MainWater affordability is a growing challenge in high-income countries across the globe1,2,3,4,5. Over the past two decades, water rates have risen three times faster than inflation6,7, driven primarily by deferred maintenance and ageing infrastructure. If historical trends continue, more than one-third of US households could face unaffordable water bills within a decade1. However, emerging contaminants requiring advanced treatment and climate change necessitating new supplies are compounding longstanding cost pressures for utilities1,8,9. Unaffordable water can limit access for drinking and hygiene2, force difficult trade-offs with other necessities such as healthcare3,10 and pose negative health impacts11, disproportionately impacting low-income and minority communities2.In many cities, climate change challenges utilities’ ability to provide reliable, affordable water access12,13,14 by straining supplies12,13 and driving demands higher as temperatures rise15,16. On the supply side, climate change can alter the frequency, severity and duration of droughts17,18 and cause drier baseline conditions14,19. Utilities respond by either reducing demands or expanding supplies20, often through costly infrastructure projects that increase rates for customers8,21. Utilities manage demands using curtailment measures during droughts or water-use efficiency programmes, increasing water rates through surcharges needed to pay for lost volumetric revenue or investments in efficiency upgrades and monitoring equipment20,22.Recent studies document growing affordability challenges1,2,4 but offer limited insight into future trends. Water affordability projections typically extrapolate historical trends but without analysing specific cost drivers such as climate that may change in the future1. In addition, social and institutional factors such as demand patterns and rate structures affect affordability23, but estimates typically neglect these factors5,24. One study highlights cost–demand interactions, demonstrating how low-income rate assistance programmes, which decrease household water costs, can increase use7. In addition, most empirical studies assess affordability at one point in time3,5, neglecting seasonal and temporal variability that may exacerbate short-term challenges4,23,25,26. High-quality projections of future water affordability challenges are essential to developing effective and sustainable policy solutions at the city, state and federal levels.Drought and climate change may prompt costly water infrastructure investments, but their affordability impacts are poorly understood. Extensive research assesses water supply infrastructure needs due to climate change, but focuses on utility-scale costs and performance27,28,29,30 rather than household-scale affordability31. In addition, infrastructure financing and rate design are typically analysed from a utility rather than household cost perspective31,32. Recent work advances understanding of climate-related infrastructure impacts on low-income households but does not consider demand feedbacks or financing mechanisms33. Other work models affordability impacts of rate increases due to droughts but focuses on short-term impacts, not climate change22,34. To our knowledge, no study has comprehensively assessed the integrated climate, infrastructure and demand drivers to quantify climate change impacts on water affordability.Our research objective is to quantify the impact of climate change on urban water affordability. We develop a rigorous city-scale modelling framework that captures the integrated climate, utility adaptation decisions and demand drivers shaping household affordability burdens in US cities. We combine a hydrological and water supply system model with infrastructure financing and rate design models, and an econometric model of household water demands. We assess the affordability impacts of utility adaptation decisions needed to maintain reliable water supply in future climates. This can include new infrastructure investments, financing mechanisms, rate structures, assistance, curtailment policies and efficiency programmes. We hypothesize that climate change could create water affordability hotspots, where existing affordability challenges are exacerbated in cities with water supply vulnerability and limited demand management opportunities.Here we complete a city-scale assessment of climate change impacts on water affordability using Santa Cruz, California as a case study and focusing on mid-century impacts. Santa Cruz has low residential water use owing to the city’s previous drought experiences35, limiting the scope for low-cost adaptation options such as demand reduction via curtailment or efficiency policies. It is also reliant on local surface water, making it vulnerable to drought36. By examining a city that has largely exhausted lower-cost options for climate adaptation such as demand-side management, we focus on costly, long-term infrastructure decisions and highlight challenges probably faced by other cities with similar climate, behavioural and regulatory conditions.ResultsAffordability implications of climate adaptationWe present a modelling framework to quantify the water affordability impacts of climate change and related adaptation measures on water supply and demand. Our model captures the interacting climate, utility adaptation decisions, financing and demand drivers that determine household affordability (Fig. 1a). First, we input future climate scenarios, developed by combining a climate model ensemble with a stochastic weather generator, into water supply systems models, which simulate river flows at the reservoir and diversions, and water allocations from the diversions to the water treatment plants. Second, we model utility adaptation decisions using a risk-of-failure (ROF) approach in which utilities take action in response to declining storage levels. Third, we assess the water rate impacts of those interventions using an infrastructure financing and rate design model. Using Santa Cruz as a case study, we focus on utility decision-making around infrastructure investments, evaluating multiple planning scenarios that determine when and what new infrastructure is developed. Fourth, we simulate household water use patterns as a function of demographics, climate, water costs and housing characteristics, capturing dynamic feedback between household behaviours, water supplies and utility decisions. For example, when the city builds new infrastructure, they pay for that infrastructure by increasing water rates; when water bills rise, households use less water, altering how much water the system needs. Finally, our framework results in estimated water bill and affordability burdens (the percentage of household income spent on water bills) across income groups.Fig. 1: Model framework and example simulations.Full size imagea, Affordability modelling framework, which integrates the water supply system with infrastructure financing, rate design and household water demand patterns. The boxes shaded in yellow indicate utility decisions, with patterns indicating decision variables assessed in this work (solid) and parameters chosen to match the case study and then tested in the sensitivity analysis (stippling). Blue boxes indicate model objectives and grey boxes indicate exogenous and endogenous parameters. Capex, capital cost; Opex, operational cost. b, An illustrative example to visualize the mechanisms through which infrastructure development impacts affordability, including reservoir storage levels under sample simulated dry, hot and moderate, cool climates and ROF values for the dry, hot scenario. c, Sample low- and high-income annual household affordability by climate scenario. We highlight the period from 2020 to 2050 for illustrative purposes, since all climate simulations are stationary. Households were randomly sampled from those below the poverty line and those in the upper quartile for income. The dashed vertical lines indicate infrastructure planning and deployment for the dry, hot climate simulation conditional on reservoir storage, total water demands and previously deployed infrastructure. The time between infrastructure planning and deployment shows the construction time, which is incorporated into our modelling framework.Using this framework, we assess how different climate scenarios drive affordability impacts. Figure 1b,c shows two illustrative, contrasting climate simulations selected to visualize the mechanisms through which infrastructure development affects affordability for two sample households. In one plausible moderate, cool climate simulation, no new infrastructure is needed, although the low-income household is already spending more than the 2.5% affordability threshold recommended by the US Environmental Protection Agency (EPA)1,3,37 (Fig. 1c). In one plausible dry, hot climate simulation, declining reservoir storage levels trigger construction of a four million gallon per day (MGD) desalination plant, leading to rate increases and higher water bills. In this example, the affordability burden for a sample low-income household increases from 4% to 6%, while a sample high-income household has negligible impacts. These examples demonstrate how climate stress can exacerbate underlying affordability burdens for low-income households.Unaffordable water under a drier climateNext, we present aggregate results across many stochastic simulations. In Santa Cruz, our scenarios show that climate change could nearly double water bills, which could leave an additional 7–16% of households with unaffordable water (Fig. 2). We test four plausible scenarios that combine contrasting climates and infrastructure planning adaptation strategies: first, our ‘baseline’ scenario with no new infrastructure and a moderate, cool climate, similar to the present-day; second, a ‘moderate climate with adaptation’ in which new infrastructure is built as needed under the same moderate, cool climate; third, a ‘dry climate with adaptation’, in which new infrastructure is built as needed under a dry, hot climate; and fourth, an ‘all climate simulations’ scenario, with the entire range of climate simulations. We choose the climate scenarios to be contrasting but within the range of CMIP6 (Supplementary Fig. 1). For each climate, we develop 50-year stochastic climate simulations, developed using precipitation and temperature changes in the CMIP6 ensemble to inform a stochastic weather generator. In the scenarios with adaptation, new infrastructure is deployed as needed using a ROF approach, in which the utility develops new infrastructure when the projected risk of storage falling below a critical threshold within 2 years exceeds a defined level (Methods; Supplementary Fig. 2). The planning strategy presented here first builds a desalination plant, although alternative strategies are compared later. We analyse the periods where infrastructure investments lead to the largest rate increases, using these periods to estimate monthly reservoir storage, added water supply costs for new infrastructure, monthly water bills and affordability burdens across households, resulting in a distribution of outcomes for each scenario.Fig. 2: Projected climate change and adaptation impacts on household affordability.Full size imagea–d, Cumulative distribution functions (CDFs) across four projected climate scenarios, each with multiple stochastic climate simulations: a baseline similar to the present-day, a moderate climate with adaptation, a dry climate with adaptation and all climate simulations for total city reservoir storage (a), new utility supply costs (b), household water bills (c) and household affordability ratios (d). Horizontal axes in c and d are truncated for visual clarity.We assess the impact of climate change-driven infrastructure investments on household bills and affordability burdens in the absence of policy interventions. In the moderate climate with adaptation, building more infrastructure leads to greater water availability via higher reservoir storage compared with the baseline scenario (Fig. 2a). By contrast, the dry climate with adaptation results in decreased and more variable reservoir storage. Low reservoir levels risk supply shortfalls, triggering new infrastructure investments. Differences in reservoir storage across scenarios with adaptation drive varying supply costs, impacting household water bills (Fig. 2b). In the moderate climate with adaptation, 60% of months require no new supply costs, whereas in the dry climate with adaptation, over 50% of months require over US$1 M in additional costs to maintain reliable supply. As we assume these costs are fully passed on to households without policy intervention, infrastructure investments translate directly into substantial increases in household water bills (Fig. 2c). The 50th percentile bills increase from US$64 to US$80 (moderate climate with adaptation) or US$120 (dry climate with adaptation). The 80th percentile bills rise from US$100 to US$148 (moderate climate with adaptation) or US$204 (dry climate with adaptation). Under the dry climate with adaptation, median water bills could nearly double from current levels.Rising bills increase the proportion of households paying more than the EPA’s affordability threshold (Fig. 2d). Currently, 19% of households exceed this threshold, highlighting the existing affordability challenges in Santa Cruz. This share could rise to 26% under a moderate climate and to 35% under a dry climate, when additional infrastructure is built for reliability. In the dry climate more than one-third of households in Santa Cruz could struggle to afford water, highlighting the scale of potential impacts without any additional policy interventions.Low-income unaffordability under climate changeNext, we analyse how simulated demands, bills and affordability trends for low-income households compare with the remaining population in Santa Cruz, finding that while low-income demands and bills are lower, their affordability burdens are substantially higher (Fig. 3). First, we find that water demand differences across income groups and scenarios are small. Across the scenarios, average low-income water use is about 0.4 hundred cubic feet (ccf) less than other households, with 80th percentile demands increasing to around 0.6 ccf lower than all other households (Fig. 3a). Small demand differences across income groups align with previous work that finds widespread demand hardening in Santa Cruz due to household efficiency and conservation behaviours36. Under the moderate climate with adaptation, building infrastructure decreases demands by about 0.5 ccf, primarily due to increased water prices and household price elasticity responses. Between the moderate and dry climates with adaptation, differences in demands are negligible (