the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 27 Jul 2023
A new drought index fitted to clay shrinkage induced subsidence over France: benefits of interactive leaf area index
Abstract. Clay shrinkage, which consists of a reduction in the volume of clay soils during dry periods, can affect buildings and cause subsidence damage. In France, losses due to subsidence are estimated at more than 16 billion € for the period 1989–2021 (CCR, 2021), and are expected to increase under the effect of climate warming. This work aims to improve the current understanding of the conditions triggering subsidence by proposing an innovative drought index. We use a daily Soil Wetness Index (SWI) to develop a new annual drought index that can be related to subsidence damage. The SWI is derived from simulations of soil moisture profiles from the Interactions between Soil, Biosphere, Atmosphere (ISBA) land surface model developed by Météo-France. The ability of the drought index to correlate with insurance claims data is assessed by calculating the Kendall rank correlation over twenty municipalities in France. The insurance data, aggregated by year and municipality, are provided by the Caisse Centrale de Réassurance (CCR). A total of 1200 configurations of the drought index are considered. They are generated by combining different calculation methods, ISBA simulation settings, soil model layers, and drought percentile thresholds. The analysis includes a comparison with the independent claim data of six additional municipalities, and to a record of official “CatNat” decrees, useful for the analysis. The best results are obtained for drought magnitudes based on SWI values of the 0.8 m to 1.0 m deep soil layer, an ISBA simulation with interactive leaf area index (LAI), and consideration of low drought SWI percentile thresholds. Comparison with claim data shows that drought magnitude is able to identify subsidence events while being spatially consistent. This drought magnitude index provides more insight into subsidence triggers while benefiting from advanced land surface modeling schemes (interactive LAI, multi-layer soil). This work paves the way for more reliable damage estimates.
-
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
(2079 KB)
-
Supplement
(377 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(2079 KB) - Metadata XML
-
Supplement
(377 KB) - BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2023-1366', Anonymous Referee #1, 18 Sep 2023
The paper is very interesting, the topic is important, and the methodology considered is appropriate.
My main concern is about the data used here, partially described in section 2.2. As explained in https://doi.org/10.5194/nhess-22-2401-2022 (Charpentier, James and Ali 2022) on a similar topic, the French system has a very specific design, where claims within a “town” (or “commune” or “municipality”) need first a national recognition before beeing accepted as "legitimate claims" (and then paid by the insurance company). In Charpentier, James and Ali (2022), it is observed that models are good to predict town that will claim losses, but the national recognition stage is much harder. Which data are used in this study ? Those obtained initially, from towns claiming losses, or those obtained after censoring, by national recognition ? In the first case, the paper is ok, and could be published. Otherwise, there is a major selection bias in the study that should, somehow, be considered.
Citation: https://doi.org/10.5194/egusphere-2023-1366-RC1 - AC1: 'Reply on RC1', Jean-Christophe Calvet, 13 Dec 2023
-
RC2: 'Comment on egusphere-2023-1366', Anonymous Referee #2, 13 Oct 2023
The paper proposes a model based drought index that can be used to quantify subsidence hazards across a number of regions in France. I have the following comments
1. Check spelling but in line 139
2. Section 4.2. lines 332 the naming conventions are very confusing - can't you just refer to depths? For example why do you refer to the surface to 20cm as LAI and deeper as SW?
3. Do you mean December? Year - line 333 Section 4.2.
4. Section 4.2 lines 336 - 338 support with evidence from the literature
5. In Section 4.2.2 line 423 you mention that household claims are the only available evidence of subsidence. Might you consider other sources such as InSAR which should work at the scale of postcode...we use this to monitor, for example, subsidence from mining operations.
6. Section 4.4.4. This is a valid and important observation i.e. the claim may be made years after the problem started to occur. To put it slightly differently, the damage may the result of a cumulation of years of movement (shrink swell) in the soil or it may be the result of a once off event. perhaps you can support this discussion point with some further references supporting your choice of one year timescale OR giving us a better idea of what the uncertainty may look like.
7. Section 4.4.5. lines 449 - 454 - Can you suggest how this problem might be overcome?
8. Overall comment: This may be a slightly naive question, but would it be possible to validate the model by comparing to locations where you have subsidence data - or even cross reference with InSar data? You are basically using the claims data as a proxy for subsidence, as pointed out earlier, there may be other sources of data both point and remote sensing data that is publically available, that can be used.
Citation: https://doi.org/10.5194/egusphere-2023-1366-RC2 - AC2: 'Reply on RC2', Jean-Christophe Calvet, 13 Dec 2023
-
RC3: 'Comment on egusphere-2023-1366', Anonymous Referee #3, 26 Oct 2023
Comments to the Author:
The paper deals with an interesting numerical approach to calculate a drought index fitted to clay shrinkage induced subsidence over France.
The reviewer asks the authors to give first more attention to the following general remarks in order to correct them:
- I’m not really convicted that we can talk about a “new drought index” in this paper and choose it as a title! This can be misleading. The reuse of existing parameter SWI derived from ISBA simulation method, with an extended vegetation representation, is in my opinion not enough to name this parameter new drought index. I suggest to the authors to modify the title according to a “new approach” or an “adaptation” of an existing parameter.
- Add a legend for Figure 5 and specify the correspondence of each color bar.
Questions and remarks to be commented:
R1: page 4 and lines 100-101, moisture variations depend also on the mineralogy of clays and their saturated and unsaturated hydraulic conductivity, the initial soil suction and how water flows depending on its hydromechanical properties. Can the authors give more details on the choice of not taking soil parameters and behaviour into account in this study?
R2: page 4 and lines 102-103, in the ISBA model, it is considered that texture is homogeneous and represented by some clay, sand and silt contents. This cannot reflect the reality when we know the heterogeneity of clayey soils in France, at the kilometre resolution and including at the same plot of the house. Thus, calculations made and improved based on the ISBA model and derived versions is a tool to have an idea to estimate the top surface soil moisture but it is still complex to deduce any real state of hydromechanical behaviour of clayey soils without considering their mineralogy, heterogeneity and hydromechanical properties such as soil water characteristic curve (SWCC).
R3: page 4 and lines 111-113, analysis of this study were based on four model layers until 1.0 m depth. One of the direct consequences of climate change is the propagation of soil desiccation in depth under severe and recurrent drought. This can reach 3.0 m depth and more depending on the close environment configuration. It would be interesting if the authors try to take into account this climate change effect through new calculations.
R4: page 5 and line 127, what do the authors mean by “volumetric soil moisture”? Is it possible to explain how simulation can provide this physical property of the soil?
R5: page 5 and lines 129-131, can the authors clarify better the “conversion” of the volumetric soil moisture to soil wetness indices (SWI) to justify considering a single definition of drought?
R6: page 6 and lines 179-181, it appears that this study is mainly based on SWI outputs of the ISBA model. I’m not convicted that these calculations are the most reliable tools for studying soil moisture variations as mentioned.
R7: page 6 and lines 189-190, I’m not sure that it is possible to assume that results based on in situ observations in the USA and Canada can be applicable to France especially under climate change context. Many assumptions are considered in this study, which show the complexity to approach the soil water content and its variations without taking into account its hydromechanical properties at a given initial state.
Citation: https://doi.org/10.5194/egusphere-2023-1366-RC3 - AC3: 'Reply on RC3', Jean-Christophe Calvet, 13 Dec 2023
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2023-1366', Anonymous Referee #1, 18 Sep 2023
The paper is very interesting, the topic is important, and the methodology considered is appropriate.
My main concern is about the data used here, partially described in section 2.2. As explained in https://doi.org/10.5194/nhess-22-2401-2022 (Charpentier, James and Ali 2022) on a similar topic, the French system has a very specific design, where claims within a “town” (or “commune” or “municipality”) need first a national recognition before beeing accepted as "legitimate claims" (and then paid by the insurance company). In Charpentier, James and Ali (2022), it is observed that models are good to predict town that will claim losses, but the national recognition stage is much harder. Which data are used in this study ? Those obtained initially, from towns claiming losses, or those obtained after censoring, by national recognition ? In the first case, the paper is ok, and could be published. Otherwise, there is a major selection bias in the study that should, somehow, be considered.
Citation: https://doi.org/10.5194/egusphere-2023-1366-RC1 - AC1: 'Reply on RC1', Jean-Christophe Calvet, 13 Dec 2023
-
RC2: 'Comment on egusphere-2023-1366', Anonymous Referee #2, 13 Oct 2023
The paper proposes a model based drought index that can be used to quantify subsidence hazards across a number of regions in France. I have the following comments
1. Check spelling but in line 139
2. Section 4.2. lines 332 the naming conventions are very confusing - can't you just refer to depths? For example why do you refer to the surface to 20cm as LAI and deeper as SW?
3. Do you mean December? Year - line 333 Section 4.2.
4. Section 4.2 lines 336 - 338 support with evidence from the literature
5. In Section 4.2.2 line 423 you mention that household claims are the only available evidence of subsidence. Might you consider other sources such as InSAR which should work at the scale of postcode...we use this to monitor, for example, subsidence from mining operations.
6. Section 4.4.4. This is a valid and important observation i.e. the claim may be made years after the problem started to occur. To put it slightly differently, the damage may the result of a cumulation of years of movement (shrink swell) in the soil or it may be the result of a once off event. perhaps you can support this discussion point with some further references supporting your choice of one year timescale OR giving us a better idea of what the uncertainty may look like.
7. Section 4.4.5. lines 449 - 454 - Can you suggest how this problem might be overcome?
8. Overall comment: This may be a slightly naive question, but would it be possible to validate the model by comparing to locations where you have subsidence data - or even cross reference with InSar data? You are basically using the claims data as a proxy for subsidence, as pointed out earlier, there may be other sources of data both point and remote sensing data that is publically available, that can be used.
Citation: https://doi.org/10.5194/egusphere-2023-1366-RC2 - AC2: 'Reply on RC2', Jean-Christophe Calvet, 13 Dec 2023
-
RC3: 'Comment on egusphere-2023-1366', Anonymous Referee #3, 26 Oct 2023
Comments to the Author:
The paper deals with an interesting numerical approach to calculate a drought index fitted to clay shrinkage induced subsidence over France.
The reviewer asks the authors to give first more attention to the following general remarks in order to correct them:
- I’m not really convicted that we can talk about a “new drought index” in this paper and choose it as a title! This can be misleading. The reuse of existing parameter SWI derived from ISBA simulation method, with an extended vegetation representation, is in my opinion not enough to name this parameter new drought index. I suggest to the authors to modify the title according to a “new approach” or an “adaptation” of an existing parameter.
- Add a legend for Figure 5 and specify the correspondence of each color bar.
Questions and remarks to be commented:
R1: page 4 and lines 100-101, moisture variations depend also on the mineralogy of clays and their saturated and unsaturated hydraulic conductivity, the initial soil suction and how water flows depending on its hydromechanical properties. Can the authors give more details on the choice of not taking soil parameters and behaviour into account in this study?
R2: page 4 and lines 102-103, in the ISBA model, it is considered that texture is homogeneous and represented by some clay, sand and silt contents. This cannot reflect the reality when we know the heterogeneity of clayey soils in France, at the kilometre resolution and including at the same plot of the house. Thus, calculations made and improved based on the ISBA model and derived versions is a tool to have an idea to estimate the top surface soil moisture but it is still complex to deduce any real state of hydromechanical behaviour of clayey soils without considering their mineralogy, heterogeneity and hydromechanical properties such as soil water characteristic curve (SWCC).
R3: page 4 and lines 111-113, analysis of this study were based on four model layers until 1.0 m depth. One of the direct consequences of climate change is the propagation of soil desiccation in depth under severe and recurrent drought. This can reach 3.0 m depth and more depending on the close environment configuration. It would be interesting if the authors try to take into account this climate change effect through new calculations.
R4: page 5 and line 127, what do the authors mean by “volumetric soil moisture”? Is it possible to explain how simulation can provide this physical property of the soil?
R5: page 5 and lines 129-131, can the authors clarify better the “conversion” of the volumetric soil moisture to soil wetness indices (SWI) to justify considering a single definition of drought?
R6: page 6 and lines 179-181, it appears that this study is mainly based on SWI outputs of the ISBA model. I’m not convicted that these calculations are the most reliable tools for studying soil moisture variations as mentioned.
R7: page 6 and lines 189-190, I’m not sure that it is possible to assume that results based on in situ observations in the USA and Canada can be applicable to France especially under climate change context. Many assumptions are considered in this study, which show the complexity to approach the soil water content and its variations without taking into account its hydromechanical properties at a given initial state.
Citation: https://doi.org/10.5194/egusphere-2023-1366-RC3 - AC3: 'Reply on RC3', Jean-Christophe Calvet, 13 Dec 2023
Peer review completion
Journal article(s) based on this preprint
Data sets
Exposure of detached houses to clay shrinkage over France by municipality MTES https://www.statistiques.developpement-durable.gouv.fr/nouveau-zonage-dexposition-au-retrait-gonflement-des-argiles-plus-de-104-millions-de-maisons
Model code and software
SURFEX modelling platform V. Masson, P. Le Moigne, E. Martin, S. Faroux, A. Alias, R. Alkama, S. Belamari, A. Barbu, A. Boone, F. Bouyssel, P. Brousseau, E. Brun, J.-C. Calvet, D. Carrer, B. Decharme, C. Delire, S. Donier, K. Essaouini, A.-L. Gibelin, H. Giordani, F. Habets, M. Jidane, G. Kerdraon, E. Kourzeneva, M. Lafaysse, S. Lafont, C. Lebeaupin Brossier, A. Lemonsu, J.-F. Mahfouf, P. Marguinaud, M. Mokhtari, S. Morin, G. Pigeon, R. Salgado, Y. Seity, F. Taillefer, G. Tanguy, P. Tulet, B. Vincendon, V. Vionnet, and A. Voldoire https://www.umr-cnrm.fr/surfex/
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 508 | 239 | 30 | 777 | 57 | 13 | 15 |
- HTML: 508
- PDF: 239
- XML: 30
- Total: 777
- Supplement: 57
- BibTeX: 13
- EndNote: 15
Viewed (geographical distribution)
| Country | # | Views | % |
|---|---|---|---|
| France | 1 | 228 | 30 |
| United States of America | 2 | 184 | 24 |
| Germany | 3 | 49 | 6 |
| Netherlands | 4 | 36 | 4 |
| United Kingdom | 5 | 32 | 4 |
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
- 228
Cited
Sophie Barthélémy
Bertrand Bonan
Jean-Christophe Calvet
Gilles Grandjean
David Moncoulon
Dorothée Kapsambelis
Séverine Bernardie
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(2079 KB) - Metadata XML
-
Supplement
(377 KB) - BibTeX
- EndNote
- Final revised paper