Evaluation of the Impact of Changing from Rainfed to Irrigated Agriculture in a Mediterranean Watershed in Spain

Round 1

Reviewer 1 Report

The manuscript presents work on the impact of changing from rainfed to irrigated agriculture.The SWAT model was used in this study to simulate stream-flow and nitrate load. The model was set up with meteorological data and agricultural management information. The four statistical indicators were used to evaluate the model in this study. The research is of practical interest and could provide guidance for farmers and managers about proper nitrate pollution control and management measures. Specific comments and suggestions are listed below.

 1  Figure 3: Why the max simulated streamflow of 2018 was small than 2013, while the precipitation was bigger?

2  Observed streamflow in Figure 4  Inconsistency with in Figure 3, for example of 2003 and 2013.

3  Section 3.3: Many factors cause the increase of drainage nitrate concentration in the basin. For example Change in land use mode, increase in fertilizer use, etc. Reasonable irrigation can promote the efficient use of soil fertilizers and reduce nitrogen pollution. It is suggested that the author to further analyze the irrigation method and related SWAT model parameters in this watershed.

4  Figure 6: Average annual nitrate concentration reached the lowest point at 2011, then increase sharply from 2011 to 2013. The cause of this phenomenon needs to be analyzed.

5  Supplementary basic information, including the area of crops planted in the region, irrigation water consumption and Irrigation water utilization rate , crop yield over the years, etc.

Author Response

Point 1:  Figure 3: Why the max simulated streamflow of 2018 was small than 2013, while the precipitation was bigger?

Response 1: The precipitation data represented in Figure 3 is only for one station, Olite meteorological station, whereas the SWAT model used precipitation from several meteorological/weather stations for its streamflow simulation, assigning each subbasin the weather station that was closest to its centroid. Due to the high spatial variability in precipitation within this area, the upstream precipitation in early 2018 (March and April) was less compared to the downstream at Olite; hence the resultant simulated streamflow was significantly less than in 2013 despite having higher precipitation. However, the observed streamflow for both years is almost equal. Moreover, streamflow is not only dependent on precipitation; other factors such as soil moisture content, evapotranspiration, and soil and crop states and condition could result in very different amounts of streamflow despite having similar precipitation.

Point 2:  Observed streamflow in Figure 4 Inconsistency with in Figure 3, for example of 2003 and 2013.

Response 2: The visible inconsistency in Figure 4 (now Figure 3) observed streamflow compared to Figure 3 (now Figure 2) was because Figure 4 (now Figure 3) was plotted using only the months when the nitrate data was available, and thus some months are missing, hence the inconsistency. The streamflow in Figure 4 (now Figure 3) has been updated to coincide with that in Figure 3 (now Figure 2) to avoid this inconsistency.

Point 3:  Section 3.3: Many factors cause the increase of drainage nitrate concentration in the basin. For example Change in land use mode, increase in fertilizer use, etc. Reasonable irrigation can promote the efficient use of soil fertilizers and reduce nitrogen pollution. It is suggested that the author to further analyze the irrigation method and related SWAT model parameters in this watershed.

Response 3:  The main SWAT model parameters used in this study for streamflow and nitrate load modeling have been discussed in section 3.1.1. It should be noted that the simulation was only carried out under rainfed conditions hence no irrigation parameters were involved. Section 3.2 discusses the irrigation dynamics in the watershed from 2017-2020. This post-irrigation observed data was then compared with the simulated rainfed data to ascertain the irrigation impact using the discussed methodology in section 2.5. The irrigation method commonly used in the watershed is sprinkler irrigation accounting for over 70% with annual irrigation water efficiencies of over 85%; this has been discussed in the second paragraph of section 3.2.

Point 4:  Figure 6: Average annual nitrate concentration reached the lowest point at 2011, then increase sharply from 2011 to 2013. The cause of this phenomenon needs to be analyzed.

Response 4: The period between 2009-2012 was when the irrigation infrastructure was being developed; hence it was a transition period where very little agricultural activity happened in the area as most of the agricultural land was uncultivated. By 2013, about 90% of the current irrigated area had been converted from rainfed to irrigation resulting in more intensive cultivation within the newly modified system, which involved the application of more fertilization because of the high-value crops cultivated and a change in the cropping cycle where cultivation was now possible even during the summer unlike before 2011 where it only happened during winter. As suggested, we have added some lines in section 3.3 to explain this phenomenon.

Point 5:  Supplementary basic information, including the area of crops planted in the region, irrigation water consumption and Irrigation water utilization rate, crop yield over the years, etc.

Response 5: Attached is the table of the requested supplementary material of the area crops planted in the watershed, annual irrigation water consumption, and crop yields.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments:

Line 23. The word in parentheses is unnecessary, please present the concept more clearly, such as: “……comparison of the rainfed and post-irrigation periods….”.

Lines 123-129. The contents of this section are already reiterated in the sentences above regarding EU directives and are better suited in the discussion and conclusions section. I suggest removing them also to lighten the introductory section.

Line 136. north and south; in capital letters, as above.

Line 166. Figure 2. This figure is not particularly interesting or informative as the information on land conversion is given in the text and is easy to understand. I suggest removing it.

Line 173. Part of the sentence is missing, please correct it.

Line 197. Input data. Nothing is stated about local point sources of pollutants. Can information be provided about the input of pollutants from point sources such as wastewater treatment plants? Could these affect the loads shown in the manuscript?

Line 203. Meteorological stations. Using data from 25 weather stations appears challenging, especially for the data preparation phase. In Figure 1, 10/11 weather stations are visible, while the remaining appear to be located outside the study area. Please indicate in the text the reason for using such a large number of stations, as well as the criterion used to discriminate the maximum distance from the study area

Lines 241-242. The FAO's Penman-Monteith method. Please revise this sentence since 2 other methods are available in SWAT to calculate ET.

Line 518. Conclusion. In this form the conclusions are merely reiterating what was presented in the R&D. I suggest rewriting the conclusions, limiting the numerical results part to the essential (i.e. streamflow, nitrate), following the outline below:

- reiterate the central question; - summarize the themes featured recap the defence and   reasoning; - present the results and offer answers to the problems raised; - highlight the contribution made by this study, as well as its limits; - offer future leads of research.

Author Response

Point 1: Line 23. The word in parentheses is unnecessary, please present the concept more clearly, such as: “……comparison of the rainfed and post-irrigation periods….”.

Response 1: Corrected as suggested by the reviewer

Point 2: Lines 123-129. The contents of this section are already reiterated in the sentences above regarding EU directives and are better suited in the discussion and conclusions section. I suggest removing them also to lighten the introductory section.

Response 2: Removed as suggested by the reviewer and added to the conclusion section

Point 3: Line 136. north and south; in capital letters, as above.

Response 3: Corrected as suggested by the reviewer

Point 4: Line 166. Figure 2. This figure is not particularly interesting or informative as the information on land conversion is given in the text and is easy to understand. I suggest removing it.

Response 4: Corrected as suggested by the reviewer

Point 5: Line 173. Part of the sentence is missing, please correct it.

Response 5: Corrected as required

Point 6: Line 197. Input data. Nothing is stated about local point sources of pollutants. Can information be provided about the input of pollutants from point sources such as wastewater treatment plants? Could these affect the loads shown in the manuscript?

Response 6: The point source data from wastewater effluents were not used in this research since the study scope was limited to effects from agricultural areas. A previous study in the watershed by Merchan et al. (2020) indicated that the contribution from wastewater treatment plants in this area was negligible as it only accounted for about 1.5% of the N loads. Thus, it was the authors' view to neglect this aspect as it would have further complicated the model setup despite having almost no effect. A statement explaining this point has been added in the manuscript for clarity.

Point 7: Line 203. Meteorological stations. Using data from 25 weather stations appears challenging, especially for the data preparation phase. In Figure 1, 10/11 weather stations are visible, while the remaining appear to be located outside the study area. Please indicate in the text the reason for using such a large number of stations, as well as the criterion used to discriminate the maximum distance from the study area

Response 7: For the study area, we obtained 25 weather stations from within and outside the watershed from the Government of Navarra. Generally, the SWAT model automatically selects the weather station closest to the centroid of each subbasin for its simulation. Therefore, due to the watershed’s high spatial variability in precipitation, it was necessary that we adequately represent the variation as much as possible, and this was made possible by using numerous stations. Additionally, we felt that since we had readily available data for these stations, it would only be appropriate to use them to improve the model’s accuracy. As suggested, a sentence has been added in the text to explain the use of many weather stations.

Point 8: Lines 241-242. The FAO's Penman-Monteith method. Please revise this sentence since 2 other methods are available in SWAT to calculate ET.

Response 8: SWAT has three options for calculating evapotranspiration: FAO Penman-Monteith, Hargreaves, and the Priestly-Taylor method. It is not possible to use all these methods, and one must choose only one for the simulation (that is, 0 for the FAO Penman-Monteith method, 1 for the Hargreaves method, and 2 for the Priestly-Taylor method). For this study, we selected option 0, which is the FAO Penman-Monteith method.

Point 9: Line 518. Conclusion. In this form the conclusions are merely reiterating what was presented in the R&D. I suggest rewriting the conclusions, limiting the numerical results part to the essential (i.e. streamflow, nitrate), following the outline below:

- reiterate the central question; - summarize the themes featured recap the defence and   reasoning; - present the results and offer answers to the problems raised; - highlight the contribution made by this study, as well as its limits; - offer future leads of research

Response 9: The conclusion section has been revised as follows:

This paper examined the impact of changing from rainfed to irrigated agriculture on streamflow, nitrate load, and nitrate concentration in a Mediterranean watershed in northern Spain by simulating the rainfed conditions using the SWAT model and compared them to the current post-irrigation period. The results indicate a significant increase in the annual streamflow, nitrate load, and nitrate concentration at the watershed outlet in the post-irrigation period. Higher irrigation impact was observed during summer and autumn when irrigation was at its peak than in winter and spring. The increase in streamflow was explained by additional water coming from irrigation, whereas the increase in nitrate export and concentration was attributed to increased fertilization from the cultivation of high nitrogen consuming crops. The implementation of irrigation and subsequent agricultural intensification resulted in changing cropping patterns and doubling of nitrate concentrations at the outlet, exceeding the Nitrate Directive thresholds recommended by the European Commission. Therefore, nitrate minimization practices such as efficient nitrogen fertilizer application and the creation of nitrogen buffer zones along the river’s riparian zone should be considered to control nitrate exportation and pollution from cultivated lands into the river. Despite this study's valuable and significant findings, more data is needed to further analyze and assess the impact of irrigation especially during summer and autumn, which was modified following irrigation. The methodology and findings from this study can be applied to other areas with similar conditions, allowing a more comprehensive assessment of the effect of changing from rainfed to irrigated agriculture on streamflow and nitrate pollution. These findings could assist farmers, water experts, and policy/decision-makers in improving water resources management at the watershed level and be useful in guiding the development of new irrigation systems, thereby improving sustainable agriculture.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

Accept in present form.