[10] Teixeira AC, Bakon M, Lopes D, et al. A systematic review on soil moisture estimation using remote sensing data for agricultural applications[J]. Science of Remote Sensing, 2025, 12: 100328.

[11] Ben Abbes A, Jarray N, Farah IR . Advances in remote sensing based soil moisture retrieval: applications, techniques, scales and challenges for combining machine learning and physical models[J]. Artificial Intelligence Review, 2024, 57(9): 224.

[12] Senanayake IP, Pathira Arachchilage KRL, Yeo IY, et al. Spatial downscaling of satellite-based soil moisture products using machine learning techniques: a review[J]. Remote Sensing, 2024, 16(12): 2067.

[13] Abbaszadeh P, Moradkhani H, Zhan XW . Downscaling SMAP radiometer soil moisture over the CONUS using an ensemble learning method[J]. Water Resources Research, 2019, 55(1): 324-344.

[14] Massoud EC, Lee H, Gibson PB, et al. Bayesian model averaging of climate model projections constrained by precipitation observations over the contiguous United States[J]. Journal of Hydrometeorology, 2020, 21(10): 2401-2418.

[15] 马柱国, 魏和林, 符淙斌. 土壤湿度与气候变化关系的研究进展与展望[J]. 地球科学进展, 1999, 14(3): 299-305.

[16] Ma ZG, Wei HL, Fu CB. Progress in the research on the relationship between soil moisture and climate change[J]. Advances in Earth Science, 1999, 14(3): 299-305.

[17] 鲍琳. 多源遥感数据协同解译地表土壤湿度变化研究[D].西安:长安大学, 2023.

[18] Bao L. Collaborative interpretation of surface soil moisture changes with multi-source remote sensing data[D]. Xi'an: Changan University, 2023.

[19] Meng XJ, Zeng JY, Yang YB, et al. High-resolution soil moisture mapping through passive microwave remote sensing downscaling[J]. The Innovation Geoscience, 2024, 2(4): 100105.

[20] Singh A, Niranjannaik M, Gaurav K . Downscaling SMAP soil moisture using a hybrid machine-learning algorithm[J]. Applied Soft Computing, 2026, 193: 114766.

[21] Wang J, Xue HZ, Dong GT, et al. Downscaling of remote sensing soil moisture products that integrate microwave and optical data[J]. Applied Sciences, 2024, 14(24): 11875.

[22] Yang N, Xiang F, Zhang HJ . The characterization of the vertical distribution of surface soil moisture using ISMN multilayer in situ data and their comparison with SMOS and SMAP soil moisture products[J]. Remote Sensing, 2023, 15(16): 3930.

[23] Mosca E, Szigeti F, Tragianni S, et al. SHAP-based explanation methods: a review for NLP interpretability[C]// Proceedings of the 29th International Conference on Computational Linguistics. Gyeongju, Republic of Korea Association for Computational Linguistics, 2022: 4593-4603.

基金资助

国家自然科学基金(U1304402)

国家自然科学基金(41977284)

河南省水利科技攻关项目(GG202412)

河南省自然资源科研项目(2019-379-16)