Quantifying the response of water and carbon balances to land cover and climate extremes across Germany

Pyarali, Karim; Zhang, Lulu; Liu, Ning; Al-Qubati, Abdulhakeem; Sun, Ge

doi:https://doi.org/10.5194/egusphere-2025-1629

Preprints

https://doi.org/10.5194/egusphere-2025-1629

Preprints

04 Aug 2025

| 04 Aug 2025

Quantifying the response of water and carbon balances to land cover and climate extremes across Germany

Karim Pyarali, Lulu Zhang, Ning Liu, Abdulhakeem Al-Qubati, and Ge Sun

Abstract. Land cover and extreme weather events are closely connected to water yield and carbon sequestration. Understanding the tradeoffs between carbon and water and how they respond to human disturbances is critical for quantifying ecosystem services. The monthly scale ecosystem model, WaSSI, was tested and applied across Germany for mapping carbon and water balances from 2001–2019. We estimated that Germany generates 84.86 billion m3 of discharge and sequesters 106.03 Tg of carbon annually. The eastern states were comparatively drier than the rest of the country, as most of their precipitation was lost as evapotranspiration. Croplands, urban areas and Evergreen Needle Forests (ENF) provide 82.5 % of the water yield, while the forests sequester the major share of carbon (56.3 %) altogether. The results highlight the importance of sparse land covers (e.g. wetlands) in carbon sequestration. Findings also suggest that national water yield and carbon balances are sensitive to extreme events. In 2002 and 2013, due to high precipitation, the stocks of key ecosystem services were notably higher. Similarly, during the drought years of 2003 and 2018, the services were reduced drastically, but we found that buffers from the previous year played an important role in mitigating negative impacts. This study highlights that, when integrated with local data, a relatively simple modelling approach is adequate to answer questions of coupled water and carbon responses to climatic variability at a large scale. We conclude that land management of both forests and croplands are vital to sustain ecosystem services under a changing climate.

Received: 13 Apr 2025 – Discussion started: 04 Aug 2025

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.

Download & links

Preprint (PDF, 3634 KB)

Supplement (686 KB)

Download & links

Karim Pyarali, Lulu Zhang, Ning Liu, Abdulhakeem Al-Qubati, and Ge Sun

Status: final response (author comments only)

RC1:
'Comment on egusphere-2025-1629', Anonymous Referee #1, 29 Sep 2025
Pyarali et al. simulate how different land cover types and climate variability – wet and dry years – modulate multiple ecosystem processes and services using the Water Supply Stress Index (WaSSI) model in Germany. The study contributes to understanding the spatio-temporal variability of the provision of key ecosystem services such as water yield and carbon sequestration, as well as their anomalies during extreme wet and dry years, at a national scale, using an ecosystem model that simulates the coupling between water and carbon cycles. Additionally, the manuscript can provide new insights into the potential loss of ecosystem services provision under future climate projections in Germany.

While the manuscript is well written and organized, I see some issues that, in my view, need to be addressed before it can be considered for publication in HESS. Specifically: (i) I found it difficult to follow the description of the model evaluation given the limited level of detail provided on validation; (ii) the manuscript would be strengthened by a more explicit discussion of how the areal extent of each land cover type influences the provision of ecosystem services; (iii) the characterization of extreme wet and dry years is based only on precipitation anomalies, which I find too narrow.

Major comments:

Lines 204 and 229: I recommend describing the performance of streamflow (Q) and evapotranspiration (ET) validation for the same watershed(s), as this would provide a more complete view of how the model represents the water balance. In addition, you could extend the validation analysis by describing how the model captures ET and Gross Primary Productivity (GPP) at sites where both variables are available.

Could you provide a potential explanation for why the model tends to overestimate ET during winter? What are the implications for your results?

Since the analysis is focused on extreme dry–hot and wet–cold years, how does the model perform during drier years (e.g., MAP < 25^th percentile) and wetter (e.g., MAP > 75^th percentile) years?

Figure 6: Are the simulated GPP and NEP variations (i.e., decreases and increases) consistent with observed changes from eddy covariance towers and satellite observations? This may provide a better indication of whether the model accurately captures the sensitivity of GPP and NEP to precipitation and temperature anomalies.

I recommend adding the model parameterization to the supplementary material, as this would make the study easier to reproduce.

Line 277: I suggest conducting the analysis while accounting for the area of each land cover type. The high-water supply attributed to croplands may largely reflect their dominant extent in Germany. Similarly, the ecohydrological role of forests could be underestimated due to their smaller extent in the country. I encourage authors to present the results considering the mean and variability contribution of each land cover (mean ± SD m³ ha^-1 yr^-1). This approach would also allow you to discuss the potential implications of land-cover change (e.g., forest to cropland) on the provision of ecosystem services.

Lines 264 and 289: The definition of drought and flood events typically incorporates not only precipitation anomalies but also temperature, along with other indicators such as soil moisture or vapour pressure deficit. Additionally, since both precipitation and temperature are inputs into the WaSSI model, I suggest that you characterize both precipitation and temperature anomalies, including temperature anomaly maps in your analysis (e.g., Figs. 8 and 9). In addition, while droughts often extend over large areas, flood-affected regions are more likely to be concentrated along river channels and flat terrains, whereas your analysis is at the country scale. Therefore, the terms ‘hot-dry’ and ‘wet-cold’ years may provide a clearer description of the extreme climate events.

Lines 314-396: You provide a lot of valuable insights in the discussion section. However, you should consider focusing more clearly on the three main results of the manuscript (i) model validation, including comparisons with previous studies; (ii) water yield and carbon sequestration associated with each land cover and the land cover change implications; and (iii) the response of these ecosystem services to extreme climatic conditions, including comparisons with previous studies and their implications under future climate projections in Germany.

Minor comments:

Line 15: Please add the full name of the WaSSI model

Line 16: I suggest replacing “from 2001-2019” with “from 2001 to 2019”

Line 52: Add space

Lines 85-88: I suggest adding a paragraph in the introduction to describe the generality of the model, particularly its previous applications.

Line 109: Please add the corresponding reference(s) to the eddy flux dataset.

Lines 115-116: Please specify if, at the monthly scale, the correlation between GPP and ET is linear or not.

Line 119: I am particularly familiar with the WaSSI model, but I am curious if this includes any calibration process.

Line 124: Is R² the coefficient of determination? Please specify.

Line 125: Which MODIS product? Add the corresponding reference.

Line 126: Please clarify why you computed the monthly deviations and describe the interpolation process, as this is not clear in the methodology.

Line 127: Which MODIS product? Add the corresponding reference to the MODIS and CGLS products.

Line 129: Please add the description of the extent of the included watersheds for Q validation

Line 160: What is meant by anthropogenic influence? Does this refer to dams, land cover change, or other factors?

Line 191: I suggest presenting the results of the main figures first (e.g., Figure 2) and then supporting your analysis with supplementary Figures and tables.

Figure 2: Add the performance metrics in each panel. All you can order the panels following a criterion such as watershed area, natural/transformed land cover area, or annual precipitation

Line 199: I am curious if the model exhibits any dependence on watershed area or dominant land cover types regarding its performance?

Line 205: I think this corresponds to the methods section.

Figure 3: Please add performance metrics. Please clarify in the scatterplot labels and lines legend which correspond to observations and simulations.

Line 224: I consider that this corresponds to the methods section

Line 253: I notice that in the results section, you use past and present tenses. I suggest using the present tense.

Lines 267-269: You state that Q and NEP are more sensitive to changes in precipitation than ET and GPP. You should also consider how climatic conditions before wet or dry years affect the responses of water and carbon fluxes, as well as their responses after dry or wet years, to examine potential lag effects. In this context, it is not clear how groundwater buffers drought impacts based on the results presented in Fig. 8 (Lines 373–375).

Figure 6: Precipitation is not typically considered an ecosystem flux

Figures 8 and 9: I suggest presenting the anomalies as percentages or standardized anomalies to better account for spatial variability across Germany.

Line 288: What are the implications of crop irrigation on water yield and carbon sequestration response to dry/wet years?

Lines 411-413: I think this sentence does not correspond to the conclusions section.

Technical corrections:

Line 52: Add space.

Line 108: I guess the correct reference is “Fan et al. (2016)”

Line 139: Add a space in “of 2018” and “100 m”. This typo is found throughout the document.

Please check whether the figure colors are suitable for color-blind readers.
Citation: https://doi.org/10.5194/egusphere-2025-1629-RC1
RC2:
'Comment on egusphere-2025-1629', Anonymous Referee #2, 01 Oct 2025
In this manuscript, the authors validate the WaSSI (Water Supply Stress Index) model for the case of Germany. Since this model has typically been applied to other study areas (i.e., the conterminous United States, Mexico, Rwanda, and Burundi), this application represents an innovative contribution. The study is also valuable in clarifying some of the dynamics behind the spatial and temporal variations of the ecosystem services considered in the study area. Nevertheless, a number of gaps are present, and taken together, they make it difficult to fully understand the methodologies and the results described in this work. In particular, some points and logical steps would benefit from a more detailed explanation in order to strengthen the overall clarity and robustness of the study. My review mainly focuses on these aspects.
In conclusion, I consider the manuscript suitable for publication, provided that the general revision suggested are carefully addressed.

General comments
While reading the manuscript, I had the impression that the term “flood” is sometimes used to refer to a “high precipitation event” (e.g., line 372) or an “extremely wet year” (e.g., line 297). However, the term “flood” typically refers to the inundation resulting from an excess of water that causes normally dry areas to be submerged. It does not simply indicate heavy rainfall events or periods of above-average precipitation. I would therefore suggest reviewing the manuscript to ensure that the term “flood” is used correctly and, where necessary, replacing it with more precise terms or expressions.

I think it would be important to consistently include references to the datasets used in the text. I would therefore suggest checking this aspect throughout the manuscript—for example, at line 164 (FLUXNET2015) and line 177 (MOD17A2HGF).

Referring to Figure 4c, I found it a bit difficult to understand how, in some areas, actual evapotranspiration could exceed potential evapotranspiration. For readers who are not experts on the topic, the intuitive expectation would be that actual evapotranspiration is less than or equal to potential evapotranspiration. It might be worthwhile to explain and comment on this difference between the two variables, as this could help readers better follow the results.

I believe that certain essential details are missing in the description of the procedures, particularly in some steps that are crucial for understanding the methodology. Specifically, the following aspects would benefit from further elaboration:
Lines 145-146: a more detailed explanation of what the quality checks consisted of could improve the clarity of both the procedure and the results.

Lines 158-160: for transparency and clarity, it could be helpful to provide further details about why those specific stations were selected, why not more were included, and how the choice of different stations might affect the results.

Lines 164-165: a short explanation of how corrections and gap filling were carried out could be useful.

Results and Discussion Section lack some important steps and information. These missing elements are essential to fully grasp the reasoning that supports the results and to properly assess their applicability in different contexts:
In the Discussion section, I did not find any direct reference to Figure 6. It might be useful to elaborate on and comment on this figure, since it is the only one showing the temporal evolution of the variables considered.

I think that from the presentation of the results, it is not clear how “buffers developed from the previous year can play a significant role in mitigating this effect”, particularly when discussing the sensitivity of ecosystem services to extreme events (lines 318–319). I believe it would be important to elaborate on this point in the Results section and, if possible, refer to the corresponding figures in the Discussion chapter.

Line 320: it may be useful to further clarify why the model can be applied to the whole Central Europe. As it is currently written, it is not entirely clear how the procedures and conclusions could also be applied to areas beyond the specific study region (i.e., Germany).

Specific comments
I am not very familiar with the WaSSI model, and I found it challenging to fully understand what it is and how it works. While some information is provided, it appears scattered across different parts of the manuscript. I think that an initial, more comprehensive presentation of the model—beyond what is already included—would be very helpful to better frame the analyses and to facilitate the interpretation of the results. This could include its main aim, the study areas considered, the way it operates, and other relevant details.

In the abstract, a lot of results are described and explained. I think it might be helpful to more clearly highlight the main novelties and contributions introduced by this work.

Please consider adding a section describing the study area (climate, land use, etc.)

Line 15: you may consider briefly clarifying what is meant by “ecosystem services” for readers less familiar with the concept.

It would improve readability if acronyms were defined when they first appear in the text. For example, “MODIS” and “CGLS” (lines 125 and 127) are introduced without prior definition.

Line 141: providing a short explanation of why exactly 10 classes were chosen, and whether you considered different numbers, might help clarify the procedure.

Line 145: including the number of removed pixels could be an informative detail.

Lines 165-169: this section might be difficult to follow for readers less familiar with the topic. Expanding the explanation with more details could make it clearer.

In Section 3.1, I think it would be very helpful to include a comparison between the performance metrics obtained for the different variables (discharge, ET, GPP) in this study and those reported in the literature from other studies applying the same or similar models (if available). Providing this context could help to better clarify the results and support the applicability of the WaSSI model to new study areas.

Figure 2: adding a legend to indicate the meaning of the two colours could improve the clarity of the plots.

Figure 4: you may consider including in the figure the names of the regions mentioned in the text, or finding a way to highlight them. This would make the reading and the analysis of the figure much easier.

Lines 316-319: referring explicitly to the relevant tables/figures in the text could make the discussion clearer.

Lines 324-327 and 333-336: a more detailed explanation of the influence of the different factors could add clarity.

Line 339: explaining why results are weaker for the Lackenberg Forest station, or what the reasons can be, could improve transparency.

Line 340 and 341: the meaning of the terms “complex” and “complicated” in this context is not entirely clear; considering alternative synonyms or more specific wording may help.

Technical comments
I warmly suggest to check that the figures are accessible also for readers with colour-vision deficiencies. Alternative colour scales might improve clarity.

Figure S1: you might consider modifying the colour scale (as noted above) and/or using different marker shapes for each category to increase readability.

Please double-check punctuation when separating sentences across the manuscript. In some cases, a full stop is missing (e.g., line 36, between “interconnected reasons” and “two of them”), while in others a full stop is used where a comma or similar punctuation might be more appropriate (e.g., line 53, before the sentence starting with “Which negatively affects”). In my opinion, this would strongly improve the readability of the manuscript.
Citation: https://doi.org/10.5194/egusphere-2025-1629-RC2

Karim Pyarali, Lulu Zhang, Ning Liu, Abdulhakeem Al-Qubati, and Ge Sun

Supplement

https://doi.org/10.5194/egusphere-2025-1629-supplement

Karim Pyarali, Lulu Zhang, Ning Liu, Abdulhakeem Al-Qubati, and Ge Sun

Viewed

Total article views: 1,310 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
1,245	43	22	1,310	29	31	32

HTML: 1,245
PDF: 43
XML: 22
Total: 1,310
Supplement: 29
BibTeX: 31
EndNote: 32

Views and downloads (calculated since 04 Aug 2025)

Month	HTML	PDF	XML	Total
Aug 2025	646	19	5	670
Sep 2025	528	11	8	547
Oct 2025	71	13	9	93

Cumulative views and downloads (calculated since 04 Aug 2025)

Month	HTML	PDF	XML	Total
Aug 2025	646	19	5	670
Sep 2025	528	11	8	547
Oct 2025	71	13	9	93

Viewed (geographical distribution)

Total article views: 1,237 (including HTML, PDF, and XML) Thereof 1,237 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 25 Oct 2025

Short summary

An ecosystem services model was applied across Germany to estimate water supply and carbon sequestration. The results showed that total annual water discharge and carbon sequestration for Germany is 85 billion m3 and 106 TgC, respectively. Furthermore, we found that croplands provide the largest amount of water, deciduous broadleaf forests sequester most of the carbon, and wetlands are very effective in absorbing carbon. During extreme events, we noticed a real impact on both services.


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