Throughout 2021, 2022, and 2023, the volumes of irrigation water used, precipitation and the amount of water drained were measured. Complete water analyses were also carried out periodically in order to determine the main quality parameters.

A total of six salt balances were carried out. Three were obtained for sub-basin AB1 and another three for sub-basin AB2. They were obtained for the irrigation periods in the years 2021, 2022, and 2023: from 5th March to 15th October in each year.

The salt balance ( Δ SB) equation was defined by the difference between salt inputs (SI) and outputs (SO). This is a common methodology and has been used in other studies conducted in the Ebro basin, such as (Andrés and Cuchí, 2014; Causapé et al., 2004a).

In the case of the present study, the main salt inputs were irrigation ( SI I ) and rainfall ( SI R ). In the case of irrigation, total dissolved ions (TDI, mg· L − 1 ) were obtained from the relationship between EC and the sum of all the ions obtained in the laboratory analysis. In that of precipitation TDI, there was no locally available data. To fill this gap, we used average data from other salt balance studies conducted in the Ebro basin in recent decades. All of these concluded that the contribution of rain salts to the total balance was low, and never greater than 5% [(Andrés and Cuchí, 2014; Causapé et al., 2004b; Quílez, 1985)].

It should be noted that salt exports (SO) correspond exclusively to the salt load of the drainage water, as other components of the balance, such as salt removal by crop uptake, are considered negligible. In the case of groundwater, there was no evidence of the aquifers in the sub-basins studied (Agència Catalana de l’Aigua, 2013; Instituto Geológico y Minero de España and Dirección General del Agua, 2010; Ministerio para la Transición Ecológica y el Reto Demográfico, 2022). In these cases, the resulting salt balance would be provided by Equation 1. Δ S B = S I I + S I R − S O (1)

Water and salt balance calculations were conducted for the entire basin. The results were then divided by the irrigated area, yielding the average salt balance per field, expressed in kg. ha − 1 . year − 1 . In this way, it was easier to make comparisons between both of the sub-basins studied and with other similar studies.

We also performed a separate balance calculation for each ion in order to know which ions were exported through drainage water and which remained in the soil.

Figure 4 presents the Piper diagram (Piper, 1944) for the four sub-basins studied, along with the irrigation water used in the Algerri-Balaguer.

The Piper diagram is a graphic procedure for geochemical interpretation of water analysis. It is useful for observing and describing possible differences between samples and sources, in terms of their chemistry. It shows the relative proportions of cations and anions and their hydrochemical characteristics.

The Piper diagram illustrates the differences in ionic composition between the irrigation water and the drainage water. In the case of the cations, the irrigation water exhibited a predominance of calcium. In the drainage water, there is no dominant cation type. This discrepancy was primarily due to the elevated levels of Mg 2 + and Na + in drainage waters and its absence in irrigation (see Table 2). In the case of anions, it is clear that different types predominate depending on the water source. While bicarbonate is the primary ion present in irrigation water, sulphates predominate in drainage water.

The irrigation water showed characteristics of a calcium-magnesium bicarbonate type. Conversely, the drainage water exhibited a composition dominated by calcium, magnesium, chlorine, and sulfate ions. However, in the case of AB6, there was a mixed type, attributable to the presence of elevated levels of Na⁺ and Cl⁻. These ions are present in the soils classified as Xeric Torriorthents that predominate in this sub-basin (Figure 1).The differences in composition between the irrigation and drainage waters was a result of the dissolution of salts present in the soil; this process is intensified with higher water inputs, such as the application of irrigation.

On the other hand, we established a relationship between the total dissolved ions − TDI, mg· L − 1 (Hart, 1974; McNeil and Cox, 2000) − , and the electrical conductivity (EC, dS· m − 1 ) of the water. TDI considers the sum of major ions in the water, expressed in mg. L − 1 . In most water, these ions correspond to C a 2 + , M g 2 + , N a + , C l − , S O 4 2 − and H C O 3 − . This serves as an initial step for determining the salt input through irrigation and salt output through drainage. The relationship between TDI and EC is generally linear, according to Equation 4, with values of k ranging between 350 and 1,000 in most waters. T D I m g · L − 1 = k · E C d S · m − 1 (4)

When assessing most salt balances, it is common to use total dissolved solids (TDS) for this purpose, establishing the relationship at 640 mg· L − 1 to 1 dS· m − 1 (Rhoades et al., 1999; USSL, 1954) for continental waters, and commonly reaching values of 890 mg. L − 1 (Rusydi, 2018). However, total dissolved solids (TDS) includes not only ionic dissolved solids but also organic solids. As our goal was to determine salt inputs and outputs, we used the TDI-EC relationship. Moreover, in waters from xeric regions, originating in calcareous soils with a prevalence of gypsum, it is advisable to establish an in situ relationship. Applying the relationship proposed by the USSL may lead to an underestimation of the salt content in drainage water and its environmental impact.

In Figure 5 we show the TDI-EC relationship for water in the four sub-basins studied and for the irrigation water.

All the relationships were higher than 640 mg· L − 1 , ranging from 804 (irrigation) to 958 (AB2). The relationships for each site were:

• T D I ( A B 1 ) = 926 · E C ( R adj 2 = 0.87 , p − v a l u e < 0.001 )

In Table 2, we also indicate the SAR (sodium adsorption ratio) and EC (electrical conductivity) of each sample site. It is important to note that in AB1, AB,2 and AB5, the SAR values did not exceed 3, while the TDI values ranged between 1,000 and 4,500 mg· L − 1 . According to the guidelines proposed by Ayers and Wescot (1985), reusing this drainage water for irrigation would not be highly recommended, as they could increase the EC e . However, as seen in Table 2 and in Figure 4 the majority of these TDI corresponds to calcium-sulfate ( Ca 2 SO 4 ) and magnesium-sulfate ( Mg 2 SO 4 ) salts, which are not very soluble and do not affect the osmotic potential of the water in the soil.

This, combined with the relatively low EC values—especially for drainage waters—suggests that in certain cases, this water could be reused for irrigation, with appropriate irrigation management. However, in the case of AB6, we observed higher SAR values (6.9) accompanied by slightly higher EC values (4.6 dS· m − 1 ). Reusing the drainage water from outlet AB6 for irrigation would lead to further complications. The nature of the dissolved ions would cause issues related to the effects of sodium on both soil and plants. This highlights the importance of knowing about the soil and geological material in a given area. We could putatively observe striking differences between sub-basins receiving waters of the same irrigation quality.

These data were compared to a previous study conducted in the area between 2006 and 2008 by Villar et al. (2015). They monitored outlets AB1 and AB2 when the irrigated area had been newly stablished (and covered 4,700 ha in the AB district compared to the present 7,000 ha). In Table 3 we present some of the water quality results obtained in their study, and compared with the present one for sub-basin AB2. In Table 3 we can observe a reduction in CE and in SAR values, accompanied with a reduction of concentration in Na⁺ and Mg 2 + ions. However, Ca 2 + concentration remains stable. Similar patterns were observed for AB1. This suggests that the application of high-quality irrigation water over the past 15 years has resulted in the leaching of salts and the improvement of soils in the area.

Figure 6 illustrates the daily dynamics of irrigation water inputs, drainage water amounts, and the estimated salt export from the irrigable area in the two studied sub-basins (AB1 and AB2). In 2021 and 2022, the impact of irrigation on the quantity of drainage water and the salts exported is clearly evident. As expected for the type of crops in the area, the majority of the irrigation water was used during the summer months, with average values ranging from 50 to 80 m 3 · ha − 1 · day − 1 . These high irrigation inputs led to an increase in drainage flow during these months; this was particularly evident at the AB2 outlet.

A delay was observed between irrigation application and drainage discharge. Despite the highest irrigation water usage occurring in July and August, peak drainage and salt discharge were recorded in September. This suggests that at the onset of the “second” irrigation period for maize, the soil was initially dry, resulting in minimal drainage. Over time, as the soil moisture increased, excessive drainage occurred from July to September, aligning with the period of peak irrigation demand. This indicates that irrigation management may have been less effective during periods of high irrigation demand compared to periods of lower demand.

Regarding the response to precipitation events, it was observed (dark blue bars in Figure 6) that the influence in AB1 was much smaller than in AB2, with the former having a significantly larger drainage basin (2587 ha vs. 114 ha). In AB2, there was a clear increase in the salt load in the days leading up to a rainfall event. In AB2, salt load values ranged from 20 to 30 kg salt· ha − 1 · day − 1 , while in AB1, it was lower, at 15 kg· ha − 1 · day − 1 . These values were consistent with those reported in other earlier salt balance studies conducted in the River Ebro basin (Abrahao et al., 2011; Andrés et al., 2020; Causapé et al., 2004a; García-Garizábal et al., 2009; Isidoro et al., 2006; Merchán et al., 2013; 2018).

The main differences between these studies were related to the geological and soil types, and also to the irrigation types (surface or sprinkler). This complicates the comparison of results across basins beyond salt exports and highlights the importance of studying each irrigated area individually. It also underscores the potential for improvements in water management and their overall impact.

In 2023, there was a noticeable decrease in irrigation water application compared to 2021: by 40% in AB1 and by 32% in AB2. The reduction in irrigation water application was mandated by the basin administrator (Confederación Hidrográfica del Ebro, CHE) in response to the hydrological drought experienced in the basin in 2023 Confederación Hidrográfica del Ebro (2024). Water availability for irrigation across the entire irrigable area was reduced to 75% compared to a normal year. However, through optimized water management at the irrigation district level, the actual irrigation water applied amounted to 70% of the amount applied in the previous 2 years (personal communication, 2023).

We observed the effects of reduced irrigation water availability in both sub-basins. As shown in Figure 6, peak irrigation application occurred during the summer months in 2021 and 2022 but this level was not reached in 2023. In that year, maximum average daily irrigation inputs did not exceed 40 m 3 · ha − 1 . This shift could primarily be attributed to changes in irrigation management and land use by farmers. They relied more on irrigation for barley and left some acreage fallow, without sowing maize. This change in behavior led to decreased crop yields. However, the reduction in irrigation also led to a 66% reduction in drainage in AB1 and a 79% reduction in AB2, compared to 2021.

Moreover, during the first irrigation period application in 2023 (spring months), which corresponded to the winter cereal cycle, we observed a lower drainage discharge despite a higher water input through irrigation compared to 2021 and 2022, as seen in Figure 6. This suggests that when restrictions were imposed in the irrigation district, farmers improved their water management, possibly due to greater concerns about water usage.

An increase in irrigation during the spring months did not result in any corresponding increase in drainage during these months. Furthermore, despite a greater irrigation input, we observed a reduction in the leaching fraction in AB2 in 2022. This decrease in drainage was attributed to higher reference evapotranspiration ( ET o ) in that year, likely influenced by higher temperatures compared to 2021 or 2023. This led to higher crop evapotranspiration ( ET c ) values and increased irrigation demand.

At the same time, the reduction observed in 2023 resulted in a lower LF, with the value approaching the optimal level (LR), with this being even lower in the case of AB2 (Figure 6). The LF decreased from 15.6% in AB1 and 14.8% in AB2, to 7.5% and 3.9%. These results suggest that most of the excess drainage occurred during the summer months, coinciding with peak irrigation activity.

In has been observed that the salt balances varied across the 3 years analyzed (Table 4), with the amount of irrigation water applied being the determining factor. During the periods of unrestricted irrigation in 2021 and 2022, the salt balance shows that salt outputs through drainage were consistently high, ranging between 996 kg· ha − 1 and 2,193 kg· ha − 1 . Since these amounts are less than the salt supplied by irrigation water, it is conclusive that salts are leaching from soils in the area. In contrast, in 2023, with irrigation inputs below average, salt outputs were significantly reduced. The salt balance (the difference between salts added by irrigation and rainfall and those removed through drainage, as indicated in Equation 1) was slightly negative in AB1 and positive in AB2. This outcome appears to confirm the validity of the Rhoades model (1974) for leaching requirements, as the only negative balance occurred in 2023 in the AB2 sub-basin, where the leaching fraction (LF) was lower than the leaching requirement (LR).

It is important to note that the drainage measures discussed in this research are those registered exclusively during the irrigation periods of the three study years. Consequently, references to salt exports pertain specifically to those occurring within this time-frame. Although there is reduction in drainage water flow during the winter period, salt exports still occurred during this season; however, these are not reflected in this study.

Individual assessments of the predominant ions were conducted for each irrigation period within the two sub-basins studied. Figure 7 provides a synthesis of the acquired data. The figure indicates that salt is exported in excess in most cases, except for sub-basin AB2 in 2023, where salt is retained in the soil. Sulfate emerged as the primary ion exported in substantial quantities, averaging −1,200 kg. ha − 1 · year − 1 , likely due to the prevalence of gypsum in the local soils. Additionally, the bicarbonate ( HCO 3 − ) supplied by irrigation water exceeded the amount exported by drainage, suggesting possible reactions with soil calcium to form carbonate precipitates (as per Equation 5). This process likely facilitated the sequestration of inorganic carbon.

Considering Equation 5, and based on an average annual increase of 416 kg· ha − 1 of HCO 3 − in the soil, approximately 41 kg C· ha − 1 · year − 1 would be sequestered in the form of inorganic carbon via carbonate precipitates. C a 2 + + 2 H C O 3 − → C a C O 3 ↓ + H 2 O + C O 2 ↑ (5)

As noted by Entry et al. (2004), irrigation can serve as a tool to augment the sequestration of carbon within the soil, working in conjunction with organic C sequestration. Furthermore, this mechanism provides a valuable means of enhancing carbon sequestration in semi-arid regions, where soil organic matter (SOM) is scarce and challenging to increase. This is particularly important in arid and semi-arid regions with low soil organic matter (SOM) levels.

Further research could focus on accounting the total amount of C sequestrated through irrigation in a certain irrigated area.

In view of the results obtained in the present study, some remedial measures can be carried out in the study area, which are described below.

From one side, an improvement in irrigation management is necessary to reduce the existing leaching fraction and reduce the downstream impact. As shown in Table 4, in years with normal irrigation (i.e. 2021-2022), leaching fractions between 9.8% and 17% were obtained during the irrigation period. This indicates that LF during the whole year exceeds these values, being five times higher than those necessary to maintain soil quality. To understand the excess LF in the study area, it is necessary to know the existing irrigation dynamics.

As indicated in Table 1, double crop (barley + maize) is the main land use in all the studied sub-basins, followed with alfalfa and maize (as in the whole studied irrigation district as well as in others irrigation districts in the Ebro basin).

Those crops are usually classified as summer crops, as its growing cycle occurs mainly during summer months. This fact, together with the reduced rainfall during this season, makes irrigation indispensable to obtain profitable productions.

At the same time, Salvador et al. (2011) observed that summer crops in the Ebro basin are usually over-irrigated. In sprinkler irrigation, they reported 20% and 25% higher irrigation water use with respect to the net irrigation requirement values in alfalfa and maize, respectively. This indicates an irrigation water use that may be leached on excess.

This results are similar to ones obtained in a previous study developed in the sub-basin AB5. Altés et al. (2023) found that maize and double crop fields were being over-irrigated between a 9%–12% with respect to the gross irrigation requirements, according to Allen et al. (1998).

One of the strategies to reduce LF is to adjust the irrigation depths to the irrigation requirements of the crops. As can be seen in Figure 6, September accounts for the highest drainage volume discharge. It is in this late period of the crop cycle when irrigation requirements begin to decrease. However, according to Altés et al. (2024), irrigation depths in the area are not reduced correspondingly, being responsible for this excess leaching.

On the other hand, water quality parameters indicate that three sub-basins drainage water can be easily reused (AB1, AB2, AB5). According to Ayers and Wescot (1985), these are medium-low quality water in therms of EC, but with low sodium content. This, combined with the use of good quality water (as is currently being used), can be a tool to reduce the total use of irrigation water diverted upstream, and in turn reduce the impact of the irrigable area downstream, without a reduction in the soil quality of the irrigable area. At the same time, it increases water resources downstream available for other users.