the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 27 Jul 2023
A comparison of RCMs and meteorological time series (1950–1996) of southern Italy as a fine calibration for hydrogeological scenarios
Abstract. Nowadays the phenomenon of Global Warming is unequivocal, as confirmed by the latest reports of the IPCC and studies of the climate-change impacts on ecosystems, global economy, and populations. Among these analyses the effect of climate change on groundwater is a very relevant task especially for regions whose economic and social development depends chiefly on groundwater availability, as for the southern Italy. In such a territorial framework, this research was focused on analyzing: i) comparison of precipitation and air temperature obtained by Regional Climate Models (RCMs) and meteorological time series recorded in a part (1950–1996) of the “historical experiment” period (1950–2005); ii) effects of climate change on scenarios of air temperature (T) and precipitation (P) and, consequently, on scenarios of actual evapotranspiration (ETR) and effective precipitation Pe (P – ETR). The latter was considered as a proxy of groundwater recharge of the principal aquifer systems of the region, represented chiefly by the karst aquifers.
To achieve a detailed hydro-climatological characterization, an Ensemble of 15 RCMs (E15) derived from the European Coordinated Regional Downscaling Experiment (EURO-CORDEX), at a spatial resolution of 0.11° (~12 km), was analyzed. Specifically, two IPCC Representative Concentration Pathways of greenhouse gases (RCP4.5 and RCP8.5) were considered. The E15 was calibrated in the validation period (1950–1996) by a statistical comparison with data observed by the regional meteorological network managed by the former National Hydrological Service (SIMN), Department of Naples, which was active in the period 1921–1999.
As a principal result, the E15 was found with a statistical structure very similar to those of observed annual precipitation (OBSP) and mean annual air temperature (OBST), characterized by a very similar frequency distribution. Accordingly, an inferential statistical approach was performed for calibrating E15 precipitation (E15P) and air temperature (E15T) based on the compensation of the difference with OBSP (+7 %) and OBST (-16 %). The E15 projects a reduction in precipitation and an increase in air temperature under both RCPs, with a divergence point between the two scenarios occurring by about 2040. As a principal result, Pe shows declining trends for both RCP scenarios, reaching a decrease of the 11-yrs moving average down to -20 %, for RCP4.5, and -50 %, for RCP8.5, even if characterized by relevant inter-annual fluctuations.
This preprint has been withdrawn.
-
Withdrawal notice
This preprint has been withdrawn.
-
Preprint
(3627 KB)
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2023-1505', Anonymous Referee #1, 07 Sep 2023
Dear Authors,
I have some comments on your work, which, although it addresses an interesting topic, is weak and lacking in some points. I would like to give you the opportunity to revise it, but it requires a considerable amount of work.
In general, I find some issues both in the organization of the text and in the analyses you have produced. Regarding the organization of the text, I see some unbalanced sections and some superfluous information, while others are missing.
The introduction needs a thorough revision. It lacks references to your considerations, and the purpose of the work is not clearly stated.
Regarding the data used, it is not clear to me (so it should be further developed in the text) whether you have always considered full years for a single station or years in which there may be gaps of some days/months. This is important, especially for precipitation, which, being cumulative, could be underestimated. It would also be appropriate to highlight how the CORDEX models were selected. There is a decent amount of recent literature on this topic and on metrics for constructing the ensemble mean. There are metrics such as Equilibrium Climate Sensitivity that help build ensembles, perhaps by investigating greater variability in the GCM or having models closer to the mean. It is also important not to consider too many warm and/or too many cold models, as they could influence the results.
Another consideration is the choice to analyze the RCP8.5 scenario. For research purposes, it is still a representative scenario used in various applications. However, in the introduction, you mention support for adaptation plans/actions. In this case, RCP8.5 (the BAU scenario) is an extreme but unlikely case (see the paper https://www.nature.com/articles/d41586-020-00177-3). I would suggest analyzing the RCP2.6 scenario as well and explicitly stating the interest from a research perspective only, leaving out the RCP8.5 scenario.
The most critical points concern bias correction and, in general, the Turc relationship. Regarding bias correction, if I understand correctly, you corrected it considering periods longer than 30 years for training. This is a limitation, especially for temperature, as above 30 years, there could be a clear climatic signal. Furthermore, it would be appropriate to state the limitations of bias correction approaches, especially regarding the assumption of stationarity. As for the Turc method, based on my knowledge, it is an approach for estimating reference evapotranspiration (saturated soil and independent of the crop type). Essentially, you are analyzing the Climatic Water Budget (CWB), which is independent of soil characteristics. At this point, permeability plays a key role in understanding whether potential evaporation can be confused with actual evaporation. In some way, a comparison with Fig. 2 might help (otherwise, I would not understand its utility except for describing a case study).
As a very minor point, I would adjust the curly braces in Fig. 7 and Fig. 8.I hope I have been of assistance. Good luck!
Citation: https://doi.org/10.5194/egusphere-2023-1505-RC1 -
AC1: 'Reply on RC1', Daniele Lepore, 15 Sep 2023
Reply to RC1: 'Comment on egusphere-2023-1505', Anonymous Referee #1, 07 Sep 2023
We thank very much the Anonymous Referee #1 for suggesting a series of changes to the manuscript aimed at its improvement.
Dear Authors,
I have some comments on your work, which, although it addresses an interesting topic, is weak and lacking in some points. I would like to give you the opportunity to revise it, but it requires a considerable amount of work. In general, I find some issues both in the organization of the text and in the analyses you have produced. Regarding the organization of the text, I see some unbalanced sections and some superfluous information, while others are missing. The introduction needs a thorough revision. It lacks references to your considerations, and the purpose of the work is not clearly stated.
In the Introduction section, we will consider new pertinent references and a clearer explanation of the purpose of the work.
Regarding the data used, it is not clear to me (so it should be further developed in the text) whether you have always considered full years for a single station or years in which there may be gaps of some days/months. This is important, especially for precipitation, which, being cumulative, could be underestimated.
We considered only data of those rain gauge stations which worked continuously during a single year (full year). Therefore, we reconstructed time series of annual cumulative precipitation which, even not continuous from year to year, represents the total amount of precipitation for each year.This aspect will be further refined and explained more clearly in the Methods Section to avoid misunderstandings.
It would also be appropriate to highlight how the CORDEX models were selected. There is a decent amount of recent literature on this topic and on metrics for constructing the ensemble mean. There are metrics such as Equilibrium Climate Sensitivity that help build ensembles, perhaps by investigating greater variability in the GCM or having models closer to the mean. It is also important not to consider too many warm and/or too many cold models, as they could influence the results.
The Mediterranean region is a complex area with a variety of climatic conditions for which an ensemble of 15 representative models can help reducing systematic errors of single models. Therefore, the models used in this study were selected among those used by previous studies for accurately simulating the effects of climate change in the Mediterranean region (Table 1). Moreover, the creation of an ensemble of 15 models (E15) was meant to compensate errors of individual models and provide a more accurate estimate of the impacts of climate change.As observed by the Anonymous Referee#1, criteria applicable for the selection of model performance and the construction of ensemble, based on bias or Transient Climate Response (TCR) and Equilibrium Climate Sensitivity (ECS), are known in the literature (Hausfather Z. and Peters G.P., 2020). Accordingly, we have verified that the selected models have ECS values exceeding 3.5 K, except for the NorESM1-M model, which has an average value of 2.8 K (Andrews et al. 2012; Flynn & Mauritsen, 2020), therefore they are highly sensitive to climate change and can provide a more realistic view of the impacts of climate change on the Mediterranean region.As a result of our selection, the Ensemble mean obtained (E15) is proved to be consistent by the low difference of precipitation and air temperature from data recorded by the regional meteorological network in the historical period (1950-1996) (Tables 3, 4 and 5).Thanks to the observation of Anonymous Referee #1, we will expand the Data and Method and the Discussion sections by considering and explaining the points indicated as well as references indicated.
Another consideration is the choice to analyze the RCP8.5 scenario. For research purposes, it is still a representative scenario used in various applications. However, in the introduction, you mention support for adaptation plans/actions. In this case, RCP8.5 (the BAU scenario) is an extreme but unlikely case (see the paper https://www.nature.com/articles/d41586-020-00177-3). I would suggest analyzing the RCP2.6 scenario as well and explicitly stating the interest from a research perspective only, leaving out the RCP8.5 scenario.
The selection of scenarios applied in this study (RCP 4.5 and RCP 8.5) was made considering the analysis of previous researches carried out in the Mediterranean area and aimed at assessing the effects of climate changes on groundwater resources (Table 1). As proved by these studies and others, developed with a general purpose and in different geographical frameworks (Bucchignani et al. 2014; Bucchignani et al. 2016; Bucchignani et al. 2018; Spinoni M. et al. 2020; Zittis et al. 2021), the RCP 8.5 scenario is commonly used because representing the most extreme conditions, even if being the least likely. Under the conditions projected by the RCP8.5 scenario, the impacts of climate change on the Mediterranean region would be particularly severe with an increase in the average air temperature of 2-3 °C by 2100. This will lead to a range of negative impacts extensively discussed in the literature cited. Based on this assumption, it is admissible to consider the RCP 8.5 scenario as representing the potential worst (albeit less likely) impact of climate change, thus being consistent with the precautionary principle adopted by the European community (Ricci & Sheng, 2013). Therefore, the impact of the RCP 8.5 scenario on groundwater recharge of principal aquifers systems of the central-southern Italy is expected helping the scientific and technical communities to be aware of the worst conditions, therefore allowing to plan proper countermeasures.The RCP2.6 scenario is certainly interesting; however, its use entails several practical difficulties in obtaining model data and would lead to an entirely new and distinct work from the present one. In fact, the RCP 2.6 scenario is not available for all CORDEX models; only 7 out of 15 selected models have the data available (https://cordex.org/data-access/regional-climate-change-simulations-for-cordex-domains/).Even in this case, thanks to the observation of Anonymous Referee #1, we will expand the Data and Method and the Discussion sections by considering and explaining the points indicated, appropriately citing the reference sources.
The most critical points concern bias correction and, in general, the Turc relationship. Regarding bias correction, if I understand correctly, you corrected it considering periods longer than 30 years for training. This is a limitation, especially for temperature, as above 30 years, there could be a clear climatic signal. Furthermore, it would be appropriate to state the limitations of bias correction approaches, especially regarding the assumption of stationarity.
The paper introduces a simple and novel method of bias correction for climate models based on inferential analysis of precipitation and air temperature data aggregated at the annual and at regional scales, because covering a significant part of the central-southern Italy territory (Figure 6). Both datasets used, RCMs and time series of the regional meteorological network (OBS) exhibit temporal and spatial discontinuities (Figure 4). For instance, some RCMs start from 1950, while others from 1970 (Table 2). Therefore, a training period slightly exceeding 30 years (from 1950 to 1996) was considered, aiming at developing the most robust analysis possible and minimizing potential climate signals.Even if considering generally appropriate the observation of the Anonymous Referee #1 regarding the duration of the training period, we recognize that using fewer years can lead to more uncertain outcomes of the future climate conditions. These aspects will be further discussed in both the methodology and discussion sections, emphasizing the limitations of the methodology and particularly the concept of temporal stationarity of the bias correction.
As for the Turc method, based on my knowledge, it is an approach for estimating reference evapotranspiration (saturated soil and independent of the crop type). Essentially, you are analyzing the Climatic Water Budget (CWB), which is independent of soil characteristics. At this point, permeability plays a key role in understanding whether potential evaporation can be confused with actual evaporation. In some way, a comparison with Fig. 2 might help (otherwise, I would not understand its utility except for describing a case study).
The empirical Turc (1954) formula is a widely used empirical approach for estimating actual evapotranspiration (ETR) based on annual values of precipitation and air temperature. The formula was reconstructed based on water balance of 254 catchments from all over the world and aimed at estimating annual evapotranspiration (ETR) from precipitation (P) and runoff (Q) at the annual scale. Therefore, the method incorporates empirically the hydrological role of the coupling soil and vegetation on the evapotranspiration regime.In general, the Turc method can be considered valid for estimating actual evapotranspiration under temperate climatic conditions and uniform vegetation cover, therefore it has been proved to be validly applicable in the Mediterranean region as it resulted from the comparison with Coutagne, Thorntwaite-Matther and MODIS estimates of actual evapotranspiration (Ruggeri et al., 2021). We will discuss this point in the Data and Methods section.
As a very minor point, I would adjust the curly braces in Fig. 7 and Fig. 8.
We will improve the Figures 7 and 8.
Citation: https://doi.org/10.5194/egusphere-2023-1505-AC1
-
AC1: 'Reply on RC1', Daniele Lepore, 15 Sep 2023
-
RC2: 'Comment on egusphere-2023-1505', Anonymous Referee #2, 19 Sep 2023
- The abstract section of the manuscript is poorly written. The author should primarily describe what work was done, how it was done, what conclusions were reached, and suggest the author to rewrite the abstract.
- It is recommended to change the phrase 'In this work' in line 58 to 'In this study.
- The introduction section is written in a chaotic manner, with almost no logical flow and inadequate explanation of the research background, current state of research, and the scientific problems to be addressed in this study. It lacks thorough literature review. From line 41 to line 48, there is not a single reference. Are the author's statements in these lines supported by any evidence?
- I couldn't find the literature Turc, 1955. Does the formula (5) for calculating evapotranspiration from precipitation based on empirical formulas consider spatial heterogeneity? Did the author determine if this formula is applicable to all grid cells in the study area?
- Exploring the impact of climate change on groundwater recharge using RCP data is a novel topic. The author primarily focuses on comparing RCP data with site data and bias correction of RCP data (perhaps my understanding is not accurate enough), rather than addressing the main research points that readers are concerned about. Therefore, I am extremely puzzled about what the innovation of this article is.
- Although the authors have conducted a certain amount of research work, this draft paper is generally disorganized. Many parts do not conform to the writing standards for scientific papers, and it fails to highlight the main research work of the paper. Therefore, I recommend the authors to thoroughly revise it before resubmitting. I suggest rejecting and resubmitting this manuscript.
Citation: https://doi.org/10.5194/egusphere-2023-1505-RC2 -
AC2: 'Reply on RC2', Daniele Lepore, 23 Sep 2023
We thank the Anonymous Reviewer #2 for the suggestions which we acknowledge to be aimed at improving the overall coherence and consistency of the manuscript.
1) The abstract section of the manuscript is poorly written. The author should primarily describe what work was done, how it was done, what conclusions were reached, and suggest the author to rewrite the abstract.
In our opinion the abstract, in this form, already contains all the elements indicated by the Anonymous Reviewer n.2. However, we recognize that with its rewriting we will obtain a more linear, compact and functional structure.
2) It is recommended to change the phrase 'In this work' in line 58 to 'In this study.
We appreciate your suggestion and we will apply the correction.
3) The introduction section is written in a chaotic manner, with almost no logical flow and inadequate explanation of the research background, current state of research, and the scientific problems to be addressed in this study. It lacks thorough literature review. From line 41 to line 48, there is not a single reference. Are the author's statements in these lines supported by any evidence?
The introduction intends to focus the reader's attention on the relevance of groundwater resources in Southern Italy, on which economic and social development strictly depends. Furthermore, it focuses on the current lack of studies predicting the effects of climate change on groundwater recharge and the future availability of groundwater resources. In its current form, the introduction follows a progressive logic, incorporating the state of the art and references supporting the topic in dedicated sections, respectively dedicated to RCMs (Regional Climate Models), par. 2, and description of the study area, par. 3.1. However, thanks to the observations of Anonymous Reviewer #2, we recognize that this section of the manuscript can be improved by eliminating some redundancies and non-functional information as well as by including additional appropriate references (e.g., lines 41 to 48).
4) I couldn't find the literature Turc, 1955. Does the formula (5) for calculating evapotranspiration from precipitation based on empirical formulas consider spatial heterogeneity? Did the author determine if this formula is applicable to all grid cells in the study area?
Turc (1955) is a widely used empirical formula to estimate annual values of actual evapotranspiration (ETR). It is based on the annual precipitation (P) and mean annual air temperature (T). It was reconstructed by the application of the water balance equation to precipitation and runoff data of 254 drainage basins of Europe, Africa, America and the East Indies. Therefore, the Author has demonstrated that the formula could be applied to all different climates, either humid, arid, hot or cold. The Turc formula is commonly considered in many hydrology textbooks (e.g. Shaw E.M. et al., 2010 - CRC Press - ISBN-10: 0415370418) along with other empirical formulas used for the ETR estimation. Since it is dated 1955, the Turc's original paper does not have a DOI but it’s available at https://www.persee.fr/doc/jhydr_0000-0001_1955_act_3_1_3278The applicability of the Turc formula to the climate conditions of southern Italy has already been positively tested through a comparison with the results obtained by the other empirical formulas of Coutagne and Thorntwaite as well as with MODIS satellite estimations (doi: 10.3390/w13020118).Since the Turc formula depends on the annual precipitation (P) and mean annual air temperature (T), it intrinsically incorporates the spatial heterogeneity of these two hydrological variables. Therefore, the formula has been applied to all grid cells for which P and T data were reconstructed by the RCMs for each year of the time series.
5) Exploring the impact of climate change on groundwater recharge using RCP data is a novel topic. The author primarily focuses on comparing RCP data with site data and bias correction of RCP data (perhaps my understanding is not accurate enough), rather than addressing the main research points that readers are concerned about. Therefore, I am extremely puzzled about what the innovation of this article is.
Groundwater recharge is strictly dependent on the effective precipitation (Pe), i.e. the difference between precipitation (P) and actual evapotranspiration (ETR). Therefore, the aim of our study is to apply RCMs data, in the RCP 8.5 and RCP 4.5 scenarios, to assess the impact of the climate changes on the effective precipitation. Accordingly, the methodological approach proposed is twofold. The first part concerns the comparison, in the so-called historical period (1950-2005), of annual precipitation and mean annual temperature data, derived both by RCMs and time series of the regional meteorological network. This comparison allowed the correction of systematic errors and the validation of RCM data at a regional scale using an innovative approach (section 3.4). The second part regards the use of the bias-corrected RCMs data, of the RCP 8.5 and RCP 4.5 scenarios, in the period 2006-2100 for the assessment of annual precipitation (P), actual evapotranspiration (ETR) and effective precipitation (Pe). The time series of the latter is to be considered the most relevant result of our research because it represents the impact of climate change on the groundwater recharge process of principal aquifer systems of southern Italy.
6) Although the authors have conducted a certain amount of research work, this draft paper is generally disorganized. Many parts do not conform to the writing standards for scientific papers, and it fails to highlight the main research work of the paper. Therefore, I recommend the authors to thoroughly revise it before resubmitting. I suggest rejecting and resubmitting this manuscript.
We understand the request of the Anonymous Reviewer #2 but we are confident that the manuscript could be improved due to the keen suggestions and observations of both the Reviewers as well as the recommendation that the Editor will provide.
Citation: https://doi.org/10.5194/egusphere-2023-1505-AC2
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2023-1505', Anonymous Referee #1, 07 Sep 2023
Dear Authors,
I have some comments on your work, which, although it addresses an interesting topic, is weak and lacking in some points. I would like to give you the opportunity to revise it, but it requires a considerable amount of work.
In general, I find some issues both in the organization of the text and in the analyses you have produced. Regarding the organization of the text, I see some unbalanced sections and some superfluous information, while others are missing.
The introduction needs a thorough revision. It lacks references to your considerations, and the purpose of the work is not clearly stated.
Regarding the data used, it is not clear to me (so it should be further developed in the text) whether you have always considered full years for a single station or years in which there may be gaps of some days/months. This is important, especially for precipitation, which, being cumulative, could be underestimated. It would also be appropriate to highlight how the CORDEX models were selected. There is a decent amount of recent literature on this topic and on metrics for constructing the ensemble mean. There are metrics such as Equilibrium Climate Sensitivity that help build ensembles, perhaps by investigating greater variability in the GCM or having models closer to the mean. It is also important not to consider too many warm and/or too many cold models, as they could influence the results.
Another consideration is the choice to analyze the RCP8.5 scenario. For research purposes, it is still a representative scenario used in various applications. However, in the introduction, you mention support for adaptation plans/actions. In this case, RCP8.5 (the BAU scenario) is an extreme but unlikely case (see the paper https://www.nature.com/articles/d41586-020-00177-3). I would suggest analyzing the RCP2.6 scenario as well and explicitly stating the interest from a research perspective only, leaving out the RCP8.5 scenario.
The most critical points concern bias correction and, in general, the Turc relationship. Regarding bias correction, if I understand correctly, you corrected it considering periods longer than 30 years for training. This is a limitation, especially for temperature, as above 30 years, there could be a clear climatic signal. Furthermore, it would be appropriate to state the limitations of bias correction approaches, especially regarding the assumption of stationarity. As for the Turc method, based on my knowledge, it is an approach for estimating reference evapotranspiration (saturated soil and independent of the crop type). Essentially, you are analyzing the Climatic Water Budget (CWB), which is independent of soil characteristics. At this point, permeability plays a key role in understanding whether potential evaporation can be confused with actual evaporation. In some way, a comparison with Fig. 2 might help (otherwise, I would not understand its utility except for describing a case study).
As a very minor point, I would adjust the curly braces in Fig. 7 and Fig. 8.I hope I have been of assistance. Good luck!
Citation: https://doi.org/10.5194/egusphere-2023-1505-RC1 -
AC1: 'Reply on RC1', Daniele Lepore, 15 Sep 2023
Reply to RC1: 'Comment on egusphere-2023-1505', Anonymous Referee #1, 07 Sep 2023
We thank very much the Anonymous Referee #1 for suggesting a series of changes to the manuscript aimed at its improvement.
Dear Authors,
I have some comments on your work, which, although it addresses an interesting topic, is weak and lacking in some points. I would like to give you the opportunity to revise it, but it requires a considerable amount of work. In general, I find some issues both in the organization of the text and in the analyses you have produced. Regarding the organization of the text, I see some unbalanced sections and some superfluous information, while others are missing. The introduction needs a thorough revision. It lacks references to your considerations, and the purpose of the work is not clearly stated.
In the Introduction section, we will consider new pertinent references and a clearer explanation of the purpose of the work.
Regarding the data used, it is not clear to me (so it should be further developed in the text) whether you have always considered full years for a single station or years in which there may be gaps of some days/months. This is important, especially for precipitation, which, being cumulative, could be underestimated.
We considered only data of those rain gauge stations which worked continuously during a single year (full year). Therefore, we reconstructed time series of annual cumulative precipitation which, even not continuous from year to year, represents the total amount of precipitation for each year.This aspect will be further refined and explained more clearly in the Methods Section to avoid misunderstandings.
It would also be appropriate to highlight how the CORDEX models were selected. There is a decent amount of recent literature on this topic and on metrics for constructing the ensemble mean. There are metrics such as Equilibrium Climate Sensitivity that help build ensembles, perhaps by investigating greater variability in the GCM or having models closer to the mean. It is also important not to consider too many warm and/or too many cold models, as they could influence the results.
The Mediterranean region is a complex area with a variety of climatic conditions for which an ensemble of 15 representative models can help reducing systematic errors of single models. Therefore, the models used in this study were selected among those used by previous studies for accurately simulating the effects of climate change in the Mediterranean region (Table 1). Moreover, the creation of an ensemble of 15 models (E15) was meant to compensate errors of individual models and provide a more accurate estimate of the impacts of climate change.As observed by the Anonymous Referee#1, criteria applicable for the selection of model performance and the construction of ensemble, based on bias or Transient Climate Response (TCR) and Equilibrium Climate Sensitivity (ECS), are known in the literature (Hausfather Z. and Peters G.P., 2020). Accordingly, we have verified that the selected models have ECS values exceeding 3.5 K, except for the NorESM1-M model, which has an average value of 2.8 K (Andrews et al. 2012; Flynn & Mauritsen, 2020), therefore they are highly sensitive to climate change and can provide a more realistic view of the impacts of climate change on the Mediterranean region.As a result of our selection, the Ensemble mean obtained (E15) is proved to be consistent by the low difference of precipitation and air temperature from data recorded by the regional meteorological network in the historical period (1950-1996) (Tables 3, 4 and 5).Thanks to the observation of Anonymous Referee #1, we will expand the Data and Method and the Discussion sections by considering and explaining the points indicated as well as references indicated.
Another consideration is the choice to analyze the RCP8.5 scenario. For research purposes, it is still a representative scenario used in various applications. However, in the introduction, you mention support for adaptation plans/actions. In this case, RCP8.5 (the BAU scenario) is an extreme but unlikely case (see the paper https://www.nature.com/articles/d41586-020-00177-3). I would suggest analyzing the RCP2.6 scenario as well and explicitly stating the interest from a research perspective only, leaving out the RCP8.5 scenario.
The selection of scenarios applied in this study (RCP 4.5 and RCP 8.5) was made considering the analysis of previous researches carried out in the Mediterranean area and aimed at assessing the effects of climate changes on groundwater resources (Table 1). As proved by these studies and others, developed with a general purpose and in different geographical frameworks (Bucchignani et al. 2014; Bucchignani et al. 2016; Bucchignani et al. 2018; Spinoni M. et al. 2020; Zittis et al. 2021), the RCP 8.5 scenario is commonly used because representing the most extreme conditions, even if being the least likely. Under the conditions projected by the RCP8.5 scenario, the impacts of climate change on the Mediterranean region would be particularly severe with an increase in the average air temperature of 2-3 °C by 2100. This will lead to a range of negative impacts extensively discussed in the literature cited. Based on this assumption, it is admissible to consider the RCP 8.5 scenario as representing the potential worst (albeit less likely) impact of climate change, thus being consistent with the precautionary principle adopted by the European community (Ricci & Sheng, 2013). Therefore, the impact of the RCP 8.5 scenario on groundwater recharge of principal aquifers systems of the central-southern Italy is expected helping the scientific and technical communities to be aware of the worst conditions, therefore allowing to plan proper countermeasures.The RCP2.6 scenario is certainly interesting; however, its use entails several practical difficulties in obtaining model data and would lead to an entirely new and distinct work from the present one. In fact, the RCP 2.6 scenario is not available for all CORDEX models; only 7 out of 15 selected models have the data available (https://cordex.org/data-access/regional-climate-change-simulations-for-cordex-domains/).Even in this case, thanks to the observation of Anonymous Referee #1, we will expand the Data and Method and the Discussion sections by considering and explaining the points indicated, appropriately citing the reference sources.
The most critical points concern bias correction and, in general, the Turc relationship. Regarding bias correction, if I understand correctly, you corrected it considering periods longer than 30 years for training. This is a limitation, especially for temperature, as above 30 years, there could be a clear climatic signal. Furthermore, it would be appropriate to state the limitations of bias correction approaches, especially regarding the assumption of stationarity.
The paper introduces a simple and novel method of bias correction for climate models based on inferential analysis of precipitation and air temperature data aggregated at the annual and at regional scales, because covering a significant part of the central-southern Italy territory (Figure 6). Both datasets used, RCMs and time series of the regional meteorological network (OBS) exhibit temporal and spatial discontinuities (Figure 4). For instance, some RCMs start from 1950, while others from 1970 (Table 2). Therefore, a training period slightly exceeding 30 years (from 1950 to 1996) was considered, aiming at developing the most robust analysis possible and minimizing potential climate signals.Even if considering generally appropriate the observation of the Anonymous Referee #1 regarding the duration of the training period, we recognize that using fewer years can lead to more uncertain outcomes of the future climate conditions. These aspects will be further discussed in both the methodology and discussion sections, emphasizing the limitations of the methodology and particularly the concept of temporal stationarity of the bias correction.
As for the Turc method, based on my knowledge, it is an approach for estimating reference evapotranspiration (saturated soil and independent of the crop type). Essentially, you are analyzing the Climatic Water Budget (CWB), which is independent of soil characteristics. At this point, permeability plays a key role in understanding whether potential evaporation can be confused with actual evaporation. In some way, a comparison with Fig. 2 might help (otherwise, I would not understand its utility except for describing a case study).
The empirical Turc (1954) formula is a widely used empirical approach for estimating actual evapotranspiration (ETR) based on annual values of precipitation and air temperature. The formula was reconstructed based on water balance of 254 catchments from all over the world and aimed at estimating annual evapotranspiration (ETR) from precipitation (P) and runoff (Q) at the annual scale. Therefore, the method incorporates empirically the hydrological role of the coupling soil and vegetation on the evapotranspiration regime.In general, the Turc method can be considered valid for estimating actual evapotranspiration under temperate climatic conditions and uniform vegetation cover, therefore it has been proved to be validly applicable in the Mediterranean region as it resulted from the comparison with Coutagne, Thorntwaite-Matther and MODIS estimates of actual evapotranspiration (Ruggeri et al., 2021). We will discuss this point in the Data and Methods section.
As a very minor point, I would adjust the curly braces in Fig. 7 and Fig. 8.
We will improve the Figures 7 and 8.
Citation: https://doi.org/10.5194/egusphere-2023-1505-AC1
-
AC1: 'Reply on RC1', Daniele Lepore, 15 Sep 2023
-
RC2: 'Comment on egusphere-2023-1505', Anonymous Referee #2, 19 Sep 2023
- The abstract section of the manuscript is poorly written. The author should primarily describe what work was done, how it was done, what conclusions were reached, and suggest the author to rewrite the abstract.
- It is recommended to change the phrase 'In this work' in line 58 to 'In this study.
- The introduction section is written in a chaotic manner, with almost no logical flow and inadequate explanation of the research background, current state of research, and the scientific problems to be addressed in this study. It lacks thorough literature review. From line 41 to line 48, there is not a single reference. Are the author's statements in these lines supported by any evidence?
- I couldn't find the literature Turc, 1955. Does the formula (5) for calculating evapotranspiration from precipitation based on empirical formulas consider spatial heterogeneity? Did the author determine if this formula is applicable to all grid cells in the study area?
- Exploring the impact of climate change on groundwater recharge using RCP data is a novel topic. The author primarily focuses on comparing RCP data with site data and bias correction of RCP data (perhaps my understanding is not accurate enough), rather than addressing the main research points that readers are concerned about. Therefore, I am extremely puzzled about what the innovation of this article is.
- Although the authors have conducted a certain amount of research work, this draft paper is generally disorganized. Many parts do not conform to the writing standards for scientific papers, and it fails to highlight the main research work of the paper. Therefore, I recommend the authors to thoroughly revise it before resubmitting. I suggest rejecting and resubmitting this manuscript.
Citation: https://doi.org/10.5194/egusphere-2023-1505-RC2 -
AC2: 'Reply on RC2', Daniele Lepore, 23 Sep 2023
We thank the Anonymous Reviewer #2 for the suggestions which we acknowledge to be aimed at improving the overall coherence and consistency of the manuscript.
1) The abstract section of the manuscript is poorly written. The author should primarily describe what work was done, how it was done, what conclusions were reached, and suggest the author to rewrite the abstract.
In our opinion the abstract, in this form, already contains all the elements indicated by the Anonymous Reviewer n.2. However, we recognize that with its rewriting we will obtain a more linear, compact and functional structure.
2) It is recommended to change the phrase 'In this work' in line 58 to 'In this study.
We appreciate your suggestion and we will apply the correction.
3) The introduction section is written in a chaotic manner, with almost no logical flow and inadequate explanation of the research background, current state of research, and the scientific problems to be addressed in this study. It lacks thorough literature review. From line 41 to line 48, there is not a single reference. Are the author's statements in these lines supported by any evidence?
The introduction intends to focus the reader's attention on the relevance of groundwater resources in Southern Italy, on which economic and social development strictly depends. Furthermore, it focuses on the current lack of studies predicting the effects of climate change on groundwater recharge and the future availability of groundwater resources. In its current form, the introduction follows a progressive logic, incorporating the state of the art and references supporting the topic in dedicated sections, respectively dedicated to RCMs (Regional Climate Models), par. 2, and description of the study area, par. 3.1. However, thanks to the observations of Anonymous Reviewer #2, we recognize that this section of the manuscript can be improved by eliminating some redundancies and non-functional information as well as by including additional appropriate references (e.g., lines 41 to 48).
4) I couldn't find the literature Turc, 1955. Does the formula (5) for calculating evapotranspiration from precipitation based on empirical formulas consider spatial heterogeneity? Did the author determine if this formula is applicable to all grid cells in the study area?
Turc (1955) is a widely used empirical formula to estimate annual values of actual evapotranspiration (ETR). It is based on the annual precipitation (P) and mean annual air temperature (T). It was reconstructed by the application of the water balance equation to precipitation and runoff data of 254 drainage basins of Europe, Africa, America and the East Indies. Therefore, the Author has demonstrated that the formula could be applied to all different climates, either humid, arid, hot or cold. The Turc formula is commonly considered in many hydrology textbooks (e.g. Shaw E.M. et al., 2010 - CRC Press - ISBN-10: 0415370418) along with other empirical formulas used for the ETR estimation. Since it is dated 1955, the Turc's original paper does not have a DOI but it’s available at https://www.persee.fr/doc/jhydr_0000-0001_1955_act_3_1_3278The applicability of the Turc formula to the climate conditions of southern Italy has already been positively tested through a comparison with the results obtained by the other empirical formulas of Coutagne and Thorntwaite as well as with MODIS satellite estimations (doi: 10.3390/w13020118).Since the Turc formula depends on the annual precipitation (P) and mean annual air temperature (T), it intrinsically incorporates the spatial heterogeneity of these two hydrological variables. Therefore, the formula has been applied to all grid cells for which P and T data were reconstructed by the RCMs for each year of the time series.
5) Exploring the impact of climate change on groundwater recharge using RCP data is a novel topic. The author primarily focuses on comparing RCP data with site data and bias correction of RCP data (perhaps my understanding is not accurate enough), rather than addressing the main research points that readers are concerned about. Therefore, I am extremely puzzled about what the innovation of this article is.
Groundwater recharge is strictly dependent on the effective precipitation (Pe), i.e. the difference between precipitation (P) and actual evapotranspiration (ETR). Therefore, the aim of our study is to apply RCMs data, in the RCP 8.5 and RCP 4.5 scenarios, to assess the impact of the climate changes on the effective precipitation. Accordingly, the methodological approach proposed is twofold. The first part concerns the comparison, in the so-called historical period (1950-2005), of annual precipitation and mean annual temperature data, derived both by RCMs and time series of the regional meteorological network. This comparison allowed the correction of systematic errors and the validation of RCM data at a regional scale using an innovative approach (section 3.4). The second part regards the use of the bias-corrected RCMs data, of the RCP 8.5 and RCP 4.5 scenarios, in the period 2006-2100 for the assessment of annual precipitation (P), actual evapotranspiration (ETR) and effective precipitation (Pe). The time series of the latter is to be considered the most relevant result of our research because it represents the impact of climate change on the groundwater recharge process of principal aquifer systems of southern Italy.
6) Although the authors have conducted a certain amount of research work, this draft paper is generally disorganized. Many parts do not conform to the writing standards for scientific papers, and it fails to highlight the main research work of the paper. Therefore, I recommend the authors to thoroughly revise it before resubmitting. I suggest rejecting and resubmitting this manuscript.
We understand the request of the Anonymous Reviewer #2 but we are confident that the manuscript could be improved due to the keen suggestions and observations of both the Reviewers as well as the recommendation that the Editor will provide.
Citation: https://doi.org/10.5194/egusphere-2023-1505-AC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 693 | 120 | 37 | 850 | 44 | 28 |
- HTML: 693
- PDF: 120
- XML: 37
- Total: 850
- BibTeX: 44
- EndNote: 28
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1