Page 69 - 《高原气象》2026年第3期
P. 69
3 期 刘文博等:基于Bi-LSTM模型和多源数据融合的玛曲地区土壤水分模拟研究 677
交换特征分析[J]. 高原气象, 42(3): 575-589. DOI: 10. 7522/ (13): 68-76. DOI: 10. 11975/j. issn. 1002-6819. 202312128.
j. issn. 1000-0534. 2022. 00083. Zhang G, Lai X, Liu K, 2023. 赵伟, 文凤平, 蔡俊飞, 2022. 被动微波土壤水分遥感产品空间降尺
Characteristics of surface water and heat exchange during soil 度研究: 方法、 进展及挑战[J]. 遥感学报, 26(9): 1699-1722.
freezing and thawing of Maqu Station in the source area of the Zhao W, Wen F P, Cai J F, 2022. Methods, progresses, and chal‐
Yellow River[J]. Plateau Meteorology, 42(3): 575-589. DOI: lenges of passive microwave soil moisture spatial downscaling[J].
10. 7522/j. issn. 1000-0534. 2022. 00083. National Remote Sensing Bulletin, 26(9): 1699-1722.
张蕾, 郭安红, 宋迎波, 等, 2024. 土壤水分数据融合及其在旱涝灾 郑涵, 王海峰, 2022. 土壤水分计量技术的研究进展[J]. 计量科学
害多维度评估中的应用[J]. 农业工程学报, 40(13): 68-76. 与 技 术 , 66(11): 31-36, 40. DOI: 10. 12338/j. issn. 2096-
DOI: 10. 11975/j. issn. 1002-6819. 202312128. Zhang L, Guo A 9015. 2022. 0169. Zheng H, Wang H F, 2022. Research progress
H, Song Y B, et al, 2024. Assessing drought and waterlogging of soil moisture measurement technology[J]. Metrology Science
in multiple dimensions using data fusion of soil water[J]. Trans‐ and Technology, 66(11): 31-36, 40. DOI: 10. 12338/j. issn.
actions of the Chinese Society of Agricultural Engineering, 40 2096-9015. 2022. 0169.
Soil Moisture Simulation in Maqu Using a Bi-LSTM Model
and Multi-Source Data Fusion
LIU Wenbo , LI Chunbin , WU Jing , MA Yuanyuan 2
1
1
1
(1. College of Resources and Environment, Gansu Agricultural University, Lanzhou 730070, Gansu, China;
2. State Key Laboratory of Cryospheric Science and Frozen Soil Engineering, Northwest Institute of Eco-Environment
and Resources, Chinese Academy of Sciences, Lanzhou 730000, Gansu, China)
Abstract: Soil moisture is the key variable of the land-atmosphere interactions and ecosystem dynamics, so ob‐
taining the soil moisture data with high spatiotemporal resolution is significant for simulating the hydrological
process and managing the resource. The study is Maqu County ,which is located on the eastern edge of the Qing‐
hai-Tibet Plateau. We choose the SMAP daily soil moisture products and hourly in situ observations from ISMN
as the base data. we build a Bi-LSTM model by using the GEE technology to construct the data in temporal
scale, and by utilizing the random forest regression to downscale the soil moisture from 9 km to 250 m with the
multi-high-resolution data (NDVl, DEM, and LST). The results show that Bi-LSTM model with a ratio con‐
strained correction can estimate the hourly soil moisture effectively, have a well performance in multi-site valida‐
tion with a maximum R of 0. 8735, and the random forest model give us a distribution characterization with
2
more refined spatial scale. In this study, we refine the remotely sensed soil moisture in space-time dimension syn‐
chronously, and overcome the limitations of traditional approaches which build model by downscaling spatial
scale only.
Key words: soil moisture; Bi-LSTM; random forest; spatiotemporal downscaling

