Journal of Geo-information Science >
Construction and Application of Active Layer Thickness Retrieval Model in Permafrost Regions Considering Soil Water and Heat Changes
Received date: 2024-09-05
Revised date: 2024-12-13
Online published: 2025-03-06
Supported by
National Natural Science Foundation of China(41964008)
National Natural Science Foundation of China(42364003)
[Objectives] Interferometric Synthetic Aperture Radar (InSAR) technology has been widely used to retrieve Active Layer Thickness (ALT) in recent years, however, existing studies have less consideration for the effects of freeze-thaw on surface deformation and soil pore hydrothermal changes, therefore, in this paper, an ALT retrieval model is constructed that considering soil water and heat changes. [Methods] The surface parameters were obtained using the InSAR technique and the CNN-BiLSTM-AM model, and the active layer thickness inversion model was constructed by taking into account the deformation of the active layer and the changes in soil pore space and water under the freeze-thaw drive. Initially, surface deformation time series from both ascending and descending Sentinel-1 data were derived using the SBAS-InSAR method, and two-dimensional decomposition was applied to obtain vertical surface deformation across the study area. Subsequently, a CNN-BiLSTM-AM model was constructed using multi-source remote sensing data, with Convolutional Neural Networks (CNN) employed to extract features from the input data. A Bidirectional Long Short-Term Memory network (BiLSTM) was used to predict these features, and a multi-head self-attention layer (AM) was added to enhance the model’s extraction of critical information, ultimately providing predicted soil moisture levels under multi-feature constraints. The vertical surface deformation time series was then used as a primary parameter to characterize changes in the active layer. By integrating soil porosity and moisture data, an ALT retrieval model was developed, yielding a detailed spatiotemporal distribution of active layer thickness along the Lanzhou-Urumqi High-speed Railway in the permafrost region. [Results] The RMSE of the model estimates validated against the measured data of the Eboling Mountain is 0.065 m, while the RMSE validated against the active layer change data of the Qinghai-Tibetan Plateau from 2018 to 2020 is 0.697 m, 0.639 m, and 0.776 m, respectively, which shows that the model has a high degree of accuracy. [Conclusions] The model presented here, based on both surface and internal soil changes in spatial structure and moisture under freeze-thaw effects, offers a novel approach for monitoring ALT in permafrost regions and for characterizing permafrost changes in a more nuanced manner.
ZHANG Delong , WEI Guanjun . Construction and Application of Active Layer Thickness Retrieval Model in Permafrost Regions Considering Soil Water and Heat Changes[J]. Journal of Geo-information Science, 2025 , 27(3) : 750 -765 . DOI: 10.12082/dqxxkx.2025.240498
表1 Sentinel-1数据参数Tab. 1 Sentinel-1 data parameter |
| Sentinel-1 | 时间范围 | 影像数/景 | 轨道方向 | 入射角/° | 方位角/° |
|---|---|---|---|---|---|
| Path 128 Frame 119 | 2017-03-20—2021-12-24 | 141 | 升轨 | 34.173 167 | -13.242 437 |
| Path 33 Frame 467 | 2017-02-05—2021-12-29 | 143 | 降轨 | 34.096 453 | -166.681 229 |
利益冲突:Conflicts of Interest 所有作者声明不存在利益冲突。
All authors disclose no relevant conflicts of interest.
| [1] |
|
| [2] |
郑度, 姚檀栋. 青藏高原形成演化及其环境资源效应研究进展[J]. 中国基础科学, 2004, 6(2):17-23.
[
|
| [3] |
罗栋梁, 雷汶杰, 康建芳, 等. 高山多年冻土区地面温度研究进展[J]. 草业科学, 2023, 40(4):942-964.
[
|
| [4] |
姚檀栋. 青藏高原水-生态-人类活动考察研究揭示“亚洲水塔”的失衡及其各种潜在风险[J]. 科学通报, 2019, 64(27):2761-2762.
[
|
| [5] |
罗栋梁, 金会军, 吴青柏, 等. 天然状态下多年冻土区活动层厚度研究进展与展望[J]. 冰川冻土, 2023, 45(2):558-574.
[
|
| [6] |
程国栋, 赵林, 李韧, 等. 青藏高原多年冻土特征、变化及影响[J]. 科学通报, 2019, 64(27):2783-2795.
[
|
| [7] |
|
| [8] |
|
| [9] |
罗京, 牛富俊, 林战举, 等. 青藏高原多年冻土区热融滑塌发育特征及规律[J]. 冰川冻土, 2022, 44(1):96-105.
[
|
| [10] |
闫家倩, 王雯杰, 韩晨. 青藏高原多年冻土活动层厚度探测方法综述[J]. 地质论评, 2021, 67(S01):209-211.
[
|
| [11] |
|
| [12] |
|
| [13] |
|
| [14] |
|
| [15] |
张中琼, 吴青柏. 气候变化情景下青藏高原多年冻土活动层厚度变化预测[J]. 冰川冻土, 2012, 34(3): 505-511.
[
|
| [16] |
徐晓明, 吴青柏, 张中琼. 青藏高原多年冻土活动层厚度对气候变化的响应[J]. 冰川冻土, 2017, 39(1):1-8.
[
|
| [17] |
|
| [18] |
|
| [19] |
|
| [20] |
赵蓉. 基于SBAs-InSAR的冻土形变建模及活动层厚度反演研究[D]. 长沙: 中南大学, 2014.
[
|
| [21] |
|
| [22] |
|
| [23] |
|
| [24] |
|
| [25] |
|
| [26] |
张成才, 吴泽宁, 余弘婧. 遥感计算土壤含水量方法的比较研究[J]. 灌溉排水学报, 2004, 23(2):69-72.
[
|
| [27] |
陈鹏. 基于深度学习的多光谱视频目标跟踪方法研究[D]. 无锡: 江南大学, 2023.
[
|
| [28] |
付平凡, 杨晓静, 苏志诚, 等. 基于集成学习的土壤含水量预测研究——以辽西地区为例[J]. 土壤, 2023, 55(3):671-681.
[
|
| [29] |
耿庆田, 刘植, 李清亮, 等. 基于一种深度学习模型的土壤湿度预测[J]. 吉林大学学报(工学版), 2023, 53(8):2430-2436.
[
|
| [30] |
|
| [31] |
|
| [32] |
毛雪松, 胡长顺, 窦明健, 等. 正冻土中水分场和温度场耦合过程的动态观测与分析[J]. 冰川冻土, 2003, 25(1):55-59.
[
|
| [33] |
|
| [34] |
彭晨阳, 盛煜, 吴吉春, 等. 祁连山区多年冻土空间分布模拟[J]. 冰川冻土, 2021, 43(1):158-169.
[
|
| [35] |
|
| [36] |
|
| [37] |
周纪, 王子卫, 丁利荣, 等. 青藏高原冰冻圈温度遥感观测、反演与应用[J]. 测绘学报, 2024, 53(5): 835-847.
[
|
| [38] |
|
| [39] |
王庆锋, 金会军, 张廷军, 等. 祁连山区黑河上游高山多年冻土区活动层季节冻融过程及其影响因素[J]. 科学通报, 2016, 61(24):2742-2756.
[
|
| [40] |
张凤, 范成彦, 牟翠翠, 等. 积雪对祁连山区黑河上游活动层热状态的影响研究[J]. 冰川冻土, 2021, 43(6):1628-1640.
[
|
| [41] |
|
| [42] |
杜冉, 彭小清, 金浩东, 等. 祁连山俄博岭地区热融洼地与冻胀草丘活动层融化深度差异性对比研究[J]. 冰川冻土, 2022, 44(1):188-202.
[
|
| [43] |
|
/
| 〈 |
|
〉 |