the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 22 Feb 2024
Generalized drought index: A novel multi-scale daily approach for drought assessment
Abstract. Drought is a complex climatic phenomenon characterized by water scarcity recognized as the most widespread and insidious natural hazard, posing significant challenges to ecosystems and human society. In this study, we propose a new daily based index for characterizing droughts, which involves standardizing precipitation and/or precipitation minus potential evapotranspiration data. The performance of this new index is assessed with data from the evaluation runs of the Coordinated Regional Climate Downscaling Experiment for the European domain and the observational data from the Iberian Gridded Dataset, covering the period from 1989 to 2009. Comparative assessments are conducted against the daily Standardized Precipitation Index (SPI), the Standardized Precipitation Evapotranspiration Index (SPEI), and a simple Z-Score standardization of climatic variables. Seven different accumulation periods are considered (7, 15, 30, 90, 180, 360, and 720 days) with three drought levels: moderate, severe, and extreme. The evaluation focuses mainly on the direct comparison amongst indices, added value assessment using the Distribution Added Value and a simple bias difference for drought characteristics. Results reveal that not only does the new index allow the characterization of flash droughts, but also demonstrates added value when compared to SPI and SPEI, especially for longer accumulation periods. In comparison to the Z-Score, the new index shows slightly greater gains, particularly for extreme drought events at lower accumulation periods. Furthermore, an assessment of the spatial extent of drought for the 2004–2005 event is performed using the observational dataset. All three indices generally provide similar representations, except for the Z-Score, which exhibits limitations in capturing extreme drought events at lower accumulation periods. Overall, the findings suggest that the new index offers improved performance and adds value comparatively to similar indices with a daily time step.
- Preprint
(3404 KB) - Metadata XML
-
Supplement
(3570 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on gmd-2024-9', Anonymous Referee #1, 25 Mar 2024
General Comments:
A new daily based drought index is proposed in this study, estimated over the Iberian Peninsula and evaluated his added value with respecto to otgher existing drought index. To this aim, an observational gridded datataset and regional models have been used. The paper is well written and the results are well explained, including figures and tables. Considering the region of analysis, and its expected challenges in a climate change context, the work is very relevant, timely and important to improve the drought monitorization over the Iberian Peninsula. The main critical point I have identified could be the use of another observational dataset as reference to better define the frmaework. Based on these points wy recommendation is accepted with minor changes or, if the new observational reference is added, major changes.
Comments:
- Why has the added value evaluated with the Evaluation experiment of CORDEX? I understand that this analysis is relevant in order to apply this index in a context of climate change but I am not sure that the added value justifies the increment of the paper and analysis.
- Are 20 years enough to make a proper normalization?
- Lines 47-50: However is a proper link between these two sentences?
“… Nevertheless, the consideration of just precipitation by SPI could be a limiting factor depending on the climatic dominating conditions in certain regions. However, with anthropogenic climate change, rising temperatures and the subsequent increases in evapotranspiration can also significantly increase the impact of drought events (Hu and Wilson, 2000; Vicente-Serrano et al., 2010)….”
- Line 77: Something is missing here? “.. and is also considered in the by the Reconnaissance Drought Index (Tsakiris and Vangelis, 2004).”
- There is a global SPI and SPEI gridded observational dataset publicly distributed (Vicente-Serrano et al 2010a and 2010b), that could be used as reference and/or to analyze the obvservational uncertainty of the results. However, the authors decided to use the evaluation experiment of CORDEX, is there any reason? I didn’t see it (probably I missed the section in which the authors explain it) in the manuscript and I am not sure what is the advantage when a new index is evaluated.
Vicente-Serrano S.M., Beguería S., López-Moreno J.I., 2010 (a): A Multi-scalar drought index sensitive to global warming: The Standardized Precipitation Evapotranspiration Index – SPEI. Journal of Climate 23(7), 1696-1718, DOI: 10.1175/2009JCLI2909.1.-- http://digital.csic.es/handle/10261/22405.
Vicente-Serrano S.M., Beguería S., López-Moreno J.I., Angulo M., El Kenawy A. (2010(b)) A global 0.5° gridded dataset (1901-2006) of a multiscalar drought index considering the joint effects of precipitation and temperature. Journal of Hydrometeorology 11(4), 1033-1043, DOI: 10.1175/2010JHM1224.1.-- http://digital.csic.es/handle/10261/23906
- Line 187: Which is the final period? 1989-2009 or 1989-2008
- Lines 185-189: EuroCORDEX also contains future projections with different scenarios.
- Line 190: The evaluation experiment is partially synchronized as it inherit the temporal correspondence of the reanalysis.
- Lines 224-225: The Iberian Peninsula presents some regions with arid and semiarid climates so this sentence could be modified accordingly: instead “Therefore, in the context of climate change, this equation is not the best option for computing PET (Beguería et al., 2014).” the authors could write something like “Therefore, in the context of climate change and the Iberian Peninsula, with arid and semiarid regions, this equation is not the best option for computing PET (Beguería et al., 2014).” which is more explicit and reflects the particular problem of the Iberian Peninsula.
- Subsection 2.4. GDI evaluation: If the GDI is obtained with an histogram estimation, could the evaluation based on histograms bias the results in favour to this index?
- How are the possitive percentages of DAV translated to added value with respect to the corresponding index?
- Table 2 is a figure? Why had the autors not used an standar table?
Citation: https://doi.org/10.5194/gmd-2024-9-RC1 -
RC2: 'Comment on gmd-2024-9', Anonymous Referee #2, 01 Apr 2024
- Abstract: The region of study (Iberian Peninsula) is not very clearly stated. In fact, as EuroCORDEX is mentioned, it could be understood that the whole Europe is the area of analysis
- Abstract: Some kind of numerical cuantification of how the proposed GDI improves previous indices could be indicated, as a summary of the discusion and conclusions section, although there not many numbers are shown, but if the authors can give some more specific quantification of such improvement, it could improve abstract paragraph. The fact that daily scales are used is also quite relevant, but it is just stated on the last phrase there. Perhaps it could be stated before and more clearly.
- Introduction, lines 91-95: More clear statements about the differences between model and observed data should be made, as they are quite different, and there they seem to be two similar types of climate information. A comment that both are gridded data, for example, should be more clearly stated, and the need of having daily resolution. When talking about observational gridded data, perhaps a comment about other gridded daily precipitation databases could be made (E-OBS, GPCP or CPC ones)?.
- Introduction/objectives, around line 126: Why IB region is chosen, some arguments could be made to justify such region, which, of course is a good one to analyze drought features. Also state at that point which time period is proposed would be fine there, although it is detailed later on methods. Nevertheless, one of my main concerns, partly indicated before, is about how Regional climate modelling EuroCORDEX material is presented. No doubt about its importance and relevance to be used for climate studies such as the one presented here. But it is confusing to me the way they are employed. Here it is a paper about a new drought index proposal, which is welcomed and quite interesting, and not an evaluation of RCMs when dealing with this index proposal. The first part about how it compares with other indices when using observations is fine. So RCMs are used not to see how they represent each index, but to obtain a potential "added value'' of this index compared with the others in terms of how models compare with observations using the Perkins skill score. In that sense, is strange not to present how EuroCORDEX models describe each drought index. At least a more detailed description of how precipitation is described from them could be of interest, as just some very brief remarks are made around lines 205-210. I would be more clear and precise when talking about "evaluation'' RCM simulations (that is, reanalysis or perfect boundary conditions forced simulations), to clarify that their usage is reasonable in the frame that is proposed here. In summary, an effort to clarify and differenciate observations and regional climate models interst and usage in the paper would be important for the study to be more precise and robust.
- Data and methods. Line 151: Only Soares et al., 2023b is cited as a reference for future IB droughts, but some others could be added, such as Quintana-Segui et al, 2016: https://hal.science/hal-01401386/, for example. I guess there are several other studies that could be named here. Even more specific studies, such as Sanchez et al., 2011 (doi: 10.1007/s10584-011-0114-9), using dry spells index, or other similar ones, could fit here.
- Data and methods, lines around 160s. The selected time period for the analysis is not totally clear there, as IB01 ranges from 1971 to 2015, 1989-2008 is the ERainterim-forced period for the 12 RCMs, but no clear statement is made until the end of the section, on line 389, and then on the final section, in line 607. Maybe it would be better to state it more clear and earlier on the text, even on the objectives paragraph, as pointed before.
- When describing EuroCORDEX ensemble, please be careful with expressions such as "daily synchronized climate data" (line 190), which should be more clearly explained, as it is not clear there if they refer to reanalysis or GCM forced simulations, just mentioned before. In any case, that statement is not usual, even if it referred to reanalysis-forced simulations, so a more clear explanation should be made there.
- In equation 19, GTI should be GDI in the subindex?
- Related to previous comments about observations/models, a more precise description of some subsection titles could be considered. I mean: 3.1 GDI general performance "using gridded observations", or 3.3. GDI added value using RCM ensembles, or something like this?
- My other main concern is about figures' representation or description, that could be clarified or better described, to my opinion. For example, I am not sure to fully understand colors in figure 2. Do they represent the same as the X axis, that is, the percentile?. It is therefore redundant?. And numbers on Y axis are probabilities of normal distribution?. Naming and describing properly axis on the figures and on the captions is essential for a good understanding of figures, together with the idea of being as simple and clear as possible using the minimum/optimum amount of information.
- It happens to me again on figure 3: RMSE and time correlation (described in line 341) are obtained from the average of all the models, or for each model, and then the average of correlations in some other way?. This was my first thought, before realizing that all these section 3.1 was made only from IB01 gridded observational database. Please, make figures and section more clear with respect to the used information, for a reader to have clear which data is being used.
- In that figure, and in general in others where all the cells/locations are included, all of them are considered as equal. I mean, there lines represent each IB point, but this approach misses some insight about subregions, coastal points, mountain areas that could add information about spatial features. Did the authors make a thought of maybe using rough subregions with homogeneous climatic conditions to performe at least some of the computations?. This idea also came to me when looking at boxplots for all the cells and models on figure 6. Besides, on that figure 3, density values are strange to me, in the sense that their units, or their maximum value, could be stated or indicated in some way, for a better understanding of such numbers?
-
Table 2: What blue and red colors mean?. How drought intensity, mean decadal frequency and mean duration are defined?.
-
Table 3: I am not sure what "aggregation of the results obtained for all models'' mean. As the percentage of land points with a certain DAV, it means that each cell has 12 DAV values, one for each model, and so the percentage is obtained from those 12 combinations of model against observations score?. The sum of the numbers at each column equals 100% then?. A title on Y axis would also be nice to be shown, following my request to a better description of figures and tables.
-
One final typographic remark, related to resolution numbers of gridded data: degrees should be shown as superscripts, but this is not made in lines 150-160
Citation: https://doi.org/10.5194/gmd-2024-9-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 1,425 | 56 | 23 | 1,504 | 19 | 13 | 16 |
- HTML: 1,425
- PDF: 56
- XML: 23
- Total: 1,504
- Supplement: 19
- BibTeX: 13
- EndNote: 16
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1