Integrating Sustainable Development Goals into the nexus approach

The need to ensure the long-term sustainability of societal development and ecosystem services has inspired new approaches to improve the management of natural resources, such as the nexus which focuses on sustainable and intelligent development by identifying and understanding the interrelations among earth resources (Liu et al. 2018).

Despite the increasing interest in the nexus approach, several authors have reported limitations in integrating multiple elements and variables, which is crucial in this approach (Endo et al. 2020). For example, the Nexus Solutions Tool (NEST) (Vinca et al. 2020) analyzes nexus trade-offs by combining a hydrological model with a resource supply planning model at the basin scale. NEST's applicability is difficult due to the data requirements and the complexity of the programming environment. In a more user-friendly platform, but focused on food systems, Sadegh et al. (2020) developed a tool to quantify, at a country level, the amount of water and energy needed for food production, the emissions related to these processes, and the potential use of those crops in bioethanol production.

A review by Simpson & Jewitt (2019) argued that the nexus faces several challenges to be adopted because most studies are focused on macro-scale resource security, ignoring local needs. The authors concluded that it is crucial to link nexus assessments with Sustainable Development Goals (SDGs) to guide policy development and governance, especially in developing countries, such as Brazil. The 17 SDGs undertaken by almost 200 nations are long-term objectives and combine multiple interests, which makes changes to traditional siloed thinking imperative (Estoque 2023).

More than 20 years have passed since the historic global mobilization against poverty, which considered economic development and environmental sustainability (Sachs 2012). In 2000, the Millennium Development Goals were proposed and later replaced by the SDGs. The last assessment report highlighted that the progress achieved is insufficient since many targets are far from being attained or considered worse than the 2015 baseline conditions (UN 2023). To achieve the SDGs, the Brazilian government has developed policies in various sectors. However, the disconnection between these actions and the differences in the amount of investment have had an impact on maintaining inequalities in important sectors, such as access to goods and services (Souza et al. 2019).

Brazil is a key player in the global commodities market. Agribusiness in the country has experienced a vigorous expansion in the last three decades, and its irrigated area is among the top 10 largest in the world. Projections indicate that irrigated areas are expected to increase by 4.2 million hectares between 2019 and 2040 (ANA 2021). In addition, Brazil is highly dependent on hydropower plants: about 52.2% of its total energy supply comes from more than 150 hydropower plants (ONS 2023). Although Brazil has a large availability of freshwater resources, severe droughts have affected the long-term sustainability of ecosystem services in several of its basins (Cunha et al. 2019).

Despite the dependence on water resources in key economic sectors, nexus studies are rare in Brazil and generally focus on specific issues (Dalla Fontana et al. 2020). For instance, the expansion of biofuel production motivated Rodriguez et al. (2018) to analyze the effects of sugarcane and soy plantations on water and land exploitation. Silva & Moraes (2018) combined a steady-state flow ecohydrological model with an economic land use model to simulate the water availability to the São Francisco River in Brazil to better understand water trade-offs for future policymaking and efficient water management. Considering the SDG monitoring, Arcoverde et al. (2023) proposed a participatory methodology for constructing indicators and indexes for the sustainability of biomes. The indicators are mostly based on census data and provide a diagnosis of the current situation, but they are not suitable for projection scenarios. In another study, Ribeiro et al. (2021) adapted an organizational performance indicator to guide stakeholders in the development and implementation of food production policies, indicating paths for achieving the SDG targets.

Located in the Southeast region of Brazil, the Paraíba do Sul Basin (PSB) plays a fundamental role in ensuring the human water needs of 9 million people in Rio de Janeiro's Metropolitan Region through the Guandu water transfer (CEIVAP 2018) and for São Paulo's Metropolitan Region through another diversion to the Cantareira system. Nevertheless, around 24% of its municipalities are located in areas with both quality or quantity water restrictions, and human activities have already impacted 51.3% of the protection zones along the rivers (CEIVAP 2021).

The rapid increase in water demand to satisfy multiple uses in the PSB and the complex interrelationships among those needs require the development of a holistic approach that includes the trade-off among water needs for food, utilities, and the environment in plausible future scenarios. Thus, this study analyzed the anthropogenic pressure on the water resources of the PSB through the nexus perspective by integrating climate, land use, agricultural, industrial, and human projections by 2050. The findings were then used to identify the weaknesses in achieving SDGs and provide guidance to overcome limitations.

In contrast to most studies in Brazil that address nexus linkages over geographic domains shaped by economic, social, and political considerations, the study area of this research is delimited by the river basin, considering that the Brazilian Water Law (Brasil 1997), which defines the basin as the fundamental spatial unit for the development and management of water, land, and related resources. In the literature, few studies have quantified future changes on nexus indicators in relation to the SDGs at the basin scale: for instance, Ma et al. (2024) with focus on food security and ecological restoration and Papadopoulou et al. (2022) in Greece. In Brazil, however, this type of study at the basin scale has yet to be developed.

Based on these principles, this study proposes indicators to represent PSB features related to water, energy, and food security and uses this information to assess the impact of current public policies on water resource availability while considering future climate, land use, and water demand projections. Following Simpson & Jewitt (2019), the selection of indicators was guided by the SDG. Considering the competitive uses of water that combine agricultural, human industrial, and hydropower in a heavily populated area, the PSB offers a unique opportunity for assessing the long-term sustainability of implemented public policies involving natural resources.

The study area includes several hydropower plants with an installed hydropower capacity of 454.16 MW (technical information about the hydropower plants and details about reservoir operational outflow restrictions can be found in Supplementary Table S1). This represents 28.82% of the installed hydropower in the entire PSB according to the Brazilian Electricity Regulatory Agency (ANEEL) data.

The average coverage of the water supply and sewer collection services (collective or individual) is approximately 98.3% (IBGE 2010) and 95.5% (ANA 2017a) of the population, respectively. Nevertheless, 69.2% of the population is served with wastewater treatment (ANA 2017a).

The historical and future land use scenarios adopted in this study were those used by Paiva et al. (2020). The future projections of this study considered three trajectories – optimistic, pessimistic, and business-as-usual – depending on the formulation and enforcement of environmental laws regarding the protection of natural vegetation and the implementation of mitigation policies. Details of the historical and future land use and land cover (LULC) change scenarios are presented in Supplementary Material S2.1.

Based on LULC scenarios and with the public policy targets in accordance with the SDGs, including the goals for the improvement of the water and sewage services established by the new Brazilian sanitation legal framework (Brasil 2020), the water demand scenarios proposed in this study were as follows:

• (a) Optimistic: societal concerns about environmental sustainability lead to investments in educational programs and improvements in technology and infrastructure to promote water conservation and increase the efficiency of the water supply system. All population is connected to the water supply, wastewater collection, and treatment systems.

• (b) Pessimistic: faster economic growth and the intensification of LULC changes in combination with an increase in per capita water consumption and the expansion of sanitation without achieving full-service coverage.

• (c) Business-as-usual: projections based on the current level of water consumption and expansion in water supply and sewage systems according to current rates.

Projections of human and industrial water per capita demand were based on water consumption measured by the São Paulo State Sanitation Company (SABESP) and the National Water and Sanitation Data System (SNIS) (Borges et al. 2022).

All data were clustered according to the per capita water demand of each municipality, and projections were calculated based on each group's historical variation. Then, these per capita projections were multiplied by the population projected for 2030 and 2050, according to the IBGE – Brazilian Institute of Geography & Statistics (2018). The per capita water consumption values obtained for each cluster are presented in Supplementary Table S2.

The total withdrawals of each scenario were computed by summing the losses in the water supply system to the water consumption projections. The percentage of water loss in the optimistic and pessimistic scenarios was based on estimations from the Rio de Janeiro Water Resources State Plan (PERHI – RJ) (COPPETEC 2014). For the business-as-usual scenario, water losses were extracted from the SNIS. The goals of each scenario are listed in Table 1. The indexes used in the business-as-usual scenario, which represent the actual sanitation trend, are presented as the aggregated value for the study region, while the indexes for each municipality are found in Supplementary Table S3.

The agricultural water consumption was focused on the paddy rice demand because it is the main consumer in the study area (ANA 2021). A recent study by Martins et al. (2023) applied the crop productivity model AQUACROP (Raes et al. 2009) to determine paddy rice's water demand under current and future climate conditions, considering the climate change (item 3.2.1) and LULC scenarios (item 3.1.1) applied in this study. Two irrigation strategies were explored: (i) assuming the continued use of actual practices in the region, the continuous ponding scheme was considered in the pessimistic and the business-as-usual scenarios, and (ii) projecting the spread of more efficient irrigation techniques, the intermittent flooding, for which re-flooding is applied when ponded water disappears at the surface and is recommended as a water-saving strategy, was incorporated in the optimistic scenario.

Consequently, to assess the effects of irrigation on water availability, the water fluxes estimated by the crop productivity model AQUACROP from Martins et al. (2023) were included in the hydrological model (item 3.2.2) to calculate the watershed's water balance.

Baseline and future climate projections used in this study are the dynamically downscaled simulations of the Eta Regional Climate Model (RCM) (Chou et al. 2014), with 20 km spatial resolution, further refined to a 5 km grid using the bilinear transformation, for two representative concentration pathways (RCPs) for carbon emissions, RCP 4.5 and 8.5, forced by three global climate models from the CMIP5: the Model for Interdisciplinary Research on Climate – version 5 (MIROC5), the Hadley Global Environment Model 2 – Earth System (HadGem2-ES), and the Canadian Earth System Model (CANESM2).

Simulations were subdivided into three periods (time slices), in agreement with Paiva et al. (2024): baseline (1990–2015), F1 (2016–2035), and F2 (2036–2055).

Future land and water uses (human, industrial, and agricultural) were combined into the climate projections produced by the three downscaled global climate models, considering the two RCP scenarios of carbon emissions. This combination resulted in 12 possible future scenarios for each climate model, which was used as the input data of the hydrologic model simulations.

The MHD-INPE (Rodriguez & Tomasella 2016) distributed the hydrological model was used to quantify the multiple water uses. The MHD-INPE is a regular grid-based hydrological model that has been used for several applications in Brazil such as climate change studies (Von Randow et al. 2019) and land use change studies (Oliveira et al. 2022).

The MHD-INPE model was also used in the PSB by Paiva et al. (2024) to assess how future climate and land use change scenarios might affect water availability and hydropower, without taking into account the influence of public policies. Additional details of the application and overall performance of the model can be found in Paiva et al. (2024).

Paddy rice water demand generated by the crop productivity model (AQUACROP) was included in the hydrological model (MHD-INPE). The AQUACROP model findings refer to the daily water balance with soil evaporation, plant transpiration, surface runoff, and deep drainage for each grid. To achieve this, the AQUACROP water balance simulations for each grid were assimilated into the corresponding regular grid cells of the MHD-INPE hydrological model, considering the fraction of planted cells occupied by paddy rice. These fluxes were aggregated considering the other uses within the cells.

To analyze energy production in the reservoirs, the concept of firm energy, which is the maximum energy produced by a hydropower plant during the most severe drought in the historical period, was used (ANEEL 2005). To calculate firm energy, it is necessary to estimate the regularized flow, which indicates the outflow to be continuously released by the reservoir during the entire historical period (Studart & Campos 2001). This calculation assumed a 5% probability of failure, which is the tolerable level of risk adopted by the national electric system. The firm energy corresponds to the power output when the regularized flow is released through the turbines. Details on reservoir operation and energy generation are provided in Supplementary Material S2.3.

To estimate future energy consumption within the PSB basin, the database of companies responsible for energy distribution in the region was used (ANEEL 2021). Projections were calculated according to historical per capita electric power consumption, considering an increase in power demand in the pessimistic scenario; decreased consumption in the optimistic scenario as a result of an improvement in the efficient use, in combination with residential and industrial solar generation; and maintaining the current growth rate in the business-as-usual scenario. Supplementary Table S4 shows per capita projections for each cluster and the three scenarios.

The nexus approach considers multiple links among different variables and stakeholders who often have distinct competing interests (Ringler et al. 2013). To represent this diversity of information, indicators that express the water, energy, and food security conditions were used and divided into two aspects: quality and quantity. Three SDGs were used to guide the selection of the indicators: SDG 2 (Zero hunger), SDG 6 (Clean water and sanitation), and SDG 7 (Affordable and clean energy).

In Brazil, a deficit in sewage collection and treatment is the main factor responsible for the substantial pollutant load released into rivers (ANA 2017a). Although agricultural activity can also affect water quality, research carried out in the study area showed that the impact of irrigated areas on water quality in the study site is limited since the organic load is low (Andrade et al. 2010). The food quality indicator (FQI) was estimated based on the city's organic sewage load diluted in the river, specifically the biochemical oxygen demand (BOD). This parameter is an indication of the quality of the water body, and it is also used in irrigation, especially for raw consumed crops.

The Eta-MIROC5 future scenarios were slightly different from the baseline period. In the case of the RCP 4.5 emission scenarios, outflows were out of the limits for a smaller percentage of time than in the baseline period, except for the Jaguari and Paraibuna reservoirs in the second time slice (2036–2055), which showed an increase in the percentage. In the RCP 8.5 scenario, except for the Paraibuna Reservoir, the reservoirs remain for a longer time than in the baseline period out of the operating limits, during the second period.

For the Eta-HADGEM2-ES model scenarios, a significant increase was observed in the percentage of time when outflows are out of the limits in all reservoirs, except Paraibuna, especially during the first time slice (2016–2035).

In the second time slice (2036–2055) in RCP 8.5, the Eta-MIROC5 projections indicated a lower percentage of failure in achieving the water target than in Eta-CANESM2 projections in most reservoirs (Figure 4).

The BOD concentration showed significant differences among the pessimistic, optimistic, and business-as-usual scenarios. There were differences in sewage collection and treatment among the trajectories (Table 1). The highest percentage of time, when the BOD concentration exceeded the class II limit, was in the business-as-usual scenario.

Despite the differences among climate model projections, most scenarios indicate a reduction in the availability of water resources, with impacts on the security of multiple resources and extreme events, such as river floods and droughts. Moreover, the most important driver of quantitative changes in water resource availability was climate change, whereas the adoption of policies to improve water management had a minor impact on the results.

A potential explanation for these results is that land transformations in the study area have occurred since the sixteenth century when the Europeans arrived (Carriello et al. 2016). Although urban regions hold 90% of the study area's population (IBGE 2010), migration from rural to urban areas is an ongoing process, which allows forest recovery in abandoned agricultural and pastureland areas (Silva et al. 2017). However, considering that most land uses are permanent, particularly in urban areas, there is almost no room for major changes in future trajectories.

The nexus indicators proposed in this study suggest that the increase in future water variability can trigger water, energy, and food supply failures in the PSB, which could be detrimental to progress on SDGs 6, 7, and 2, respectively. In terms of hydropower, the results of this study are of the same order as those reported by Paiva et al. (2024). Because this study includes projections of human and industrial consumption based on current trajectories of per capita consumption for different future scenarios (Table 1), together with future rice irrigation demand, which were not considered in Paiva et al. (2024), it is possible to conclude that the relative contribution of these components is not significant. Thus, considering SDG 6, universal access to drinking water and increasing the efficiency of water use by reducing water losses in distribution systems and adopting an intermittent irrigation technique have not impacted future water availability. Results indicated that the percentage of time when water demand is not met in the optimist scenario varies by only 5% compared with the pessimist one. Consequently, even if policies for water use efficiency are implemented, additional adaptation measures will be required.

This result was consistent with those of Marques et al. (2022) for the lower PSB. That is, the reduction in water consumption – predicted even for the optimistic scenarios – was not effective in ensuring the balance between demand and supply, particularly in the water transfer to the metropolitan area of RJ (Guandu System).

Regarding energy security, more than half of the scenarios indicated an increase in the percentage of time in which hydropower was lower than firm energy, mainly in 2036–2055. In addition, hydroelectricity, which accounts for only 19% of the basin's total consumption in the baseline period, is expected to be further reduced to an average of 15% of the future power demand. This trend restricts the reliability and efficiency of the energy systems and also undermines broad access to energy, as considered in SDG 7. Impacts of energy production are explained by the fact that the reservoir operational rules prioritize human and industrial demands, and the need to maintain outflows within maximum and minimum limits to ensure the maintenance of the ecosystem services and flood control.

Fulfilling these multiple uses may become even more difficult in the future, as the reservoirs are more likely to operate out of their outflow limits. In the case of the Jaguari reservoir, where the diversion to the Cantareira system is located, the frequency of this type of event increased to 85% of the time in the worst-case climate change scenarios compared with the baseline period.

Brazil's power generation and transmission system is interconnected and mostly comprising hydropower plants, which allows the transfer of energy among Brazil's regions, taking advantage of the difference among hydrological regimes. Consequently, a future reduction in energy generation in the PSB, as suggested by the nexus indicators, might be compensated by the energy contribution of other basins.

However, because several studies have suggested that most hydropower plants in Brazil are likely to be negatively affected, mainly in the North and Northeast basins (MCTI 2021), developing other sources of renewable energy is necessary to adapt to adverse scenarios.

In this context, an attractive alternative for expanding the clean energy supply infrastructure, related to SDG 7, is the distributed generation based on solar energy (installed capacity below 5 MW), which has increased in the last decade: in 2015, there was 0.06 MW installed capacity of solar energy in the study area, and that number reached 122.56 MW in 2022. The scenarios proposed by the Ministry of Mines and Energy (MME 2020) indicate that the installed capacity in the study area might achieve 73.36 MW by 2030. Therefore, combining different power generation sources and synergistic rules can ensure long-term energy security. Moreover, under critical scenarios, the operational rules of reservoirs can be adapted to reduce the risk of water supply failure rather than to maximize hydropower.

Regarding water quality, future scenarios, either pessimistic or optimistic, indicated lower BOD values than the business-as-usual scenario. While this scenario maintains the current levels of sanitation coverage services, the two projected scenarios reflect the government's commitment to the SDGs, which assume broader coverage than the business-as-usual levels do. Thus, the indicator revealed that the actual level of sanitation led to a continuous decline in water quality. For instance, in the worst-case climate change scenario, BOD values exceeded the quality standards for 90% of the time. In the business-as-usual scenario, 11 of the 42 cities whose sewage was collected in the basin area removed less than 60% of the BOD (Brasil 2011), and in 19 of them, more than 50% of the population did not have their sewage treated. This lack of adequate sanitation services highlights the need to implement sewage collection and proper treatment for all. This is imperative to ensure access to water, especially in times of water scarcity (SDG 6). The current situation affects activities that use raw water, such as agriculture, and increases the possibility of food contamination (SDG 2).

This study presents a broader perspective on the future impacts of climate and land use changes on water, energy, and food security in the upper and middle PSB than has been presented in the literature. It identifies trends of future water availability by applying indicators that consider multiple interests and resource security issues. The use of the nexus approach considering the water, energy, and food components, which are closely related to SDGs 2, 6, and 7, makes it possible to identify what are the main impacts of future changes in water availability on those goals, indicating which policies should be prioritized by water managers.

The scenarios revealed a decline in water availability, regardless of the public policies and restrictions on land use and cover changes, especially in scenarios with high carbon emissions (RCP 8.5). In at least one scenario of all the sub-basins, an increase in the frequency of failures to meet water demand is observed compared with the baseline period. However, in the basins with water transfers, Jaguari and Santa Cecília, the average frequency of this event is 39%, while in the other basins, it is only 11%. The limited impact of the public policies analyzed may be related to the extended period of human occupation and the high levels of access to water and energy, which restrain their expansion in future scenarios. Regarding the agricultural sector, which has the highest water demand, future scenarios indicate low impacts due to the adoption of a more sustainable irrigation technology.

Besides the higher variability in water discharges in future scenarios, water quality is likely to be affected during long periods of low discharges. Therefore, improving sanitation service coverage in the basin is imperative in terms of water supply, sewage collection, and treatment. There are also opportunities for improvement in reducing water loss in distribution systems, improving water treatment efficiency, implementing new rules for water reuse, and increasing the variability of the renewable energy source. All these actions are crucial for ensuring the preservation of the basin's water quality under future water scarcity scenarios. The comparison between the business-as-usual and optimistic future scenarios shows that, in the worst-case climate change scenario, improving access to sewage collection and the quality of treatment could reduce the frequency at which BOD values exceed recommended standards by up to 85% of the time.

It is clear that resource management policies need to be diversified to minimize the adverse impacts of land use and climate change.

Among the limitations of this study, it is worth mentioning that projections of future human and industrial demand are based on trends of the last two decades. However, water resource demand depends on a variety of uncertainties, related to the cultural and economic drivers. Besides this, potential synergies of hydropower with solar generation, which has experienced a rapid expansion in the basin, were not considered in this study. Finally, although the MHD-INPE simulates the hydrological processes in each cell of the grid that represents the study area, the results are aggregated according to the sub-basins. Locating water treatment plants, both for potable and wastewater treatments, within the correspondent model grid cell, can provide a more realistic picture of water quantity and quality.

E.A.C.: conceptualization, model simulation, writing − original draft; A.C.E.P.: development of land use and land cover change maps, calibration and validation of the hydrological model, and final review; M.A.M.: quantification of crop's water requirements and final review; D.M.S.: water consumption data provision; V.M.: conceptualization and final review; and J.T.: conceptualization, code improvement, result assessment, and manuscript review.

This work was supported by the Higher Education Personnel Improvement Coordination – CAPES, Brazil, financing Code 001 and the National Council for Scientific and Technological Development – CNPq, Brazil, under Grant Agreement Codes 306846/2017-9, 304695/2020-3, and 428995/2018-7.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.