Interpreting Conv-LSTM for Spatio-Temporal Soil Moisture Prediction in China
, Yongkun Zhang 1, Ye Zhang 1, Wei Shangguan 1,*, Qingliang Li 2, Lu Li 1 and Shijie Jiang 3
Round 1
Reviewer 1 Report
1. Line 108: You mentioned SM variability was high over QTP. Did you mean spatial variability?
2. Line 115: source domain? Or source dataset?
3. Line 124: Reframe sentence. Typos occurred.
4. How did you select the predictors? Did you perform any sensitivity experiment?
Author Response
Point 1: Line 108: You mentioned SM variability was high over QTP. Did you mean spatial variability?
Response 1: Thank you so much for the comment. Yes, we mean spatial variability and we revised it in Line 110.
Point 2: Line 115: source domain? Or source dataset?
Response 2: We feel sorry for the inconvenience brought to the reviewer. We used the inappropriate word. In fact, we mean the source dataset. And it has been revised in Line 117 as follows:
In this study, we used ERA5-Land as the source dataset for establishing a model to estimate SM in China.
Point 3: Line 124: Reframe sentence. Typos occurred.
Response 3: Thank you so much for your careful check. The sentence was revised in line 127-128, as follows:
The variables are effective because they influence the dynamics of soil moisture through natural processes such as evaporation, transpiration, infiltration, and runoff.
Point 4: How did you select the predictors? Did you perform any sensitivity experiment?
Response 4: First of all, we selected the predictors based on the previous data driven works (Fang and Shen, 2020; Han et al. 2023, Zhang et al., 2023). The meteorological forcing variables in our work are total precipitation (P), 10 m U-wind component (U), 10 m V-wind component (V), 2 m temperature (TA), net surface solar radiation (SR), and net surface thermal radiation (TR). They are also commonly used in physical-based modeling (Sheffield et al., 2006; Balsamo et al., 2009). And the static variables included soil properties (sand (SAND), silt (SILT), clay (CLAY) content, bulk density (BULK), land cover type (LAND) and digital elevation model (DEM). Many works have proven their inevitable functions in predicting soil moisture (Zhang et al., 2023). Secondly, we did not do sensitivity experiment ourselves but referred to the sensitivity experiment in O. and Orth’s work (2021), in which the permutation importance result shows that these static variables are main contributor to predict the surface soil moisture.
Balsamo, G. et al. A revised hydrology for the ECMWF model: Verification from field site to terrestrial water storage and impact in the Integrated Forecast System. J. Hydrometeorol. 10, 623–643, (2009). https://doi.org/10.1175/2008JHM1068.1 .
Fang, K., & Shen, C. Near-Real-Time Forecast of Satellite-Based Soil Moisture Using Long Short-Term Memory with an Adaptive Data Integration Kernel, Journal of Hydrometeorology, 21(3), 399-413. (2020). https://doi.org/10.1175/JHM-D-19-0169.1
Han, Q., et al. Global long term daily 1 km surface soil moisture dataset with physics informed machine learning. Sci Data 10, 101 (2023). https://doi.org/10.1038/s41597-023-02011-7
O., S., & Orth, R. Global soil moisture data derived through machine learning trained with in-situ measurements. Sci Data 8, 170 (2021). https://doi.org/10.1038/s41597-021-00964-1
Sheffield, J., et al. Development of a 50-year high-resolution global dataset of meteorological forcings for land surface modeling. J. Clim. 19, 3088–3111, (2006). https://doi.org/10.1175/JCLI3790.1.
Zhang Y., et al. Real-Time Forecast of SMAP L3 Soil Moisture Using Spatial–Temporal Deep Learning Model with Data Integration. Remote Sensing. 15(2) 366, (2023). https://doi.org/10.3390/rs15020366
Author Response File: 
Author Response.pdf
Reviewer 2 Report
The paper entitled 'Interpreting Conv-LSTM for spatio-temporal soil moisture pre- 2 diction in China' is an interesting and well-structured work.
The topic is relevant both for what concerns the state of soil moisture, in a climate change perspective it is a factor of enormous importance, and for deep learning technology. As far as I am concerned, I find it a good work.
Author Response
Point 1: The paper entitled 'Interpreting Conv-LSTM for spatio-temporal soil moisture prediction in China' is an interesting and well-structured work.
The topic is relevant both for what concerns the state of soil moisture, in a climate change perspective it is a factor of enormous importance, and for deep learning technology. As far as I am concerned, I find it a good work.
Response 1: Many thanks for your positive comments. It is our great honours receiving your recommendation.
Author Response File: Author Response.pdf
Reviewer 3 Report
Appreciate the efforts of the authors. The presentation and clarity of the paper is good. I suggest to work on,
1. Enhance the quality of the images.
2. More explanations on the interpretability aspects.
Author Response
Point 1: Enhance the quality of the images.
Response 1: Thank you very much for your comments and suggestions and for helping us to improve the manuscript. All the images were improved and replaced. The quality has been enhanced.
Point 2: More explanations on the interpretability aspects.
Response 2: Thank you for your rigorous consideration. We improved the explanation about the interpretability in the introduction section. We revised the manuscript as follows:
In Lines 52-55:
Interpretability refers to a passive characteristic of a model referring to the level at which a given model makes sense for a human observer [17]. Enhancing interpretability means that we can extract relevant information from an AI model regarding relationships either contained in the data or learned by the model [15].
In Lines 546-549:
- Barredo Arrieta, A.; Díaz-Rodríguez, N.; Del Ser, J.; Bennetot, A.; Tabik, S., Barbado, A.; Garcia, S.; Gil-Lopez, S.; Molina, D.; Benjamins, R.; Chatila, R.; Herrera, F. Explainable Artificial Intelligence (XAI): Concepts, taxonomies, opportunities and challenges toward responsible AI. Inf. Fusion 2020, 58, 82–115, doi:10.1016/j.inffus.2019.12.012.
Author Response File: Author Response.pdf
