The suitability of a hybrid framework including data driven approaches for hydrological forecasting

Hauswirth, Sandra M.; Bierkens, Marc F. P.; Beijk, Vincent; Wanders, Niko

doi:https://doi.org/10.5194/hess-2022-89

Preprints

https://doi.org/10.5194/hess-2022-89

Preprints

31 Mar 2022

The suitability of a hybrid framework including data driven approaches for hydrological forecasting

Sandra M. Hauswirth¹, Marc F. P. Bierkens^1,2, Vincent Beijk³, and Niko Wanders¹ Sandra M. Hauswirth et al.

Show author details

¹Utrecht University, Department of Physical Geography, Princetonlaan 8a, Utrecht, The Netherlands
²Deltares, Daltonlaan 600, 3584 BK Utrecht, The Netherlands
³Rijkswaterstaat, Water, Verkeer en Leefomgeving, Griffioenlaan 2, Utrecht, The Netherlands

Received: 02 Mar 2022 – Discussion started: 31 Mar 2022

Abstract. Hydrological forecasts are important for operational water management and near future planning, even more so in light of increased occurrences of extreme events such as floods and droughts. Having a flexible forecasting framework that can deliver this information in fast and computational efficient manner is critical. In this study, the suitability of a hybrid forecasting framework, combining data-driven approaches and seasonal (re)forecasting information to predict hydrological variables was explored. Target variables include discharge and surface water levels for various stations at national scale with the Netherlands as focus. Five different ML models, ranging from simple to more complex and trained on historical observations of discharge, precipitation, evaporation and sea water levels, were run with seasonal (re)forecast data (EFAS and SEAS5) of these driver variables in a hindcast setting. The results were evaluated using the evaluation metrics Anomaly Correlation Coefficient (ACC), Continuous Ranked Probability (Skill) Score (CRPS and CRPSS), and Brier Skill Score (BSS) in comparison to a climatological reference hindcast. Aggregating results of all stations and ML models revealed that the hindcasting framework outperformed the climatological reference forecasts by roughly 60 % for discharge predictions (80 % for surface water level predictions). Skilful prediction for the first lead month, independently of initialization month, can be made for discharge. The skill extends up to 2–3 months for spring months due to snow melt dynamics captured in the training phase of the model. Surface water levels hindcasts showed similar skill and skilful lead times. While the different ML models showed differences in performance during a testing and training phase using historical observations, running the ML framework in a hindcast setting showed only minor differences between the models, which is attributed to the uncertainty in seasonal forecasts. However, despite being trained on historical observations, the hybrid framework used in this study shows similar skilful predictions as previous large scale forecasting systems. With our study we show that a hybrid framework is able to bring location specific skilful seasonal forecast information with global seasonal forecast inputs. At the same time our hybrid approach is flexible and fast, and as such a hybrid framework could be adapted to make it even more interesting to water managers and their needs, for instance a part of a fast model-predictive control framework.

Download & links

Preprint (PDF, 5577 KB)

Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
Preprint (5577 KB)

Download & links

The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.

Preprint (5577 KB)
BibTeX
EndNote

Journal article(s) based on this preprint

Sandra M. Hauswirth et al.

Interactive discussion

Status: closed

RC1:
'Comment on hess-2022-89', Anonymous Referee #1, 05 May 2022
The authors present a hybrid forecasting framework combining data driven approaches (using local, in-situ observations) and seasonal reforecasting information from large scale models to predict hydrological variables. The authors show that skillful predictions can be obtained with this hybrid framework. Although the idea of this framework is innovative and deserves publication a major revision is required (see comments below).

As suggested in the title and also throughout the manuscript a heavy focus for the assessment of this hybrid framework is based on the prediction of the variables river discharge and surface water levels, which seems to refer mostly to river water levels (it is actually not specified whether surface water levels refer to river water levels, sea level or even lake levels). However in the introduction and in section 2.2 the usage of sea water levels is mentioned. Furthermore, Fig A2 as well as Fig. 3 suggest indeed that sea water level observations are being considered. In the remainder of the manuscript the authors do not distinguish between sea level and river water levels but only mention surface water level measurements and it seems that in some of the analysis water level measurements from rivers as well as sea levels are mixed together (e.g. Fig. 4, Fig. A1, Fig. A3, Fig. A4). The mixed results are then used to derive general conclusions about the predictive skill of the hybrid framework. For example, in line 202 it is mentioned that Fig. A1 representing the CRPSS for surface water levels shows even better performance than the one for river discharge (Fig. 2). If sea and river water levels have been merged together it is however not possible to do such a comparison as the underlying processes that drive changes in sea level and river water levels are different. In addition to the mixing of those two variables, conclusions are made throughout the manuscript which are mostly only applicable to river water levels and river discharge. For example, lines 225-243 describe the results for the station Hagenstein Boven and state that the increase in skill in the early spring months are due to the fact of snow melt dynamics. Obviously this conclusion is not valid for the results obtained from sea level stations. However, no further analysis is provided for the skill observed in sea level stations. The same is true for the section 3.2 on hydrological low flows which also is not applicable to sea level measurements. Instead, section 3.3 mentions surface water level predictions (and it is not clear whether this refers only to river levels or to sea levels or to both) and makes some general conclusions but does not provide any further detail. Even in the introduction the manuscript provides primarily references in relation to streamflow forecasting and fresh water management but does not make any reference to coastal water level predictions.

In my view, the authors have two options to improve this issue: 1.) either you focus your analysis only on fresh water, i.e. only on river discharge and river water level predictions and remove from the analysis all sea level predictions or 2.) the authors clearly separate the results and their analysis for sea level predictions from river discharge/water levels predictions expanding the manuscript with the relevant sections and presenting separate conclusions/discussions for sea level and river flow/level predictions.

Introduction: Whereas the introduction mentions various examples for streamflow predictions no example is mentioned for sea level/coastal predictions that would support the integration of sea water levels into this analysis

Materials and Methods: Please add a section that describes the number of observation stations, its locations, its observation record, the variables used (river discharge, river water level, sea level) that have been used in the manuscript for training the ML models and that have been used for the analysis. Figs 3 and 4, and Figs. A2 and A3 show different station locations and it is totally unclear which observations have been used in this manuscript.

Figures 3 and A2 are not readable! Please increase the legends!

Figure 4 shows a station along the coast but is showing the ACC for discharge hindcasts. How is that possible? Or is there actually a small river, which is not shown in the Figure, flowing into the sea for that station?

Section 3.2: It is stated that “BSS confirms the earlier findings and shows the same trend of increased performance in the first lead weeks….”. I disagree with this finding as Fig. 6 shows clearly that the BSS for Feb/May and Sept is low in contrast to the findings for the general performance.

Section 2.2.: Please add a very brief explanation to the lagged times series approach as most of the readers will not be familiar with this approach.

Lines 205-212: It is stated that only minor differences are observed for the different ML models. Please analyse better the reason for this. One would assume that advanced DL methods such as LSTM would perform better than multiple linear regression

Figure A1: Does this figure show combined results of sea level and river levels? If yes, please separate these two variables.
Citation: https://doi.org/10.5194/hess-2022-89-RC1
- AC1: 'Reply on RC1', Sandra Margrit Hauswirth, 19 Jun 2022
  
  Anonymous Referee #1
  Referee comment on "The suitability of a hybrid framework including data driven approaches for hydrological forecasting" by Sandra M. Hauswirth et al., Hydrol. Earth Syst. Sci. Discuss., https://doi.org/10.5194/hess-2022-89-RC1, 2022
  General comment:
  The authors present a hybrid forecasting framework combining data driven approaches (using local, in-situ observations) and seasonal reforecasting information from large scale models to predict hydrological variables. The authors show that skillful predictions can be obtained with this hybrid framework. Although the idea of this framework is innovative and deserves publication a major revision is required (see comments below).
  
  Response: We want to thank the reviewer for taking the time to review our manuscript and acknowledging the innovative aspects of our research. We will address the comments, suggestions and open questions point by point below.
  
  Major comment:
  As suggested in the title and also throughout the manuscript a heavy focus for the assessment of this hybrid framework is based on the prediction of the variables river discharge and surface water levels, which seems to refer mostly to river water levels (it is actually not specified whether surface water levels refer to river water levels, sea level or even lake levels). However in the introduction and in section 2.2 the usage of sea water levels is mentioned. Furthermore, Fig A2 as well as Fig. 3 suggest indeed that sea water level observations are being considered. In the remainder of the manuscript the authors do not distinguish between sea level and river water levels but only mention surface water level measurements and it seems that in some of the analysis water level measurements from rivers as well as sea levels are mixed together (e.g. Fig. 4, Fig. A1, Fig. A3, Fig. A4). The mixed results are then used to derive general conclusions about the predictive skill of the hybrid framework. For example, in line 202 it is mentioned that Fig. A1 representing the CRPSS for surface water levels shows even better performance than the one for river discharge (Fig. 2). If sea and river water levels have been merged together it is however not possible to do such a comparison as the underlying processes that drive changes in sea level and river water levels are different. In addition to the mixing of those two variables, conclusions are made throughout the manuscript which are mostly only applicable to river water levels and river discharge. For example, lines 225-243 describe the results for the station Hagestein Boven and state that the increase in skill in the early spring months are due to the fact of snow melt dynamics. Obviously this conclusion is not valid for the results obtained from sea level stations. However, no further analysis is provided for the skill observed in sea level stations. The same is true for the section 3.2 on hydrological low flows which also is not applicable to sea level measurements. Instead, section 3.3 mentions surface water level predictions (and it is not clear whether this refers only to river levels or to sea levels or to both) and makes some general conclusions but does not provide any further detail. Even in the introduction the manuscript provides primarily references in relation to streamflow forecasting and fresh water management but does not make any reference to coastal water level predictions. In my view, the authors have two options to improve this issue: 1.) either you focus your analysis only on fresh water, i.e. only on river discharge and river water level predictions and remove from the analysis all sea level predictions or 2.) the authors clearly separate the results and their analysis for sea level predictions from river discharge/water levels predictions expanding the manuscript with the relevant sections and presenting separate conclusions/discussions for sea level and river flow/level predictions.
  
  Response: We appreciate the reviewer’s input and remarks on the interpretation of our results and the combination of our datasets included. However, we believe that there is a misunderstanding regarding the model framework and its set up, which is leading to the misunderstanding in the results interpretation.
  The modelling framework is based on a previous study by the main author and we acknowledge that the description in the current manuscript might not have been sufficient to fully follow without consulting the previous study. Because of this, the term “surface water levels” appears not to be properly coined in the current manuscript. This, in turn, may have led to the issue raised by the reviewer that surface water levels not being clearly enough defined and leading to an erroneous interpretation of the results. When we refer to “surface water levels” in our paper we mean water levels of rivers, streams and lakes Thus, we focus only on fresh water flows and fresh water levels as forecasting targets.
  The results for the different locations (as shown in Fig A2 and A3) thus pertain to forecasts of fresh surface water levels that are based on the machine learning models trained and validated for that specific location. As input data for the machine learning model the following variables were considered: discharge of the two main rivers entering the Netherlands, precipitation and evapotranspiration of one meteorological station in the centre of the Netherlands, as well as sea level observations close to one of the major dam systems at the coast of the Netherlands. Thus, sea water levels are an input variable to our machine learning based predictions and are not predicted themselves. When training of the machine learning models, observation of both input variables (discharge of the two main rivers entering the Netherlands, precipitation and evapotranspiration of one meteorological station in the centre of the Netherlands) and output variables (river discharge and river water levels) were used. When forecasting the trained machine learning models were forced with forecasts of the input variables. So all conclusions and interpretations pertain to forecasted fresh (river, streams) surface water levels
  
  We will pay special attention to clarify this during the revision of our manuscript to prevent future confusion. This would also be in line of the reviewers first suggestion.
  
  Other comments:
  Introduction: Whereas the introduction mentions various examples for streamflow predictions no example is mentioned for sea level/coastal predictions that would support the integration of sea water levels into this analysis
  Response: In line with the previous explanation we will revisit the introduction and adjust it to make it more comprehensible
  
  Materials and Methods: Please add a section that describes the number of observation stations, its locations, its observation record, the variables used (river discharge, river water level, sea level) that have been used in the manuscript for training the ML models and that have been used for the analysis. Figs 3 and 4, and Figs. A2 and A3 show different station locations and it is totally unclear which observations have been used in this manuscript.
  Response: In addition to the current link to the previous study where the modelling framework was developed, trained and tested, we plan to expand the explanation on the modelling framework so that it will also be followable for this study
  
  Figures 3 and A2 are not readable! Please increase the legends!
  Response: We will revisit the figure and improve the readability
  
  Figure 4 shows a station along the coast but is showing the ACC for discharge hindcasts. How is that possible? Or is there actually a small river, which is not shown in the Figure, flowing into the sea for that station?
  Response: The reviewer did see this correctly. The current figure only shows the main river network, however there are many smaller rivers and streams that are unfortunately not depicted. The station shown close to the coast is a measuring site placed at one of the sluices, which is also connected by to the main rivers by a smaller river branch. We will make this information more clear in the manuscript.
  
  Section 3.2: It is stated that “BSS confirms the earlier findings and shows the same trend of increased performance in the first lead weeks….”. I disagree with this finding as Fig. 6 shows clearly that the BSS for Feb/May and Sept is low in contrast to the findings for the general performance.
  Response: We do acknowledge and explain the low performance in the following sentences: “The BSS confirms the earlier findings and shows the same trend of increased performance in the first lead weeks, additional skill of several weeks is found for early spring and early summer months. However, tiles with lower performance throughout long lead periods, late summer and winter months can be spotted for this station. Some of these weeks appear to be more difficult to predict compared to early months in the year. This is likely due to unequal distribution of low flow occurrences throughout the year: where during summer months low flows can be more common and therefore chances to not fully capturing every event are higher, the low flows during winter are less common and captured relatively well with the snow melt dynamic as seen in previous scores.”
  
  We acknowledge that the first sentence might be misleading and we will adapt it.
  
  Section 2.2.: Please add a very brief explanation to the lagged times series approach as most of the readers will not be familiar with this approach.
  Response: We will include a brief explanation of the lagged time series approach together with the extended explanation of the modelling framework. However, we would also like to highlight the reference to the previous paper, in which the development and further detail of the modelling framework are shown.
  
  Lines 205-212: It is stated that only minor differences are observed for the different ML models. Please analyse better the reason for this. One would assume that advanced DL methods such as LSTM would perform better than multiple linear regression
  Response: This is indeed a remarkable result, since when used with observed input data, the differences between the different ML models were larger. The most likely explanation is that the forecasting skill is very much dependent on the skill by which the input variables are forecasted, which apparently make the differences in skill between the ML models insignificant in comparison. We will add an explanation for this in the manuscript.
  
  Figure A1: Does this figure show combined results of sea level and river levels? If yes, please separate these two variables..or
  Response: Figure A1 only shows results of forecasted fresh (river, streams, lakes) surface water levels
  
  Citation: https://doi.org/10.5194/hess-2022-89-AC1
RC2:
'Comment on hess-2022-89', Anonymous Referee #2, 18 May 2022
The study is building heavily on results published elsewhere (Hauswirth et al., 2021). I guess that is fine and always a difficult call to decide just how much information to provide so that a paper becomes a stand-alone piece of work without unnecessary repetition. However, in places I would have liked a little more info in this paper so I would not necessarily have to read the previous paper. For example, lines 114-117; this seems important and a bit more in-depth description of the datasets (SEAS5) and the ‘lagged time series approach’ which I am not familiar with.

Perhaps I have overlooked something, but I am not sure how the all the forecasts made by the different models and different ensembles are aggregated into one CDF (as per Fig.2)? Also, If there is only a minor difference in the performance between the five different ML models (line 205) then what is the advantage of using all of them rather than selecting the ‘best’ or most credible model for a particular site and use that? Might it be useful to add a section highlighting the performance of each of the five ML models in contrast to the aggregate performance of the entire system?

Minor comments:

The acronym LSTM is not defined?

Figs.2-6 did not come off well in my black and white copy, but ok for online viewing (which in fairness is probably the most common by now).

Line 253: Why the 20^th percentile? Or why only the 20^th percentile? I could imagine that the ability to assess low-flow across a range of severities would be of interest?

Line 253: Do any of your models include the effect of human interventions and their potential impacts on low flow? For example, water restrictions, operation of control structures to manage low flows etc? If not, is this likely to important in a highly regulated system such as the Netherland’s water ways? I think there is so mentioning of this in lines 220-225, but seems to me this is particularly important during low flow?

Line 293: I am not sure I understand how you incorporated water management into your models? Was this done in this study, or is that something that was part of the Hauswirth et al. (2021) study? I think perhaps more detail on this could be included in this manuscript as this seems interesting and important (even if you did not find a strong effect).
Citation: https://doi.org/10.5194/hess-2022-89-RC2
- AC2: 'Reply on RC2', Sandra Margrit Hauswirth, 19 Jun 2022
  
  Anonymous Referee #2
  Referee comment on "The suitability of a hybrid framework including data driven approaches for hydrological forecasting" by Sandra M. Hauswirth et al., Hydrol. Earth Syst. Sci. Discuss., https://doi.org/10.5194/hess-2022-89-RC2, 2022
  The study is building heavily on results published elsewhere (Hauswirth et al., 2021). I guess that is fine and always a difficult call to decide just how much information to provide so that a paper becomes a stand-alone piece of work without unnecessary repetition. However, in places I would have liked a little more info in this paper so I would not necessarily have to read the previous paper. For example, lines 114-117; this seems important and a bit more in-depth description of the datasets (SEAS5) and the ‘lagged time series approach’ which I am not familiar with.
  Perhaps I have overlooked something, but I am not sure how the all the forecasts made by the different models and different ensembles are aggregated into one CDF (as per Fig.2)? Also, If there is only a minor difference in the performance between the five different ML models (line 205) then what is the advantage of using all of them rather than selecting the ‘best’ or most credible model for a particular site and use that? Might it be useful to add a section highlighting the performance of each of the five ML models in contrast to the aggregate performance of the entire system?
  Response: We want to thank the reviewer for taking the time to review our manuscript and sharing thoughts and suggestions for improvement. We will address the comments, suggestions and open questions point by point below.
  We agree that our study is heavily building up on a previous study by the main author, which was a conscious decision that was made. The idea was and is to use the base modelling framework for different experiments. Therefore, we wanted to fully focus on the model development aspect in the first publication and refer to it for future work. We acknowledge that in this case, the explanation of the modelling framework was too short (as also indicated by the other reviewer). We plan to extend the description of the base modelling framework, including the lagged time series approach. Furthermore, we will extend the description of the new introduced datasets such as SEAS5 and EFAS as recommended.
  
  Regarding the comment Fig 2 and how the different ensembles were aggregated into one CDF, we would like to refer to line 196 (‘This was done by computing all the individual daily CRPSS results of all the hindcasts of every station and method, before aggregating the individual CRPSS scores to different temporal scales (weekly in Fig. 2).) and line 155 for the specific calculation procedure (‘Equation 1 …was used and the CRPS computed over all ensemble members for each lead day of every hindcast before aggregating it to other temporal scales’). For creating Fig 2 the different CRPSS scores for the different station and method combinations were additionally aggregated by lead day and the average was used for plotting.
  
  The reviewer poses a good question regarding what is the advantage of using different ML models, while they only show minor differences in their results. While performing the experiments, we were not expecting to only observe minor differences, while predictions using observed input data did show differences (Hauswirth et al., 2021). The most likely explanation is that the forecasting skill is very much dependent on the skill by which the input variables are forecasted, which apparently make the differences in skill between the ML models insignificant in comparison. We will add an explanation for this in the manuscript.
  
  To keep the manuscript easy to follow we decided to focus on one method in more detail. However, we will add a section on this as this is an interesting aspect to compare to the previous study (Hauswirth et al., 2021), where there were differences found between the models.
  
  Minor comments:
  The acronym LSTM is not defined?
  
  Response: We will define the acronym LSTM (Long Short Term Memory Model) to prevent confusion.
  Figs.2-6 did not come off well in my black and white copy, but ok for online viewing (which in fairness is probably the most common by now).
  
  Response: We will revisit the figures 2 - 6 and see if we can improve the readability for printed version.
  Line 253: Why the 20th percentile? Or why only the 20th percentile? I could imagine that the ability to assess low-flow across a range of severities would be of interest?
  
  Response: We chose the 20^th percentile as it is a commonly chosen threshold for low flows and was also interesting to our collaboration partners.
  
  Line 253: Do any of your models include the effect of human interventions and their potential impacts on low flow? For example, water restrictions, operation of control structures to manage low flows etc? If not, is this likely to important in a highly regulated system such as the Netherland’s water ways? I think there is so mentioning of this in lines 220-225, but seems to me this is particularly important during low flow?
  
  Response: The Netherlands is indeed a country with high water management infrastructures and plans. We therefore are working with the National Water Authority to include expert knowledge where possible and available. In our previous study we did include a separate water management scenario where we included operational plans of the major water infrastructures. We did see that there was an improvement in our simulations, albeit very minor in terms of our modelling results (Hauswirth et al., 2021).
  Line 293: I am not sure I understand how you incorporated water management into your models? Was this done in this study, or is that something that was part of the Hauswirth et al. (2021) study? I think perhaps more detail on this could be included in this manuscript as this seems interesting and important (even if you did not find a strong effect).
  
  Response: In line with the previous comment about water management above: We did a run with water management influence using the same approach as in Hauswirth et al. (2021). This run includes operational rules of main infrastructures which are related to the Rhine discharge at Lobith (one of our main input variables) for two specific input locations and two additional observation records of locations based at smaller infrastructures. We were therefore able to use the same approach regarding the operational rules for the main infrastructures, as these are based on the Rhine discharge we obtain from the EFAS dataset. For the two other additional timeseries climatology was used as operational plans were not available.
  
  We acknowledge that the information given in this manuscript related to water management aspects are very limited and heavily based on the previous study. We will extend the explanation so that the reader will be able to follow better.
  
  Citation: https://doi.org/10.5194/hess-2022-89-AC2

Peer review completion

AR: Author's response | RR: Referee report | ED: Editor decision

ED: Reconsider after major revisions (further review by editor and referees) (16 Jul 2022) by Thomas Kjeldsen

AR by Sandra Margrit Hauswirth on behalf of the Authors (10 Aug 2022) Author's response Author's tracked changes Manuscript

ED: Referee Nomination & Report Request started (05 Sep 2022) by Thomas Kjeldsen

RR by Anonymous Referee #2 (05 Dec 2022)

RR by Louise Arnal (05 Jan 2023)

Suggestions for revision or reasons for rejection

This manuscript presents a hybrid framework for ensemble seasonal hydrological forecasting in The Netherlands, forced with seasonal hydro-meteorological hindcasts from EFAS Seasonal and SEAS5. It shows the performance of the framework for discharge and surface water level hindcasts at several stations and discusses advantages of using a hybrid framework for hydrological forecasting, compared to large-domain dynamical systems. This manuscript is overall well thought-through and presents interesting original findings. However, a few key elements are missing from the analysis to address the objective of the paper more fully. Below are some comments which will hopefully help guide the manuscript revision.

Main comments:
- One of the objectives of this work is to explore the suitability of a hybrid framework for ensemble seasonal hydrological forecasting in The Netherlands, compared to more traditional large-domain dynamical systems. As mentioned in the Discussion, “hybrid forecasting systems have the ability to provide more local information compared to large-scale physically based systems”. To truly show this, it would be interesting to compare the performance of your system to the information already contained in the EFAS Seasonal hindcasts. To do this, two ideas are:
o Using EFAS Seasonal hindcasts as the reference instead of the observed climatology for skill score calculations. The EFAS hindcast values at grid cells that overlap the stations could be used or regional averages of the EFAS hindcast values could be calculated.
o Evaluating the performance of stations downstream of the input stations (Lobith and Eijsden) compared to the performance at those input stations.
- The quality of the EFAS and SEAS5 hindcasts compared to the observations used to train the models is a key element discussed throughout the manuscript. This warrants showing results of the analysis of the pre-processed inputs compared to the station observations.
- Here, you apply different ML models, which appear to give similar results for discharge hindcasts and further results are shown for a single model only. Would you recommend that water managers use a single model or would an ensemble of models be more appropriate. If the latter, could a super-ensemble be constructed from all or a sub-selection of the models? A discussion on this point would be interesting to read in the manuscript.
- Did you 1) calculate weekly and monthly averages from the variables prior to calculating verification scores or did you 2) average the daily score values to weekly and monthly timescales? P6 L144-145 and the “pre-processing” step on Fig. 1 suggests approach 1) was used, but P6 L155-158 suggests approach 2) was used. Please clarify in the manuscript. If approach 2) was used, this suggests that you expect to capture the daily variability with a seasonal system. I would expect to see a lot of noise and the final scores to be an average of that noise - but this is just a feeling. If you indeed used approach 2) and changing the method is not an option, could you maybe explore how this affects one set of results for an example station and show results in the Appendix?

Specific comments:
- Title: Consider adding the terms “seasonal” and perhaps also “ensemble” in front of “hydrological forecasting”.
- Abstract:
o P1 L2: Please introduce the term “flexible” and what it means here.
o P1 L4: Clarify that the seasonal (re)forecasting information is obtained from dynamical models to highlight the hybrid nature of this framework from the start.
o P1 L6-7: Please spell out/introduce the acronyms ML, EFAS and SEAS5 here.
- Introduction:
o P2 L30-31: Instead of “open source” do you mean “openly available data” from these systems? Open source suggests that the source codes used to run these systems are available. Same comment on P4 L106.
o P2 L37-40: I can see why you wanted to mention the ESP given the relevance for seasonal hydrological forecasting, but I would suggest removing this sentence completely as it doesn’t appear relevant to your paper specifically.
o P2 L46: Please elaborate on how a hybrid system offers more flexibility and is less intensive regarding data use. In doing so, I would suggest moving some content from P3 L63-68 closer to the hypothesis, as it provides some useful context that would be valuable to have earlier on.
o P3 L69: I would suggest removing “data-driven” as these steps are general to all forecasting systems and not specific to a data-driven system only.
o P3 L71: Perhaps add “in near real-time” at the end of the sentence to clarify the distinction with step one?
o P3 L72: Here and throughout the manuscript you use the term “data-driven forecasting framework” to refer to the framework you developed. Would the term “hybrid framework” be more appropriate, as is used in the title?
o P3 L74-76: Please consider removing this sentence as it is not essential here and a bit distracting to the flow of the paragraph, and it is already mentioned in the Methods.
o P3 L76-77: I would suggest rephrasing to “based on their historical performance to forecast discharge and …”.
o P3 L82: Please consider changing “predictions” to “conditions”.
- Material and Methods:
o P4 L99: Could you please mention how many stations were used here?
o P4 L104-105: Please briefly discuss in the manuscript why these specific variables were chosen and not others, such as soil moisture or groundwater.
o P4 L109: How was the station data extracted from this gridded product? Did you do any downscaling? Same question for SEAS5. Did you extract specific grid cells or calculate spatial averages?
o P5 L111: Please mention here what input data is used by this module.
o P5 L114-117: I would find it useful to have a bit more information in the manuscript about the data-processing as I don’t understand this step very well.
o P5 L121-142: It would be great to have a bit more information about the model setup. How were the input data used? Was it a mix of data prior to, on the initialization date and over the forecasting period? And with what temporal and spatial resolutions?
o P5 L124-125: Please briefly explain how water management aspects are incorporated here.
o P5 L125-126: How many years of observations were the models trained on?
o P5 L127: Add the station names here. I am assuming that these are the Lobith and Eijsden stations mentioned on P9 L234?
o P5 L127-129: Could you please briefly explain in this manuscript why you tested multiple ML methods instead of a single one? This links back to my 3rd main comment.
o P5 L140-141: Please add the number of ensemble members of the final outputs here. At what lead times were the hindcasts produced?
o P6 L144-145: Mention daily time scales too. I think “bi-weekly” should instead be “weekly”.
o P6 L158: I suggest changing “baseline” to “reference” in the text to match the equation terminology or modifying the equation to use the word “baseline”.
o P6 L161: Please mention the range and best value for the CRPSS. Same comment for the other scores.
o P6 L169: It is not clear to me from the Methods that the threshold chosen was indeed the 20th percentile and how it was calculated. Is it calculated from the observations?
o P7 L174-175: Why wasn’t the climatology used as a reference like for the CRPSS?
o P7 L176-184: Please consider providing the ACC formula.
o P7 L178-180: This is a bit hard to follow. Could you please clarify by giving an example?
o P7 L180: By “compared to the normal correlation” do you mean “relative to the climate values”?
o P7 L184: Could you please also mention how many sample points were available for the scores’ calculations, especially for the BS.
- Results:
o P7 L186-187: Please mention which target variable here.
o P7 L195-196: Mention that this is for discharge hindcasts.
o P8 L200: Is the convergence of skill a feature of the ML models? Just curious.
o Fig. 3: Could you maybe add the main rivers (Rhine and Meuse, which you often mention in the text) to the map for readers not very familiar with The Netherlands. The plots to the right of the map are too small, please make them larger and consider adding the station names at the top of each column. You could also highlight the station Hagenstein Boven on the map using a different symbol or by providing the station code in the text.
o P9 L226-228: It would be interesting to comment on the noticeable differences between the initialization months from Fig. 4 too, before talking about the differences in more detail on Fig. 5.
o P9 L236: What is the performance of the seasonal meteorological hindcasts for this time of year? You could perhaps refer to results from the literature looking at the SEAS5 precipitation and evaporation hindcasts quality over The Netherlands/Europe.
o Fig. 4: Consider making the rivers blue on the maps to distinguish them from other lines. The main rivers Rhine and Meuse could have thicker lines to differentiate them further. You could also display all larger river stations on one side of the map and the smaller river stations on the other side of the map to make the distinction clearer visually and the results more easily interpretable. Why does the colorbar start at 0.4 instead of 0? Same comment for other ACC plots in the manuscript.
o Fig. 5: Add “initialization month” as an x-label or in the caption, and mention that this shows the ACC for discharge hindcasts.
o Fig. 6: Same comments as above. I would also suggest removing the leftmost ACC heatmap as it is already shown above. Why is the ACC the only heatmap with empty values? Were all results significant for the CRPSS and BSS? P7 L184 suggests that the same significance criteria were applied to all scores.
o P12 L249: The rank histogram is more of a measure of the hindcast spread compared to that of the observations. The CRPS is a combined measure of the spread of the hindcast and its accuracy relative to the observed value. Please consider rephrasing.
o P12 L249-250: Is this shown on any specific figure?
o P12 L257: Except for hindcasts initialized in January.
o P12 L258-261: Did you consider defining season-specific thresholds instead?
o P12 L261-262: It would be interesting to see a map of the BSS as well.
o P12 L263: Please briefly mention Fig. A2 here to discuss results from different models, as the section title suggests.
- Discussion:
o P13 L298: Consider rephrasing to “For hindcasts initialized in the early spring …”.
o P13 L304: It would be interesting to get an indication of how fast these ML models are for generating such hindcasts.
o P14 L316: This framework still relies on outputs from dynamical systems, which will have to be run (albeit by a different entity/forecasting centre) beforehand. I think it would be worth being more specific and clarifying that your framework offers a computationally frugal way to provide more localized forecast information, which a large-domain system could only do at such a high resolution with a higher computational demand. But that it doesn’t cut the computational costs of running EFAS Seasonal and SEAS5 in the first place.

Hide

ED: Publish subject to minor revisions (review by editor) (05 Jan 2023) by Thomas Kjeldsen

AR by Sandra Margrit Hauswirth on behalf of the Authors (14 Jan 2023) Author's response Author's tracked changes Manuscript

ED: Publish as is (15 Jan 2023) by Thomas Kjeldsen

AR by Sandra Margrit Hauswirth on behalf of the Authors (16 Jan 2023) Author's response Manuscript

Interactive discussion

Status: closed

RC1:
'Comment on hess-2022-89', Anonymous Referee #1, 05 May 2022
The authors present a hybrid forecasting framework combining data driven approaches (using local, in-situ observations) and seasonal reforecasting information from large scale models to predict hydrological variables. The authors show that skillful predictions can be obtained with this hybrid framework. Although the idea of this framework is innovative and deserves publication a major revision is required (see comments below).

As suggested in the title and also throughout the manuscript a heavy focus for the assessment of this hybrid framework is based on the prediction of the variables river discharge and surface water levels, which seems to refer mostly to river water levels (it is actually not specified whether surface water levels refer to river water levels, sea level or even lake levels). However in the introduction and in section 2.2 the usage of sea water levels is mentioned. Furthermore, Fig A2 as well as Fig. 3 suggest indeed that sea water level observations are being considered. In the remainder of the manuscript the authors do not distinguish between sea level and river water levels but only mention surface water level measurements and it seems that in some of the analysis water level measurements from rivers as well as sea levels are mixed together (e.g. Fig. 4, Fig. A1, Fig. A3, Fig. A4). The mixed results are then used to derive general conclusions about the predictive skill of the hybrid framework. For example, in line 202 it is mentioned that Fig. A1 representing the CRPSS for surface water levels shows even better performance than the one for river discharge (Fig. 2). If sea and river water levels have been merged together it is however not possible to do such a comparison as the underlying processes that drive changes in sea level and river water levels are different. In addition to the mixing of those two variables, conclusions are made throughout the manuscript which are mostly only applicable to river water levels and river discharge. For example, lines 225-243 describe the results for the station Hagenstein Boven and state that the increase in skill in the early spring months are due to the fact of snow melt dynamics. Obviously this conclusion is not valid for the results obtained from sea level stations. However, no further analysis is provided for the skill observed in sea level stations. The same is true for the section 3.2 on hydrological low flows which also is not applicable to sea level measurements. Instead, section 3.3 mentions surface water level predictions (and it is not clear whether this refers only to river levels or to sea levels or to both) and makes some general conclusions but does not provide any further detail. Even in the introduction the manuscript provides primarily references in relation to streamflow forecasting and fresh water management but does not make any reference to coastal water level predictions.

In my view, the authors have two options to improve this issue: 1.) either you focus your analysis only on fresh water, i.e. only on river discharge and river water level predictions and remove from the analysis all sea level predictions or 2.) the authors clearly separate the results and their analysis for sea level predictions from river discharge/water levels predictions expanding the manuscript with the relevant sections and presenting separate conclusions/discussions for sea level and river flow/level predictions.

Introduction: Whereas the introduction mentions various examples for streamflow predictions no example is mentioned for sea level/coastal predictions that would support the integration of sea water levels into this analysis

Materials and Methods: Please add a section that describes the number of observation stations, its locations, its observation record, the variables used (river discharge, river water level, sea level) that have been used in the manuscript for training the ML models and that have been used for the analysis. Figs 3 and 4, and Figs. A2 and A3 show different station locations and it is totally unclear which observations have been used in this manuscript.

Figures 3 and A2 are not readable! Please increase the legends!

Figure 4 shows a station along the coast but is showing the ACC for discharge hindcasts. How is that possible? Or is there actually a small river, which is not shown in the Figure, flowing into the sea for that station?

Section 3.2: It is stated that “BSS confirms the earlier findings and shows the same trend of increased performance in the first lead weeks….”. I disagree with this finding as Fig. 6 shows clearly that the BSS for Feb/May and Sept is low in contrast to the findings for the general performance.

Section 2.2.: Please add a very brief explanation to the lagged times series approach as most of the readers will not be familiar with this approach.

Lines 205-212: It is stated that only minor differences are observed for the different ML models. Please analyse better the reason for this. One would assume that advanced DL methods such as LSTM would perform better than multiple linear regression

Figure A1: Does this figure show combined results of sea level and river levels? If yes, please separate these two variables.
Citation: https://doi.org/10.5194/hess-2022-89-RC1
- AC1: 'Reply on RC1', Sandra Margrit Hauswirth, 19 Jun 2022
  
  Anonymous Referee #1
  Referee comment on "The suitability of a hybrid framework including data driven approaches for hydrological forecasting" by Sandra M. Hauswirth et al., Hydrol. Earth Syst. Sci. Discuss., https://doi.org/10.5194/hess-2022-89-RC1, 2022
  General comment:
  The authors present a hybrid forecasting framework combining data driven approaches (using local, in-situ observations) and seasonal reforecasting information from large scale models to predict hydrological variables. The authors show that skillful predictions can be obtained with this hybrid framework. Although the idea of this framework is innovative and deserves publication a major revision is required (see comments below).
  
  Response: We want to thank the reviewer for taking the time to review our manuscript and acknowledging the innovative aspects of our research. We will address the comments, suggestions and open questions point by point below.
  
  Major comment:
  As suggested in the title and also throughout the manuscript a heavy focus for the assessment of this hybrid framework is based on the prediction of the variables river discharge and surface water levels, which seems to refer mostly to river water levels (it is actually not specified whether surface water levels refer to river water levels, sea level or even lake levels). However in the introduction and in section 2.2 the usage of sea water levels is mentioned. Furthermore, Fig A2 as well as Fig. 3 suggest indeed that sea water level observations are being considered. In the remainder of the manuscript the authors do not distinguish between sea level and river water levels but only mention surface water level measurements and it seems that in some of the analysis water level measurements from rivers as well as sea levels are mixed together (e.g. Fig. 4, Fig. A1, Fig. A3, Fig. A4). The mixed results are then used to derive general conclusions about the predictive skill of the hybrid framework. For example, in line 202 it is mentioned that Fig. A1 representing the CRPSS for surface water levels shows even better performance than the one for river discharge (Fig. 2). If sea and river water levels have been merged together it is however not possible to do such a comparison as the underlying processes that drive changes in sea level and river water levels are different. In addition to the mixing of those two variables, conclusions are made throughout the manuscript which are mostly only applicable to river water levels and river discharge. For example, lines 225-243 describe the results for the station Hagestein Boven and state that the increase in skill in the early spring months are due to the fact of snow melt dynamics. Obviously this conclusion is not valid for the results obtained from sea level stations. However, no further analysis is provided for the skill observed in sea level stations. The same is true for the section 3.2 on hydrological low flows which also is not applicable to sea level measurements. Instead, section 3.3 mentions surface water level predictions (and it is not clear whether this refers only to river levels or to sea levels or to both) and makes some general conclusions but does not provide any further detail. Even in the introduction the manuscript provides primarily references in relation to streamflow forecasting and fresh water management but does not make any reference to coastal water level predictions. In my view, the authors have two options to improve this issue: 1.) either you focus your analysis only on fresh water, i.e. only on river discharge and river water level predictions and remove from the analysis all sea level predictions or 2.) the authors clearly separate the results and their analysis for sea level predictions from river discharge/water levels predictions expanding the manuscript with the relevant sections and presenting separate conclusions/discussions for sea level and river flow/level predictions.
  
  Response: We appreciate the reviewer’s input and remarks on the interpretation of our results and the combination of our datasets included. However, we believe that there is a misunderstanding regarding the model framework and its set up, which is leading to the misunderstanding in the results interpretation.
  The modelling framework is based on a previous study by the main author and we acknowledge that the description in the current manuscript might not have been sufficient to fully follow without consulting the previous study. Because of this, the term “surface water levels” appears not to be properly coined in the current manuscript. This, in turn, may have led to the issue raised by the reviewer that surface water levels not being clearly enough defined and leading to an erroneous interpretation of the results. When we refer to “surface water levels” in our paper we mean water levels of rivers, streams and lakes Thus, we focus only on fresh water flows and fresh water levels as forecasting targets.
  The results for the different locations (as shown in Fig A2 and A3) thus pertain to forecasts of fresh surface water levels that are based on the machine learning models trained and validated for that specific location. As input data for the machine learning model the following variables were considered: discharge of the two main rivers entering the Netherlands, precipitation and evapotranspiration of one meteorological station in the centre of the Netherlands, as well as sea level observations close to one of the major dam systems at the coast of the Netherlands. Thus, sea water levels are an input variable to our machine learning based predictions and are not predicted themselves. When training of the machine learning models, observation of both input variables (discharge of the two main rivers entering the Netherlands, precipitation and evapotranspiration of one meteorological station in the centre of the Netherlands) and output variables (river discharge and river water levels) were used. When forecasting the trained machine learning models were forced with forecasts of the input variables. So all conclusions and interpretations pertain to forecasted fresh (river, streams) surface water levels
  
  We will pay special attention to clarify this during the revision of our manuscript to prevent future confusion. This would also be in line of the reviewers first suggestion.
  
  Other comments:
  Introduction: Whereas the introduction mentions various examples for streamflow predictions no example is mentioned for sea level/coastal predictions that would support the integration of sea water levels into this analysis
  Response: In line with the previous explanation we will revisit the introduction and adjust it to make it more comprehensible
  
  Materials and Methods: Please add a section that describes the number of observation stations, its locations, its observation record, the variables used (river discharge, river water level, sea level) that have been used in the manuscript for training the ML models and that have been used for the analysis. Figs 3 and 4, and Figs. A2 and A3 show different station locations and it is totally unclear which observations have been used in this manuscript.
  Response: In addition to the current link to the previous study where the modelling framework was developed, trained and tested, we plan to expand the explanation on the modelling framework so that it will also be followable for this study
  
  Figures 3 and A2 are not readable! Please increase the legends!
  Response: We will revisit the figure and improve the readability
  
  Figure 4 shows a station along the coast but is showing the ACC for discharge hindcasts. How is that possible? Or is there actually a small river, which is not shown in the Figure, flowing into the sea for that station?
  Response: The reviewer did see this correctly. The current figure only shows the main river network, however there are many smaller rivers and streams that are unfortunately not depicted. The station shown close to the coast is a measuring site placed at one of the sluices, which is also connected by to the main rivers by a smaller river branch. We will make this information more clear in the manuscript.
  
  Section 3.2: It is stated that “BSS confirms the earlier findings and shows the same trend of increased performance in the first lead weeks….”. I disagree with this finding as Fig. 6 shows clearly that the BSS for Feb/May and Sept is low in contrast to the findings for the general performance.
  Response: We do acknowledge and explain the low performance in the following sentences: “The BSS confirms the earlier findings and shows the same trend of increased performance in the first lead weeks, additional skill of several weeks is found for early spring and early summer months. However, tiles with lower performance throughout long lead periods, late summer and winter months can be spotted for this station. Some of these weeks appear to be more difficult to predict compared to early months in the year. This is likely due to unequal distribution of low flow occurrences throughout the year: where during summer months low flows can be more common and therefore chances to not fully capturing every event are higher, the low flows during winter are less common and captured relatively well with the snow melt dynamic as seen in previous scores.”
  
  We acknowledge that the first sentence might be misleading and we will adapt it.
  
  Section 2.2.: Please add a very brief explanation to the lagged times series approach as most of the readers will not be familiar with this approach.
  Response: We will include a brief explanation of the lagged time series approach together with the extended explanation of the modelling framework. However, we would also like to highlight the reference to the previous paper, in which the development and further detail of the modelling framework are shown.
  
  Lines 205-212: It is stated that only minor differences are observed for the different ML models. Please analyse better the reason for this. One would assume that advanced DL methods such as LSTM would perform better than multiple linear regression
  Response: This is indeed a remarkable result, since when used with observed input data, the differences between the different ML models were larger. The most likely explanation is that the forecasting skill is very much dependent on the skill by which the input variables are forecasted, which apparently make the differences in skill between the ML models insignificant in comparison. We will add an explanation for this in the manuscript.
  
  Figure A1: Does this figure show combined results of sea level and river levels? If yes, please separate these two variables..or
  Response: Figure A1 only shows results of forecasted fresh (river, streams, lakes) surface water levels
  
  Citation: https://doi.org/10.5194/hess-2022-89-AC1
RC2:
'Comment on hess-2022-89', Anonymous Referee #2, 18 May 2022
The study is building heavily on results published elsewhere (Hauswirth et al., 2021). I guess that is fine and always a difficult call to decide just how much information to provide so that a paper becomes a stand-alone piece of work without unnecessary repetition. However, in places I would have liked a little more info in this paper so I would not necessarily have to read the previous paper. For example, lines 114-117; this seems important and a bit more in-depth description of the datasets (SEAS5) and the ‘lagged time series approach’ which I am not familiar with.

Perhaps I have overlooked something, but I am not sure how the all the forecasts made by the different models and different ensembles are aggregated into one CDF (as per Fig.2)? Also, If there is only a minor difference in the performance between the five different ML models (line 205) then what is the advantage of using all of them rather than selecting the ‘best’ or most credible model for a particular site and use that? Might it be useful to add a section highlighting the performance of each of the five ML models in contrast to the aggregate performance of the entire system?

Minor comments:

The acronym LSTM is not defined?

Figs.2-6 did not come off well in my black and white copy, but ok for online viewing (which in fairness is probably the most common by now).

Line 253: Why the 20^th percentile? Or why only the 20^th percentile? I could imagine that the ability to assess low-flow across a range of severities would be of interest?

Line 253: Do any of your models include the effect of human interventions and their potential impacts on low flow? For example, water restrictions, operation of control structures to manage low flows etc? If not, is this likely to important in a highly regulated system such as the Netherland’s water ways? I think there is so mentioning of this in lines 220-225, but seems to me this is particularly important during low flow?

Line 293: I am not sure I understand how you incorporated water management into your models? Was this done in this study, or is that something that was part of the Hauswirth et al. (2021) study? I think perhaps more detail on this could be included in this manuscript as this seems interesting and important (even if you did not find a strong effect).
Citation: https://doi.org/10.5194/hess-2022-89-RC2
- AC2: 'Reply on RC2', Sandra Margrit Hauswirth, 19 Jun 2022
  
  Anonymous Referee #2
  Referee comment on "The suitability of a hybrid framework including data driven approaches for hydrological forecasting" by Sandra M. Hauswirth et al., Hydrol. Earth Syst. Sci. Discuss., https://doi.org/10.5194/hess-2022-89-RC2, 2022
  The study is building heavily on results published elsewhere (Hauswirth et al., 2021). I guess that is fine and always a difficult call to decide just how much information to provide so that a paper becomes a stand-alone piece of work without unnecessary repetition. However, in places I would have liked a little more info in this paper so I would not necessarily have to read the previous paper. For example, lines 114-117; this seems important and a bit more in-depth description of the datasets (SEAS5) and the ‘lagged time series approach’ which I am not familiar with.
  Perhaps I have overlooked something, but I am not sure how the all the forecasts made by the different models and different ensembles are aggregated into one CDF (as per Fig.2)? Also, If there is only a minor difference in the performance between the five different ML models (line 205) then what is the advantage of using all of them rather than selecting the ‘best’ or most credible model for a particular site and use that? Might it be useful to add a section highlighting the performance of each of the five ML models in contrast to the aggregate performance of the entire system?
  Response: We want to thank the reviewer for taking the time to review our manuscript and sharing thoughts and suggestions for improvement. We will address the comments, suggestions and open questions point by point below.
  We agree that our study is heavily building up on a previous study by the main author, which was a conscious decision that was made. The idea was and is to use the base modelling framework for different experiments. Therefore, we wanted to fully focus on the model development aspect in the first publication and refer to it for future work. We acknowledge that in this case, the explanation of the modelling framework was too short (as also indicated by the other reviewer). We plan to extend the description of the base modelling framework, including the lagged time series approach. Furthermore, we will extend the description of the new introduced datasets such as SEAS5 and EFAS as recommended.
  
  Regarding the comment Fig 2 and how the different ensembles were aggregated into one CDF, we would like to refer to line 196 (‘This was done by computing all the individual daily CRPSS results of all the hindcasts of every station and method, before aggregating the individual CRPSS scores to different temporal scales (weekly in Fig. 2).) and line 155 for the specific calculation procedure (‘Equation 1 …was used and the CRPS computed over all ensemble members for each lead day of every hindcast before aggregating it to other temporal scales’). For creating Fig 2 the different CRPSS scores for the different station and method combinations were additionally aggregated by lead day and the average was used for plotting.
  
  The reviewer poses a good question regarding what is the advantage of using different ML models, while they only show minor differences in their results. While performing the experiments, we were not expecting to only observe minor differences, while predictions using observed input data did show differences (Hauswirth et al., 2021). The most likely explanation is that the forecasting skill is very much dependent on the skill by which the input variables are forecasted, which apparently make the differences in skill between the ML models insignificant in comparison. We will add an explanation for this in the manuscript.
  
  To keep the manuscript easy to follow we decided to focus on one method in more detail. However, we will add a section on this as this is an interesting aspect to compare to the previous study (Hauswirth et al., 2021), where there were differences found between the models.
  
  Minor comments:
  The acronym LSTM is not defined?
  
  Response: We will define the acronym LSTM (Long Short Term Memory Model) to prevent confusion.
  Figs.2-6 did not come off well in my black and white copy, but ok for online viewing (which in fairness is probably the most common by now).
  
  Response: We will revisit the figures 2 - 6 and see if we can improve the readability for printed version.
  Line 253: Why the 20th percentile? Or why only the 20th percentile? I could imagine that the ability to assess low-flow across a range of severities would be of interest?
  
  Response: We chose the 20^th percentile as it is a commonly chosen threshold for low flows and was also interesting to our collaboration partners.
  
  Line 253: Do any of your models include the effect of human interventions and their potential impacts on low flow? For example, water restrictions, operation of control structures to manage low flows etc? If not, is this likely to important in a highly regulated system such as the Netherland’s water ways? I think there is so mentioning of this in lines 220-225, but seems to me this is particularly important during low flow?
  
  Response: The Netherlands is indeed a country with high water management infrastructures and plans. We therefore are working with the National Water Authority to include expert knowledge where possible and available. In our previous study we did include a separate water management scenario where we included operational plans of the major water infrastructures. We did see that there was an improvement in our simulations, albeit very minor in terms of our modelling results (Hauswirth et al., 2021).
  Line 293: I am not sure I understand how you incorporated water management into your models? Was this done in this study, or is that something that was part of the Hauswirth et al. (2021) study? I think perhaps more detail on this could be included in this manuscript as this seems interesting and important (even if you did not find a strong effect).
  
  Response: In line with the previous comment about water management above: We did a run with water management influence using the same approach as in Hauswirth et al. (2021). This run includes operational rules of main infrastructures which are related to the Rhine discharge at Lobith (one of our main input variables) for two specific input locations and two additional observation records of locations based at smaller infrastructures. We were therefore able to use the same approach regarding the operational rules for the main infrastructures, as these are based on the Rhine discharge we obtain from the EFAS dataset. For the two other additional timeseries climatology was used as operational plans were not available.
  
  We acknowledge that the information given in this manuscript related to water management aspects are very limited and heavily based on the previous study. We will extend the explanation so that the reader will be able to follow better.
  
  Citation: https://doi.org/10.5194/hess-2022-89-AC2

Peer review completion

AR: Author's response | RR: Referee report | ED: Editor decision

ED: Reconsider after major revisions (further review by editor and referees) (16 Jul 2022) by Thomas Kjeldsen

AR by Sandra Margrit Hauswirth on behalf of the Authors (10 Aug 2022) Author's response Author's tracked changes Manuscript

ED: Referee Nomination & Report Request started (05 Sep 2022) by Thomas Kjeldsen

RR by Anonymous Referee #2 (05 Dec 2022)

RR by Louise Arnal (05 Jan 2023)

Suggestions for revision or reasons for rejection

Hide

ED: Publish subject to minor revisions (review by editor) (05 Jan 2023) by Thomas Kjeldsen

AR by Sandra Margrit Hauswirth on behalf of the Authors (14 Jan 2023) Author's response Author's tracked changes Manuscript

ED: Publish as is (15 Jan 2023) by Thomas Kjeldsen

AR by Sandra Margrit Hauswirth on behalf of the Authors (16 Jan 2023) Author's response Manuscript

Journal article(s) based on this preprint

Sandra M. Hauswirth et al.

Viewed

Total article views: 870 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
646	208	16	870	5	5

HTML: 646
PDF: 208
XML: 16
Total: 870
BibTeX: 5
EndNote: 5

Views and downloads (calculated since 31 Mar 2022)

Month	HTML	PDF	XML	Total
Mar 2022	121	28	3	152
Apr 2022	200	52	3	255
May 2022	118	22	5	145
Jun 2022	52	18	5	75
Jul 2022	28	10	0	38
Aug 2022	24	19	0	43
Sep 2022	29	13	0	42
Oct 2022	16	5	0	21
Nov 2022	27	14	0	41
Dec 2022	17	10	0	27
Jan 2023	14	17	0	31

Cumulative views and downloads (calculated since 31 Mar 2022)

Month	HTML	PDF	XML	Total
Mar 2022	121	28	3	152
Apr 2022	200	52	3	255
May 2022	118	22	5	145
Jun 2022	52	18	5	75
Jul 2022	28	10	0	38
Aug 2022	24	19	0	43
Sep 2022	29	13	0	42
Oct 2022	16	5	0	21
Nov 2022	27	14	0	41
Dec 2022	17	10	0	27
Jan 2023	14	17	0	31

Viewed (geographical distribution)

Total article views: 798 (including HTML, PDF, and XML) Thereof 798 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 26 Jan 2023

Download

The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.

Preprint (5577 KB)
Metadata XML

Short summary

Forecasts on water availability are important for water managers. We test a hybrid framework based on machine learning models and global input data for generating seasonal forecasts. Our evaluation shows that our discharge and surface water level predictions are able to create reliable forecasts up to 2 months ahead. We show that a hybrid framework, developed for local purpose, combined and rerun with global data is able to create valuable information similar to large scale forecasting models.


Total:	0
HTML:	0
PDF:	0
XML:	0