The influence of soil moisture on the heavy precipitation event in July 2021 in Western Europe

Fohrmann, Till; Szemkus, Svenja; Heuser, Oliver; Valmassoi, Arianna; Friederichs, Petra

doi:10.5194/egusphere-2026-2424

Preprints

https://doi.org/10.5194/egusphere-2026-2424

Preprints

20 May 2026

| 20 May 2026

The influence of soil moisture on the heavy precipitation event in July 2021 in Western Europe

Till Fohrmann, Svenja Szemkus, Oliver Heuser, Arianna Valmassoi, and Petra Friederichs

Abstract. Soil moisture-precipitation feedback is an important factor in the water and energy cycles. But how important is it on the time scale of an atmospheric extreme precipitation event? We are investigating this question using the example of heavy precipitation in July 2021, which led to destructive flash floods in Western Europe. To quantify the importance of land-atmosphere coupling and continental moisture sources for the precipitation, we perform numerical simulations with wet, dry and normal soil moisture conditions over Europe. Ensembles of simulations are performed using a global set up of the ICON numerical weather prediction model with a grid refinement over Europe. To account for both the limited predictability and the delayed response of the atmosphere to changes in soil moisture, we use data assimilation to steer the system’s development toward the extreme event, but only to the extent necessary so that our interventions in the soil are not undone. We find that the moisture supply of the event crucially depends on continental moisture sources. This result is further confirmed using moisture tracking. Conversely, increased soil moisture only leads to slight precipitation increases, since surface moisture fluxes are energy-limited. Moisture is also important for the development of the near surface low pressure system, which had a central role in the event. Our ensemble simulations also show that there is potential for more devastating events, i.e., more precipitation.

Received: 27 Apr 2026 – Discussion started: 20 May 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

Till Fohrmann, Svenja Szemkus, Oliver Heuser, Arianna Valmassoi, and Petra Friederichs

Status: final response (author comments only)

RC1:
'Comment on egusphere-2026-2424', Vikki Thompson, 19 Jun 2026
This study investigates the importance of soil moisture on the western European rainfall (July 2021). Simulations of the event with altered soil moisture over Europe are undertaken, and the results analysed. The processes are investigated, and evidence presented that soil moisture influenced the event – with greater soil moisture leading to greater rainfall magnitudes. The authors identify that the moisture levels play a role in the development of the low pressure system. The manuscript is clearly written, relevant, and a useful contribution. However, I feel a little more analysis of the dynamical aspects would greatly strengthen the conclusions - and some figures could be improved.

Major comments:

Role of moisture on low pressure system: I think this could be investigated more thoroughly. I am concerned that Fig.5 is misleading. I took the main message as the differences between the CTL, WET, and DRY ensembles – but this is not what is most obvious in the way the figure is presented. I am not sure the linear fits are useful – as a majority are not significant fits. Perhaps instead showing the difference in the spatial patterns of SLP and Z500 (in a similar way to Fig. 3) would be more useful. I would also be interested in seeing the difference in vertical velocity between the ensemble. In Line 213 processes are hypothesised – this could be tested more quantitively. Further assessment of the differences between SLP, Z500, and vertical velocity in the three ensembles could strengthen the conclusions.

A statement about the wider implications of the findings of this study in relation to our current changing climate would be useful. As summer soil moisture over Europe is projected to decrease, this study would suggest that we should see less extreme rainfall - contrary to much other literature. Comments about what the results mean for society / policy makers, and greater explanation of outstanding research questions, would be beneficial. This could fit both within the conclusions, and to better motivate the study within the introduction.

Minor comments:

Consider the use of ‘rainfall’ rather than precipitation throughout, as in this case study it is only rainfall, in situations of snow the conditions/processes may differ.

Abstract – line 14: ‘more precipitation’, this could be more accurately stated – greater magnitude (as opposed to greater area / greater timescale).

Introduction – line 20/23: Tradowsky et al (2023) and Kreienkamp et al (2021) are essentially the same study – just one is the rapid WWA report and one a slower, extended, study. In the text it makes it appear as if the two are more independent than they truly are.

Introduction line 37 - Relevant reference (apologies for self-citation): Thompson et al 2026 The role of subtle dynamical changes in the European extreme rainfall of July 2021 | Climate Dynamics | - we included moisture tracking using LAGRANTO. This study includes analysis of the case study with several different model set-ups (ensemble boosting, PGW, ECMWF forecasts) so may be relevant elsewhere too – but we did not assess the role of soil moisture itself.

Fig.2 caption: it’s should be ‘its’.

Fig.2 It took me a moment to realise 1.7 etc. Meant 1 Jul, perhaps label as ‘date’? Or re-label as ‘number of days before event’

Line 146 : its source region (no apostrophe)

Line 168 – As the ensemble is initialised from a specific state it will not fully incorporate the plausible range of natural variability. I do agree that including multiple ensemble members is very useful, but not for the reason outlined here – more because very small differences in the dynamics could cause larger differences in precip.

Line 199: grammar (colon followed by capital)
Citation: https://doi.org/10.5194/egusphere-2026-2424-RC1
RC2:
'Comment on egusphere-2026-2424', Anonymous Referee #2, 17 Jul 2026
This paper discusses the influence of soil moisture on the rainfall on the mid July 2021 floods in Europe. The authors find that very dry soils substantially decreases the amount of rain, and tracked the source of moisture back to the European continent. The paper is (reasonably) well written, well understandable and provides useful information. I also liked the results about the dynamical adjustment to soil moisture. Yet, the statistics presented are sometimes rather abstract (examples below) and I think the paper could improve substantially when these abstract quantities are expressed in terms of more usual or useful statistics. Also I found that in some places the physical interpretation could be improved. Finally, while I liked the setup of the experiments, I thought that the soil moisture perturbations were quite crude, and it would be perhaps better to do a moderate soil moisture perturbation (or being more clear that these are large perturbations). I think the paper can be published with some relatively modest adjustments and clarifications.
Specific examples of statistics:
While I understand the use of the total accumulated rain over both time and space (giving a number order 10^13) -- since the aim is to track the total amount of water to it's source -- it is also quite difficult to understand and since I could not find the area size, I was also unable to estimate how much rainfall that is in mm units. Despite that observations are uncertain, I still think it is good to give an observational estimate.

I could not interpreted the soil moisture values in Fig 4d. A value of 1 kg/m2 corresponds to 1 mm of water. With typical volumetric values of soil moisture capacity (10-40%), this means you are looking a 1 cm of soil at most. Why not plot a soil wetness index of the first 1m of the soil, which is a typical parameter used in many studies.

I could not figure out / understand the units of evaporation in Figure 4. I would expect the number to be much larger since if I multiply this by time I do not get the accumulated area summed rainfall amounts (order 10*13).

Specific examples of physics/physical interpretation:
I am unsure about the interpretation of the wet experiment. Setting soil moisture to saturation (in the whole root zone?) is in typical soil schemes not so different than setting it to the field capacity. In both cases, there is no stress from the plants, so you are in a energy-limited regime. The question is how far is the control experiment from field capacity, since if it is close than you would not expect a large increase in evaporation from the wet experiment. So, as stated above a soil wetness index (normalised soil moisture between field capacity and wilting point) could be more useful.

While I liked the moisture tracking in the paper, as an outsider to this field, I am also a bit confused on how it is actually done and what it means as explained in section 2.3. What are the main assumptions there? Somehow, I do not understand how you get from rainfall accumulation (something like 100 mm) to tagged moisture (a portion of the actual moisture?), to tagged moisture transport, and evaporation from the source area. What do you assume here ? I also tried to look at the reference to method, but that did not help me very much without going to the details of that paper. So, if you could provide a bit more informatation that may guide the reader who is not familiar with the method. I guess that the procedure assumes the same dynamics (a suggested on line 114), meaning that subtle changes in dynamics could strong alter the result. In that sense, the moisture leading to precipitation should come from somewhere, but if you remove that source of moisture the atmosphere will subtle adjust to get the moisture from a different area (e.g. because of dryer air leading to more evaporation elsewhere or subtle dynamical adjustment). In that sense I liked the comparison between the moisture tracking showing that the EU continent is a main source, and the model experiment showing indeed large response of rainfall to dryer soil conditions.

I think it would be benificial to also show a few more normal statistics, like evaporation, for instance by showing a map of evaporation alongside Figure 6 (or in the supplement). Figures of sensible heat flux and temperature would also help to understand the results better as well as perhaps humidity. A 5% drop in relative humidity and a 2 degree warming would likely be associated to increases in absolute humidity (not decrease), so I was also a bit unsure here on how the results looked.

line 253-257. see also first remark. Given that you do not know how far the control spoil moisture is from field capacity it is hard to make an interpretation here.

Minor remarks:
Table 1. The table is really detailed (perhaps too), but hardly discussed in the text, and is basically very similar to the integral of Figure 6 over de different PRUDENCE areas (except that is percentage of the total). Do the areas include both land and sea ?

Figure 3. I am not a particular fan of the colour coding here as it is so hard to see differences in the order of a factor 2. And difference plots mainly show a spatial shift for the wet versus control. As I understand the box area is adjusted to compensate for a small shift as stated on line 141, but perhaps good to remind the reader here because it looks like that in the wet scenario the peak moves outside the box area.

Figure 4. I think it is quite interesting that the event basically comes into two burst on the 13'th and 14'th of July, and that in the second burst the WET simulation produces quite a bit more precipitation than the CTL (and DRY). Is that explainable in term of tagged moisture or perhaps just the normal evaporation, or can you difference in the dynamics (pressure) on the 14th? In this respect it is also good to remark the value of the ensemble. Some of the members are actually closer the observation, but the mean is rather low, which corresponds to the large sensitvity of system to relatively small pertrbations as other studies have found.

L 262. I think the logic here is that it is energy limited since increasing the soil moisture has no effect. If it is already in partly soil moisture limited regime, removing more soil moisture will further decrease evaporation.

L 35. Related also to the discussion on the method. The suggestion that moisture comes from the American continent is hard to physically understand, and makes you doubt these methods (although you show later that this is not the case and results look perfectly fine). Also I do not understand the connection between discussing the role of American forrest, and finding no increase in evaporation in Europe.
Citation: https://doi.org/10.5194/egusphere-2026-2424-RC2

Till Fohrmann, Svenja Szemkus, Oliver Heuser, Arianna Valmassoi, and Petra Friederichs

Data sets

Simulation output Till Fohrmann https://doi.org/10.5281/zenodo.19818501

Interactive computing environment

Visualization code Till Fohrmann https://github.com/tfohrmann/july21_eval

Till Fohrmann, Svenja Szemkus, Oliver Heuser, Arianna Valmassoi, and Petra Friederichs

Viewed

Total article views: 287 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
197	72	18	287	18	15

HTML: 197
PDF: 72
XML: 18
Total: 287
BibTeX: 18
EndNote: 15

Views and downloads (calculated since 20 May 2026)

Month	HTML	PDF	XML	Total
May 2026	118	34	10	162
Jun 2026	18	10	3	31
Jul 2026	58	20	5	83
Aug 2026	3	8	0	11

Cumulative views and downloads (calculated since 20 May 2026)

Month	HTML	PDF	XML	Total
May 2026	118	34	10	162
Jun 2026	18	10	3	31
Jul 2026	58	20	5	83
Aug 2026	3	8	0	11

Viewed (geographical distribution)

Total article views: 265 (including HTML, PDF, and XML) Thereof 265 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 15 Aug 2026

Short summary

The heavy rain over Western Europe in mid July 2021 caused destructive flash floods. Previous research shows that soil moisture played an important role in amplifying the impacts. But how important is soil moisture as a water source for the rain itself? We investigated this question by simulating scenarios with especially wet and dry European soils. We find that soils were integral in supplying water to the event and in strengthening the low pressure system “Bernd” important to the event.


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