the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Drought intensity-duration-frequency curves based on deficit in precipitation and streamflow for water resources management
Yonca Cavus
Kerstin Stahl
Hafzullah Aksoy
Abstract. Drought estimates in terms of physically measurable variables such as precipitation deficit or streamflow deficit are key knowledge for an effective water management. How these deficits vary with the drought event severity indicated by commonly used standardized indices is often unclear. Drought characteristics assigned the same value in index are not necessarily the same in different regions, and in different months of the same region. We investigate drought to remove this disadvantage of the index-based drought IDF curves and develop intensity-duration-frequency (IDF) curves in terms of the associated deficit. In order to study the variation of deficits, we use the link between precipitation and streamflow, and the associated indices, standardized precipitation index (SPI) and standardized streamflow index (SSI). More specifically, the analysis relies on frequency analysis combined with the total probability theorem applied to the critical drought severity. The critical drought has varying durations and it is extracted from dry periods. IDF curves in terms of precipitation and streamflow deficits for the most severe drought of each drought duration in each year are then subject to comparison of statistical characteristics of droughts for different return periods. Precipitation and streamflow data from two catchments, the Seyhan River (Turkey) and the Kocher River (Germany) provide examples for two climatically and hydrologically different cases. A comparison of the two cases allows to test a similar method in different hydrological conditions. We found that precipitation and streamflow deficits vary systematically reflecting seasonality and the magnitude of precipitation and streamflow characteristics of the catchments. Deficits change from one month to another at a given station. Higher precipitation deficits were observed in winter months compared to summer months. Additionally, we assessed observed past major droughts experienced in both catchments on the IDF curves which show that the major droughts have return periods at the order of 100 years at short durations. This coincides with the observation in the catchments and shows the applicability of the IDF curves. The IDF curves can be considered a tool for using in a range of specific activities of agriculture, ecology, industry, energy, water supply etc. This is particularly important to end-users and decision-makers to act against the drought quickly and precisely in a more physically understandable manner.
Yonca Cavus et al.
Status: open (until 29 Jun 2023)
-
RC1: 'Comment on hess-2023-107', Anonymous Referee #1, 28 May 2023
reply
General Comments
Drought has been widely evaluated by using different standardized drought indices in the literature. Some studies take the drought indices to develop drought intensity-duration-frequency (IDF) curves as reviewed by the authors in this paper. It is emphasized that the drought indices are dimensionless, and it is therefore difficult to interpret them physically, meaning that precipitation and streamflow deficits cannot be quantified directly from the drought index-based drought IDF curves. This is a true argument. At this point, this paper aims to develop drought IDF curves in terms of precipitation and streamflow deficits to overcome this difficulty. The proposed methodology was applied on two different climatic regions, one from Turkey and another from Germany. The applicability of the IDF curves has been shown by using historical droughts experienced in both regions. Another remarkable result in this study is Table 3 which represents the practical value of IDF curves for various sectors. In my opinion, the deficit-IDF curves developed in this study will be useful tools in establishing proper drought mitigation strategies. Therefore, I recommend publication of the paper after minor revision based on the following specific comments and technical corrections.
Specific Comments (in the order they appear in the text)
- Lines 12-13: Revise to clarify the statement “Drought characteristics assigned the same value in index are not necessarily the same in different regions, and in different months of the same region” as it is not clear.
- Line 96: There is no need to emphasize the ratio of the agricultural lands (51%) in Kocher. The paper in not agriculture-oriented and results are not affected by this particular ratio. The previous sentence good enough to state that the catchment has importance for agriculture as it was stated for Seyhan previously in this paragraph.
- Lines 135-137: It is known that Gamma distribution is used for SPI but it is advised to make a clear reference to McKee et al. (1993). How did you find that the General Extreme Value (GEV) is the best-fit probability distribution function for the standardized streamflow index (SSI)? Among which distributions?
- Line 170: Can you give the name of any other curve tested for establishing functional relationship between the drought indicators and indices?
- Line 183: It is not clear what DId in Eq. 2 is. Please make the statement clear.
- Figure 6: The 12-month threshold, which is a constant value over the year is not given in Figure 6. It can be added to Figure 6 as a line or be indicated in the figure caption.
- Lines 243-266: Precipitation threshold and streamflow threshold (in Figure 6) are used in the development of IDF curves (in Figure 5). Thus, I suggest that Figure 6 comes before Figure 5. The text about Figures 5 and 6 needs a thorough check. Please revise your text by considering that Figure 6 tells the story about the within year variability of precipitation and streamflow thresholds and Figure 5 is a tool based on this story. Also, probably because of the disorganization of Figures 5 and 6, I found some statements unclear or redundant: (a) Omit statement ‘Comparisons between the …’ starting in Line 253. It is a sentence hard to understand and to me again the statement about the drought duration is not true. The second part coming after ‘although’ is confusing. (b) Revise statement ‘Short and more …’ starting in Line 257. While revising pay attention to replace ‘drought in precipitation deficit’ with ‘drought in precipitation’. (c) Delete statement ‘With the deficit …’ starting in Line 262. It reflects an opinion about the IDF curves rather than being result. This is well fit to the Discussion Section (subsection 5.1) where this fact has already been emphasized in Lines 312-314. There is no need to have it here.
Technical Corrections
- Lines 80-85: So many “we” in this paragraph.
- Line 116 (Caption of Table 1): Put the name of countries in parenthesis as (Turkey) and (Germany) after each catchment to make the table caption self-explanatory.
- Line 118 (Figure 1): For a better visualization, name of the regions (Seyhan and Kocher) should be centered at the top their group of graphs instead of giving them in the precipitation graph only.
- Line 177: DIc should be italic.
- Line 225 (Caption of Figure 4): Please add the following sentence to the caption to clarify horizontal axis: ’Horizontal axis indicate SPI and SSI values at the top of each scatter diagram.’
- Line 257 (end): Make plural as “… intense droughts …”
- Lines 378-380: The footnote is related to Table 3 but there is no “*” in the table. The authors may consider putting it into the Table caption instead of keeping it as a footnote to avoid such confusion.
Citation: https://doi.org/10.5194/hess-2023-107-RC1 -
AC1: 'Reply on RC1', Yonca Cavus, 02 Jun 2023
reply
Please find our responses to the Comments of Reviewer 1 in the PDF attached.
-
RC2: 'Comment on hess-2023-107', Anonymous Referee #2, 04 Jun 2023
reply
General comments
The manuscript presents a methodology for development of intensity-duration-frequency (IDF) curves based on the frequency analysis of severity of the critical (the most severe) droughts of different fixed durations based on two non-dimensional indices, SPI and SSI, and their consequent transformation into corresponding dimensional quantities, deficits of precipitation and streamflow. In this way, final IDF curves are easier to interpret in various contexts in which drought analysis is needed.
The overall presentation of the methodology in the paper is clear, with a minor shortcoming that I describe as the first specific comment. The motivation for proposing this methodology is laid out well. The part of the discussion related to the potential for the use of the IDF curves provides an excellent overview. In my opinion, the discussion and conclusion sections in the paper lacks two aspects that I describe as the second and the third specific comment.
I find that the proposed approach, which considers different fixed drought durations, provides a more rigorous structural description of droughts than other approaches where duration is treated as an additional variable. I therefore commend the authors on devising this methodology.
Overall, I propose minor revision of the manuscript in accordance with the comments below.
Specific comments
- Description of Step 4 of the methodology (around L170) lacks specification of data used in developing logistic regression between SPI/SSI and precipitation/streamflow. Is this based only on the critical droughts or on all droughts? What was the time scale of precipitation/streamflow used in the regression analysis (matching the time scale of SPI/SSI or something more general, like annual precipitation)? I suggest that this is precisely described so that the methodology can be fully comprehended and potentially replicated for other sites by other authors.
- The methodology builds on droughts defined using the thresholds SPI = 0 and SSI = 0. However, specific areas for which the droughts need to be analysed may require different types of thresholds. For example, in agriculture, different crops have different water requirements in terms of precipitation. How would your methodology fit to that type of thresholds? Furthermore, can the proposed methodology be easily adapted to other indices than SPI and SSI?
- In my opinion, the conclusions lack a brief overview of open questions and potential for further research. These open questions could be regional IDF curves, estimation at ungauged sites etc.
Minor comments and technical corrections
L72-74: Sentence starting with “Because…” is unclear or unfinished. Please revise.
L113: Replace “likely to observe” with “likely to be observed”.
L153: Remove “for each year” from this sentence. Frequency analysis is done for the series of annual maximum severities, but not for each year.
L175: Please check this part: “drought INDICATORS were converted to precipitation and streamflow deficits using the relation between the drought indicators and indices”. Looks like the first “indicators” should be replaced by “indices”.
L319: Section 5.2 heading is incomplete: “… in different climatic” what? (maybe conditions?)
L393: I suggest replacing “variability” with “temporal variability” (since one could also possibly discuss spatial variability of droughts).
Citation: https://doi.org/10.5194/hess-2023-107-RC2 -
AC2: 'Reply on RC2', Yonca Cavus, 06 Jun 2023
reply
Please find our responses to the Comments of Reviewer 2 in the PDF attached.
Yonca Cavus et al.
Yonca Cavus et al.
Viewed
HTML | XML | Total | BibTeX | EndNote | |
---|---|---|---|---|---|
442 | 128 | 17 | 587 | 3 | 4 |
- HTML: 442
- PDF: 128
- XML: 17
- Total: 587
- BibTeX: 3
- EndNote: 4
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1