Potential of the Coupled WRF/WRF-Hydro Modeling System for Flood Forecasting in the Ouémé River (West Africa)
, Joël Arnault 3, Nana Ama Browne Klutse 1,4, Zhenyu Zhang 3, Harald Kunstmann 3 and Philip G. Oguntunde 5Round 1
Reviewer 1 Report
A model based on the Weather Research and Forecasting had been evaluated for flood modeling. One of the major drawbacks would be lack of comparative analysis with data-driven methods. Thus please elaborate on the research limitation.
The methodology behind WRF must be explained with more details.
Validation methodology can be more elaborated.
How the integration of WRF-Hydro done?
How the proposed model can perform in other study cases? what is the generalization ability of the proposed method? how it can be evaluated or improved?
How the model can be improved?
What are the pros and cons of the proposed method?
Please elaborate on the potential of the model vs. machine learning forecasting systems.
Manuscript is too long with the lack of consistency. Please consider improve the writing and organization.
Author Response
A model based on the Weather Research and Forecasting had been evaluated for flood modeling. One of the major drawbacks would be lack of comparative analysis with data-driven methods.
Response: The goal of the present study is to provide a reliable forecasting method for recurrent flooding events recorded in the study area.
In the introduction and from line 57 (in the previous manuscript) we provided resources to address this question (it can be seen in file water-1641180_tracking.pdf from lines 82 to 96). We summarized the various methodologies already applied, among them the data-driven method, and stated their limitations: “To determine an accurate flood forecasting ... in the performance of the forecasting”.
Thus please elaborate on the research limitation.
Response: As a limitation, the methodology proposed is costly in terms of computational time and resources.
The methodology behind WRF must be explained with more details.
Response: We explained the methodology behind WRF in section 2.2 (line 230-262 in file water-1641180_tracking.pdf). In fact, WRF is an atmospheric model which solves the equations of atmospheric motions. The atmospheric state variables are the wind components, pressure, temperature, humidity, and hydrometeor mixing ratios. The lower boundary of the atmospheric variables is forced with a land surface model. In the case of the present study, the supplementary hydrological model extension WRF-Hydro, which is a complex land surface model, with an explicit description of river streamflow, is employed.
Validation methodology can be more elaborated.
Response: The validation methodology used for the model has been presented in section 2.3 (line 280 downward in the tracking file) as well as the efficiency criteria: “For calibrating the model WRF-Hydro 4.0, we focus … where P1 324 and P2 are the shared periods containing into the whole study period named P (2008-2010)”.
How the integration of WRF-Hydro done?
Response: WRF-Hydro affords a set of supplementary hydrological components to the WRF model, which enables the capability of lateral redistribution of the hydrological condition at the land surface. Then, in the fully coupled model, WRF/WRF-Hydro is handled as an extension of the land surface model in WRF.
How the proposed model can perform in other study cases?
Response: The proposed model (fully coupled WRF/WRF-Hydro) allows for simulating streamflow from global meteorological data. As such, the coupled model can easily be applied to any region of the world, which is an advantage of the method used. Nevertheless, the application of WRF/WRF-Hydro (referred to as WRF-H in the manuscript) to another region may require a different calibration to optimize the quality of the modeled streamflow results over the concerned area. Additionally, the global forcing data to drive the model are available.
what is the generalization ability of the proposed method?
Response: The current applied methodology (coupled hydrological-atmosphere modeling system) takes into account the interaction (feedback) of the Land surface-atmosphere which plays important role in the rainfall producing systems; they should be more sensitive to any change in the atmospheric dynamics and/or the land surface state. In addition, we stated in the current manuscript (line 1200-1226 in file water-1641180_tracking.pdf) that the uncertainty of streamflow modeling comes from the uncertainties in the model structure, model parameters, and model driving data. Here we focus on the uncertainty in the driving precipitation data using the fully coupled WRF-H proposed.
how it can be evaluated or improved?
Response: We stated in the conclusion that the model has some shortcomings. The human impacts (factors) on streamflow, such as irrigation and dam management, are not yet considered in WRF-Hydro, which we see as a future and necessary model improvement for flood risks management purposes (line 1259-1267 in file water-1641180_tracking.pdf).
How the model can be improved?
Response: We stated above some limitations of our model, in addition to these, we can rise that, human impacts on streamflow, such as irrigation and dam management, are not yet considered in WRF-Hydro, which we see as a future and necessary model improvement for flood risks management purposes.
What are the pros and cons of the proposed method?
Pro: The driving data are available up to real-time and also for forecasting up to 10 days of lead time from some projects.
Cons: anthroponotic factors are not yet considered in the model
Please elaborate on the potential of the model vs. machine learning forecasting systems.
The machine learning approach is much more computationally efficient than the model approach, which is of course a serious advantage for a forecasting system. It is more practical than other techniques due to its ability to handle complex nonlinear systems. It has been applied for various purposes such as precipitation estimation, hydrological modeling, flood forecasting, flood inundation, etc. However, a process-based model may be more suitable to predict the non-linear response of the land-atmosphere system to a changing climate. It has many advantages, such as convenient calculation and high efficiency, easy to expand, transplant, and maintain in time, which can establish a general model structure for weather forecasting and studying the atmosphere; thus, it can be widely applied to weather forecasting and simulation in most regions.
Manuscript is too long with the lack of consistency. Please consider improve the writing and organization.
The manuscript has been well organized and each section has been well linked to others.
Globally, the presentation of the results has been more reorganized compared to the previous manuscript. We have presented in:
3.1. the evaluation of the WRF-only simulated precipitation,
3.2. the results of the calibration method used to optimize the WRF-Hydro model in offline mode are presented.
3.3. the simulation of the optimized model WRF-H to simulate the precipitation and the discharge is presented.
3.4. the evaluation of the soil moisture content by the optimized model WRF-H has been done.
3.5. as a model (which is not the reality), we investigate the uncertainty of the model WRF-H to simulate the precipitation and discharge by activating a stochastic kinetic-energy backscatter scheme (SKEBS).
Reviewer 2 Report
The article concerns the flood forecasting in the Ouémé-River. It may be interesting for readers of Water. In general, this manuscript is well organized and written, with detailing the framework approach of the study, clearly stated methodology and nicely presented findings. The following requests/suggestions should be taken into account to improve the quality of the manuscript.
- Protecting people from floods is an important issue for every country. Please refer to the flood risk. It is determined by, among others population density, the use of river valleys and floodplains, technical and communication infrastructure, etc. Flood risk management aims to reduce the potential negative effects of floods on human health, the environment, cultural heritage and economic activity.
- Of course, it is impossible to eliminate all the causes (it is impossible, for example, to stop heavy rainfall), but the effects of floods can be limited. Are there any anti-flood structures in the analyzed area?
- Flood forecasting should be combined with an early warning system. Is it possible to warn residents in the analyzed area?
- Please correct the text formatting, eg lines 558-563.
Author Response
The article concerns the flood forecasting in the Ouémé-River. It may be interesting for readers of Water. In general, this manuscript is well organized and written, with detailing the framework approach of the study, clearly stated methodology and nicely presented findings. The following requests/suggestions should be taken into account to improve the quality of the manuscript.
- Protecting people from floods is an important issue for every country. Please refer to the flood risk. It is determined by, among others population density, the use of river valleys and floodplains, technical and communication infrastructure, etc. Flood risk management aims to reduce the potential negative effects of floods on human health, the environment, cultural heritage, and economic activity.
Response: In the introduction and from line 46 (in the manuscript) we provided resources to address this question (it can be seen with the tracking file from lines 46 to 67): “Floods happen in geomorphologically … of the water level upward of the river could spread downstream”.
Ouémé-river is used for irrigation, hydroelectricity, transport, water resource management, which led to the construction of hydrological structures. That construction exposes the residents on the riverfronts to flooding risks.
- Flood forecasting should be combined with an early warning system. Is it possible to warn residents in the analyzed area?
Response: The manuscript has been updated to address this topic in the:
Introduction (line 59-67 in file water-1641180_tracking.pdf): “To prevent these damages during the rainy season, … of the water level upward of the river could spread downstream”
Introduction (line 59-67 in file water-1641180_tracking.pdf): The introduction of flood forecasting … should be appropriate actions to plan.
Conclusion (line 77-81 in file water-1641180_tracking.pdf):” Further studies are needed to an EWS in West Africa regions “.
- Of course, it is impossible to eliminate all the causes (it is impossible, for example, to stop heavy rainfall), but the effects of floods can be limited. Are there any anti-flood structures in the analyzed area?
Response: There are not yet anti-flood structures established in Benin. But some projects are working on that aspect, especially the OMiDelta project and PAPC (programme d’assainissement pluvial de la ville de Cotonou)
Please correct the text formatting, eg lines 558-563.
Round 2
Reviewer 1 Report
Comments had been addressed. It can be accepted.
