the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 09 Jan 2023
A Study on the Fully Coupled Atmosphere-Land-Hydrology Process and Streamflow Simulations over the Source Region of the Yellow River
Abstract. The Source Region of the Yellow River (SRYR) is known as the "Water Tower of the Yellow River", which is the most important water conservation area in the upper reaches of the Yellow River. The streamflow of the SRYR makes an important contribution to the water resources in the Yellow River basin. Based on the Weather Research and Forecasting Model Hydrological modeling system (WRF-Hydro) model, by using meteorological, hydrological observations and reanalysis data, the key variables of the coupled atmosphere-land-hydrological processes over the SRYR during the 2013 rainy season (May–August) are analyzed, and the simulation results of the fully coupled WRF-Hydro with those of the standalone WRF are compared, whose aim is to assess the impact of hydrological coupling on the regional atmospheric model settings. The results show that the WRF-Hydro model has ability to depict the characteristics of streamflow over the SRYR with a Nash Efficiency Coefficient (NSE) of 0.44 during the calibration period from June 1st, 2012 to September 30th, 2012 and a NSE of 0.61 during the validation period from May 1st, 2013 to August 31st, 2013. Compared with the standalone WRF model, the fully coupled model tends to show better performance with respect to temperature, downward longwave radiation, downward shortwave radiation, latent heat, sensible heat and soil temperature and moisture. Although the wet bias of the coupled simulated precipitation slightly increases (2.51 mm vs. 2.50 mm) due to the consideration of lateral flow of soil water, the simulation results of the land-atmosphere water-heat exchange fluxes and soil heat fluxes are comparably improved. Compared with the observations, the mean Root Mean Square Error (RMSE) of latent and sensible heat is reduced to 32.27 W∙m-2 and 24.91 W∙m-2, and of surface soil temperature and moisture is reduced to 4.22 K and 0.06 m3/m3. Besides, the fully coupled model is able to capture the variation characteristics of streamflow with a NSE of 0.33, which indicates that the fully coupled WRF-Hydro model has great potential for characterizing coupled atmosphere-land-hydrological processes and streamflow simulation in the cold climatical and complex topographic regions.
- Preprint
(10169 KB) - Metadata XML
- BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on hess-2022-409', Anonymous Referee #1, 03 Feb 2023
A very interesting study, I am impressed by the involved simulation experiments. My recommendation is that the paper be published in HESS provided the following points are addressed by the authors.
How to prove the conclusion is robust since that the study only focused on the 2013 rainy season. If we look at different periods, whether the results and related conclusion change.
For the NSE values in calibration and validation periods, it is a little strange that the model has better skill in validation period, which differs from our normally expection. Why?
Just as pointed out in the introduction that several researchers had already applied the coupled WRF-hydro in simulating hydrological process in source regions of the Three River, so, what’s the novelty in this study? Are there any research gaps that were still not addressed, or did the author get some new findings?
Figure 1, it is suggested to involve much more hydrological observation in this study. Such as the streamflow JiMai, MaQu…
Figure 5, please supply the captions for temperature and precipitation in the figure title.
Citation: https://doi.org/10.5194/hess-2022-409-RC1 -
AC1: 'Reply on RC1', Jun Wen, 04 Feb 2023
Dear Anonymous Reviewer,
Thanks for your comments for the manuscript entitled "A study on the fully coupled atmosphere-land-hydrology process and streamflow simulations over the Source Region of the Yellow River". These suggestions are quite valuable and helpful to improve the quality of our manuscript. We have carefully read the comments and made necessary corrections or revisions. A point-to-point response to your comments is presented following.
Sincerely.
Jun Wen on behalf all authors
REVIEWER COMMENTS
A very interesting study, I am impressed by the involved simulation experiments. My recommendation is that the paper be published in HESS provided the following points are addressed by the authors.
Response: Thanks for your recognition and constructive comments to this research, the manuscript was carefully revised with reference to your suggestions.
Q1. How to prove the conclusion is robust since that the study only focused on the 2013 rainy season. If we look at different periods, whether the results and related conclusion change.
Response: Thanks for your question, the above-mentioned two questions are to be addressed as following separately.
- Firstly, the capabilities of WRF-Hydro model in long-term hydrological process simulation over the Source Region of the Yellow River has been proved in author’sanother research (Chen et al., 2022). Besides, the aim of this study is to investigate the effects of climate change on land surface and water cycle processes and the feedback of land surface hydrological cycle to precipitation. On the basis of the reliable driving data from coupling model, the simulation of the hydrological process is also reliable. Therefore, the study only focuses on the 2013 rainy season is reasonable and the conclusion is robust.
- Since the calibrated model is capable of characterizing the fully coupled atmosphere-land-hydrology process overthe Source Region of the Yellow River, the relevant results may have differences in values when paying attention to different time periods, but within the acceptable range, which will not affect the qualitative conclusions.
Q2. For the NSE values in calibration and validation periods, it is a little strange that the model has better skill in validation period, which differs from our normally expection. Why?
Response: Thanks for the questions you raised. In fact, we also noticed this fact during the analyzing the results. By checking the driven data and the sensitivity parameters of the model, as well as consulting the relevant references of daily streamflow simulation over the Source Region of the Yellow River, it can be found that the model is very sensitive to precipitation, the large value of the precipitation data lead to the large simulated streamflow during the calibration period and the appearance of the simulated streamflow peak later with a smaller NSE value, which also appears in other relevant references (Zhang et al., 2017; Gu et al., 2021). However, the variation of precipitation hydrograph during the validation period is relatively gentle and consistent with the streamflow hydrograph. So it could be concluded that the model has somewhat better skill during the validation period.
Q3. Just as pointed out in the introduction that several researchers had already applied the coupled WRF-Hydro in simulating hydrological process in source regions of the Three River, so, what’s the novelty in this study? Are there any research gaps that were still not addressed, or did the author get some new findings?
Response: Thanks for your comments. Just as pointed out in the introduction session, the researches on the coupled/uncoupled WRF-Hydro in simulating hydrological process mainly focus on the short-term flood events in small and medium-scale watersheds which mostly located in plain areas with the single underlying surface conditions, the calibration of the sensitive parameters is relatively easy. So one of the novelties in this study is that the study region is a large-scale watershed with complex underlying surface and climate conditions, and there are large challenges in the calibration of the sensitive parameters and the calculation of the model. The other novelty is that the fully coupled atmosphere-land-hydrology process is explored and the water-heat exchange process between the atmosphere and land surface in this research is quantitatively studied which is rarely considered in other studies.
Besides, compared with other studies which focused on the hydrological process over the Source Regions of the Three River by using WRF-Hydro model, this research gets a higher NSE in streamflow simulation and proves that the fully coupled WRF-Hydro model has ability to reproduce the daily streamflow over the Source Region of the Yellow River, which is a large improvement compared with the research results of Li et al (2021).
Q4. Figure 1, it is suggested to involve much more hydrological observation in this study. Such as the streamflow JiMai, MaQu…
Response: Thanks for your suggestions. Due to data security and other reasons, it is very difficult to obtain the observed streamflow data. At present, the daily streamflow data over the Source Region of the Yellow River which can be obtained from the Yellow River Water Conservancy Bureau is only Tangnaihai hydrological station. Therefore, the applicability of the model in other hydrological stations might be potentially verified in the future research.
Q5. Figure 5, please supply the captions for temperature and precipitation in the figure title.
Response: Thanks for your reminding. The captions for temperature and precipitation have been supplied in the figure title in the revised manuscript. (Line 261, page 13)
References:
Chen, Y. L., Wen, J., Yang, C. G., Long, T. P., Li, G. W., Jia, H. J., and Liu, Z.: Analysis on the applicability of different precipitation products and WRF-Hydro model over the Source Region of the Yellow River, Chinese Journal of Atmospheric Sciences, [preprint], https://doi.org/10.3878/j.issn.1006-9895.2205.22057, 2020.
Gu, T. W., Chen, Y. D., Gao, Y. F., Qin, L. Y., Wu, Y. Q., and Wu, Y. Z.: Improved streamflow forecast in a small-medium sized river basin with coupled WRF and WRF-Hydro: Effects of radar data assimilation, Remote Sensing, 13, 3251, https://doi.org/10.3390/rs13163251, 2021.
Li, G. W., Meng, X. H., Blyth, E., Chen, H., Shu, L. L., Li, Z. G, Zhao, L., and Ma, Y. M.: Impact of fully coupled hydrology-atmosphere processes on atmosphere conditions: Investigating the performance of the WRF-Hydro model in the Three River Source Region on the Tibetan Plateau, China, Water, 13, 3409, https://doi.org/10.3390/w13233409, 2021.
Zhang, A., Li, T. J., Fu, W., and Wang, Y. T.: Model simulation of flood season runoff in the headwaters of the Yellow River Basin using satellite-ground merged precipitation data, Journal of Basic Science and Engineering, 25, 1-16, https://doi.org/10.16058/j.issn.1005-0930.2017.01.001, 2017.
-
AC1: 'Reply on RC1', Jun Wen, 04 Feb 2023
-
RC2: 'Comment on hess-2022-409', Anonymous Referee #2, 12 Mar 2023
This study applied two models (fully coupled WRF-Hydro and standalone WRF, the only difference between them is the parameterization of lateral hydrological processes) to the Source Regions of the Yellow River. The sensitivity of simulated meteorological factors, heat fluxes, soil temperature, and soil moisture to the consideration of the lateral flow of soil water was compared in terms of their temporal variability and spatial distribution. The key conclusion is that the performance of the coupled model (WRF-Hydro) is better than WRF in simulating air temperature, downward longwave radiation, downward shortwave radiation, heat fluxes, and soil moisture/temperature. However, in general, I find this study less interesting, and no significant new findings were made. The major comments are:
(1) there were no attempts to modify or improve the model parametrizations, given that the NSE of the fully coupled model was very low (0.33). Instead, only the differences in the simulations between WRF-Hydro and WRF were compared. This comparison of course can show that the lateral soil water flow is important for land-atmosphere water-heat exchange, but it makes me hard to know how to improve the performance of the model. In case of such poor performance of the model, does it make sense to make such a comparison?
(2) the reasons why the fully coupled model is better were not diagnosed. What are the dominant processes in WRF-Hydro? Why do they work?
(3), the model validation is not complete. 3-month daily streamflow and two-day heat fluxes (latent and sensible) were used to validate the model. I think it is not complete to fully demonstrate the ability of the model to capture the seasonal and interannual variability.
Minor comments:
Line 43, please check if the streamflow is dominated by glacial meltwater in this basin.
Line 28, in the Introduction section, no clear scientific question was addressed and the motivation of the study is not clear.
Line 127, where does the vegetation type data come from?
Line 159 and 170, the two titles are the same.
Line 180-181, I think the processes related to the key parameters can be introduced in the context.
Line 213, considering the low NSE value, I don’t think the model has the ability to produce a realistic hydrological regime.
Line 233, same as above.
Line 397-400, I don’t know how you design the sensitivity analysis. More details should be provided.
Citation: https://doi.org/10.5194/hess-2022-409-RC2 -
AC2: 'Reply on RC2', Jun Wen, 24 Mar 2023
Dear Anonymous Reviewer,
Many thanks for taking the time to review our manuscript and provide constructive suggestions and comments. These suggestions and comments are quite valuable and helpful to improve the quality of our manuscript. We have made our best efforts to respond your questions and made necessary corrections or revisions. The relevant references for this revision and a point-to-point response to your suggestions and comments are presented following.
With best wishes,
Jun Wen on behalf all authors
REVIEWER COMMENTS
This study applied two models (fully coupled WRF-Hydro and standalone WRF, the only difference between them is the parameterization of lateral hydrological processes) to the Source Regions of the Yellow River. The sensitivity of simulated meteorological factors, heat fluxes, soil temperature, and soil moisture to the consideration of the lateral flow of soil water was compared in terms of their temporal variability and spatial distribution. The key conclusion is that the performance of the coupled model (WRF-Hydro) is better than WRF in simulating air temperature, downward longwave radiation, downward shortwave radiation, heat fluxes, and soil moisture/temperature. However, in general, I find this study less interesting, and no significant new findings were made.
Response: Thank you for reading our manuscript and your sincere and constructive suggestions. We will revise the manuscript according to your suggestions.
The major comments are:
1. There were no attempts to modify or improve the model parametrizations, given that the NSE of the fully coupled model was very low (0.33). Instead, only the differences in the simulations between WRF-Hydro and WRF were compared. This comparison of course can show that the lateral soil water flow is important for land-atmosphere water-heat exchange, but it makes me hard to know how to improve the performance of the model. In case of such poor performance of the model, does it make sense to make such a comparison?
Response: Thanks for the questions you raised. We believe it is meaningful to compare the simulation results of standalone WRF and coupled WRF-Hydro model.
Firstly, before carrying out the fully coupled simulations, the relevant sensitive parameters and parameterization schemes in the uncoupled WRF-Hydro model that affect variations of the land surface water-heat exchange process and streamflow had been calibrated. The result shows that the Correlation Coefficient (R) between simulated and observed streamflow is 0.81 and the Nash Efficiency Coefficient (NSE) is 0.61 during the validation period. Therefore, it is concluded that WRF-Hydro model has potential to reasonably characterize the variation of streamflow over the Source Region of the Yellow River. On the basis, the parameterization schemes in the WRF model were optimized, especially those affect the performance of the precipitation simulation, in order to get a more accurate hydrological driven data. Then, the fully coupled simulations were carried out.
The reason for comparing the differences in the simulations between WRF-Hydro and WRF not only to demonstrate the importance of the lateral soil water flow, but to explore the most significant factors influencing the simulated streamflow in the fully coupled model. In other words, it is to find out the reason for the low NSE (only 0.33), which contributes to improving the simulation performance in the future research.
2. The reasons why the fully coupled model is better were not diagnosed. What are the dominant processes in WRF-Hydro? Why do they work?
Response: Thanks for your comments. The reasons why the fully coupled WRF-Hydro model is better are the consideration of the computation of the lateral redistribution and re-infiltration of the water, the coupled model contributes to a better simulation of the soil moisture content (Gochis et al., 2015). The improved simulation of the soil moisture affects the computation of the sensible and latent heat fluxes, which influence humidity and temperature in the lower atmosphere and consequently precipitation. Therefore, the physical process of the coupling of land-atmosphere is expected to improve the forecast skill of precipitation, this in turn improves spatiotemporal distribution soil moisture content which can eventually determine the magnitude of the surface and channel runoff.
The above-mentioned other two questions are to be addressed as following separately.
1) The dominant processes in the WRF-Hydro model include five parts: land surface process, subsurface flow routing process, overland flow routing process, baseflow process and channel and reservoir routing process.
2) The steps of the WRF-Hydro model run can be divided as follows: Firstly, the 1-dimensional column land surface model (LSM) calculates the vertical fluxes of energy, moisture and soil thermal and moisture states. Then these parameters are subsequently disaggregated from the 1D LSM grid to a high resolution. Then subsurface lateral flow in WRF-Hydro is calculated prior to the routing of overland flow to allow exfiltration from fully saturated grid cells to be added to the infiltration excess calculated by the LSM. Next, the WRF-Hydro calculates the water table depth according to the depth of the top of the saturated soil layer that is nearest to the surface. Then overland flow is defined. Finally, the baseflow and channel routing processes have also been implemented.
3. The model validation is not complete. 3-month daily streamflow and two-day heat fluxes (latent and sensible) were used to validate the model. I think it is not complete to fully demonstrate the ability of the model to capture the seasonal and interannual variability.
Response: Thanks for your question. The ability of WRF-Hydro model in long-term hydrological process simulation over the Source Region of the Yellow River has been validated in our another research (Chen et al., 2022). Besides, duo to the national restrictions on data security protection, the availability on the heat fluxes is extremely limited and the quality of the observed data also needs to be considered. Furthermore, the aim of this study is to investigate the effects of climate change on land surface and water cycle processes and the feedback of land surface hydrological cycle to precipitation. On the basis of the reliable driving data from coupling model, the simulation of the hydrological process is also reliable. Therefore, the study only focuses on the 2013 rainy season is reasonable. And the ability of the model to capture the seasonal and inter-annual variability of the heat fluxes is supposed to be carried out in the future research.
Minor comments:
Q1. Line 43, please check if the streamflow is dominated by glacial meltwater in this basin.
Response: Thanks for your reminding and the manuscript has been revised according to the latest studies of Wang et al (2020) which found that precipitation and evaporation contribute 87.9% to the streamflow variation over the Source Region of the Yellow River.
Q2. Line 28, in the Introduction section, no clear scientific question was addressed and the motivation of the study is not clear.
Response: Thanks for your comments. In fact, the aims and motivation of this study have been illustrated in the Line 81-83 of the manuscript and the scientific question which needs to be addressed in this study is added in the revised manuscript.
Q3. Line 127, where does the vegetation type data come from?
Response: Thanks for your question, the vegetation type data in this study is from the Geographical Static Data of the Preprocessing System (WPS) of the Weather Research and Forecasting (WRF) model.
Q4. Line 159 and 170, the two titles are the same.
Response: We are sorry for this mistake, and the corresponding changes have been made in the revised manuscript.
Q5. Line 180-181, I think the processes related to the key parameters can be introduced in the context.
Response: Thanks for your constructive suggestions. The calibration processes related to the key parameters have been added in the revised manuscript.
Q6. Line 213, considering the low NSE value, I don’t think the model has the ability to produce a realistic hydrological regime.
Response: Thanks for your comments. In fact, it can be found from the several references (Li et al., 2021; Sarkar and Himesh, 2021) that the evaluation indices for the hydrological processes simulations of the large-scale watershed in complex terrain areas are not very high, the simulation results of our research with a R of 0.81 and a NSE of 0.61 are better than before. Additionally, the variation of the simulated streamflow hydrograph is in agreement with the observation despite the errors of peak flow, so it can be thought that the model has the ability to produce a relatively realistic hydrological regime.
Q7. Line 233, same as above.
Response: Thanks, one of the reasons for lower R and higher RMSE on the typical cloudy day is the complexity of weather conditions and physical processes. The other is that there is a unique “Lake Climate” near the flux observation station (i.e. the Eling Lake station), which leads to uncertainties in the simulation results but within an acceptable range.
Q8. Line 397-400, I don’t know how you design the sensitivity analysis. More details should be provided.
Response: Sorry for this confusion. The details of the sensitivity analysis have been added in the revised version of the manuscript.
References:
Chen, Y. L., Wen, J., Yang, C. G., et al. Analysis on the applicability of different precipitation products and WRF-Hydro model over the Source Region of the Yellow River [J], Chinese Journal of Atmospheric Sciences, [preprint], https://doi.org/10.3878/j.issn.1006-9895.2205.22057, 2020. (In Chinese with English abstract)
Gochis, D. J., Yu, W., & Yates, D. N. The WRF‐Hydro model technical description and user's guide, version 3.0. NCAR Technical Document. 120 pages. Retrieved from http://www.ral.ucar.edu/projects/wrf_hydro/, 2015.
Li, G. W., Meng, X. H., Blyth, E., et al. Impact of fully coupled Hydrology-Atmosphere processes on atmosphere conditions: Investigating the performance of the WRF-Hydro model in the Three River Source Region on the Tibetan Plateau, China [J]. Water, 13: 3409, https://doi.org/10.3390/w13233409, 2021.
Sarkar, S., &Himesh, S. Evaluation of the skill of a fully-coupled atmospheric-hydrological model in simulating extreme hydrometeorological event: A case study over Cauvery River Catchment [J]. Pure and Applied Geophysics, 178, 1063-1086, https://doi.org/10.1007/s00024-021-02684-4, 2021.
Wang, D. X., Tian, S. M., Jiang, S. Q., et al. Research Progress of the Evolution of Runoff in the Source Area of the Yellow River[J]. Yellow River, 42(9): 90-95, https://doi.org/10.3969/ j.issn.1000-1379.2020.09.017, 2020.
Citation: https://doi.org/10.5194/hess-2022-409-AC2
-
AC2: 'Reply on RC2', Jun Wen, 24 Mar 2023
Status: closed
-
RC1: 'Comment on hess-2022-409', Anonymous Referee #1, 03 Feb 2023
A very interesting study, I am impressed by the involved simulation experiments. My recommendation is that the paper be published in HESS provided the following points are addressed by the authors.
How to prove the conclusion is robust since that the study only focused on the 2013 rainy season. If we look at different periods, whether the results and related conclusion change.
For the NSE values in calibration and validation periods, it is a little strange that the model has better skill in validation period, which differs from our normally expection. Why?
Just as pointed out in the introduction that several researchers had already applied the coupled WRF-hydro in simulating hydrological process in source regions of the Three River, so, what’s the novelty in this study? Are there any research gaps that were still not addressed, or did the author get some new findings?
Figure 1, it is suggested to involve much more hydrological observation in this study. Such as the streamflow JiMai, MaQu…
Figure 5, please supply the captions for temperature and precipitation in the figure title.
Citation: https://doi.org/10.5194/hess-2022-409-RC1 -
AC1: 'Reply on RC1', Jun Wen, 04 Feb 2023
Dear Anonymous Reviewer,
Thanks for your comments for the manuscript entitled "A study on the fully coupled atmosphere-land-hydrology process and streamflow simulations over the Source Region of the Yellow River". These suggestions are quite valuable and helpful to improve the quality of our manuscript. We have carefully read the comments and made necessary corrections or revisions. A point-to-point response to your comments is presented following.
Sincerely.
Jun Wen on behalf all authors
REVIEWER COMMENTS
A very interesting study, I am impressed by the involved simulation experiments. My recommendation is that the paper be published in HESS provided the following points are addressed by the authors.
Response: Thanks for your recognition and constructive comments to this research, the manuscript was carefully revised with reference to your suggestions.
Q1. How to prove the conclusion is robust since that the study only focused on the 2013 rainy season. If we look at different periods, whether the results and related conclusion change.
Response: Thanks for your question, the above-mentioned two questions are to be addressed as following separately.
- Firstly, the capabilities of WRF-Hydro model in long-term hydrological process simulation over the Source Region of the Yellow River has been proved in author’sanother research (Chen et al., 2022). Besides, the aim of this study is to investigate the effects of climate change on land surface and water cycle processes and the feedback of land surface hydrological cycle to precipitation. On the basis of the reliable driving data from coupling model, the simulation of the hydrological process is also reliable. Therefore, the study only focuses on the 2013 rainy season is reasonable and the conclusion is robust.
- Since the calibrated model is capable of characterizing the fully coupled atmosphere-land-hydrology process overthe Source Region of the Yellow River, the relevant results may have differences in values when paying attention to different time periods, but within the acceptable range, which will not affect the qualitative conclusions.
Q2. For the NSE values in calibration and validation periods, it is a little strange that the model has better skill in validation period, which differs from our normally expection. Why?
Response: Thanks for the questions you raised. In fact, we also noticed this fact during the analyzing the results. By checking the driven data and the sensitivity parameters of the model, as well as consulting the relevant references of daily streamflow simulation over the Source Region of the Yellow River, it can be found that the model is very sensitive to precipitation, the large value of the precipitation data lead to the large simulated streamflow during the calibration period and the appearance of the simulated streamflow peak later with a smaller NSE value, which also appears in other relevant references (Zhang et al., 2017; Gu et al., 2021). However, the variation of precipitation hydrograph during the validation period is relatively gentle and consistent with the streamflow hydrograph. So it could be concluded that the model has somewhat better skill during the validation period.
Q3. Just as pointed out in the introduction that several researchers had already applied the coupled WRF-Hydro in simulating hydrological process in source regions of the Three River, so, what’s the novelty in this study? Are there any research gaps that were still not addressed, or did the author get some new findings?
Response: Thanks for your comments. Just as pointed out in the introduction session, the researches on the coupled/uncoupled WRF-Hydro in simulating hydrological process mainly focus on the short-term flood events in small and medium-scale watersheds which mostly located in plain areas with the single underlying surface conditions, the calibration of the sensitive parameters is relatively easy. So one of the novelties in this study is that the study region is a large-scale watershed with complex underlying surface and climate conditions, and there are large challenges in the calibration of the sensitive parameters and the calculation of the model. The other novelty is that the fully coupled atmosphere-land-hydrology process is explored and the water-heat exchange process between the atmosphere and land surface in this research is quantitatively studied which is rarely considered in other studies.
Besides, compared with other studies which focused on the hydrological process over the Source Regions of the Three River by using WRF-Hydro model, this research gets a higher NSE in streamflow simulation and proves that the fully coupled WRF-Hydro model has ability to reproduce the daily streamflow over the Source Region of the Yellow River, which is a large improvement compared with the research results of Li et al (2021).
Q4. Figure 1, it is suggested to involve much more hydrological observation in this study. Such as the streamflow JiMai, MaQu…
Response: Thanks for your suggestions. Due to data security and other reasons, it is very difficult to obtain the observed streamflow data. At present, the daily streamflow data over the Source Region of the Yellow River which can be obtained from the Yellow River Water Conservancy Bureau is only Tangnaihai hydrological station. Therefore, the applicability of the model in other hydrological stations might be potentially verified in the future research.
Q5. Figure 5, please supply the captions for temperature and precipitation in the figure title.
Response: Thanks for your reminding. The captions for temperature and precipitation have been supplied in the figure title in the revised manuscript. (Line 261, page 13)
References:
Chen, Y. L., Wen, J., Yang, C. G., Long, T. P., Li, G. W., Jia, H. J., and Liu, Z.: Analysis on the applicability of different precipitation products and WRF-Hydro model over the Source Region of the Yellow River, Chinese Journal of Atmospheric Sciences, [preprint], https://doi.org/10.3878/j.issn.1006-9895.2205.22057, 2020.
Gu, T. W., Chen, Y. D., Gao, Y. F., Qin, L. Y., Wu, Y. Q., and Wu, Y. Z.: Improved streamflow forecast in a small-medium sized river basin with coupled WRF and WRF-Hydro: Effects of radar data assimilation, Remote Sensing, 13, 3251, https://doi.org/10.3390/rs13163251, 2021.
Li, G. W., Meng, X. H., Blyth, E., Chen, H., Shu, L. L., Li, Z. G, Zhao, L., and Ma, Y. M.: Impact of fully coupled hydrology-atmosphere processes on atmosphere conditions: Investigating the performance of the WRF-Hydro model in the Three River Source Region on the Tibetan Plateau, China, Water, 13, 3409, https://doi.org/10.3390/w13233409, 2021.
Zhang, A., Li, T. J., Fu, W., and Wang, Y. T.: Model simulation of flood season runoff in the headwaters of the Yellow River Basin using satellite-ground merged precipitation data, Journal of Basic Science and Engineering, 25, 1-16, https://doi.org/10.16058/j.issn.1005-0930.2017.01.001, 2017.
-
AC1: 'Reply on RC1', Jun Wen, 04 Feb 2023
-
RC2: 'Comment on hess-2022-409', Anonymous Referee #2, 12 Mar 2023
This study applied two models (fully coupled WRF-Hydro and standalone WRF, the only difference between them is the parameterization of lateral hydrological processes) to the Source Regions of the Yellow River. The sensitivity of simulated meteorological factors, heat fluxes, soil temperature, and soil moisture to the consideration of the lateral flow of soil water was compared in terms of their temporal variability and spatial distribution. The key conclusion is that the performance of the coupled model (WRF-Hydro) is better than WRF in simulating air temperature, downward longwave radiation, downward shortwave radiation, heat fluxes, and soil moisture/temperature. However, in general, I find this study less interesting, and no significant new findings were made. The major comments are:
(1) there were no attempts to modify or improve the model parametrizations, given that the NSE of the fully coupled model was very low (0.33). Instead, only the differences in the simulations between WRF-Hydro and WRF were compared. This comparison of course can show that the lateral soil water flow is important for land-atmosphere water-heat exchange, but it makes me hard to know how to improve the performance of the model. In case of such poor performance of the model, does it make sense to make such a comparison?
(2) the reasons why the fully coupled model is better were not diagnosed. What are the dominant processes in WRF-Hydro? Why do they work?
(3), the model validation is not complete. 3-month daily streamflow and two-day heat fluxes (latent and sensible) were used to validate the model. I think it is not complete to fully demonstrate the ability of the model to capture the seasonal and interannual variability.
Minor comments:
Line 43, please check if the streamflow is dominated by glacial meltwater in this basin.
Line 28, in the Introduction section, no clear scientific question was addressed and the motivation of the study is not clear.
Line 127, where does the vegetation type data come from?
Line 159 and 170, the two titles are the same.
Line 180-181, I think the processes related to the key parameters can be introduced in the context.
Line 213, considering the low NSE value, I don’t think the model has the ability to produce a realistic hydrological regime.
Line 233, same as above.
Line 397-400, I don’t know how you design the sensitivity analysis. More details should be provided.
Citation: https://doi.org/10.5194/hess-2022-409-RC2 -
AC2: 'Reply on RC2', Jun Wen, 24 Mar 2023
Dear Anonymous Reviewer,
Many thanks for taking the time to review our manuscript and provide constructive suggestions and comments. These suggestions and comments are quite valuable and helpful to improve the quality of our manuscript. We have made our best efforts to respond your questions and made necessary corrections or revisions. The relevant references for this revision and a point-to-point response to your suggestions and comments are presented following.
With best wishes,
Jun Wen on behalf all authors
REVIEWER COMMENTS
This study applied two models (fully coupled WRF-Hydro and standalone WRF, the only difference between them is the parameterization of lateral hydrological processes) to the Source Regions of the Yellow River. The sensitivity of simulated meteorological factors, heat fluxes, soil temperature, and soil moisture to the consideration of the lateral flow of soil water was compared in terms of their temporal variability and spatial distribution. The key conclusion is that the performance of the coupled model (WRF-Hydro) is better than WRF in simulating air temperature, downward longwave radiation, downward shortwave radiation, heat fluxes, and soil moisture/temperature. However, in general, I find this study less interesting, and no significant new findings were made.
Response: Thank you for reading our manuscript and your sincere and constructive suggestions. We will revise the manuscript according to your suggestions.
The major comments are:
1. There were no attempts to modify or improve the model parametrizations, given that the NSE of the fully coupled model was very low (0.33). Instead, only the differences in the simulations between WRF-Hydro and WRF were compared. This comparison of course can show that the lateral soil water flow is important for land-atmosphere water-heat exchange, but it makes me hard to know how to improve the performance of the model. In case of such poor performance of the model, does it make sense to make such a comparison?
Response: Thanks for the questions you raised. We believe it is meaningful to compare the simulation results of standalone WRF and coupled WRF-Hydro model.
Firstly, before carrying out the fully coupled simulations, the relevant sensitive parameters and parameterization schemes in the uncoupled WRF-Hydro model that affect variations of the land surface water-heat exchange process and streamflow had been calibrated. The result shows that the Correlation Coefficient (R) between simulated and observed streamflow is 0.81 and the Nash Efficiency Coefficient (NSE) is 0.61 during the validation period. Therefore, it is concluded that WRF-Hydro model has potential to reasonably characterize the variation of streamflow over the Source Region of the Yellow River. On the basis, the parameterization schemes in the WRF model were optimized, especially those affect the performance of the precipitation simulation, in order to get a more accurate hydrological driven data. Then, the fully coupled simulations were carried out.
The reason for comparing the differences in the simulations between WRF-Hydro and WRF not only to demonstrate the importance of the lateral soil water flow, but to explore the most significant factors influencing the simulated streamflow in the fully coupled model. In other words, it is to find out the reason for the low NSE (only 0.33), which contributes to improving the simulation performance in the future research.
2. The reasons why the fully coupled model is better were not diagnosed. What are the dominant processes in WRF-Hydro? Why do they work?
Response: Thanks for your comments. The reasons why the fully coupled WRF-Hydro model is better are the consideration of the computation of the lateral redistribution and re-infiltration of the water, the coupled model contributes to a better simulation of the soil moisture content (Gochis et al., 2015). The improved simulation of the soil moisture affects the computation of the sensible and latent heat fluxes, which influence humidity and temperature in the lower atmosphere and consequently precipitation. Therefore, the physical process of the coupling of land-atmosphere is expected to improve the forecast skill of precipitation, this in turn improves spatiotemporal distribution soil moisture content which can eventually determine the magnitude of the surface and channel runoff.
The above-mentioned other two questions are to be addressed as following separately.
1) The dominant processes in the WRF-Hydro model include five parts: land surface process, subsurface flow routing process, overland flow routing process, baseflow process and channel and reservoir routing process.
2) The steps of the WRF-Hydro model run can be divided as follows: Firstly, the 1-dimensional column land surface model (LSM) calculates the vertical fluxes of energy, moisture and soil thermal and moisture states. Then these parameters are subsequently disaggregated from the 1D LSM grid to a high resolution. Then subsurface lateral flow in WRF-Hydro is calculated prior to the routing of overland flow to allow exfiltration from fully saturated grid cells to be added to the infiltration excess calculated by the LSM. Next, the WRF-Hydro calculates the water table depth according to the depth of the top of the saturated soil layer that is nearest to the surface. Then overland flow is defined. Finally, the baseflow and channel routing processes have also been implemented.
3. The model validation is not complete. 3-month daily streamflow and two-day heat fluxes (latent and sensible) were used to validate the model. I think it is not complete to fully demonstrate the ability of the model to capture the seasonal and interannual variability.
Response: Thanks for your question. The ability of WRF-Hydro model in long-term hydrological process simulation over the Source Region of the Yellow River has been validated in our another research (Chen et al., 2022). Besides, duo to the national restrictions on data security protection, the availability on the heat fluxes is extremely limited and the quality of the observed data also needs to be considered. Furthermore, the aim of this study is to investigate the effects of climate change on land surface and water cycle processes and the feedback of land surface hydrological cycle to precipitation. On the basis of the reliable driving data from coupling model, the simulation of the hydrological process is also reliable. Therefore, the study only focuses on the 2013 rainy season is reasonable. And the ability of the model to capture the seasonal and inter-annual variability of the heat fluxes is supposed to be carried out in the future research.
Minor comments:
Q1. Line 43, please check if the streamflow is dominated by glacial meltwater in this basin.
Response: Thanks for your reminding and the manuscript has been revised according to the latest studies of Wang et al (2020) which found that precipitation and evaporation contribute 87.9% to the streamflow variation over the Source Region of the Yellow River.
Q2. Line 28, in the Introduction section, no clear scientific question was addressed and the motivation of the study is not clear.
Response: Thanks for your comments. In fact, the aims and motivation of this study have been illustrated in the Line 81-83 of the manuscript and the scientific question which needs to be addressed in this study is added in the revised manuscript.
Q3. Line 127, where does the vegetation type data come from?
Response: Thanks for your question, the vegetation type data in this study is from the Geographical Static Data of the Preprocessing System (WPS) of the Weather Research and Forecasting (WRF) model.
Q4. Line 159 and 170, the two titles are the same.
Response: We are sorry for this mistake, and the corresponding changes have been made in the revised manuscript.
Q5. Line 180-181, I think the processes related to the key parameters can be introduced in the context.
Response: Thanks for your constructive suggestions. The calibration processes related to the key parameters have been added in the revised manuscript.
Q6. Line 213, considering the low NSE value, I don’t think the model has the ability to produce a realistic hydrological regime.
Response: Thanks for your comments. In fact, it can be found from the several references (Li et al., 2021; Sarkar and Himesh, 2021) that the evaluation indices for the hydrological processes simulations of the large-scale watershed in complex terrain areas are not very high, the simulation results of our research with a R of 0.81 and a NSE of 0.61 are better than before. Additionally, the variation of the simulated streamflow hydrograph is in agreement with the observation despite the errors of peak flow, so it can be thought that the model has the ability to produce a relatively realistic hydrological regime.
Q7. Line 233, same as above.
Response: Thanks, one of the reasons for lower R and higher RMSE on the typical cloudy day is the complexity of weather conditions and physical processes. The other is that there is a unique “Lake Climate” near the flux observation station (i.e. the Eling Lake station), which leads to uncertainties in the simulation results but within an acceptable range.
Q8. Line 397-400, I don’t know how you design the sensitivity analysis. More details should be provided.
Response: Sorry for this confusion. The details of the sensitivity analysis have been added in the revised version of the manuscript.
References:
Chen, Y. L., Wen, J., Yang, C. G., et al. Analysis on the applicability of different precipitation products and WRF-Hydro model over the Source Region of the Yellow River [J], Chinese Journal of Atmospheric Sciences, [preprint], https://doi.org/10.3878/j.issn.1006-9895.2205.22057, 2020. (In Chinese with English abstract)
Gochis, D. J., Yu, W., & Yates, D. N. The WRF‐Hydro model technical description and user's guide, version 3.0. NCAR Technical Document. 120 pages. Retrieved from http://www.ral.ucar.edu/projects/wrf_hydro/, 2015.
Li, G. W., Meng, X. H., Blyth, E., et al. Impact of fully coupled Hydrology-Atmosphere processes on atmosphere conditions: Investigating the performance of the WRF-Hydro model in the Three River Source Region on the Tibetan Plateau, China [J]. Water, 13: 3409, https://doi.org/10.3390/w13233409, 2021.
Sarkar, S., &Himesh, S. Evaluation of the skill of a fully-coupled atmospheric-hydrological model in simulating extreme hydrometeorological event: A case study over Cauvery River Catchment [J]. Pure and Applied Geophysics, 178, 1063-1086, https://doi.org/10.1007/s00024-021-02684-4, 2021.
Wang, D. X., Tian, S. M., Jiang, S. Q., et al. Research Progress of the Evolution of Runoff in the Source Area of the Yellow River[J]. Yellow River, 42(9): 90-95, https://doi.org/10.3969/ j.issn.1000-1379.2020.09.017, 2020.
Citation: https://doi.org/10.5194/hess-2022-409-AC2
-
AC2: 'Reply on RC2', Jun Wen, 24 Mar 2023
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 822 | 267 | 44 | 1,133 | 44 | 39 |
- HTML: 822
- PDF: 267
- XML: 44
- Total: 1,133
- BibTeX: 44
- EndNote: 39
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1