Assessing the effect of land cover on ISBA snow water equivalent simulations over Europe

Rojas-Munoz, Oscar; Ardilouze, Constantin; Bonan, Bertrand; Tzanos, Diane; Ghent, Darren; Lamarche, Céline; Nagler, Thomas; Calvet, Jean-Christophe

doi:10.5194/egusphere-2026-65

Preprints

https://doi.org/10.5194/egusphere-2026-65

Preprints

26 Jan 2026

| 26 Jan 2026

Assessing the effect of land cover on ISBA snow water equivalent simulations over Europe

Oscar Rojas-Munoz, Constantin Ardilouze, Bertrand Bonan, Diane Tzanos, Darren Ghent, Céline Lamarche, Thomas Nagler, and Jean-Christophe Calvet

Abstract. An accurate representation of the land surface is essential for simulating the exchange of energy, water and carbon between the land and the atmosphere. This study evaluates the impact of land cover representation on snow simulations in the Interactions Between Soil, Biosphere and Atmosphere (ISBA) land surface model in Europe between 2010 and 2022. The study employs the European Centre for Medium-Range Weather Forecasts (ECMWF) ERA5 atmospheric forcing dataset. Offline simulation experiments were conducted using two different versions of the model to prescribe land cover. The most recent version uses the latest land cover data from the European Space Agency's (ESA) Climate Change Initiative (CCI). The model's ability to reproduce snow dynamics was evaluated through a comparison of the simulations with ESA CCI satellite snow water equivalent (SWE) retrievals and ERA5 snow analyses. The ERA5 analysis shows the highest level of agreement with satellite observations of SWE at the domain scale. On average, both the ERA5 and ISBA simulations tend to overestimate SWE compared to the CCI SWE. However, it is also possible that the CCI SWE product underestimates the actual SWE. This bias is particularly large during the warm winter of 2020, while the scaled SWE anomalies are comparable to those observed by ESA CCI and ERA5. Using ESA CCI land cover data reduces the ISBA SWE bias by around 33 %, with this reduction being observed over most of the domain. These findings emphasise the importance of accurate land cover data for improving snow representation in land surface models and highlight the need for updated vegetation information in future snow-related applications.

Received: 05 Jan 2026 – Discussion started: 26 Jan 2026

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, 1992 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 (1992 KB)

Supplement (420 KB)

Download & links

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

Journal article(s) based on this preprint

24 Jul 2026

Assessing the effect of land cover on ISBA snow water equivalent and land surface temperature simulations over Europe

Oscar Rojas-Munoz, Constantin Ardilouze, Bertrand Bonan, Diane Tzanos, Darren Ghent, Céline Lamarche, Thomas Nagler, and Jean-Christophe Calvet

The Cryosphere, 20, 4099–4115, https://doi.org/10.5194/tc-20-4099-2026,https://doi.org/10.5194/tc-20-4099-2026, 2026

Short summary

Oscar Rojas-Munoz, Constantin Ardilouze, Bertrand Bonan, Diane Tzanos, Darren Ghent, Céline Lamarche, Thomas Nagler, and Jean-Christophe Calvet

Interactive discussion

Status: closed

RC1:
'Comment on egusphere-2026-65', Anonymous Referee #1, 17 Feb 2026
Review of “Assessing the effect of land cover on ISBA snow water equivalent simulations over Europe” by O. Rojas-Munoz.
In this manuscript, the authors are presenting different simulations of snow water equivalent (SWE) across Europe with the ISBA model using two different land cover databases. They compared their simulations against ERA5 snow analyses as well as against the ESA CCI SWE databases. They showed that their simulation with the newest land cover database provided on average better SWE estimates over the full domain, but still some discrepancies with ERA5 or the SWE database could be observed. I found the paper well written and clear. I think this manuscript is fit for publication after a few improvements. Additional analysis is needed to examine how land cover changes between the two databases affect SWE estimates, which may explain why performance has declined in regions like Northern Europe.
Major comments:
I think additional analysis should be done to explain the degradation in SWE in norther Europe (Figs. 4,5, and L. 243-246). I think the degradation should be put in perspective with the new land cover in this area from temperate grasslands to bare soil with no vegetation (Fig. 1). Could this degradation be caused by a poor representation of certain processes in this new land classification?

Overall, the manuscript should place greater emphasis on comparing how SWE biases change spatially between the old and new land cover datasets for each vegetation type, as this represents the central contribution of the study.
Section 2.1: I think it would be interesting to have a more complete description of the ISBA model, including the interaction of the snow with the atmosphere, the forest, and the ground. I don’t find the description of ISBA-A-gs pertinent for this study.

96, 113, 125. It is mentioned that the product is regridded/interpolated/resampled to a coarse resolution. What kind of regridded/interpolation method is used?

The modelled surface temperature is compared to observations (CCI-LST). The authors need to explain how they calculated grid-averaged surface temperatures (between soil, vegetation, snow, …) that they compared to CCI-LST. Also, this evaluation of LST should maybe be mentioned in the title of the manuscript and in the abstract. As it is right now, this evaluation feels out of place in the manuscript.

171-172: why was the comparison of the surface temperature done in a sub-domain? Can this subdomain be shown on Fig. 1?

Minor comments:
160-161: it is not clear to me how the LAI is calculated, and it should be explained why it matters in this study.

187: How are the anomalies calculated? The LAI anomalies also need to be further explained.

204: ‘narrowing the gap with the observations”. I think another term than ‘gap’ should be used here. Bias? And I am not sure the CCI SWE product should be called ‘observations’?

205: ‘discrepancies’. This should be explicitly explained in the manuscript rather than leaving it for the reader to interpret.

The quality of some figures (e.g. Fig. 2) should be improved.

61: I am surprised to see some big correlation differences in some areas that receive almost not snow, such as south-west of France, central Spain, and the UK. Maybe some filtering should be done?
Citation: https://doi.org/10.5194/egusphere-2026-65-RC1
- AC1: 'Reply on RC1', Jean-Christophe Calvet, 29 Apr 2026
  
  See the enclosed Supplement.
  
  Citation: https://doi.org/10.5194/egusphere-2026-65-AC1
RC2:
'Comment on egusphere-2026-65', Anonymous Referee #2, 17 Mar 2026
This investigates the impact of changing the land cover (LC) classification within ISBA (driven by ERA5 forcing) on the SWE estimates. ESA CCI SWE is used as the reference SWE dataset. Comparisons are also made with ERA5. Replacing the old LC information with one based on the ESA CCI LC v2.0.7 improved the SWE bias when compared to ESA CCI over the full temporal and spatial domain. Although New LC improved versus CCI, ERA5 often has better performance with CCI compared to the ISBA simulation. The authors also look at the impact of change in LC on LST but are unable to attribute LST differences to the changes in SWE.
The manuscript is clear and well presented however I found the analysis and the discussion to be a bit limited. I would like to see additional analysis that looks at changes in SWE according to changes in LC classification. As it stands, the analysis shows that changing the LC information alters the SWE which is an expected result but it stops short of linking the SWE changes to specific changes in LC classification.

For example:
Did all areas that had a change in LC classification experience a change in SWE? If not, which LC changes did not result in a SWE change (i.e. SWE agnostic to those different LCs).

Which types of classification change experience that largest change in SWE?

Were certain LC changes responsible for most of the change in MB?

From a qualitative perspective it appears that moving from bare soil (3) to deciduous broadleaf (9) had an impact on SWE. This should be discussed further.

One suggestion is to quantify the number of pixels that experienced a reclassification and to further categorize those according to type of change (i.e. % of pixels that changed from 13 to 7) and finally to examine the change in SWE according to these LC changes.
I am curious why areas that did not appear to have a change in LC (e.g. Sweden) seem to have more SWE in the ISBA run with the New LC compared to the old LC. This should be examined.
Overall, while the manuscript is well-written and clear I found it lacking in analysis to support the main research objective.

Specific minor comments
Figure 3 and associated text: How did you address the absence of SWE estimates from CCI when producing Figure 3? Were the areas masked in CCI also removed from ISBA and ERA5?
L308-310: Limit is closer to 200mm than 100mm, although SWE exceeding 200mm is common in the CCI SWE product (see Fig. 4.10 [March] in Barella et al. 2024 and also Luojus et al. 2021). See Figure 4.15 in for spatial map of local biases.
Citation: https://doi.org/10.5194/egusphere-2026-65-RC2
- AC2: 'Reply on RC2', Jean-Christophe Calvet, 29 Apr 2026
  
  See the enclosed Supplement.
  
  Citation: https://doi.org/10.5194/egusphere-2026-65-AC2

Interactive discussion

Status: closed

RC1:
'Comment on egusphere-2026-65', Anonymous Referee #1, 17 Feb 2026
Review of “Assessing the effect of land cover on ISBA snow water equivalent simulations over Europe” by O. Rojas-Munoz.
In this manuscript, the authors are presenting different simulations of snow water equivalent (SWE) across Europe with the ISBA model using two different land cover databases. They compared their simulations against ERA5 snow analyses as well as against the ESA CCI SWE databases. They showed that their simulation with the newest land cover database provided on average better SWE estimates over the full domain, but still some discrepancies with ERA5 or the SWE database could be observed. I found the paper well written and clear. I think this manuscript is fit for publication after a few improvements. Additional analysis is needed to examine how land cover changes between the two databases affect SWE estimates, which may explain why performance has declined in regions like Northern Europe.
Major comments:
I think additional analysis should be done to explain the degradation in SWE in norther Europe (Figs. 4,5, and L. 243-246). I think the degradation should be put in perspective with the new land cover in this area from temperate grasslands to bare soil with no vegetation (Fig. 1). Could this degradation be caused by a poor representation of certain processes in this new land classification?

Overall, the manuscript should place greater emphasis on comparing how SWE biases change spatially between the old and new land cover datasets for each vegetation type, as this represents the central contribution of the study.
Section 2.1: I think it would be interesting to have a more complete description of the ISBA model, including the interaction of the snow with the atmosphere, the forest, and the ground. I don’t find the description of ISBA-A-gs pertinent for this study.

96, 113, 125. It is mentioned that the product is regridded/interpolated/resampled to a coarse resolution. What kind of regridded/interpolation method is used?

The modelled surface temperature is compared to observations (CCI-LST). The authors need to explain how they calculated grid-averaged surface temperatures (between soil, vegetation, snow, …) that they compared to CCI-LST. Also, this evaluation of LST should maybe be mentioned in the title of the manuscript and in the abstract. As it is right now, this evaluation feels out of place in the manuscript.

171-172: why was the comparison of the surface temperature done in a sub-domain? Can this subdomain be shown on Fig. 1?

Minor comments:
160-161: it is not clear to me how the LAI is calculated, and it should be explained why it matters in this study.

187: How are the anomalies calculated? The LAI anomalies also need to be further explained.

204: ‘narrowing the gap with the observations”. I think another term than ‘gap’ should be used here. Bias? And I am not sure the CCI SWE product should be called ‘observations’?

205: ‘discrepancies’. This should be explicitly explained in the manuscript rather than leaving it for the reader to interpret.

The quality of some figures (e.g. Fig. 2) should be improved.

61: I am surprised to see some big correlation differences in some areas that receive almost not snow, such as south-west of France, central Spain, and the UK. Maybe some filtering should be done?
Citation: https://doi.org/10.5194/egusphere-2026-65-RC1
- AC1: 'Reply on RC1', Jean-Christophe Calvet, 29 Apr 2026
  
  See the enclosed Supplement.
  
  Citation: https://doi.org/10.5194/egusphere-2026-65-AC1
RC2:
'Comment on egusphere-2026-65', Anonymous Referee #2, 17 Mar 2026
This investigates the impact of changing the land cover (LC) classification within ISBA (driven by ERA5 forcing) on the SWE estimates. ESA CCI SWE is used as the reference SWE dataset. Comparisons are also made with ERA5. Replacing the old LC information with one based on the ESA CCI LC v2.0.7 improved the SWE bias when compared to ESA CCI over the full temporal and spatial domain. Although New LC improved versus CCI, ERA5 often has better performance with CCI compared to the ISBA simulation. The authors also look at the impact of change in LC on LST but are unable to attribute LST differences to the changes in SWE.
The manuscript is clear and well presented however I found the analysis and the discussion to be a bit limited. I would like to see additional analysis that looks at changes in SWE according to changes in LC classification. As it stands, the analysis shows that changing the LC information alters the SWE which is an expected result but it stops short of linking the SWE changes to specific changes in LC classification.

For example:
Did all areas that had a change in LC classification experience a change in SWE? If not, which LC changes did not result in a SWE change (i.e. SWE agnostic to those different LCs).

Which types of classification change experience that largest change in SWE?

Were certain LC changes responsible for most of the change in MB?

From a qualitative perspective it appears that moving from bare soil (3) to deciduous broadleaf (9) had an impact on SWE. This should be discussed further.

One suggestion is to quantify the number of pixels that experienced a reclassification and to further categorize those according to type of change (i.e. % of pixels that changed from 13 to 7) and finally to examine the change in SWE according to these LC changes.
I am curious why areas that did not appear to have a change in LC (e.g. Sweden) seem to have more SWE in the ISBA run with the New LC compared to the old LC. This should be examined.
Overall, while the manuscript is well-written and clear I found it lacking in analysis to support the main research objective.

Specific minor comments
Figure 3 and associated text: How did you address the absence of SWE estimates from CCI when producing Figure 3? Were the areas masked in CCI also removed from ISBA and ERA5?
L308-310: Limit is closer to 200mm than 100mm, although SWE exceeding 200mm is common in the CCI SWE product (see Fig. 4.10 [March] in Barella et al. 2024 and also Luojus et al. 2021). See Figure 4.15 in for spatial map of local biases.
Citation: https://doi.org/10.5194/egusphere-2026-65-RC2
- AC2: 'Reply on RC2', Jean-Christophe Calvet, 29 Apr 2026
  
  See the enclosed Supplement.
  
  Citation: https://doi.org/10.5194/egusphere-2026-65-AC2

Peer review completion

AR – Author's response | RR – Referee report | ED – Editor decision | EF – Editorial file upload

ED: Publish subject to revisions (further review by editor and referees) (03 May 2026) by Nora Helbig

AR by Jean-Christophe Calvet on behalf of the Authors (12 Jun 2026) Author's response Author's tracked changes Manuscript

ED: Referee Nomination & Report Request started (12 Jun 2026) by Nora Helbig

RR by Anonymous Referee #2 (25 Jun 2026)

RR by Anonymous Referee #1 (26 Jun 2026)

Suggestions for revision or reasons for rejection

In this revised manuscript, the authors considered my comments from the first round of revisions. The manuscript has substantially improved. Particularly, the additional analysis on the changes in SWE based on the changes in land cover (LC) is welcome. The authors also found a mistake in the filtering procedure for ESA CCI SWE. With this new filtering procedure, the modelled SWE values are closer to the ESA CCI SWE. However, I find that now the improvements in simulated SWE simulated with the new LC are marginal. Indeed, the authors find that the SWE RMSD is reduced by less than 2 mm using the new LC. This is a result itself, but I think the manuscript should be rephrased to emphasis that the change of LC only slightly changes the overall model performance (similar for LST). In addition, I find the SWE RMSD over the domain quite large (values between 33 and 35 mm, Table 1), when the SWE values across the domain are mostly less than 100 mm (Figure S7). I think the overall performance of the model should be better discussed and avenues to improve it in the future suggested. Based on these two comments, I would recommend major revisions of the revised manuscript.

Additional comments:
(The manuscript lines correspond to those of the track change manuscript)
1. No results on the performance of the model and the change of LC to simulate LST are discussed in the abstract. This should be added.
2. In section 2.1, I think it would be important to state here that the LST corresponds to the radiative surface temperature. It is stated later, but I think it could be introduced earlier.
3. The authors now specify that a linear interpolation is used to upscale values to the 0.25o grid. I am not sure that a linear interpolation is the most appropriate interpolation. Can this be discussed?
4. I would suggest moving section 2.4 before the current section 2.3 (“ERA5 SWE and LST data”)
5. I think a deeper discussion or introduction on the differences between the LC maps is needed. Does the new LC correspond to changes of vegetation type over time or is it more accurate than the previous one due to improved algorithms/satellite data?
6. Figure 1c, could it be specified in this caption that the grid cells where ESA CCI SWE are available correspond to the blue ones?
7. L. 185-188: These sentences stating different periods are confusing. Why are simulations repeated four times over the 1981-1989 and 2010-2019 periods when the study period is 2010-2022?
8. As before, I think the quality of the graphs should be improved. I still find them a bit blurry in the PDF. Maybe increasing the dpi to 300 could help. In addition, the texts (labels, titles), could be bigger, as well as the ticks (degree values). In figures showing the time series (e.g. Fig. 3, Fig.8), the reference data ESA CCI could be put in black to highlight that it’s the reference data.
9. Figure 7: consider doing a 2x2 plot
10. L.324-325: “Over 90 % of the domain has RMSD values below 40 mm, indicating generally low errors in snow estimation.” I think the threshold of 40 mm is quite high as the snowpack is generally under 100 mm (Figure S7). Could the normalized RMSD be looked at?
11. L. 351: ‘DBF’, ‘CNF’: consider writing them fully.
12. Figure 9: it is not clear to me if these maps are only for the winter seasons or over the whole years? I am wondering as this figure is used to relate the LST errors with the modelled SWE bias.

Hide

ED: Publish subject to revisions (further review by editor and referees) (28 Jun 2026) by Nora Helbig

AR by Jean-Christophe Calvet on behalf of the Authors (07 Jul 2026) Author's response Author's tracked changes Manuscript

ED: Referee Nomination & Report Request started (08 Jul 2026) by Nora Helbig

RR by Anonymous Referee #1 (12 Jul 2026)

ED: Publish as is (13 Jul 2026) by Nora Helbig

AR by Jean-Christophe Calvet on behalf of the Authors (15 Jul 2026) Manuscript

Journal article(s) based on this preprint

24 Jul 2026

Assessing the effect of land cover on ISBA snow water equivalent and land surface temperature simulations over Europe

Oscar Rojas-Munoz, Constantin Ardilouze, Bertrand Bonan, Diane Tzanos, Darren Ghent, Céline Lamarche, Thomas Nagler, and Jean-Christophe Calvet

The Cryosphere, 20, 4099–4115, https://doi.org/10.5194/tc-20-4099-2026,https://doi.org/10.5194/tc-20-4099-2026, 2026

Short summary

Oscar Rojas-Munoz, Constantin Ardilouze, Bertrand Bonan, Diane Tzanos, Darren Ghent, Céline Lamarche, Thomas Nagler, and Jean-Christophe Calvet

Supplement

https://doi.org/10.5194/egusphere-2026-65-supplement

Data sets

ESA CCI Global Land Cover Maps, Version 2.0.7 P. Defourny et al. https://catalogue.ceda.ac.uk/uuid/b382ebe6679d44b8b0e68ea4ef4b701c/

ESA CCI Aqua MODIS LST data, Version 4 D. Ghent et al. https://doi.org/10.5285/d56a6215ce394ddd8dff6bea5dbb0780

ESA CCI snow water equivalent, Version 3.1 K. Luojus et al. https://doi.org/10.5285/9d9bfc488ec54b1297eca2c9662f9c81

Model code and software

SURFEX, Version 9 CNRM https://www.umr-cnrm.fr/surfex/data/OPEN-SURFEX/open_surfex_v9_0_0_20231024.tar.gz

Oscar Rojas-Munoz, Constantin Ardilouze, Bertrand Bonan, Diane Tzanos, Darren Ghent, Céline Lamarche, Thomas Nagler, and Jean-Christophe Calvet

Viewed

Total article views: 1,867 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
1,141	601	125	1,867	223	108	172

HTML: 1,141
PDF: 601
XML: 125
Total: 1,867
Supplement: 223
BibTeX: 108
EndNote: 172

Views and downloads (calculated since 26 Jan 2026)

Month	HTML	PDF	XML	Total
Jan 2026	325	165	35	525
Feb 2026	309	188	55	552
Mar 2026	341	173	19	533
Apr 2026	80	32	5	117
May 2026	40	21	5	66
Jun 2026	13	10	3	26
Jul 2026	32	9	3	44
Aug 2026	1	3	0	4

Cumulative views and downloads (calculated since 26 Jan 2026)

Month	HTML	PDF	XML	Total
Jan 2026	325	165	35	525
Feb 2026	309	188	55	552
Mar 2026	341	173	19	533
Apr 2026	80	32	5	117
May 2026	40	21	5	66
Jun 2026	13	10	3	26
Jul 2026	32	9	3	44
Aug 2026	1	3	0	4

Viewed (geographical distribution)

Total article views: 1,860 (including HTML, PDF, and XML) Thereof 1,860 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 03 Aug 2026

Download

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

Preprint (1992 KB)
Metadata XML

Short summary

The impact of land cover representation on snow simulations within the ISBA land surface model in Europe between 2010 and 2022 is assessed. The ECMWF ERA5 atmospheric forcing dataset is used to conduct offline simulations with and without new land cover data. Using the new data reduces the ISBA snow water equivalent bias by around 33 %. These results highlight the importance of accurate land cover data for improving snow representation in land surface models.


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