At the beginning of 2023, Uruguay encountered a substantial drinking water shortage, the most significant of its kind in the country's history. An extreme decline in the volume of water in the main dam of the Santa Lucía river prompted OSE to extract water in a location near the Río de la Plata and pump saline water to the potabilization process for a period of six months. Conventional potabilization processes were not enough to remove salts which were consequently delivered to the population. The crisis gave rise to a number of adverse consequences, including the possibility of 'potential' repercussions on the health and economy of the population, particularly among the most vulnerable sectors (Lizbona and Delbono 2024). National authorities made amendments to the regulations to increase the maximum values of different parameters allowed in the water supply (Ferreira et al. 2024). Public demonstrations and congregations were organised to draw attention to the issue of water. Although the primary cause of the crisis was identified as the three years negative anomaly in rainfall due to a severe drought and the relation of climate changes with precipitation (Llopart et al. 2014; Menéndez et al. 2016), other contributing factors were also identified as its determinantss including lack of prevision and adequated infrastructure and use of water with other objectives thant for population drinking water (Ferreira et al. 2024). The preponderance of these factors is attributable to the fact that successive governments have not accorded adequate priority to the issue of potable water (Lizbona and Delbono 2024). The health implications of the crisis were not evaluated, nor were the underlying causes thoroughly analysed.

The objective of this study was to investigate the underlying socio-environmental causes and the subsequent impact on health of the water crisis that occurred in 2023 in the capital city of Uruguay. We evaluated how the physico-chemical parameters reflecting water quality evolved during the six-month crisis period in the water intakes compared with previos years, and evaluated the quality of water distributed to the population (i.e. concentration of chloride, sodium, trihalomethanes) including 15 health facilities. The assessment of potential health risks driven by the exposure to the contaminated water was conducted by estimating the relative proportion to the recommended percentage of sodium intake in the diet of various population groups and comparing the recorded levels of trihalometans (desinfection residuals) with international and national regulations. Finally, the historical spatio-temporal changes in land and water use from 1985 to 2023 in the Santa Lucía river catchment area were analized to evidence the role of changes in land use and its effect on water quantity and quality.

Water Water extracted from the Santa Lucía river is processed for potabilization at the Aguas Corrientes plant and supplied to over 50% of the country's population (1,750,000 people), including the capital city of Uruguay (Montevideo) and surrounding cities and villages (Ríos 2018; OSE 2023) (Fig. 1). The Santa Lucía River, Santa Lucía Grande, is the route by which water reaches Aguas Corrientes, as well as other streams and rivers such as the Santa Lucía Chico, which has the Paso Severino Dam, and the Canelón Grande stream with its dam. Aguas Corrientes has a dam that generates a reservoir with a small storage volume for water intake. The dam is vital to stop water from entering downstream, which might be salty (as it was during the crisis) (Ríos 2018; OSE 2023). The process of purifying surface water in Uruguay, carried out by the national agency (OSE; Obras Sanitarias del Estado), follows the Standard 833:2008 (UNIT 2010; OSE 2012). The treated water is stored in tanks and distributed through the supply network. Monitoring is carried out at different stages of the process to ensure that the water meets potability standards.

Figure 1. Maps showing A) South America, the location of Uruguay, and the Santa Lucía river basin. B) The Santa Lucia river basin, including its main rivers and the contours of the capital city, Montevideo, as well as the locations of the 15 health facilities within the city.

We analysed the temporal dynamics of land use changes by means of Spearman's rank correlations and linear regressions. The temporal dynamics of annual average water level was also evaluated in relation to changes in the area of land use. A linear model and piecewise segmented model were constructed and and evaluated by means of Akaike information criteria (AIC), goodnes of fit and explained variance. The segmented model allowed the estimation of the slopes and intercepts before and after a breakpoint, as well as a confidence interval for the breakpoint and was implemented in the {segmented} package with the R software (Muggeo 2008).

Physicochemical parameters at the intakes of the potabilization plant changed dramatically at the beginning of 2023, due to the mixing of estuarine and fresh water (Fig. 2). An increase in the values of chlorides, sodium, conductivity and TSS was registered. During 2023, total chlorides concentration, mainly composed of sodium chlorides, was ten-fold larger than in previous years (Fig. 3A). These variables were correlated with each other (e.g. chlorides and sodium: Spearman's rank correlation rho = 0.95; N = 487, p < 0.01). High concentration of these variables were observed in the first eight months of the year, and decreased after significant precipitations were recorded in the area (for example chloride in Fig. 2). Distributed water also presented higher conductivity, TSS, total chlorides and sodium during 2023. This was also evident in the water distributed at the taps of health facilities in Montevideo (Fig. 3B). Sodium concentration, varied between 34 and 568 mg/L, being similar among the 15 health facilities surveyed (Fig. 3B).

Figure 2. Temporal change of chloride concentration in raw water and daily accumulated precipitation during the year 2023.

Figure 3. A) Chloride concentration in the raw water at the water intakes of the Aguas Corrientes potabilization plant comparing from year 2019 to year 2023. B) Sodium concentration in tap water at the 15 studied health facilities from 8 June to 21 August. The red dotted line shows the maximum permitted sodium level in drinking water.

The concentration of THM also increased significantly in distributed water during 2023, attaining values larger than in previous years (Fig. 4). This was especially true for bromoform and dibromomethane which was reflected in a significant increase in the claculated index of THM tap water for year 2023. No differences were observed for chloroform and bromodichloromethane individually.

Figure 4. Concentration of A) bromoform, B) concentration of dibromomethane and 3) index of trihalomethanes (THM) in distributed tap water comparing the crisis year (2023) with previous years. The red dotted lines show the maximum permitted values for drinking water for each variable.

The estimated sodium intake and its proportion in relation to the recommended daily doses for different population groups varyed from 10% to 140% (Fig. 5). All groups were exposed to sodium values above the 50% of the recommended dose in the period. Hypertensive adults followed by breastfeeding women were exposed to the highest relative doses, on several occasions exceeding the suggested dietary sodium intake.

Figure 5. Variation in the % of recommended daily sodium intake for different population groups considering the whole studied period and including box plots with median, 25th and 75th percentiles and ranges.

Annual accumulated precipitation varied between 607 and 1,814 mm per year, with minimum values recorded in the years 2008 (607 mm/year), 2022 (713/year), 2023 (740 mm/year), 1989 (768 mm/year), 1987 (786 mm/year) and 2015 (860 mm/year). The last four consecutive years showed values below the total average (Fig. 6A).

A significant change in land use towards anthropogenic activities occurred in the Santa Lucía basin (Fig. 6B). This is evident in the decline of natural vegetation, which fell from 60.8% to 41.4% between 1985 and 2023, while anthropogenic use increased from 39.2% to 58.6% (Spearman's rank correlation rho = -0.999, N = 38, p < 0.001). Currently, anthropogenic activities include urbanisation, forestry, agriculture and grazing, and are widely distributed throughout the catchment area. In contrast, natural areas, including forests, grasslands and wetlands, are predominantly located in the upper regions of the basin or in close proximity to watercourses. These natural areas are also used for extensive livestock farming, which is not included in the anthropogenic land use area.

A positive correlation between the area of anthropogenic activity and the area occupied by surface water (Spearman's rank correlation rho = 0.87, N = 38, p < 0.001), which is indicative of the increase of reservoirs for irrigation was observed (see Fig. 6B). This is also evidenced in 2023, with a total of 28,123 reservoirs registered in the basin, with a combined surface area larger than 5,000 hectares (Fig. 7).

Figure 6. A) Temporal changes annual accumulated precipitation including its average value of the 39 considered years in a dotted grey line and B) annual changes in anthropogenic, natural and water areas.

Figure 7. Reservoirs and dams in the Santa Lucía river basin in year 2023 accounting a total of 28,123 water bodies with a combined surface area larger than 5,000 hectares.

Figure 8. Temporal changes in the Santa Lucía river's average water level in response to the increase in anthropogenic land area. The segmented model fitted to this relationship is shown as a red line and explained in Table 3. Five-year periods are represented by different symbols according to the figure legend, and the number of the years with particularly low water levels are included.

The increase in anthropogenic area was also related to the river's water level in two different stages (Fig. 8). This relation was well represented with two models between water level and anthropogenic land use: a linear model and a segmented model. The segmented model had the lower AIC and therefore was selected to explain how water level changes with increase in anthropogenic land use, it showed that the water level fluctuated around a stable average trend (slope = -0.03) until the anthropogenic area increased by 700,000 hectares. After this breakpoint, the water level significantly decreased with increasing anthropogenic land use (slope = -1.16), which coincided with the last decade (Fig. 8; Table 3). There was a significant negative correlation between the water level and water flow (Spearman's rank correlation rho = 0.732, p < 0.001).

The loss of water quality and quantity directly and indirectly affected health, increasing the population 's risk of illness, generating social, economic and cultural impacts. The observed concentrations of chlorides, sodium, total suspended solids, conductivity, and THMs exceeded the maximum values permitted in national and international standards (UNIT 2010; OSE 2012; World Health Organization 2022) for at least 5 months. This led to amendments in national regulations, increasing the maximum allowable values in chlorides from 250 to 720 mg/l, sodium from 200 to 440 mg/l, total suspended solids from 1000 to 1626 mg/l and conductivity from 2000 to 2981 mg/l. The THM bromoform and the THM index maximum allowed values were also increased from 100 to 350 ug/l and from 1 to 5 respectively (Ferreira et al. 2024). This change to the regulations, made to accommodate unexpected situations, disregards the safety guidelines on which the regulations were based.

Increased salinization of fresh water is a global health issue (Kaushal 2016). Although no health-related guideline values are provided by (World Health Organization 2022) for chlorides, sodium, total suspended solids and conductivity, various studies have shown the negative impact of these substances in drinking water on health (Nahian et al. 2018; Rosinger et al. 2021). The excessive consumption of sodium is associated with the development of high blood pressure, one of the main risk factors for cardiovascular disease. The main source of sodium is the diet and the presence of high salt concentrations in drinking water has been shown also as a source of illness, particularly in coastal and semi-desertic areas (Vineis et al. 2011; Shammi et al. 2019; Kalankesh et al. 2022). In coastal Bangladesh high blood pressure was significantly associated with drinking water salinity, people exposed to saline drinking water had higher chances of developing hypertension than those who consumed fresh water. Women had a 31% higher chance of being hypertensive than men, and people aged 35 years and above were 2.4 times more likely to be hypertensive compared to those below 35 years (Nahian et al. 2018).

During the crisis, the sodium concentration in drinking water frequently exceeded the recommended daily intake, with particularly high levels estimated for individuals already diagnosed with hypertension and breastfeeding mothers (Fig. 5). This issue is particularly salient in the context of the Uruguayan population, as 36.6% of adults aged 25–64 years have high blood pressure and cardiovascular disease is among the leading causes of death (Sandoya et al. 2012; Tinoco et al. 2022; Domenech Carboni and Strozzi Scala 2024). Furthermore, it is estimated that 58% of hypertensive adults remain underdiagnosed (Ministerio de Salud Pública 2013). Paediatric arterial hypertension is also an underdiagnosed disease worldwide, with a known prevalence of 2–3% (Saieh and Lagomarsino 2009). Breastfeeding mothers due to higher requirements of water were also exposed to ilness related to sodium in drinking water. Maternal sodium intake is not expected to directly influence maternal milk (Aumeistere et al. 2020) but might affect the risk of (pre)eclampsia and gestational hypertension (Khan et al. 2014). Further estimation of sodium intake should consider other factors such as the percentege of consumption of bottled water, which would reduce sodium exposure, or the use of boiled water for direct consumption or food preparation, which can further concentrate salts by up to 2.5 times, potentially increasing the associated health risks. The use of boiled water is also a common practice in the preparation of infant milk formulas.

The presence of elevated levels of organic matter in raw water, due to its mixture with the water from Río de la Plata (Tudurí et al. 2018), promoted an increase in chlorination as a step in the potabilization process. Combined with elevated concentrations of bromides present in natural waters of the Río de la Plata, led to a significant increase in brominated THM concentrations, as has been shown for different ecosystems (Ged and Boyer 2014; World Health Organization 2022). THM levels exceeded the national regulatory threshold by 53.3% to 253%, with bromoform concentrations reaching up to 288 µg/L—nearly three times the World Health Organization's guideline value of 100 µg/L (World Health Organization 2022). The effects on health of THMs result either through direct consumption or by inhalation of indoor air largely due to volatilization from drinking-water, inhalation and dermal exposure during showering or bathing. The effects are often due to long-term exposure leading to genetic problems, increased cancer risk and fetal malformations, but there is also evidence that THMs can cause symptoms of respiratory obstruction in a shorter period of time (Evlampidou et al. 2020).

Other important aspects were related to the loss of mental health. Water insecurity has been shown to generate significant physical and psychological impacts on society that have not been considered (Wutich and Ragsdale 2008). The crisis affected differentially the population, with larger impacts on people with less resources (Ferreira et al. 2024). The access to bottled water is restricted to people unable to afford the economic costs which are not able to shift to safe water. This also has a differential impact because health issues are exacerbated by other vulnerabilities, such as a lack of adequate housing and poor nutrition.

During the severe three-year drought caused by an extended La Niña event, the system was unable to provide the necessary water to sustain water treatment in the first half of 2023 (Ferreira et al. 2024). However, this was not the only cause of the water shortage. The significant increase in anthropogenic land and water use led to a tipping point at which the water level and variability reduced significantly prior to the severe drought impacting the basin. This was part of a political decision to allow the construction of reservoirs and dams to provide water for agroindustry particularly agriculture, with the provision of potable water being a secondary objective.

The direct relationship between increase in the area devoted to anthropogenic uses (mainly crops and forestation) and the water area, along with a decrease in natural area was demonstrated here for the last 40 years. After 2010, this sustained change also affected the river water level showing a non-linear response, which was evident even 10 years before the start of the three-years-long drought. The storage of water in reservoirs for irrigation, the presence of several intakes and wells for productive purposes fragmented the water courses, reducing the amount of water available in the natural ecosystem and its availability for potabilization. Also, the increase in water area caused by dams might increase evaporative losses can account up to 15.8% of the storage, posing significant challenges for water allocation and conservation strategies, with notable economic and environmental consequences (Nevermann et al. 2024) including the decreases in water discharge (Biemans et al. 2011). Further, the intense and sustained modification of soils for anthropogenic purposes has resulted in the loss of native vegetation and natural flood zones might also affect regional and local climate (Smith et al. 2023).

The 2023 water crisis was part of a complex system and actions are needed based on interdisdciplinary knoweldege (Du Plessis 2022). Actions must involve a political, interdisciplinary and participative articulation so that the proposals have institutional and civil society support and are transformed into collective actions. The hydrosocial framework should be applied to investigate the causes and consequences of drinking water quality loss as well as to work in the solution (Brand 2007; Boelens et al. 2016). Other factors contributing to the crisis were related to inadequate water management due to lack of sufficient staff and lack of maintenance of OSE's infrastructure that generates large losses of potable water. Previous symptoms and needs arising in other national spheres, such as agriculture, academy and the population (e.g. CNDAV), which in time could have prevented or at least mitigated the state of water emergency reached, were not considered (Llambías 2023). The management of water resources in Uruguay is underpinned by robust legislation and regulations, and the population is mobilised to defend this shared good. However, the current water crisis revealed gaps in the implementation and enforcement of these regulations. It is vital to assess the efficiency of existing laws and consider adjustments to address the shortcomings identified during the water crisis. We also need to understand how inequalities and the extraction of resources affect ecosystems and public health. For these reasons, it is important to combine different types of research to create a complete plan of action. In the future, we may work on projects to share information about the environmental aspects of the crisis and its likely health impacts. We will focus on teaching people how to spot symptoms and assess people's health. This could include monitoring health indicators like blood pressure, especially in areas with limited access to good quality water. Identifying preventive measures is crucial to mitigate future crises.