Climate change adaptation for drinking water and ecological flows through sustainable agricultural practices

ABSTRACT

Analysing the impacts of climate change on water resources is crucial to identify vulnerabilities and prioritize actions. We investigated climate change impacts on drinking water supply, emphasizing sustainable agriculture as an adaptation strategy, using the SWAT model in the Cávado River basin, Portugal. Our study highlights an increase in months with river discharge below ecological flow post water abstraction due to climate change. Notably, nitrate concentration was more influenced by sustainable agriculture practices than climate change. Our study highlights the vital role of adaptive strategies, especially sustainable agriculture, in securing water resources amidst challenges posed by climate change.

KEYWORDS:

• Future water supply
• best management practices
• Portuguese watershed
• SWAT
• ecological flows

Introduction

Climate change is one of the major threats to human and environmental systems (IPCC, 2021). Changes in precipitation patterns, increasing temperatures, and more frequent extreme weather events can have serious consequences for the quality and availability of drinking water, which is vital for human well-being (Delpla et al., 2009).

The impacts of climate change on water resources are complex and vary by location and region; however, there is a clear evidence that climate change is putting additional pressure on already vulnerable water resources (Leveque et al., 2021; Skaland & Wong, 2022). Higher temperatures and changes in precipitation patterns are increasing water scarcity and the frequency and severity of droughts and floods (Obeysekera et al., 2011; Schewe et al., 2014). Moreover, elevated temperatures augment the mineralization of nitrogen, phosphorus and carbon from soil organic matter. Simultaneously, increased precipitation intensity leads to higher surface runoff and soil erosion, consequently amplifying the transport of inorganic nutrients and other contaminants to surface waters that may compromise drinking water safety (Delpla et al., 2009). Conversely, increased periods of water scarcity can reduce the river dilution capacity and further increase the vulnerability of surface waters to contamination, threatening the safety of water for drinking supply and recreational activities (Gooré Bi et al., 2015; JalliffierVerne et al., 2015). Water contamination can range from excess of inorganic nutrients that feed harmful algal blooms to various types of contaminants, including bacteria, virus and emergent chemical contaminants, such as microplastics and pharmaceuticals (Delpla et al., 2009; Galindo-Miranda et al., 2019). Impaired water quality will then become a growing problem that can impact aquatic ecosystems and limit water uses for drinking, domestic, and recreational purposes (Michalak, 2016). This can lead to substantial economic losses, with an estimation of 4 billion dollars lost in the United States annually, as a result of harmful algal blooms (Kudela et al., 2015).

Agriculture is one of the most significant consumers of water resources, with irrigation accounting for about 70% of all freshwater withdrawals worldwide (Molden et al., 2011). Furthermore, agriculture is one of the main contributors to water pollution due to fertilizer application and inadequate tillage practices that increase soil erosion (Evans et al., 2019). Climate change is expected to exacerbate the contribution of agriculture to water pollution by increasing surface runoff and diffuse pollution. In this context, implementing sustainable agricultural practices that reduce water consumption and pollutants export, will be crucial to the adaptation of the drinking water sector to climate change. Hydrological models play a vital role in anticipating the impact of climate change and predicting the effectiveness of sustainable agricultural practices, as they enable us to simulate the behaviour of the water cycle and evaluate the potential impacts of different scenarios on water availability and quality (Dias et al., 2020).

The impacts of climate change on water quantity have been extensively examined (Joseph et al., 2020; Konapala et al., 2020; Schewe et al., 2014), but its effects on water quality have received less attention (Delpla et al., 2009; Li et al., 2020). Studies have addressed the role of forests and other sustainable practices in drinking water supply and treatment costs (Mulatu et al., 2021; Piaggio & Siikamäki, 2021). However, only a few studies have addressed the combined effect of climate change on both the quantity and quality of drinking water (Qiu et al., 2019).

Regions already struggling with water issues will face increased vulnerability due to climate change. Southern Europe, exemplified by Portugal, faces a unique set of challenges intensified by limited freshwater resources during critical periods (Lionello et al., 2014). This vulnerability necessitates a focused examination of the impacts of climate change on water availability, especially concerning drinking water supply, which has been less explored in previous studies (Carvalho-Santos et al., 2016; Serpa et al., 2017). Recent studies focusing on the impacts of climate change on water resources in Portugal have provided valuable insights. These investigations consistently predict a decrease in river discharge under diverse climate scenarios, emphasizing the need for region-specific studies (Serpa et al., 2015; Stefanova et al., 2015). However, amidst these valuable contributions, a critical gap remains in understanding the specific effects of climate change on water availability for drinking water supply. Despite projections of reduced river discharge (Almeida et al., 2018; Ramião et al., 2023), no comprehensive information exists regarding the direct consequences of climate change on drinking water resources.

In this study, we examined the impacts of climate change on the quantity and quality of surface water for drinking water supply, and the importance of sustainable agricultural practices as an adaptation strategy. We used the Cávado River basin (northwest Portugal) as a case study, one of the most important basins for drinking water supply in Portugal, providing water to a population of more than 700,000 individuals (Ângela Fernandes Machado, 2019). Additionally, this is one of the few studies exploring the combined effect of climate change and sustainable agricultural practices on the quantity and quality of drinking water sources. We hypothesize that climate change will reduce water availability and increase nutrient concentration, especially during the dryer months, and aggravate the impact of drinking water abstraction on the river ecological flow. We also hypothesize that sustainable agricultural practices will be important to mitigate nutrient increments under climate change, however, nutrient concentration may not surpass critical thresholds of drinking water standards. Through this study, we hope to contribute to promote the adoption of sustainable practices as a way of providing safe drinking water and proposing strategies to climate change adaptation.

Materials and methods

Study area

The study was performed in the Cávado River basin (1581 km²), located in north-west Portugal (Figure 1), between the Atlantic and Mediterranean regions, with an average annual precipitation of 1300 mm, and minimum and maximum temperatures of 3°C and 29°C, respectively (data from 1999 to 2018 from two meteorological stations provided by the Portuguese Institute for Sea and Atmosphere). The basin geology is dominated by granite, and the soil type is dominated by Umbric Leptosols and Dydtric Antrosols (Leitão et al., 2013). The upstream area of the basin (Barroso region) is characterized by scrubland, forests (Figure 1), extensive livestock farming and rainfed crops (FAO, 2018). In contrast, downstream lands are dominated by intensive maize production, high water needs, and the application of large amounts of slurry and manure on arable lands (DRAEDM, 2007).

Figure 1. Location of the Cávado River basin, land cover (DGT, 2010), calibration sites, dams (SNIRH), and the location and name of the most important water abstractions for drinking water treatment (i.e., Penide and Ponte do Bico).

Map displaying the geographic extent of the Cávado River basin in Portugal, featuring various geographic elements pertinent to water resource management. The map delineates land cover types, alongside calibration sites and dams. Additionally, prominent water abstraction points for drinking water treatment, notably Penide and Ponte do Bico, are marked. The purpose of the map is to provide a visual overview of the study area and highlight critical features relevant to water resource assessment and management amidst climate change impacts.

There are nine dams in the Cávado River basin, and most are only used for hydropower generation. The hydropower plants in the basin generated 1822 GWh in 2019, which is equivalent to 20% of the total national hydropower generation (FFMS, 2020; EDP, 2019). The Cávado River basin is also critical to supply drinking water for the region, providing water to more than 700,000 citizens from eight municipalities (APA, 2012). Surface water is responsible for 97.7% of the drinking water supply in the Cávado River basin (APA, 2012). The water quantity and quality in the Cávado River is also important to maintain several recreational activities, including swimming and boating along seven inland bathing waters. However, 45% of the river water bodies in the Cávado River basin was lower than good according to the Water Framework Directive, due to diffuse pollution from agriculture, point discharges from urban and industrial activities, and hydromorphological alterations from nine dams (APA, 2016).

Input data and SWAT setup

River discharge and nitrate concentration were simulated using the Soil & Water Assessment Tool (SWAT), version SWAT2012 rev. 670, in ArcSWAT 2012.10_5.21 interface for ArcGis (Winchell et al., 2013). SWAT is a physically based, semi-distributed and continuous timescale hydrological model that runs on a daily time step (Arnold et al., 2013). SWAT model setup and input data used to simulate river discharge and nitrate concentration in the Cávado River basin were previously described in Ramião et al. (2022) and Ramião et al. (2023).

The watershed was delineated using the Shuttle Radar Topography Mission (SRTM) 1 Arc-Second Global (USGS, 2018) as the digital elevation model, and the stream network shapefile of the Water Framework Directive (SNIAmb, 2018) as burn-in streams, to force the model to create the same sub-basins defined in the Water Framework Directive. The hydrological response units were created using a land cover map for mainland Portugal of 2010 (DGT, 2010), a data set of soil ecological value as soil map (Leitão et al., 2013), and three slope classes (i.e., 0–10%, 10–25% and >25%). The reservoirs, point sources and surface water abstractions were included in the model using data from the National Water Resources Information System (SNIRH), the Portuguese Power Company (EDP, 2019), the National Geographic Information System (SNIG), and the Portuguese Environment Agency (APA, 2012).

Parameterization of vegetation and soil was based on a former SWAT study in north-west Portugal (Carvalho-Santos et al., 2016), whereas management operations for each land cover were set based on the literature (APA, 2016; DRAEDM, 2007; FAO, 2018).

Surface runoff was simulated from daily precipitation using the curve number equation method. Evapotranspiration was simulated using the Hargreaves equation, because the Pennan–Monteith equation provided unsatisfactory results.

River discharge and nitrate calibration

River discharge and nitrate export were calibrated near the outlet of the basin (Figure 1), using a semi-automated calibration in SWAT-CUP software version 5.2.1 (Abbaspour et al., 2015). The semi-automated calibration was performed considering 3 iterations, with 200 simulations each, and the Nash–Sutcliffe as the objective function. Calibration was performed on a monthly time step from 1995 to 1997, and the calibrated parameters ranges were applied to validate the model with observed data from 1998 to 2000. The parameters considered for the semi-automated calibration were selected based on a global sensitive analysis of the most sensitive parameters in the literature (Abbaspour et al., 2015, 2007; Shrestha et al., 2016; Molina-Navarro et al., 2017).

SWAT outputs by land cover for leaf area index, evapotranspiration, total biomass, and nitrate export were manually calibrated before the semi-automated calibration in SWAT-CUP, by changing sensitive parameters in the soil (.sol), vegetation (.crop) and management databases (.mgt), and the parameters USLE_K, USLE_P and USLE_C (Panagos, Borrelli, Meusburger et al., 2015; Panagos, Borrelli, Poesen et al., 2015; Panagos et al., 2014), when comparing the SWAT outputs with expected values from the literature (ICNF, 2015; Tereso et al., 2011).

River discharge and nitrate export had already been calibrated in the Cávado River basin. Additional methodological details are available in Ramião et al. (2022) and Ramião et al. (2023).

Climate projections

The impact of climate change on river discharge and nitrate concentration in the Cávado River basin was assessed using the same methodology applied by Ramião et al. (2023). Climate data of daily precipitation (pr), minimum and maximum 2-metre air temperature (tasmin, tasmax), near-surface relative humidity (hurs), surface downwelling shortwave radiation (rsds), and near-surface wind speed (sfcWind), from the EUROCORDEX project at 0.11° resolution, were retrieved from the Earth System Grid Federation for 1976–2005 and 2031–2060, under Representative Concentration Pathway 4.5 (RCP4.5) and Representative Concentration Pathway 8.5 (RCP8.5) scenarios. Climate change impact was examined considering an ensemble of four regional climate models, based on the ability of each regional climate model to simulate precipitation, minimum and maximum temperature in the Cávado River basin, which was tested with statistical goodness-of-fit measures between observed (E-OBS) and simulated values (regional climate models). The hydroGOF package (Zambrano-Bigiarini, 2014) was used to calculate the statistical goodness-of-fit measures, and the climate4R packages (Iturbide et al., 2019) were used to analyse EURO-CORDEX and E-OBS data in R.

EURO-CORDEX data were bias corrected using E-OBS as the reference data set, and employing the linear scaling method considering the difference (additive) between the observed (E-OBS) and simulated (model historical) means in the training period (1976–2005) for temperature, and the quotient (multiplicative) between the observed and simulated means for precipitation.

The carbon dioxide concentration was defined in SWAT according to IPCC (2013) as 330 ppm for the historical period, 460 ppm for RCP4.5, and 489 ppm for RCP8.5.

The Standardized Precipitation–Evapotranspiration Index (SPEI) and the Warm Spell Duration Index (WSDI) were used to examine drought and extreme warm periods under climate change. SPEI is a meteorological drought index that considers the effect of both precipitation and potential evapotranspiration on drought, based on a monthly climatic water balance (precipitation minus potential evapotranspiration), which is adjusted using a three-parameter log–logistic probability distribution, and converted to standard deviations with respect to average values on a range of timescales from 1–48 months (Vicente-Serrano et al., 2012). Positive SPEI values indicate greater than median water balance and negative values indicate less than median water balance (Vicente-Serrano et al., 2010). The SPEI index was calculated for a 6-month period, representing a wet (October–March) and a dry season (April–September), and considering the historical period as the reference period, to test whether the climate will become wetter or drier in the future and the respective magnitude.

The WSDI is one of the climate extremes indices defined by the Expert Team on Climate Change Detection and Indices (Karl et al., 1999), and it refers to the annual count of days with at least 6 consecutive days when maximum temperature > 90th percentile. The SPEI and WSDI indices were calculated in R using the climate4R.climdex packages (Iturbide et al., 2019).

Sustainable agricultural practices

The importance of sustainable agricultural practices (SAPs) in reducing nitrate concentration for drinking water sources was examined considering two tillage operations (conventional versus conservation) and the implementation or not of filter strips. Conservation tillage and the implementation of filter strips were considered sustainable agricultural practices, whereas conventional tillage and no filter strips were the current management practices. These practices were selected based on previous studies assessing the effectiveness of sustainable agricultural practices in the Cávado River basin (Ramião et al., 2022). Filter strips were chosen for their proven effectiveness in depleting nitrate levels, whereas conservation tillage, although conventionally categorized as a sustainable agricultural practice due to its potential benefits, such as reducing soil erosion and improving water infiltration (Liu et al., 2017; Mancuso et al., 2021), had been found to increase nitrate concentration in previous research in the Cávado River basin (Ramião et al., 2022).

The different tillage operations were simulated by changing the mixing efficiency and depth of mixing (Dechmi & Skhiri, 2013; Tuppad et al., 2010), whereas the filter strips were simulated as strips of varying width in all agricultural areas within the riparian zone, according to the Portuguese plan for the Common agricultural policy 2023–2027 of the European Commission (GPP, 2021). In this sense, when the slope was lower than 10% the filter strips were 3 m wide; when the slope was between 11% and 25% the filter strips were 10 m wide; and when the slope was equal or higher than 25% the filter strips were 15 m wide. Additional methodological details can be found in Ramião et al. (2022).

Determination of the river ecological flow

The river ecological flow was determined to examine whether the drinking water abstractions could compromise the ecological flow under current and future climate conditions. It was determined using the most suitable hydrological method to examine the river ecological flow in mainland Portugal, recommended by the national protocol for river ecological flow determination (APA, 2018), which follows the guidelines from the European Commission for Ecological Flows in the Implementation of the Water Framework Directive (European Commission, 2015). A three-tiered hierarchy approach is generally recommended for ecological flow determination, including a hydrological approach, followed by a hydraulic and a holistic approach (APA, 2018). However, hydraulic and holistic approaches are highly resource-demanding, whereas the hydrological approach can be calculated based on SWAT results and it is suitable for regional planning (European Commission, 2015). Therefore, the ecological flow was determined considering only the hydrological approach.

The river ecological flow was determined on a monthly basis for the historical period (i.e., 1976–2005), considering the quartiles of the flow duration curve from daily discharge, with data simulated in SWAT without water abstraction in Penide and Ponte do Bico (i.e., before hydraulic disturbance). From February to June the quartile used to determine the river ecological was Q75, while from July to January it was Q90 (APA, 2018).

Threshold for nitrate concentration in drinking water

The drinking water quality standards of the European Drinking Water Directive (2020/2184/EC) and the guideline for drinking water quality of the World Health Organization (World Health Organization, 2017) established a maximum of 50 mg/l of NO₃ in water intended for human consumption. This threshold was used to examine the influence of climate change and sustainable agricultural practices on the number of days surpassing 50 mg/l of NO₃. The analysis was performed on a daily basis because it is the timescale for water abstraction in Penide and Ponte do Bico. There are nine sites for surface water abstraction in the Cávado River basin; however, only Penide and Ponte de Bico (water treatment plants) were considered because they are responsible for 66% and 27% of the total surface water abstracted, respectively (APA, 2012). Given the predominant reliance on surface water for drinking water supply, with surface water accounting for 97.7% of the total water abstraction in the basin (APA, 2012), our study strategically prioritizes the examination of surface water dynamics.

Results and discussion

SWAT calibration and validation

The SWAT hydrological model had been calibrated and validated for the Cávado River basin in a previous study (Ramião et al., 2022), with an overall favourable performance of the model for river discharge and nitrate export, as indicated by the Nash Sutcliffe efficiency and the coefficient of determination (R²; Moriasi et al., 2015; Table 1).

Table 1. Nash Sutcliffe efficiency, coefficient of determination (R²), P-factor and R-factor for the calibration and validation of river discharge and nitrate export in the Cávado River basin, considering monthly data from 1995–1997 for calibration, and 1998–2000 for validation. The P-factor represents the percentage of observed data enclosed by the 95PPU, and the R-factor is the average width of the 95PPU divided by the standard deviation of corresponding observed data.

	River discharge		Nitrate
	Calibration	Validation	Calibration	Validation
Nash Sutcliffe efficiency	0.72	0.62	0.38	0.44
R^2	0.76	0.73	0.57	0.55
P-factor	0.73	0.55	0.53	0.69
R-factor	0.94	0.75	0.6	0.72

Climate change indices

The Standardized Precipitation–Evapotranspiration Index (SPEI) suggests a lower than median water balance under future climate conditions, mostly from April to September (dry season) under the RCP8.5 scenario (Figure 2). This agrees with the predicted lower precipitation and higher evapotranspiration in the Cávado River basin under climate change, especially during the summer under RCP8.5 (Ramião et al., 2023). Additionally, the Warm Spell Duration Index (WSDI) predicts an increase in the number of days with at least six consecutive days when the maximum temperature >90th percentile (Figure 2). These data suggest lower water availability and an increase in the number of episodes with extreme temperatures, which will probably exacerbate the pressures on the drinking water supply sector by decreasing water availability while increasing its demand.

Figure 2. Change in 30-year average Standardized Precipitation–Evapotranspiration Index (SPEI) and Warm Spell Duration Index (WSDI) (annual count of days with at least 6 consecutive days when the maximum temperature >90th percentile) in 2031–2060 relative to 1976–2005, under RCP4.5 and RCP8.5 scenarios. The SPEI was calculated for a 6-month period, representing a dry (April–September) and a wet season (October–March). Positive SPEI values indicate greater than median water balance and negative values indicate less than median water balance. The boxplots display the dispersion among four regional climate models.

Boxplots illustrating the change in 30-year average Standardized Precipitation–Evapotranspiration Index and Warm Spell Duration Index from 2031 to 2060 relative to 1976–2005, projected under RCP4.5 and RCP8.5 scenarios. Warm Spell Duration Index is higher under RCP4.5 and RCP8.5, whereas Standardized Precipitation–Evapotranspiration Index is negative especially during the dry season.

The effect of climate change and sustainable agricultural practices on river discharge

River discharge is expected to decrease during most of the year under climate change, especially from May to September under RCP8.5, in both Penide and Ponte do Bico (Figure 3). Although previous studies have predicted a decrease in river discharge under climate change (Carvalho-Santos et al., 2016; Serpa et al., 2017), including in the Cávado River basin (Ramião et al., 2023), no previous information was available on the effects of climate change on water availability to drinking water supply. Our study provides novel information for the drinking water sector by showing that water availability at the water abstraction sites will decrease under climate change, especially during the driest months (Figure 3), when water demand is expected to be highest (Leveque et al., 2021). Moreover, we found that the implementation of sustainable agricultural practices is unlikely to have an effect on river discharge (Figure 3). This was surprising because forests play a critical role in regulating water quantity in a watershed, by increasing water infiltration into the soil and recharging groundwater, releasing water into the atmosphere through evapotranspiration, and regulating river discharge by capturing and storing the water during wet periods, and releasing it gradually during dry periods (Filoso et al., 2017). Considering the importance of forests to regulate water quantity, the negligible effects of riparian forests on river discharge under climate change (Figure 3) might be related to limitations of the SWAT model. In fact, the filter strips in SWAT reduce sediments and nutrients but do not affect surface runoff (Arnold et al., 2013), meaning that SWAT is unlikely to capture the effect of filter strips on water balance. An improved algorithm that simulates filter strips as dynamic forests that affect both the quantity and quality of water would improve the usefulness of the SWAT model.

Figure 3. Percentage of change in monthly 30-year average river discharge in 2031–2060 relative to 1976–2005, under RCP4.5 and RCP8.5 scenarios, in Ponte do Bico and Penide water treatment plants (i.e., with water abstraction for drinking water treatment), considering current management practices, filter strips, and conservation tillage. The boxplots display the dispersion among four regional climate models.

Four graphs depict the percentage change in monthly river discharge under different climate change scenarios and management practices. Each graph represents Ponte do Bico and Penide under RCP4.5 and RCP8.5 scenarios, with boxplots showing the 30-year average data for each month and management practice. The results indicate an anticipated decrease in river discharge, especially from May to September under RCP8.5, with minimal impact from sustainable agricultural practices.

The effect of tillage practices on water availability and river discharge depends on several factors, such as the type of tillage, soil type, and climate conditions (Morris et al., 2010). Tillage can increase soil erosion, which leads to a loss of soil and a reduction in its water holding capacity (Van Wie et al., 2013). Consequently, water availability for crops and groundwater recharge is expected to decrease, which contributes to lower river discharge during dry periods (Asmamaw, 2017). On the other hand, tillage practices involving minimal soil disturbance, such as conservation tillage, can improve soil structure and water infiltration, leading to lower runoff and river discharge during wet periods, and increased groundwater recharge and higher river discharge during dry periods (Asmamaw, 2017). However, in our study, tillage practices had no effect on river discharge under climate change (Figure 3), suggesting that despite redistributing nutrients, pesticide and residue in the soil profile, the effect of residue distribution on surface runoff between conventional and conservation tillage, as defined in our study, is negligible. We should also consider that the percentage of agricultural areas upstream Penide and Ponte de Bico is fairly low (Figure 3), and therefore, moving from conventional to conservation tillage may not be sufficient to detect changes in river discharge, as tillage practices mainly affects water quality (Arnold et al., 2013).

The effect of climate change and sustainable agricultural practices on ecological flow

In our study, climate change is predicted to increase the number of months where the river discharge is below the ecological flow after water abstraction in Ponte do Bico and Penide (Figure 4), meaning that water abstraction will occur at the expenses of additional environmental degradation. Ecological flows refer to the quantity and timing of water needed to sustain freshwater ecosystems and the services they provide (APA, 2018). The hydrological regime plays a key role in the structure and functioning of aquatic ecosystems, by influencing pollutants concentration, water temperature, nutrient cycling, oxygen availability, the processes shaping river channels and floodplains, and the dispersion of some aquatic organisms (European Commission, 2015). For instance, prolonged low flows may facilitate the establishment of drought-tolerant or opportunistic invasive species (Rahel & Olden, 2008), and the likelihood of eutrophication in a river system (Charlton et al., 2018). Invasive species can have significant impacts on drinking water abstraction by clogging intake structures, reducing water and increasing facility maintenance costs (Gallardo & Aldridge, 2018). The Cávado River basin is already facing serious problems due to invasive species, especially the water hyacinth (Eichhornia crassipes). Even though there is no direct information on the impact of water hyacinth on drinking water abstraction, it is possible that the dense mats formed by this species could impact water intake structures or reduce water flow, potentially affecting water availability for drinking water purposes. Therefore, it is crucial to recognize the importance of ecological flows and manage them properly for the benefit of the entire ecosystems, especially under future warming conditions.

Figure 4. Number of months (in 30 years) where the river discharge is lower than the ecological flow in Ponte do Bico and Penide water treatment plants, during the historical, RCP45 and RCP85 periods, under current management practices, the implementation of filter strips, and conservation tillage. The boxplots display the dispersion among four regional climate models.

Two graphs illustrate the number of months (in 30 years) where river discharge is below ecological flow after water abstraction at Ponte do Bico and Penide. The left graph represents Ponte do Bico, and the right graph represents Penide. Each graph depicts historical data and projections under RCP4.5 and RCP8.5 scenarios, with boxplots showing the dispersion among four regional climate models. The x-axis denotes the time period, and each period includes boxplots for current management practices, filter strips, and conservation tillage. Results show an increase in the number of months with river discharge below ecological flow due to climate change.

In recent years, several countries have included strategies to ensure an adequate water provision through environmental flows (Harwood et al., 2017; Mezger et al., 2019). The Water Framework Directive (2000/60/EC) already acknowledge the vital role of water quantity and hydromorphological dynamics in supporting freshwater ecosystems and the achievement of environmental objectives, even though it does not explicitly use the term ecological flow (Ramos et al., 2017). Despite the considerable advances made in the scientific and legislative realms concerning ecological flows, there has been little progresses in their implementation (Mezger et al., 2019; Ramos et al., 2017). Our study shows that climate change will exacerbate the need of deliberating and implementing ecological flows, by reducing river discharge (Figure 3) and the vulnerability of ecological flows to water abstraction (Figure 4). In fact, the blueprint to safeguard Europe’s water resources has already stressed the urgent need to better address over-abstraction of water, the second most common pressure on EU ecological status (European Commission, 2015), which can be aggravated under climate change. Therefore, it is vital to implement adaptation measures that could ensure sufficient water supply without compromising freshwater ecosystems and the services they provide.

The sustainable agricultural practices assessed in this study are not expected to mitigate the impact of water abstraction on ecological flow under climate change (Figure 4). However, as discussed in the previous section, the minor effects of sustainable agricultural practices on river discharge might be more related to limitations of the SWAT model. Additionally, measures to mitigate the impact of water abstraction on ecological flow under climate change should be explored in future studies. Among the measures already examined to ensure sufficient drinking water supply that could minimize the impact of water abstraction on ecological flow are: (1) promotion of water infiltration and retention, (2) increase of water storage capacity, (3) awareness raising, and (4) water savings and efficiency (Garnier & Holman, 2019). Infiltration can be increased by implementing agricultural practices that limit soil compaction and reducing soil sealing (Basche & DeLonge, 2019; Tobias et al., 2018). An increase in water storage capacity often relies on the construction of new reservoirs or the improvement of the existing ones (Garnier & Holman, 2019). However, reservoirs have high costs and a large environmental footprint (Schmutz & Moog, 2018). Raising awareness plays a significant role in regulating water consumption because consumer behaviour has a great impact on water demand, which in turn determines the amount of water abstracted and its impact on ecological flow (Garnier & Holman, 2019). Conserving water can be achieved by either reducing the demand for water or improving water efficiency, including the use of more efficient irrigation methods, the use of treated or harvested rainwater for non-potable usages, and the detection and reduction of water losses in the distribution systems (Garnier & Holman, 2019).

The effect of climate change and sustainable agricultural practices on nitrate concentration and nitrate threshold for drinking water

Our study provides evidence that nitrate concentration in surface water is likely to be more affected by the implementation of sustainable agricultural practices than by climate change (Figure 5). This means that sustainable agricultural practices have a large potential to mitigate or exacerbate the impact of climate change on nitrate concentration. In this sense, although nitrate concentration is expected to increase during most of the year under climate change when sustainable agricultural practices are not implemented, the implementation of filter strips can reduce nitrate concentration further below its historical concentrations (Figure 5). Filter strips act as natural buffers, effectively capturing sediments and reducing nutrient runoff into surface waters (Zhang et al., 2010). The intricate network of vegetation within these strips not only enhances water infiltration into the soil but also promotes groundwater recharge, thereby minimizing the transport of pollutants downstream (Mancuso et al., 2021). Although there are studies suggesting that the efficiency of filter strips in removing pollutants will decrease under climate change due to an increase in pollutant loads and lower water residence times (Yang et al., 2019), our study suggests that they might still be crucial to reduce nitrate concentration in drinking water sources.

Figure 5. Percentage of change in monthly 30-year average nitrate concentration in 2031–2060 relative to 1976–2005, under RCP4.5 and RCP8.5 scenarios, considering current management practices, filter strips, and conservation tillage. The boxplots display the dispersion among four regional climate models.

Four graphs illustrate the percentage change in monthly 30-year average nitrate concentration from 2031 to 2060 relative to 1976–2005, under RCP4.5 and RCP8.5 scenarios, considering different management practices. Each graph represents Ponte do Bico and Penide, with boxplots displaying the 30-year average data for each month and management practice. Results indicate higher nitrate concentrations with conservation tillage under climate change conditions at both sites, whereas lower concentrations are observed with the implementation of filter strips under climate change scenarios.

On the other hand, and despite being generally considered as a sustainable agricultural practice (Liu et al., 2017), conservation tillage is expected to increase nitrate concentration under climate change, especially after fertilizer application in May and October (Figure 5). The observed rise in nitrate concentration emphasizes the need for a nuanced understanding of the outcomes associated with conservation tillage. Although it contributes to the reduction of sediment export and enhances water availability for sectors like hydropower and agriculture (Asmamaw, 2017; A. Shrestha et al., 2021), the potential trade-off involves an increase in nutrient concentrations, particularly nitrate, posing challenges to sectors like drinking water supply. The fundamental reason behind this phenomenon lies in the intricate dynamics of nutrient distribution within the soil under conservation tillage (Maguire et al., 2011). The reduced soil disturbance characteristic of conservation tillage can lead to the accumulation of nutrients near the surface, contributing to elevated nitrate concentrations in surface waters (Dell et al., 2012). Our findings underscore the importance of carefully balancing the advantages of conservation tillage with potential challenges related to nutrient concentrations. Notably, incorporating fertilizer into the soil under conservation tillage was found to mitigate the observed increase in nitrate concentrations (Ramião et al., 2022), highlighting the potential for a combined approach to reduce both sediment and nutrient exports under climate change. This insight provides valuable considerations for implementing sustainable water resource management strategies in river basins affected by climate change.

Climate change and sustainable agricultural practices are expected to have a minor effect on the number of days with nitrate concentration above the maximum allowed concentration in drinking water (i.e., 50 mg/l; Figure 6). This is probably related to the very high threshold for nitrate concentration in drinking water, when compared to the typical concentration values in Penide and Ponte de Bico. However, nitrate concentrations in Penide and Ponte do Bico are extremely low compared to other river basins (Bouraoui et al., 2011), meaning that climate change may still threat the quality of drinking water where nitrate concentration is already high or where river discharge is expected to suffer more severe discharge depletions. For instance, river discharge in southern Portugal may decrease by 75%, whereas nitrogen concentrations are expected to increase by 500% (Almeida et al., 2018). Although rivers in southern Europe are expected to face severe water quality problems under climate change (Molina-Navarro et al., 2014; Rocha et al., 2020), our study suggests that nitrate concentration in the Cávado River basin may not be a direct concern for the drinking water sector.

Figure 6. Number of days when nitrate concentration is above 50 mg/l in Ponte do Bico and Penide water treatment plants, during the historical (1976–2005), RCP45 and RCP85 periods, under current management practices, the implementation of filter strips, and conservation tillage. The 50 mg/l of nitrate is the maximum concentration allowed in drinking water (European Drinking Water Directive (2020/2184/EC)). The boxplots display the dispersion among four regional climate models.

Two graphs depict the number of days with nitrate concentration exceeding 50 mg/l at Ponte do Bico and Penide water treatment plants. The left graph represents Ponte do Bico, while the right graph represents Penide. Data cover historical (1976–2005), RCP4.5, and RCP8.5 periods, considering current management practices, filter strips, and conservation tillage. Boxplots illustrate variability among four regional climate models. Results indicate minimal impact of climate change and sustainable agricultural practices on the number of days exceeding the maximum allowed concentration for drinking water.

Despite the minor effect of climate change and sustainable agricultural practices on the number of days above the nitrate threshold for drinking water, their effect will be more pronounced in Penide, where it will increase from 4 days in the historical period to 27 days under RCP8.5, when adopting current management practices (Figure 6). The implementation of filter strips will allow to reduce the number of days above 50 mg/l under RCP8.5 to 21 days, whereas the implementation of conservation tillage will increase the number of days to 33 days (Figure 6). Even though these differences between sustainable agricultural practices are small compared to the time period analysed (i.e., 30 years), they may have consequences for the management and costs at the water treatment plant of Penide. Furthermore, nitrate increments may mirror changes in other quality parameters under climate change, which may be already a concern for the drinking water sector. For example, a high concentration of aluminium is already threatening the drinking water sector in the region and might be exacerbated by lower flows under climate change. Aluminium and other quality parameters were not assessed in this study due to the lack of observed data for calibrating the model. Considering the direct and indirect effects of nitrate on drinking water quality, our study suggests that nitrate concentration may not be a direct concern for the drinking water sector under climate change, but it may have an indirect effect by exacerbating eutrophication and the proliferation of aquatic invasive species due to higher nitrate concentrations and lower river discharge.

Conclusions

In conclusion, our study offers critical insights into the complex interplay between climate change, sustainable agricultural practices, and freshwater resources. As we anticipated, the impact of climate change on the quantity of surface water for drinking water supply is substantial, particularly during the summer months, with river discharge depletions reaching 12%. The projections indicate an increase in the number of months where river discharge falls below ecological flow levels after water abstraction, signifying an increased risk of environmental degradation. Although sustainable agricultural practices assessed in this study may not fully mitigate the impact of water abstraction on ecological flow, it is crucial to explore additional measures, such as the promotion of water infiltration and retention, expansion of water storage capacity, awareness campaigns, and water efficiency initiatives. The examination of nitrate concentration revealed that sustainable agricultural practices have a more pronounced impact than climate change. Under climate change conditions, nitrate concentration may increase by 10%; however, the implementation of filter strips demonstrates a remarkable capacity to decrease nitrate concentration by 48%, underscoring their significant role in preserving water quality. Conversely, conservation tillage, while advantageous for sediment reduction, reveals an adverse effect on nitrate concentrations, potentially increasing it by 95% under climate change. In the context of the maximum allowable nitrate concentration for drinking water, our study suggests a minor effect of climate change and sustainable agricultural practices. The Cávado River basin serves as a microcosm illustrating the intricate balance required between human water needs and ecological preservation. Our study contributes valuable insights to guide future research endeavours and policy initiatives, underscoring the imperative for sustainable, collaborative, and adaptive water resource management.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the Fundação para a Ciência e a Tecnologia [2022.06375.PTDC]. This work was also supported by the project Trees4Water: Tree-based solutions for water quality improvement (2022.06375.PTDC DOI10.54499/2022.06375.PTDC), funded by national funds through the Fundação para a Ciência e a Tecnologia (FCT). This work had further support from FCT, by national funds through the strategic projects “Financiamento Programático” UIDB/04050/2020 (DOI 10.54499/UIDB/04050/2020) awarded to CBMA and LA/P/0069/2020 (doi.org/10.54499/LA/P/0069/2020) awarded to the Associate Laboratory ARNET. José Pedro Ramião was supported by Fundação para a Ciência e a Tecnologia (FCT) (SFRH/BD/141486/2018) and the European Social Fund through the “Programa Operacional Regional do Norte” of the European Commission. Claudia Carvalho-Santos was supported by the “Financiamento Programático” UIDP/04050/2020 funded by national funds through the FCT.