Climate hazards and water insecurity, by geographic and economic vulnerabilities

Climate hazards and water insecurity, by geographic and economic vulnerabilities | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Climate hazards and water insecurity, by geographic and economic vulnerabilities Indira Bose, Edward Frongillo, Claire Dooley, Thalia Sparling, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7959507/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Climate hazards (floods, storms, droughts, and rainfall anomalies) are intensifying, threatening national water security. Individual water security, however, likely varies by exposure to climate hazards and people’s geographic and economic vulnerabilities. We spatiotemporally linked climate hazards with nationally representative survey data on individual water insecurity experiences from 29 low- and middle-income countries (n=36,342). We estimated associations between each hazard and the probability of experiencing water insecurity, and examined heterogeneity by vulnerability (climate zone, urbanicity, subnational wealth and individual income). Floods, droughts and storms were associated with higher water insecurity, rainfall anomalies were not. Geographic and economic vulnerabilities modified these associations, but the directionality varied by hazard type. For example, floods and droughts were associated with higher water insecurity in rural areas, but storms were associated with higher water insecurity in urban areas. Water insecurity experiences make evident the heterogenous toll of climate hazards, and can guide appropriate climate adaptation strategies. Health sciences/Risk factors Earth and environmental sciences/Environmental social sciences/Climate-change impacts/Environmental health Earth and environmental sciences/Natural hazards Earth and environmental sciences/Hydrology Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 1. Introduction Climate change is exacerbating the intensity and frequency of extreme climate events (1). Climate events, e.g. floods, droughts, and rainfall anomalies (i.e., higher or lower rainfall than the long-term average), threaten physical water supplies at the national and watershed level (2). Water security at the individual level does not always align with the physical water supply in the area in which an individual resides (3). Individual water security hinges on opportunities to obtain water that is reliable, safe, affordable and sufficient for their basic domestic needs (4), which is shaped by water acquisition systems and management (5). Climate hazards may disrupt these opportunities, reshaping lived experiences of water security. The risks posed by climate change have been conceptualised by the working group of the Sixth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC), as a dynamic interaction between climate-related hazards (the potential occurrence of a climate event) with the exposure and vulnerability of the affected population (6). Whether a climate hazard results in water insecurity for an individual depends on both their exposure to the event and their vulnerability ( Figure 1 ). Figure 1: The IPCC’s conceptualization of the dynamic interaction between climate hazard, exposure and vulnerability (6) , adapted to the risk of experiencing water insecurity . Italicized text indicates how these concepts have been operationalized in these analyses. Figure 2: Conceptual framework showing how climate hazards may contribute to individual water insecurity. Exposure pathways (blue) describe the ways that climate hazards may affect the different dimensions of water insecurity (7). The geographic and economic factors (red) may modify the vulnerability of individuals to water insecurity. Rounded boxes outlined in black are concepts that have been measured; square boxes indicate hypotheses. The pathways linking climate hazards to water security are numerous ( Figure 2 ). For example, floods and heavy rainfall events may reduce water acceptability and use due to pathogen run-off, or destroy water infrastructure, thereby limiting availability of sufficient safe water (8-11). Prolonged droughts may reduce water availability, result in increased competition for limited resources reducing accessibility, or concentrate pathogens in water sources, reducing water quality (8-10). Climate events may also have economic impacts on the ability to buy or treat water; these events may also influence the psychological manifestations of water insecurity by causing anxiety about water availability or safety, in turn leading to behavioural changes around water use (10, 12, 13). Many factors can modify the vulnerability of an individual to experiencing water insecurity when exposed to a climate event ( Figure 2 ). Geographic factors such as climate zone, may modify the effect of climate hazards. People living in arid areas, which have higher physical water scarcity (14), may be at higher risk of experiencing water insecurity following a climate event compared to settings where there is higher groundwater storage and recharge. As for urbanicity, in many countries there are large disparities in water infrastructure between rural and urban settings, with rural areas typically lagging behind urban areas in access to improved water sources (15, 16). Rural communities may also face disproportionately large challenges because unimproved water sources such as surface water sources are often less resilient to climate events (17, 18). Rural communities are also more likely to have livelihoods dependent on agriculture, and therefore may have to make trade-offs between water for domestic consumption and crops and livestock when facing shortages (13). Additionally, because economic resources often vary within countries, there is often differential access to resilient infrastructure and water management and governance practices. This, in turn, modifies vulnerability to the effects of climate hazards and water insecurity (19-21). At the individual level, economic status can also result in disparities in water insecurity (22-24). For example, wealthier individuals may have more capacity to buffer their water access when faced with a climate event. Little is known about how climate hazards affect individual water security, and how this varies by vulnerabilities. Therefore, we investigated the associations between climate hazards and water insecurity, and if these relationships were modified by geographic and economic vulnerabilities. We first tested the hypothesis that climate hazards (reported flood, storms and drought events and rainfall anomalies) are associated with greater relative risk of experiencing individual water insecurity ( Figure 2 , H1). We then investigated the heterogeneity of these associations by geographic factors, specifically climate zones and urbanicity (H2). We hypothesized that there would be greater risk of water insecurity in arid zones and rural areas. Third, we hypothesized that wealth would modify these associations (H3), with greater risk of water insecurity among individuals living in areas of lower subnational wealth and among those with lower incomes. We tested these hypotheses using nationally representative, georeferenced survey data from 29 low- and middle-income countries (Figure S1). Individuals’ level of water insecurity was classified using a cross-country equivalent measure of experiences of water insecurity, the Individual Water Insecurity Experiences (IWISE) Scale (25). We spatiotemporally linked IWISE data with reported climate events and satellite-derived data on rainfall anomalies in the prior 12 months. Reported climate event data (floods, storms and droughts) were extracted from the International Disaster Database (EM-DAT) (26), and rainfall anomalies were based on the Standardised Precipitation Index (SPI) extracted from the Climate Engine (27, 28); individuals were classified as being exposed or not exposed to these hazards based on their geolocations. Firstly, to investigate the association between climate hazards and water insecurity ( Figure 2, H1), we ran multinomial regression models, adjusting for key covariates. Water insecurity was assessed categorically; individuals were classified as experiencing no-to marginal, low, moderate and high water insecurity. To test the hypotheses that geographic and economic vulnerabilities modified these associations, we ran stratified logistic regression models, in which we operationalised water insecurity as a binary outcome (“water insecure” experiencing moderate-to-high water insecurity vs “water secure” experiencing no-to-low water insecurity). (See Methods for further details.) 2. Results 2.1 Participant Characteristics The final pooled sample of participants included 36,342 individuals from 29 LMICs (Figure S1). Thirty one percent of individuals were classified as experiencing moderate or high water insecurity in the last 12 months (30.5%) ( Table 1 ). Floods were the climate hazard to which respondents were most commonly exposed, with 28.3% of individuals exposed to at least one flood event in the last year (see Table 1 ). Six percent (5.76%) were exposed to storms, and 3.25% were exposed to droughts. Rainfall anomalies were more common; 69.8% of individuals were exposed to abnormally wet conditions (high SPI) and 56.2% exposed to abnormally dry conditions (low SPI). Being exposed to abnormally wet conditions mostly concurred with reported floods, with 55.7% of exposure classifications concurring (Table S1). Similarly, being exposed to abnormally dry conditions mostly concurred with reported drought (Table S1). Most respondents lived in tropical climate zones (56.7%) and in urban/peri-urban areas (72.5%). Most respondents (78.4%) lived in an area of low subnational wealth (in the bottom two terciles of the international wealth index), and most were in a low-income quintile (59.4%). Table 1 Individual characteristics of the 2020 Gallup World Poll participants across 29 countries included in the pooled analytic sample (n=36,342)* Mean/% SE Primary Outcome Variables Individual Water Insecurity (%) 1 No-to-marginal 39.8% 0.005 Low 29.6% 0.004 Moderate 22.6% 0.004 High 7.91% 0.003 Primary Exposure Variables Reported Climate Events (%) 2 Reported flood events (ref no floods) 28.3% 0.004 Reported storm events (ref no storms) 5.76% 0.002 Reported drought events (ref no droughts) 3.25% 0.002 SPI (%) 3 Low (ref not low) 56.2% 0.005 High (ref not high) 69.8% 0.004 Vulnerability Variables Climate zone (%) 4 Temperate 20.9% 0.004 Arid 22.3% 0.004 Tropical 56.7% 0.005 Urbanicity (%) 5 Rural area 27.5% 0.004 Peri-Urban/Urban 72.5% 0.004 Subnational Wealth (%) 6 Low (Bottom two tertiles) 78.4% 0.004 High (Upper tertile) 21.6% 0.004 Individual Income (%) 7 Low (Lowest three brackets) 59.4% 0.004 High (Highest two brackets) 40.6% 0.004 Other Covariates 8 Gender (%) Women 50% 0.005 Age (mean) 33.9 0.138 Household size (mean) 5.88 0.033 Education (%) Primary 42.6% 0.005 Secondary 50.5% 0.005 Tertiary 6.92% 0.002 Difficulty living on present income (%) Getting by or living comfortably 43.3% 0.005 Difficult or very difficult 56.7% 0.005 Lives disrupted by COVID (%) Not at all 22.5% 0.004 Some 31.9% 0.004 A lot 45.6% 0.005 * Values have been calculated using normalised and rescaled survey weights within each country 1 Individual water insecurity category classified using IWISE-12, the 12-item Water Insecurity Experiences Scale. IWISE-12 scores range from 0-36. The categories were derived using the following cut-points: “no-to-marginal” (scores of 0-2); “low” (scores of 3-11), “moderate” (scores of 12-23); or “high” water insecurity (scores of 24-36) (29) . 2 Individuals exposed to at least one reported climate disaster over the 12 months prior to the survey, extracted from The International Disaster Database (EM-DAT) (26) . 3 Individuals exposed to high or low Standardised Precipitation Index (SPI) for at least one month over the 12 months prior to the survey, extracted from the rom the Climate Engine (27, 28) . Low SPI represents abnormally dry conditions, and high SPI represents abnormally wet conditions. 4 The climate zone of each respondent was classified using the Köppen-Geiger climate classification maps (30) . 5 The urbanicity of each respondent was classified using their DEGURBA classification (31) . 6 Subnational wealth was classified using the International Wealth Index at subnational level (32) . 7 Individual income was classified using the income quintile for each respondent in the Gallup World Poll within the income distribution for that country. 8 All other covariates were based on individual responses in the Gallup World Poll 2.2 Climate hazards and water insecurity (H1) Exposure to climate hazards was associated with higher water insecurity, as hypothesized ( Figure 3 panel A, Table S1 Model 1 ). Floods were associated with higher relative risk of low (1.25 RRR, 95% CI: 1.16-1.35), moderate (1.22 RRR, 95% CI: 1.12-1.33), and high (1.34 RRR, 95% CI:1.18-1.52) water insecurity ( Figure 3 , Table S2). Storms were similarly associated with increased relative risk of moderate (1.31 RRR, 95% CI: 1.10-1.57) and high water insecurity (1.49 RRR, 95% CI: 1.16-1.92) ( Figure 3 , Table S2). Droughts were associated with increased relative risk of high water insecurity (1.42 RRR, 95% CI: 1.02-1.98) ( Figure 3 , Table S2). Exposure to rainfall anomalies, as measured through SPI, was not associated with any category of water insecurity, in a second multinomial regression model accounting for all key covariates (Figure S2, Table S3 Model 2 ). Abnormally wet conditions (high SPI) were not associated with any water insecurity category: low (1.03 RRR, 95% CI: 0.96-1.11); moderate (0.97 RRR, 95% CI: 0.89-1.05); and high (0.93 RRR, 95% CI:0.82-1.05). Likewise, abnormally dry conditions (low SPI) were not associated with any water insecurity category: low (0.97 RRR, 95% CI: 0.90-1.04); moderate (0.99 RRR, 95% CI: 0.91-10.7); and high (1.10 RRR, 95% CI:0.98-1.25). We then adjusted for geographic and economic vulnerabilities in the aforementioned multinomial regression models. We found that climate zone, subnational wealth and individual wealth were independently associated with water insecurity; urbanicity was not ( Figure 3 panels B & C , Table S2, and Figure S2, Table S3). We observed that those living in arid zones had higher relative risk of low, moderate and high water insecurity compared to those living in temperate zones, but respondents living in tropical zones only had higher relative risk of experiencing high water insecurity compared to those living in temperate zones ( Figure 3 panel B , Table S2, and Figure S2, Table S3). Higher wealth was associated with lower water insecurity. However, high subnational wealth was only associated with lower risk of experiencing a high level of water insecurity, but high individual income was associated with lower risk of all water insecurity categories ( Figure 3 panel C , Table S2, and Figure S2, Table S3). Figure 3: Results from multinomial logistic regression models displaying the relative risk of experiencing three categories of water insecurity associated with exposure to reported climate events and moderating vulnerability factors (panels B and C) (n=36,342, covering 29 countries) * *Adjusted for perceived adequacy of household income, gender, age, household size, education level and COVID-19 pandemic-related life disruptions and country-level fixed effects. 2.3 Do geographic vulnerabilities (climate zone and urbanicity) modify the associations of climate hazards with water insecurity? (H2) To investigate hypotheses 2, that the associations between water insecurity and climate hazards were modified by geographic vulnerabilities, we ran logistic regression models where we operationalised water insecurity as a binary variable (representing moderate-and -high water insecurity, referred to henceforth as “water insecurity”). These models were stratified by climate zone and urbanicity and adjusted for key covariates. In the full, un-stratified sample, floods, storms and droughts were associated with higher odds of experiencing moderate-to-high water insecurity ( Figure 4 (grey bars) , Table S4 Model 3 ). From the logistic regression models stratified by climate zone, we found that climate zone modified the association between climate hazards and water insecurity, but these associations were not consistent across each type of hazard ( Figure 4A , Tables S6 Model 4 ). Storms were associated with higher odds of water insecurity in temperate (1.72 OR, 95% CI: 1.13-2.63) and tropical zones (1.35 OR, 95% CI: 1.01-1.79), but not in arid zones (0.37 OR, 95% CI: 0.13- 1.05). Floods were only associated with higher odds of water insecurity in tropical zones (1.20 OR, 95% CI:1.02-1.42). Droughts were associated with higher odds of experiencing water insecurity in temperate zones (1.48 OR, 95% CI: 1.05-2.09). The number of droughts reported by those living in tropical zones (n=55)) was low, so the confidence intervals were wide. From the logistic regression models stratified by urbanicity, we observed that urbanicity modified the associations between climate event and water insecurity ( Figure 4B , Tables S7 Model 5 ). Droughts and floods were associated with higher odds of experiencing water insecurity in rural areas (floods 1.33 OR, 95% CI: 1.06-1.68; droughts 2.02 OR, 95% CI:1.22-3.36) but not in urban/peri-urban areas (floods 1.01 OR, 95% CI: 0.88-1.16; droughts 1.07 OR, 95% CI:0.79-1.44; Figure 4B , Table S7). On the other hand, storms were associated with higher odds of experiencing water insecurity in urban/peri-urban areas (1.47 OR, 95% CI: 1.13-1.93) but not in rural areas (0.96 OR, 95% CI: 0.65-1.41; Figure 4B , Table S7). In the full-unstratified sample, using a logistic regression to model the association between rainfall anomalies (high and low SPI), neither high nor low SPI was associated with experiencing water insecurity ( Figure 5 A & B , Tables S9 Model 6 ). In the models stratified by climate zone and urbanicity, high nor low SPI was still not associated with experiencing water insecurity regardless of the climate zone or urbanicity of the respondent ( Figure 5 A & B , Tables S9, Models 7 & 8 ). Figure 4: Results from logistic regression models displaying the odds of experiencing moderate-high water insecurity if exposed to climatic events and the effect modification by geographic and economic vulnerabilities. Models were stratified by A) Climate Zone; B) Urbanicity; C) Subnational Wealth; and D) Individual Income (see Table S5 for the number of observations for each event within each stratified model)* *Adjusted for perceived adequacy of household income, gender, age, household size, education level and COVID-19 pandemic-related life disruptions, and country-level fixed effects; and climate zone in all except A; and urbanicity in all except B; and subnational wealth in all except C; and individual income in all except D. 2.4 Do economic vulnerabilities (subnational wealth and individual income) modify the associations of climate hazards with water insecurity? (H3) To investigate the hypotheses that the associations between water insecurity and climate hazards were modified by economic vulnerabilities, we ran logistic regression models in which we operationalised water insecurity as a binary variable. These models were stratified by subnational wealth and individual income and adjusted for key covariates. Subnational wealth modified some of the associations between reported climate events and water insecurity ( Figure 4C and Table S10 Model 9). People living in areas of low subnational wealth who were exposed to storms had higher odds of experiencing water insecurity (1.54 OR, 95% CI:1.20-1.98; Figure 4C and Table S10). Among those living in an area of high subnational wealth, however, no difference in odds of experiencing water insecurity was seen between those exposed and not exposed to storms (0.62 OR, 95% CI: 0.37-1.04; Figure 4C and Table S10). No association was found between water insecurity and exposure to the other climate events (floods and droughts), regardless of subnational wealth ( Figure 4C and Table S10). We observed only one difference between the associations of reported climate events with water insecurity when models were stratified by individual income group ( Figure 4D and Table S11 Model 10 ). For those in low-income quintiles, floods were associated with higher odds of experiencing water insecurity (1.16 OR, 95% CI: 1.00-1.35, Figure 4D and Table S11), but no other event was associated with experiencing water insecurity regardless of individual wealth ( Figure 4D and Table S11). For SPI, similar to geographic vulnerabilities, in the models stratified by economic vulnerabilities exposure to high or low SPI was not associated with experiencing water insecurity regardless of the subnational wealth or individual income of the respondent ( Figure 5C & D , Tables S9 Model 11 & 12 ). 3. Discussion In this first analysis of the relationships between climate hazards and experiences of water insecurity, we found that climate hazards (floods, storms and droughts) were associated with higher water insecurity in a pooled sample covering 29 LMICs. These analyses represent a significant increase in the resolution of our understanding of how climate hazards relate to human well-being because they provide insights into individual experiences, rather than using proxies of water insecurity based on physical water supplies at community, national or watershed levels. They also bring to light the differential effects these hazards can have depending on geographic and economic vulnerabilities. Exposure to rainfall anomalies, as measured by SPI, was not associated with water insecurity. SPI is often used in early warning and climate hazard monitoring systems; however, the lack of association found between SPI and water insecurity suggests that this metric alone may not capture well the risks associated with climate hazards. Both heavy rainfall and decreased rainfall may have implications for water availability and safety (8), but these may not be captured well in satellite-derived gridded data which average rainfall across a grid-cell (5x 5km resolution), particularly as rainfall may vary widely over small distances (33). Increasing weather station coverage that can better capture these events could enable enhanced understanding of the associated risks. As hypothesised (H2), geographic factors, specifically climate zone and urbanicity, modified the association between climate hazards and water insecurity. Whilst climate zones were found to be independently associated with water insecurity, with arid zones associated with higher risk of experiencing water insecurity compared to temperate zones, the associations between climate hazards and water insecurity differed inconsistently across these zones. For example, in temperate regions storms and droughts were associated with higher odds of being water insecure but floods were not. This suggests that we need to tailor climate adaptation strategies according to specific geographic characteristics, rather than using a generic strategy according to hazard type. Although urbanicity was not associated with water insecurity, with no difference in the relative risk of experiencing water insecurity found between those residing in rural compared to urban areas, urbanicity modified the association between climate hazards and water insecurity. In other words, whilst living in a rural compared to an urban setting was unrelated to people’s water insecurity, urbanicity was associated with their vulnerability to different hazard types. Exposure to floods and droughts were associated with higher odds of water insecurity in rural areas, whereas exposure to storms was associated with higher odds of water insecurity in urban areas. This is consistent to an extent with our expectation that climate hazards, such as floods and droughts, disproportionately affect those in rural areas due lower access to improved water sources (15, 16) and infrastructure in general. While we had not anticipated that storms would be associated with higher water insecurity in urban but not rural settings, this may be due to built-up urban areas being more vulnerable to these hazards due to their drainage systems that can become overwhelmed (34). This pattern further indicates the need for differentiated investments based on specific geographic factors, including urbanicity as well as climate zone. For example, investing in greater flood and drought resilient infrastructure may be more important in rural areas while storm-proofing is more important in urban areas. Whilst economic factors, specifically subnational wealth and individual income, were associated with lower relative risk of water insecurity, these factors did not seem to influence the vulnerability to climate hazards (H3). Although there were some exceptions; storms were associated with higher water insecurity in areas with low subnational wealth, and floods were associated with higher water insecurity among people with low individual incomes. The lack of differential associations observed for the other hazards makes it difficult to draw strong conclusions about how these economic factors affect vulnerability. Access to resilient infrastructure, better management of water systems and access to humanitarian assistance in disasters may play an important role in buffering the effect of climate hazards (19, 20, 35). Further research is required to understand how these infrastructural investments and social programs may influence vulnerability to climate hazards beyond economic status. Experiencing water insecurity has substantial implications for people’s overall health and well-being. For example, it is closely associated with higher food insecurity and higher likelihood of experiencing poor physical and mental health outcomes (7). Given the close associations found between climate hazards and water insecurity, and that they differed by geographic vulnerabilities, this study highlights the importance of investing in climate adaptation measures that are context-specific. Monitoring using experiential measures of water security can allow for a more nuanced understanding of disparities, and how these may fluctuate because of these climate hazards. This in turn can help guide strategic investments within countries to mitigate the risks brought about by these hazards. In addition to these policy implications, this study identifies future research needs. One is greater investment in climate data. Many have highlighted the challenge of defining droughts as these events do not manifest as visibly as flood events, which may lead to underreporting (36-38). Whilst alternative metrics such as the Standardized Precipitation Evapotranspiration Index (SPEI) may potentially better capture meteorological drought than reported disaster or SPI data (39), geographic coverage of this data remains limited. Droughts are also typically slow onset, so a longer timeseries of water insecurity data than one year may be required to better understand the potential effects of this hazard. Future analyses might also explore how climate hazards shape each of the four domains of water security (7). Although this study provides novel insights into the risks associated with climate hazards, using large nationally representative datasets, there are some limitations. Whilst climate hazards are exogenous and reverse causality is not plausible, we cannot infer causality as these data are cross-sectional; there may be other factors that increase vulnerability that were not captured. Furthermore, there are some temporal limitations in our ability to interpret the associations of these hazards with water insecurity, as the recall period of water insecurity was over the last 12 months, but people were classified as exposed to a climate event if they had been exposed to at least one event (or for SPI in at least one month) over that year. Collecting water insecurity experiences more frequently in longitudinal datasets using shorter recall periods, such as one month, would enable a more accurate understanding of the effect of climate events. A further limitation is that there may be inaccuracies in the reporting of climate events captured in the EM-DAT database. Individual water insecurity data could only be matched with climate event data at the administrative level, so we cannot be certain if all the respondents living within that unit were equally affected, i.e., there might be some exposure misclassification. Finally, the international wealth index was used as proxy variable, but it covers large administrative areas within which there might be variations in access to infrastructure that may influence the associations between climate hazards and water insecurity (13). In conclusion, climate hazards were associated with higher water insecurity across 29 LMICs. Data on water insecurity experiences made evident the heterogenous toll of climate hazards and can guide appropriate climate adaptation strategies. Given the anticipated increase in frequency and intensity of climate events due to climate change, efforts should be made to address vulnerabilities and disparities that may occur due to both climate hazards and geographic and economic factors. More robust (higher resolution, higher frequency) measures of both climate events and water insecurity can help us to better understand the risks of climate events and support the development of strategies to mitigate the harms of climate hazards to both water security and health in places where risks are the greatest. 4. Methods 4.1 Survey design of the Gallup World Poll Nationally representative data on individuals’ experiences with water insecurity were collected in 29 countries between September 2020 and February 2021 by the Gallup World Poll (GWP). These countries included 20 from Sub-Saharan Africa ( Benin, Burkina Faso, Cameroon, Congo Brazzaville, Cote d’Ivoire, Ethiopia, Gabon, Ghana, Guinea, Kenya, Mali, Mauritius, Namibia, Nigeria, Senegal, South Africa, Tanzania, Togo, Uganda, Zambia ), 4 from North Africa ( Algeria, Egypt, Morocco and Tunisia ), 3 from Latin America ( Brazil, Guatemala and Honduras ) and 2 from Asia ( India and Bangladesh ), see Figure S1. GWP data collection procedures are described in detail elsewhere (25). Briefly, non-institutionalized individuals aged 15 or older were eligible and were randomly sampled using stratified sampling methods depending on sampling frame (telephone vs face-to-face). Surveys were conducted predominantly via telephone due to COVID-19 restrictions, except for Mali, Senegal and two out of three of the survey waves in India that were conducted face-to-face. Respondent geolocations were determined by recording the GPS location of the primary sampling units in face-to-face surveys and via a series of respondents’ answers to location-related questions in telephone surveys. If the location of a respondent could not be determined from the telephone surveys, then the centroid of their administrative unit (level 1) was assigned (which is typically a large subdivision within a country such as a province, region, or state). In these cases, the respondent was flagged as not having an accurate geolocation and excluded from the analysis. Geolocations were dislocated by 2km in urban and 5km in rural in any direction, and an additional random 1% of rural by 10km, from the original location to ensure anonymity; all dislocation remained with the original level 1 administrative unit. 4.2 Outcome: water insecurity The primary dependent variable was experiences with water insecurity measured at the individual level using the Individual Water Insecurity Experiences (IWISE) Scale, which has been established as reliable, valid, and equivalent for making cross-country comparisons (25). The 12 items in the scale query about universal experiences with water, including limitations to water-related behaviours (e.g., unable to wash hands), psychosocial effects (e.g., worrying about water), and interruptions in water supply (Supplementary Table S12). Respondents were asked how frequently they had experienced these issues over the last 12 months, with answer options “never (score 0)”, “rarely (1-2 months, score 1)”, “sometimes (in some but not every month, score 2)” or “often/always (in almost every month, score 3)”. Responses were summed to create a score with a range of 0 to 36, with higher scores indicating higher water insecurity. Respondents missing one or more IWISE responses were excluded (n=1,483, 3.6%). Four ordinal categories of water insecurity were made using established cut-points: “no-to-marginal” (0-2); “low” (3-11), “moderate” (12-23); or “high” water insecurity (scores of 24-36) (29). For the analyses exploring how geographic and economic vulnerabilities modified the association between climate hazards and water insecurity, we operationalised a binary outcome representing moderate-to-high water insecurity (IWISE score > 12), referred to as “water insecurity” in the results (cf. 2.3 & 2.4). 4.3 Exposure variables: climate events and precipitation anomalies The primary exposure variables were reported climate disaster events (drought, flood and storms) and precipitation anomalies based on the Standard Precipitation Index (SPI). We extracted reported climate disaster events from 2019-2021 that occurred in the countries corresponding to the GWP dataset from The International Disaster Database (EM-DAT)(26). The EM-DAT data includes events that have reported at least ten deaths, at least 100 affected people and/or a response that involves either a call for international assistance or a declaration of an emergency. We extracted events classified as a “drought”, “flood” and “storm” (26). There was one glacial lake outburst that we reclassified as a flood. We linked each event to the lowest possible administrative unit using the Global Administrative Areas (GADM) administrative unit boundaries (40), based on the location names and administrative unit provided in the EM-DAT database. We assumed that an event impacted all areas of an administrative unit equally given the geographical information available in the data. We considered a respondent to be exposed to an event category if their administrative unit had experienced at least one event in the last 12 months prior to the interview. We extracted Standardised Precipitation Index (SPI) data from the Climate Engine (27, 28), which uses satellite-derived rainfall data from The Climate Hazards Group InfraRed Precipitation with Station (CHIRPS)(41)) to classify precipitation anomalies from the long-term average. SPI values are calculated using the number of standard deviations from the mean of accumulative precipitation for a given location and period. We classified areas as being exposed to high precipitation if the SPI values were > 1.5 (corresponding to extremely and very wet conditions) and low precipitation if SPI values were < -1.5 (corresponding to extremely dry or severely dry conditions (42, 43)). We derived frequencies for each respondent in the GWP by collecting monthly SPI values which give the deviation, for a given month, from the monthly mean between 1991 and 2020. These values were calculated for each CHIRPS grid cell (5x 5km resolution) corresponding to the countries within the GWP dataset. We extracted SPI data for each respondent based on dislocated geolocations. We then used these monthly classifications to count the frequency each respondents’ location had been exposed to high or low precipitation (as defined previously) in the 12 months prior to survey. For our analysis, we considered ‘exposed’ to be at least 1 month exposure to high or low SPI. 4.4 Vulnerability variables: geographic and economic factors We investigated how geographic (climate zone and urbanicity) and economic (sub-national wealth and individual income) factors might modify the risk of water insecurity associated with climatic events, specifically, cf. Figure 2. We classified the climate zone of each respondent in the GWP using the Köppen-Geiger climate classification maps of 10km spatial resolution, version 1 (30). Five groups are included in this classification: tropical, arid, temperate, cold (continental), and polar. Due to the small sample size of respondents in cold and polar regions (n=44), we re-assigned them to the most common climate zone in their country. The level of urbanicity for each respondent was provided by the GWP according to their DEGURBA classification (31). Based on these DEGURBA classifications we generated a dichotomous variable representing whether lived in an urban/ peri-urban areas (“1”) or rural areas (“0”). Subnational wealth was established using the International Wealth Index (IWI) at subnational level (32), from Global Data Lab (44, 45). This index is based on assets derived from household survey data that are comparable between countries (32). The subnational area names were matched to GADM administrative units, and IWI values were then linked to respondents based on their administrative unit. Tertile IWI brackets were generated per country and assigned to respondents based on their subnational region. These tertiles were combined to form a dichotomous variable representing low (lowest two tertiles) and high (highest tertile) subnational wealth. Individual income was operationalised based on the income quintile for each respondent in GWP within the income distribution for that country. We created a dichotomous variable for individual income, pooling the quintiles of the lowest three brackets to represent low individual income, and the highest two brackets to represent high individual income. 4.5 Other covariates We also extracted other individual covariates from the GWP dataset that have been shown to be relevant for water insecurity elsewhere, as these may be related with individual vulnerability: perceived adequacy of household (HH) income (a self-reported measure of wealth beyond what is captured using income; living comfortably or getting by on HH income “0” or finding it difficult or very difficult on HH income “1” ), gender ( women “1” or men “0 ”), age (c ontinuous variable ), household size (c ontinuous variable ), education level ( primary “0”, secondary “1 ”, tertiary “2”), and COVID-19 pandemic-related life disruptions (a self-reported measure of how the pandemic has disrupted their lives; not at all “0”, some “1 ”, a lot “2”) (24, 46). 4.6 Statistical analysis We linked GWP respondents to the exposure (climate hazards) and vulnerability variables (climate zone, sub-regional wealth) using the geolocation data. All analyses were conducted using R (47). Of a total 41,048 respondents, we excluded those with missing IWISE items (n = 1,483), country-coordinate mismatches (n = 2), flagged inaccurate geolocations (n = 1,876), or missing covariates (n = 1,345), yielding 36,342 respondents across 29 countries. Survey weights were normalised and rescaled within country for pooled analysis. To assess any exclusion bias from geolocation flags, we compared WISE category distributions between all vs non-flagged respondents per country (chi-squared; Table S13); no evidence of bias was found. Because SPI measures rainfall anomalies, with low SPI indicative of abnormally dry conditions and high SPI indicative of abnormally wet conditions, exposure to low SPI might be associated with exposure to reported droughts and exposure to high SPI might be associated with exposure to reported floods. We tested these associations using a contingency table analysis; these results are reported in the participants characteristics section in the results section (cf. 2.1, Table S1). We estimated associations of water insecurity categories with (i) reported climate events and (ii) SPI using survey-weighted multinomial logistic regression with country fixed effects, adjusting for geographic (climate zone, urbanicity) and economic (individual income and subnational wealth) vulnerabilities, and other key covariates: perceived adequacy of household income, gender, age, household size, education level and COVID-19 pandemic-related life disruptions. Because multiple hazards can occur over a 12-month period, models for reported events included all three climate event types simultaneously; SPI models included high and low SPI simultaneously. To understand if geographic factors (H2), specifically climate zone and urbanicity, modified the association between water insecurity and climate hazards (reported events and SPI), we tested for an interaction using logistic regression models. In these models we classified the outcome as a binary variable representing moderate-to-high water insecurity (IWISE score > 12) to reduce the complexity of the models and aid interpretability of the results. Because evidence of an interaction was found with some of these climate events, we ran stratified models to examine the differences in these effects, to further aid interpretability. Models were stratified by climate zone: arid, temperate and tropical. Models were also stratified by urbanicity: rural and urban (including peri-urban). Sensitivity analysis was conducted using a self-reported definition of urbanicity, which yielded similar results (Supplementary Text 1, Figure S3). To understand if economic factors (H3), specifically subnational wealth and individual income, modified the association between water insecurity and climate hazards, we ran logistic regression models stratified by subnational wealth (low and high) and individual wealth (low and high). All models included country-level fixed effects, were survey-weighted and adjusted for perceived adequacy of household income, gender, age, household size, education level and COVID-19 pandemic-related life disruptions. Models were also adjusted for geographic (urbanicity, climate zone) and economic (individual income and subnational wealth) factors, as appropriate. We report report adjusted relative risk ratios (aRRR) and adjusted odds ratios (aOR) with 95% Cis. 4.7 Ethics Gallup World Poll sought approval for their survey from theh governing bodies required in each country and obtained informed consent from all participants. For this analysis we received deidentified data from GWP. Declarations This study was based on deidentified data made available by Gallup. Gallup World Poll followed their standard protocol for obtaining consent from participants. The authors of this paper were not involved with the consent or data collection process. No participants were involved in study design, implementation, or dissemination, including the writing of this manuscript. Informed consent of all survey participants was obtained, and survey protocols were approved by Gallup’s Internal Review Board and by the governing bodies in countries where approval is required References IPCC. Climate Change 2022: Impacts, Adaptation, and Vulnerability. Contribution of Working Group II to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change Cambridge University Press. Cambridge University Press, Cambridge, UK and New York, NY, USA,; 2022. Caretta AMMA, Arfanuzzaman RBM, Morgan SMR, Kumar M. Water. In: Climate Change 2022: Impacts, Adaptation, and Vulnerability. Contribution of Working Group II to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change. 2022. Kuzma S, Saccoccia L, Chertock M. Water Resources Institute. 2023. [23/01/2025]. Available from: https://www.wri.org/insights/highest-water-stressed-countries. Jepson WE, Wutich A, Colllins SM, Boateng GO, Young SL. Progress in household water insecurity metrics: a cross-disciplinary approach. WIREs Water. 2017;4(3):e1214. Young S, Miller J, Bose I. Measuring human experiences to advance safe water for all. Evanston: Institute for Policy Research, Northwestern University. 2024. O'Neill BC, van Aalst M, Zaiton Ibrahim Z, Berrang-Ford L, Bhadwal S, Buhaug H, et al. Key risks across sectors and regions. Cambridge University Press, Cambridge, UK and New York, NY, USA,; 2022. Rosinger AY, Young SL. The toll of household water insecurity on health and human biology: Current understandings and future directions. WIREs Water. 2020;7(6):e1468. Mellor JE, Levy K, Zimmerman J, Elliott M, Bartram J, Carlton E, et al. Planning for climate change: The need for mechanistic systems-based approaches to study climate change impacts on diarrheal diseases. Science of The Total Environment. 2016;548-549:82-90. Levy K, Woster AP, Goldstein RS, Carlton EJ. Untangling the Impacts of Climate Change on Waterborne Diseases: a Systematic Review of Relationships between Diarrheal Diseases and Temperature, Rainfall, Flooding, and Drought. Environ Sci Technol. 2016;50(10):4905-22. Broyles LM, Pakhtigian EL, Rosinger AY, Mejia A. Climate and hydrological seasonal effects on household water insecurity: A systematic review. Wiley Interdisciplinary Reviews: Water. 2022;9(3):e1593. Rosinger AY. Household water insecurity after a historic flood: Diarrhea and dehydration in the Bolivian Amazon. Social Science & Medicine. 2018;197:192-202. Bose I, Dreibelbis R, Green R, Murray KA, Ceesay O, Kovats S. Coping strategies for household water insecurity in rural Gambia, mediating factors in the relationship between weather, water and health. BMC Public Health. 2024;24(1):3150. Bose I, Dreibelbis R, Green R, Murray KA, Ceesay O, Kovats S. Climate change, seasonality and household water security in rural Gambia: A qualitative exploration of the complex relationship between weather and water. PLOS Water. 2024;3(6):e0000239. Vörösmarty CJ, Green P, Salisbury J, Lammers RB. Global Water Resources: Vulnerability from Climate Change and Population Growth. Science. 2000;289(5477):284-8. Bain RES, Wright JA, Christenson E, Bartram JK. Rural: urban inequalities in post 2015 targets and indicators for drinking-water. Science of The Total Environment. 2014;490:509-13. UNICEF, WHO. Joint Monitoring Programme for Water Supply, Sanitation and Hygiene (JMP) https://washdata.org/ [ Howard G, Bartram J, World Health O, United Kingdom. Dept. for International D, United States of America. University of North Carolina at Chapel H. Vision 2030 : the resilience of water supply and sanitation in the face of climate change: technical report / Guy Howard, Jamie Bartram. Geneva: World Health Organization; 2010. MacAllister DJ, MacDonald AM, Kebede S, Godfrey S, Calow R. Comparative performance of rural water supplies during drought. Nat Commun. 2020;11(1):1099. Leal Filho W, Totin E, Franke JA, Andrew SM, Abubakar IR, Azadi H, et al. Understanding responses to climate-related water scarcity in Africa. Science of The Total Environment. 2022;806:150420. Charles KJ, Howard G, Villalobos Prats E, Gruber J, Alam S, Alamgir ASM, et al. Infrastructure alone cannot ensure resilience to weather events in drinking water supplies. Science of The Total Environment. 2022;813:151876. Geremew G, Cumming O, Haddis A, Freeman MC, Ambelu A. Rainfall and Temperature Influences on Childhood Diarrhea and the Effect Modification Role of Water and Sanitation Conditions: A Systematic Review and Meta-Analysis. International Journal of Environmental Research and Public Health. 2024;21(7):823. Meehan K, Jepson W, Harris LM, Wutich A, Beresford M, Fencl A, et al. Exposing the myths of household water insecurity in the global north: A critical review. Wiley Interdisciplinary Reviews: Water. 2020;7(6):e1486. Meehan K, Jurjevich JR, Chun NM, Sherrill J. Geographies of insecure water access and the housing–water nexus in US cities. Proceedings of the National Academy of Sciences. 2020;117(46):28700-7. Young SL, Bethancourt HJ, Ritter ZR, Frongillo EA. Estimating national, demographic, and socioeconomic disparities in water insecurity experiences in low-income and middle-income countries in 2020–21: a cross-sectional, observational study using nationally representative survey data. The Lancet Planetary Health. 2022;6(11):e880-e91. Young SL, Bethancourt HJ, Ritter ZR, Frongillo EA. The Individual Water Insecurity Experiences (IWISE) Scale: reliability, equivalence and validity of an individual-level measure of water security. BMJ Global Health. 2021;6(10):e006460. EM-DAT [Internet]. 2023 [cited February 2023]. Available from: www.emdat.be. Climate Engine, version 2.1. [Internet]. 2023 [cited 14th November 2023]. Available from: http://climateengine.org. Huntington JL, Hegewisch KC, Daudert B, Morton CG, Abatzoglou JT, McEvoy DJ, et al. Climate Engine: Cloud Computing and Visualization of Climate and Remote Sensing Data for Advanced Natural Resource Monitoring and Process Understanding. Bulletin of the American Meteorological Society. 2017;98(11):2397-410. Frongillo EA, Bethancourt HJ, Miller JD, Young SL. Identifying ordinal categories for the Water Insecurity Experiences Scales. Journal of Water, Sanitation and Hygiene for Development. 2024:washdev2024042. Beck HE, Zimmermann NE, McVicar TR, Vergopolan N, Berg A, Wood EF. Present and future Köppen-Geiger climate classification maps at 1-km resolution. Scientific data. 2018;5(1):1-12. Dijkstra L, Florczyk AJ, Freire S, Kemper T, Melchiorri M, Pesaresi M, et al. Applying the degree of urbanisation to the globe: A new harmonised definition reveals a different picture of global urbanisation. Journal of Urban Economics. 2021;125:103312. Smits J, Steendijk R. The International Wealth Index (IWI). Social Indicators Research. 2015;122(1):65-85. Colston JM, Ahmed T, Mahopo C, Kang G, Kosek M, de Sousa Junior F, et al. Evaluating meteorological data from weather stations, and from satellites and global models for a multi-site epidemiological study. Environmental Research. 2018;165:91-109. Zhou Q. A review of sustainable urban drainage systems considering the climate change and urbanization impacts. Water. 2014;6(4):976-92. Stringer LC, Mirzabaev A, Benjaminsen TA, Harris RMB, Jafari M, Lissner TK, et al. Climate change impacts on water security in global drylands. One Earth. 2021;4(6):851-64. Lloyd-Hughes B. The impracticality of a universal drought definition. Theoretical and applied climatology. 2014;117:607-11. Wilhite DA, Glantz MH. Understanding: the drought phenomenon: the role of definitions. Water international. 1985;10(3):111-20. Stanke C, Kerac M, Prudhomme C, Medlock J, Murray V. Health effects of drought: a systematic review of the evidence. PLoS Curr. 2013;5. Tirivarombo S, Osupile D, Eliasson P. Drought monitoring and analysis: Standardised Precipitation Evapotranspiration Index (SPEI) and Standardised Precipitation Index (SPI). Physics and Chemistry of the Earth, Parts A/B/C. 2018;106:1-10. GADM database of Global Administrative Areas, version 2.0 [Internet]. 2012 [cited July 2022]. Available from: https://gadm.org. Funk CC, Peterson PJ, Landsfeld MF, Pedreros DH, Verdin JP, Rowland JD, et al. A quasi-global precipitation time series for drought monitoring. Report. Reston, VA; 2014. Report No.: 832. Guerreiro MJ, Lajinha T, Abreu I. Flood analysis with the standardized precipitation index (SPI). 2007. Olanrewaju CC, Reddy M. Assessment and prediction of flood hazards using standardized precipitation index—A case study of eThekwini metropolitan area. Journal of Flood Risk Management. 2022;15(2):e12788. Smits J. GDL Area Database. Nijmegen, The Netherlands. 2016:16-101. Global Data Lab International Wealth Index, version v1.0. [Internet]. Available from: https://globaldatalab.org/wealth/. Young SL, Bethancourt HJ, Frongillo EA, Viviani S, Cafiero C. Concurrence of water and food insecurities, 25 low-and middle-income countries. Bulletin of the World Health Organization. 2023;101(2):90-101. Team RC. _R: A Language and Environment for Statistical Computing_. R Foundation for Statistical Computing, Vienna, Austria; 2023. Additional Declarations There is NO Competing Interest. Supplementary Files ClimatehazardswaterinsecurityNatureSuppFinalsubmit.docx Supplementary Files Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7959507","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":535614269,"identity":"facc97fc-55ba-4951-8bb7-22043a89cfd1","order_by":0,"name":"Indira Bose","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA8klEQVRIiWNgGAWjYBACNgYGgwNAOgGIGR88YDgAk0ggSguzQQLDAQmCWoDAAKaCTYIoLXzSzRsPV9Qw5PHzHz5WkfDnTp1uA/PDD4xtabgdJnOs4OCZYwzFkjPS0m4ktj2TMDvAZizB2JaDW4tEjsHBBjaGxA03eMxuJDYcBmphMGNgbKsgoOUfUMv5898KEv6AtLB/I6ylsQ2o5UAOG0MCG0gLD8gWfA5LKzjY2CcB8ouxRGLbYclth3mKJRLO4fa+/IzkzR8bvtmAQuzhhw9/DvObHW/f+OFDWTJOLVAggcRmZiAQkaNgFIyCUTAKCAIA2E5XGEm2YyoAAAAASUVORK5CYII=","orcid":"https://orcid.org/0000-0002-3045-8851","institution":"London School of Hygiene and Tropical Medicine","correspondingAuthor":true,"prefix":"","firstName":"Indira","middleName":"","lastName":"Bose","suffix":""},{"id":535614270,"identity":"1453b25a-fee6-411c-a110-2731911ef26f","order_by":1,"name":"Edward Frongillo","email":"","orcid":"","institution":"University of South Carolina","correspondingAuthor":false,"prefix":"","firstName":"Edward","middleName":"","lastName":"Frongillo","suffix":""},{"id":535614271,"identity":"1eaeaebb-1807-4214-9154-5bb788585a06","order_by":2,"name":"Claire Dooley","email":"","orcid":"","institution":"University College London","correspondingAuthor":false,"prefix":"","firstName":"Claire","middleName":"","lastName":"Dooley","suffix":""},{"id":535614272,"identity":"da6dda96-dd6e-4ed7-8f0c-0195ed344ddb","order_by":3,"name":"Thalia Sparling","email":"","orcid":"","institution":"London School of Hygiene and Tropical Medicine","correspondingAuthor":false,"prefix":"","firstName":"Thalia","middleName":"","lastName":"Sparling","suffix":""},{"id":535614273,"identity":"9da6c69d-b855-4b8d-acbf-6bbcd2ff716b","order_by":4,"name":"Suneetha Kadiyala","email":"","orcid":"https://orcid.org/0000-0002-9101-1471","institution":"London School of Hygiene \u0026 Tropical Medicine","correspondingAuthor":false,"prefix":"","firstName":"Suneetha","middleName":"","lastName":"Kadiyala","suffix":""},{"id":535614274,"identity":"26a4dc6b-e108-4d2b-bb99-f6011694dc33","order_by":5,"name":"Sera Young","email":"","orcid":"https://orcid.org/0000-0002-1763-1218","institution":"Northwestern University","correspondingAuthor":false,"prefix":"","firstName":"Sera","middleName":"","lastName":"Young","suffix":""}],"badges":[],"createdAt":"2025-10-27 12:50:42","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7959507/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7959507/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":95519976,"identity":"c73615ba-3117-4934-ad16-e7b20b9660b4","added_by":"auto","created_at":"2025-11-10 09:09:35","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":29272,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eThe IPCC’s conceptualization of the dynamic interaction between climate hazard, exposure and vulnerability (6), adapted to the risk of experiencing water insecurity\u003c/strong\u003e\u003cem\u003e. Italicized text indicates how these concepts have been operationalized in these analyses.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"Picture1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7959507/v1/701abce404db2c04770bb0de.jpg"},{"id":95519989,"identity":"6b0dc862-0e2b-4e53-954c-5c987c82109f","added_by":"auto","created_at":"2025-11-10 09:09:40","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":92304,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eConceptual framework showing how climate hazards may contribute to individual water insecurity. \u003c/strong\u003eExposure pathways (blue) describe the ways that climate hazards may affect the different dimensions of water insecurity (7). The geographic and economic factors (red) may modify the vulnerability of individuals to water insecurity. Rounded boxes outlined in black are concepts that have been measured; square boxes indicate hypotheses.\u003c/p\u003e","description":"","filename":"Picture2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7959507/v1/d7e202d750facbd3d292a506.jpg"},{"id":95519984,"identity":"16d8270c-378f-4c15-a82b-dac562d2c0aa","added_by":"auto","created_at":"2025-11-10 09:09:37","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":59558,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eResults from multinomial logistic regression models displaying the relative risk of experiencing three categories of water insecurity associated with exposure to reported climate events and moderating vulnerability factors (panels B and C) (n=36,342, covering 29 countries) *\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e*Adjusted for perceived adequacy of household income, gender, age, household size, education level and COVID-19 pandemic-related life disruptions and country-level fixed effects.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"Picture3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7959507/v1/c96bd9d4394a3ae8c287141d.jpg"},{"id":95519977,"identity":"1ed62028-d2a0-4ed4-81fe-17c87e4374f1","added_by":"auto","created_at":"2025-11-10 09:09:35","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":58987,"visible":true,"origin":"","legend":"\u003cp\u003eResults from logistic regression models displaying the odds of experiencing moderate-high water insecurity if exposed to climatic events and the effect modification by geographic and economic vulnerabilities. Models were stratified by A) Climate Zone; B) Urbanicity; C) Subnational Wealth; and D) Individual Income (see Table S5 for the number of observations for each event within each stratified model)*\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e*Adjusted for perceived adequacy of household income, gender, age, household size, education level and COVID-19 pandemic-related life disruptions, and country-level fixed effects; and climate zone in all except A; and urbanicity in all except B; and subnational wealth in all except C; and individual income in all except D.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"Picture4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7959507/v1/45837ef032e6c8bad3d1b7bf.jpg"},{"id":95519975,"identity":"5edcb0f3-2d82-49fa-b150-837e6d620875","added_by":"auto","created_at":"2025-11-10 09:09:34","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":56707,"visible":true,"origin":"","legend":"\u003cp\u003eResults from logistic regression models displaying the odds of experiencing moderate-high water insecurity if exposed to high and low SPI and the effect modification by geographic and economic vulnerabilities. Models were stratified by A) Climate Zone; B) Urbanicity; C) Subnational Wealth; and D) Individual Income (see Table S8 for the number of observations for each event within each stratified model)*\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e*Adjusted for perceived adequacy of household income, gender, age, household size, education level and COVID-19 pandemic-related life disruptions, and country-level fixed effects; and climate zone in all except A; and urbanicity in all except B; and subnational wealth in all except C; and individual income in all except D.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"Picture5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7959507/v1/4c97836f5fb90e9abac9117b.jpg"},{"id":97248368,"identity":"9dd27f00-2e74-4573-939a-a1d8ba3aa4a4","added_by":"auto","created_at":"2025-12-02 12:55:35","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1259917,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7959507/v1/a642a442-46f3-4fb2-ba38-fb72ae17bc88.pdf"},{"id":95519985,"identity":"82f0a7a3-8456-4d92-b773-31fb6d17ca8d","added_by":"auto","created_at":"2025-11-10 09:09:38","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":293466,"visible":true,"origin":"","legend":"Supplementary Files","description":"","filename":"ClimatehazardswaterinsecurityNatureSuppFinalsubmit.docx","url":"https://assets-eu.researchsquare.com/files/rs-7959507/v1/70e02056d73a59eca2b36a0b.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Climate hazards and water insecurity, by geographic and economic vulnerabilities","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eClimate change is exacerbating the intensity and frequency of extreme climate events (1). Climate events, e.g. floods, droughts, and rainfall anomalies (i.e., higher or lower rainfall than the long-term average), threaten physical water supplies at the national and watershed level (2). Water security at the individual level does not always align with the physical water supply in the area in which an individual resides (3). Individual water security hinges on opportunities to obtain water that is reliable, safe, affordable and sufficient for their basic domestic needs (4), which is shaped by water acquisition systems and management (5). Climate hazards may disrupt these opportunities, reshaping lived experiences of water security.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe risks posed by climate change have been conceptualised by the working group of the Sixth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC), as a dynamic interaction between \u003cem\u003eclimate-related hazards\u003c/em\u003e (the potential occurrence of a climate event) with the \u003cem\u003eexposure\u003c/em\u003e and \u003cem\u003evulnerability\u003c/em\u003e of the affected population (6). Whether a \u003cem\u003eclimate hazard\u003c/em\u003e results in water insecurity for an individual depends on both their \u003cem\u003eexposure\u003c/em\u003e to the event and their \u003cem\u003evulnerability\u003c/em\u003e (\u003cstrong\u003eFigure 1\u003c/strong\u003e).\u0026nbsp;\u003cbr\u003e\u003cstrong\u003eFigure 1: The IPCC\u0026rsquo;s conceptualization of the dynamic interaction between climate hazard, exposure and vulnerability\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003e(6)\u003c/strong\u003e\u003cstrong\u003e, adapted to the risk of experiencing water insecurity\u003c/strong\u003e\u003cem\u003e. Italicized text indicates how these concepts have been operationalized in these analyses.\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFigure 2: Conceptual framework showing how climate hazards may contribute to individual water insecurity.\u0026nbsp;\u003c/strong\u003eExposure pathways (blue) describe the ways that climate hazards may affect the different dimensions of water insecurity (7). The geographic and economic factors (red) may modify the vulnerability of individuals to water insecurity. Rounded boxes outlined in black are concepts that have been measured; square boxes indicate hypotheses.\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe pathways linking climate hazards to water security are numerous (\u003cstrong\u003eFigure 2\u003c/strong\u003e). For example, floods and heavy rainfall events may reduce water acceptability and use due to pathogen run-off, or destroy water infrastructure, thereby limiting availability of sufficient safe water (8-11). Prolonged droughts may reduce water availability, result in increased competition for limited resources reducing accessibility, or concentrate pathogens in water sources, reducing water quality (8-10). Climate events may also have economic impacts on the ability to buy or treat water; these events may also influence the psychological manifestations of water insecurity by causing anxiety about water availability or safety, in turn leading to behavioural changes around water use (10, 12, 13).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eMany factors can modify the vulnerability of an individual to experiencing water insecurity when exposed to a climate event (\u003cstrong\u003eFigure 2\u003c/strong\u003e). Geographic factors such as climate zone, may modify the effect of climate hazards. People living in arid areas, which have higher physical water scarcity (14), may be at higher risk of experiencing water insecurity following a climate event compared to settings where there is higher groundwater storage and recharge. As for urbanicity, in many countries there are large disparities in water infrastructure between rural and urban settings, with rural areas typically lagging behind urban areas in access to improved water sources (15, 16). Rural communities may also face disproportionately large challenges because unimproved water sources such as surface water sources are often less resilient to climate events (17, 18). Rural communities are also more likely to have livelihoods dependent on agriculture, and therefore may have to make trade-offs between water for domestic consumption and crops and livestock when facing shortages (13).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAdditionally, because economic resources often vary within countries, there is often differential access to resilient infrastructure and water management and governance practices. This, in turn, modifies vulnerability to the effects of climate hazards and water insecurity (19-21). At the individual level, economic status can also result in disparities in water insecurity (22-24). For example, wealthier individuals may have more capacity to buffer their water access when faced with a climate event.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eLittle is known about how climate hazards affect individual water security, and how this varies by vulnerabilities. Therefore, we investigated the associations between climate hazards and water insecurity, and if these relationships were modified by geographic and economic vulnerabilities. We first tested the hypothesis that climate hazards (reported flood, storms and drought events and rainfall anomalies) are associated with greater relative risk of experiencing individual water insecurity (\u003cstrong\u003eFigure 2\u003c/strong\u003e,\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003eH1). We then investigated the heterogeneity of these associations by geographic factors, specifically climate zones and urbanicity (H2). We hypothesized that there would be greater risk of water insecurity in arid zones and rural areas. Third, we hypothesized that wealth would modify these associations (H3), with greater risk of water insecurity among individuals living in areas of lower subnational wealth and among those with lower incomes.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe tested these hypotheses using nationally representative, georeferenced survey data from 29 low- and middle-income countries (Figure S1). Individuals\u0026rsquo; level of water insecurity was classified using a cross-country equivalent measure of experiences of water insecurity, the Individual Water Insecurity Experiences (IWISE) Scale (25). \u0026nbsp;We\u0026nbsp;spatiotemporally linked IWISE data with reported climate events and satellite-derived data on rainfall anomalies in the prior 12 months. Reported climate event data (floods, storms and droughts) were extracted from the International Disaster Database (EM-DAT) (26), and rainfall anomalies were based on the Standardised Precipitation Index (SPI) extracted from the Climate Engine (27, 28); individuals were classified as being exposed or not exposed to these hazards based on their geolocations.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFirstly, to investigate the association between climate hazards and water insecurity (\u003cstrong\u003eFigure 2,\u0026nbsp;\u003c/strong\u003eH1), we ran multinomial regression models, adjusting for key covariates. Water insecurity was assessed categorically; individuals were classified as experiencing no-to marginal, low, moderate and high water insecurity. To test the hypotheses that geographic and economic vulnerabilities modified these associations, we ran stratified logistic regression models, in which we operationalised water insecurity as a binary outcome (\u0026ldquo;water insecure\u0026rdquo; experiencing moderate-to-high water insecurity vs \u0026ldquo;water secure\u0026rdquo; experiencing no-to-low water insecurity). (See\u003cstrong\u003e\u0026nbsp;Methods\u003c/strong\u003e for further details.)\u0026nbsp;\u003c/p\u003e"},{"header":"2. Results ","content":"\u003cp\u003e\u003cem\u003e2.1 Participant Characteristics\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe final pooled sample of participants included 36,342 individuals from 29 LMICs (Figure S1). Thirty one percent of individuals were classified as experiencing moderate or high water insecurity in the last 12 months (30.5%) (\u003cstrong\u003eTable 1\u003c/strong\u003e).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFloods were the climate hazard to which respondents were most commonly exposed, with 28.3% of individuals exposed to at least one flood event in the last year (see \u003cstrong\u003eTable 1\u003c/strong\u003e). Six percent (5.76%) were exposed to storms, and 3.25% were exposed to droughts. Rainfall anomalies were more common; 69.8% of individuals were exposed to abnormally wet conditions (high SPI) and 56.2% exposed to abnormally dry conditions (low SPI).\u003c/p\u003e\n\u003cp\u003eBeing exposed to abnormally wet conditions mostly concurred with reported floods, with 55.7% of exposure classifications concurring (Table S1). Similarly, being exposed to abnormally dry conditions mostly concurred with reported drought (Table S1).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eMost respondents lived in tropical climate zones (56.7%) and in urban/peri-urban areas (72.5%). Most respondents (78.4%) lived in an area of low subnational wealth (in the bottom two terciles of the international wealth index), and most were in a low-income quintile (59.4%).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eTable 1 Individual characteristics of the 2020 Gallup World Poll participants across 29 countries included in the pooled analytic sample (n=36,342)*\u003c/em\u003e\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"696\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 500px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eMean/%\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSE\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 500px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003ePrimary Outcome Variables\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"bottom\" style=\"width: 196px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eIndividual Water Insecurity (%)\u003c/strong\u003e\u003csup\u003e1\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003eNo-to-marginal\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e39.8%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.005\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003eLow\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e29.6%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.004\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003eModerate\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e22.6%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.004\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003eHigh\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e7.91%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.003\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 500px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003ePrimary Exposure Variables\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"bottom\" style=\"width: 196px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eReported Climate Events (%)\u003csup\u003e2\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003eReported flood events (ref no floods)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e28.3%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.004\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003eReported storm events (ref no storms)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e5.76%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.002\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003eReported drought events (ref no droughts)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e3.25%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.002\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSPI (%)\u003csup\u003e3\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003eLow (ref not low)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e56.2%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.005\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003eHigh (ref not high)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e69.8%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.004\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 500px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eVulnerability Variables\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"bottom\" style=\"width: 196px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eClimate zone (%)\u003csup\u003e4\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;Temperate\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e20.9%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.004\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;Arid\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e22.3%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.004\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;Tropical\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e56.7%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.005\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eUrbanicity (%)\u003csup\u003e5\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;Rural area\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e27.5%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.004\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003ePeri-Urban/Urban\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e72.5%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.004\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSubnational Wealth (%)\u003csup\u003e6\u003c/sup\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003eLow (Bottom two tertiles)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e78.4%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.004\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003eHigh (Upper tertile)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e21.6%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.004\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eIndividual Income (%)\u003csup\u003e7\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003eLow (Lowest three brackets)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e59.4%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.004\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;High (Highest two brackets)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e40.6%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.004\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 500px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eOther Covariates\u003csup\u003e8\u003c/sup\u003e\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"bottom\" style=\"width: 196px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eGender (%)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;Women\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e50%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.005\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAge (mean)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e33.9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.138\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eHousehold size (mean)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e5.88\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.033\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eEducation (%)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003e\u0026nbsp;Primary\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e42.6%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.005\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003eSecondary\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e50.5%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.005\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003eTertiary\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e6.92%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.002\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eDifficulty living on present income (%)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 98px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;Getting by or living comfortably\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e43.3%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.005\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003e\u0026nbsp; Difficult or very difficult\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e56.7%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.005\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eLives disrupted by COVID (%)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;Not at all\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e22.5%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.004\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;Some\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e31.9%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.004\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 500px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;A lot\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e45.6%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 98px;\"\u003e\n \u003cp\u003e0.005\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cem\u003e*\u003c/em\u003e\u003cem\u003e\u0026nbsp;Values have been calculated using normalised and rescaled survey weights within each country\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u003csup\u003e1\u003c/sup\u003e\u003c/em\u003e\u003cem\u003e\u0026nbsp;Individual water insecurity category classified using IWISE-12, the 12-item Water Insecurity Experiences Scale. IWISE-12 scores range from 0-36. The categories were derived using the following cut-points: \u0026ldquo;no-to-marginal\u0026rdquo; (scores of 0-2); \u0026ldquo;low\u0026rdquo; (scores of 3-11), \u0026ldquo;moderate\u0026rdquo; (scores of 12-23); or \u0026ldquo;high\u0026rdquo; water insecurity (scores of\u003c/em\u003e 24-36)\u0026nbsp;\u003cem\u003e(29)\u003c/em\u003e\u003cem\u003e.\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u003csup\u003e2\u0026nbsp;\u003c/sup\u003e\u003c/em\u003e\u003cem\u003eIndividuals exposed to at least one reported climate disaster over the 12 months prior to the survey, extracted from\u003c/em\u003e \u003cem\u003eThe International Disaster Database (EM-DAT)\u0026nbsp;\u003c/em\u003e\u003cem\u003e(26)\u003c/em\u003e\u003cem\u003e.\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u003csup\u003e3\u0026nbsp;\u003c/sup\u003e\u003c/em\u003e\u003cem\u003eIndividuals exposed to high or low\u003c/em\u003e \u003cem\u003eStandardised Precipitation Index (SPI) for at least one month over the 12 months prior to the survey, extracted from the rom the Climate Engine\u0026nbsp;\u003c/em\u003e\u003cem\u003e(27, 28)\u003c/em\u003e\u003cem\u003e. Low SPI represents abnormally dry conditions, and high SPI represents abnormally wet conditions.\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u003csup\u003e4\u003c/sup\u003e\u003c/em\u003e\u003cem\u003e\u0026nbsp;The climate zone of each respondent was classified using the K\u0026ouml;ppen-Geiger climate classification maps\u0026nbsp;\u003c/em\u003e\u003cem\u003e(30)\u003c/em\u003e\u003cem\u003e.\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u003csup\u003e5\u003c/sup\u003e\u003c/em\u003e\u003cem\u003e\u0026nbsp;The urbanicity of each respondent was classified using their DEGURBA classification\u0026nbsp;\u003c/em\u003e\u003cem\u003e(31)\u003c/em\u003e\u003cem\u003e.\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u003csup\u003e6\u003c/sup\u003e\u003c/em\u003e\u003cem\u003eSubnational wealth was classified using the International Wealth Index at subnational level\u0026nbsp;\u003c/em\u003e\u003cem\u003e(32)\u003c/em\u003e\u003cem\u003e.\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u003csup\u003e7\u0026nbsp;\u003c/sup\u003e\u003c/em\u003e\u003cem\u003eIndividual income was classified using the income quintile for each respondent in the Gallup World Poll within the income distribution for that country.\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u003csup\u003e8\u003c/sup\u003e\u003c/em\u003e\u003cem\u003eAll other covariates were based on individual responses in the Gallup World Poll\u003c/em\u003e\u003cem\u003e\u003cbr\u003e\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e2.2 Climate hazards and water insecurity (H1)\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eExposure to climate hazards was associated with higher water insecurity, as hypothesized (\u003cstrong\u003eFigure 3\u0026nbsp;\u003c/strong\u003epanel A, Table S1 \u003cem\u003eModel 1\u003c/em\u003e). Floods were associated with higher relative risk of low (1.25 RRR, 95% CI: \u0026nbsp; 1.16-1.35), moderate (1.22 RRR, 95% CI: 1.12-1.33), and high (1.34 RRR, 95% CI:1.18-1.52) water insecurity (\u003cstrong\u003eFigure 3\u003c/strong\u003e, Table S2). Storms were similarly associated with increased relative risk of moderate (1.31 RRR, 95% CI: 1.10-1.57) and high water insecurity (1.49 RRR, 95% CI: 1.16-1.92) (\u003cstrong\u003eFigure 3\u003c/strong\u003e, Table S2). Droughts were associated with increased relative risk of high water insecurity (1.42 RRR, 95% CI: 1.02-1.98) (\u003cstrong\u003eFigure 3\u003c/strong\u003e, Table S2). \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eExposure to rainfall anomalies, as measured through SPI, was not associated with any category of water insecurity, in a second multinomial regression model accounting for all key covariates (Figure S2, Table S3 \u003cem\u003eModel 2\u003c/em\u003e). Abnormally wet conditions (high SPI) were not associated with any water insecurity category: low (1.03 RRR, 95% CI: \u0026nbsp;0.96-1.11); moderate (0.97 RRR, 95% CI: 0.89-1.05); and high (0.93 RRR, 95% CI:0.82-1.05). \u0026nbsp;Likewise, abnormally dry conditions (low SPI) were not associated with any water insecurity category: low (0.97 RRR, 95% CI: \u0026nbsp;0.90-1.04); moderate (0.99 RRR, 95% CI: 0.91-10.7); and high (1.10 RRR, 95% CI:0.98-1.25). \u0026nbsp; \u0026nbsp; \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe then adjusted for geographic and economic vulnerabilities in the aforementioned multinomial regression models. We found that climate zone, subnational wealth and individual wealth were independently associated with water insecurity; urbanicity was not (\u003cstrong\u003eFigure 3\u0026nbsp;\u003c/strong\u003e\u003cem\u003epanels B \u0026amp; C\u003c/em\u003e, Table S2, and Figure S2, Table S3). We observed that those living in arid zones had higher relative risk of low, moderate and high water insecurity compared to those living in temperate zones, but respondents living in tropical zones only had higher relative risk of experiencing high water insecurity compared to those living in temperate zones (\u003cstrong\u003eFigure 3\u003c/strong\u003e\u003cem\u003e\u0026nbsp;panel B\u003c/em\u003e, Table S2, and Figure S2, Table S3). Higher wealth was associated with lower water insecurity. \u0026nbsp;However, high subnational wealth was only associated with lower risk of experiencing a high level of water insecurity, but high individual income was associated with lower risk of all water insecurity categories (\u003cstrong\u003eFigure 3\u003c/strong\u003e\u003cem\u003e\u0026nbsp;panel C\u003c/em\u003e, Table S2, and Figure S2, Table S3).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eFigure 3: Results from multinomial logistic regression models displaying the relative risk of experiencing three categories of water insecurity associated with exposure to reported climate events and moderating vulnerability factors (panels B and C) (n=36,342, covering 29 countries) *\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e*Adjusted for perceived adequacy of household income, gender, age, household size, education level and COVID-19 pandemic-related life disruptions and country-level fixed effects. \u0026nbsp;\u003c/em\u003e\u003cem\u003e\u003cbr\u003e\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e2.3 Do geographic vulnerabilities (climate zone and urbanicity) modify the associations of climate hazards with water insecurity? (H2)\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eTo investigate hypotheses 2, that the associations between water insecurity and climate hazards were modified by geographic vulnerabilities, we ran logistic regression models where we operationalised water insecurity as a binary variable (representing moderate-and -high water insecurity, referred to henceforth as \u0026ldquo;water insecurity\u0026rdquo;). These models were stratified by climate zone and urbanicity and adjusted for key covariates. In the full, un-stratified sample, floods, storms and droughts were associated with higher odds of experiencing moderate-to-high water insecurity (\u003cstrong\u003eFigure 4\u0026nbsp;\u003c/strong\u003e\u003cem\u003e(grey bars)\u003c/em\u003e, Table S4 \u003cem\u003eModel 3\u003c/em\u003e).\u003c/p\u003e\n\u003cp\u003eFrom the logistic regression models stratified by climate zone, we found that climate zone modified the association between climate hazards and water insecurity, but these associations were not consistent across each type of hazard (\u003cstrong\u003eFigure 4A\u003c/strong\u003e, Tables S6 \u003cem\u003eModel 4\u003c/em\u003e). Storms were associated with higher odds of water insecurity in temperate (1.72 OR, 95% CI: 1.13-2.63) and tropical zones (1.35 OR, 95% CI: 1.01-1.79), but not in arid zones (0.37 OR, 95% CI: 0.13- 1.05). Floods were only associated with higher odds of water insecurity in tropical zones (1.20 OR, 95% CI:1.02-1.42). Droughts were associated with higher odds of experiencing water insecurity in temperate zones (1.48 OR, 95% CI: 1.05-2.09). The number of droughts reported by those living in tropical zones (n=55)) was low, so the confidence intervals were wide.\u003c/p\u003e\n\u003cp\u003eFrom the logistic regression models stratified by urbanicity, we observed that urbanicity modified the associations between climate event and water insecurity (\u003cstrong\u003eFigure 4B\u003c/strong\u003e, Tables S7 \u003cem\u003eModel 5\u003c/em\u003e). Droughts and floods were associated with higher odds of experiencing water insecurity in rural areas (floods 1.33 OR, 95% CI: 1.06-1.68; droughts 2.02 OR, 95% CI:1.22-3.36) but not in urban/peri-urban areas (floods 1.01 OR, 95% CI: 0.88-1.16; droughts 1.07 OR, 95% CI:0.79-1.44; \u003cstrong\u003eFigure 4B\u003c/strong\u003e, Table S7). On the other hand, storms were associated with higher odds of experiencing water insecurity in urban/peri-urban areas (1.47 OR, 95% CI: 1.13-1.93) but not in rural areas (0.96 OR, 95% CI: 0.65-1.41; \u003cstrong\u003eFigure 4B\u003c/strong\u003e, Table S7).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn the full-unstratified sample, using a logistic regression to model the association between rainfall anomalies (high and low SPI), neither high nor low SPI was associated with experiencing water insecurity (\u003cstrong\u003eFigure 5 A \u0026amp; B\u003c/strong\u003e, Tables S9 \u003cem\u003eModel 6\u003c/em\u003e). In the models stratified by climate zone and urbanicity, high nor low SPI was still not associated with experiencing water insecurity regardless of the climate zone or urbanicity of the respondent (\u003cstrong\u003eFigure 5 A \u0026amp; B\u003c/strong\u003e, Tables S9, \u003cem\u003eModels 7 \u0026amp; 8\u003c/em\u003e). \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFigure 4: Results from logistic regression models displaying the odds of experiencing moderate-high water insecurity if exposed to climatic events and the effect modification by geographic and economic vulnerabilities. Models were stratified by A) Climate Zone; B) Urbanicity; C) Subnational Wealth; and D) Individual Income (see Table S5 for the number of observations for each event within each stratified model)*\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e*Adjusted for perceived adequacy of household income, gender, age, household size, education level and COVID-19 pandemic-related life disruptions, and country-level fixed effects; and climate zone in all except A; and urbanicity in all except B; and subnational wealth in all except C; and individual income in all except D.\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e2.4 Do economic vulnerabilities (subnational wealth and individual income) modify the associations of climate hazards with water insecurity? \u0026nbsp;(H3)\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eTo investigate the hypotheses that the associations between water insecurity and climate hazards were modified by economic vulnerabilities, we ran logistic regression models in which we operationalised water insecurity as a binary variable. These models were stratified by subnational wealth and individual income and adjusted for key covariates. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSubnational wealth modified some of the associations between reported climate events and water insecurity (\u003cstrong\u003eFigure 4C\u0026nbsp;\u003c/strong\u003eand Table S10 \u003cem\u003eModel 9).\u0026nbsp;\u003c/em\u003ePeople living in areas of low subnational wealth who were exposed to storms had higher odds of experiencing water insecurity (1.54 OR, 95% CI:1.20-1.98; \u003cstrong\u003eFigure 4C\u0026nbsp;\u003c/strong\u003eand Table S10). \u0026nbsp;Among those living in an area of high subnational wealth, however, no difference in odds of experiencing water insecurity was seen between those exposed and not exposed to storms (0.62 OR, 95% CI: 0.37-1.04;\u003cstrong\u003e\u0026nbsp;Figure 4C\u003c/strong\u003e and Table S10). No association was found between water insecurity and exposure to the other climate events (floods and droughts), regardless of subnational wealth (\u003cstrong\u003eFigure 4C\u0026nbsp;\u003c/strong\u003eand Table S10).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe observed only one difference between the associations of reported climate events with water insecurity when models were stratified by individual income group (\u003cstrong\u003eFigure 4D\u003c/strong\u003e and Table S11 \u003cem\u003eModel 10\u003c/em\u003e). For those in low-income quintiles, floods were associated with higher odds of experiencing water insecurity (1.16 OR, 95% CI: 1.00-1.35, \u003cstrong\u003eFigure 4D\u003c/strong\u003e and Table S11), but no other event was associated with experiencing water insecurity regardless of individual wealth (\u003cstrong\u003eFigure 4D\u0026nbsp;\u003c/strong\u003eand Table S11).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFor SPI, similar to geographic vulnerabilities, in the models stratified by economic vulnerabilities exposure to high or low SPI was not associated with experiencing water insecurity regardless of the subnational wealth or individual income of the respondent (\u003cstrong\u003eFigure 5C \u0026amp; D\u003c/strong\u003e, Tables S9 \u003cem\u003eModel 11 \u0026amp; 12\u003c/em\u003e).\u003c/p\u003e"},{"header":"3. Discussion","content":"\u003cp\u003eIn this first analysis of the relationships between climate hazards and experiences of water insecurity, we found that climate hazards (floods, storms and droughts) were associated with higher water insecurity in a pooled sample covering 29 LMICs. These analyses represent a significant increase in the resolution of our understanding of how climate hazards relate to human well-being because they provide insights into individual experiences, rather than using proxies of water insecurity based on physical water supplies at community, national or watershed levels. They also bring to light the differential effects these hazards can have depending on geographic and economic vulnerabilities. \u003c/p\u003e\n\u003cp\u003eExposure to rainfall anomalies, as measured by SPI, was not associated with water insecurity. SPI is often used in early warning and climate hazard monitoring systems; however, the lack of association found between SPI and water insecurity suggests that this metric alone may not capture well the risks associated with climate hazards. Both heavy rainfall and decreased rainfall may have implications for water availability and safety (8), but these may not be captured well in satellite-derived gridded data which average rainfall across a grid-cell (5x 5km resolution), particularly as rainfall may vary widely over small distances (33). Increasing weather station coverage that can better capture these events could enable enhanced understanding of the associated risks.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAs hypothesised (H2), geographic factors, specifically climate zone and urbanicity, modified the association between climate hazards and water insecurity. \u0026nbsp;Whilst climate zones were found to be independently associated with water insecurity, with arid zones associated with higher risk of experiencing water insecurity compared to temperate zones, the associations between climate hazards and water insecurity differed inconsistently across these zones. For example, in temperate regions storms and droughts were associated with higher odds of being water insecure but floods were not. This suggests that we need to tailor climate adaptation strategies according to specific geographic characteristics, rather than using a generic strategy according to hazard type.\u003c/p\u003e\n\u003cp\u003eAlthough urbanicity was not associated with water insecurity, with no difference in the relative risk of experiencing water insecurity found between those residing in rural compared to urban areas, urbanicity modified the association between climate hazards and water insecurity. In other words, whilst living in a rural compared to an urban setting was unrelated to people’s water insecurity, urbanicity was associated with their vulnerability to different hazard types. Exposure to floods and droughts were associated with higher odds of water insecurity in rural areas, whereas exposure to storms was associated with higher odds of water insecurity in urban areas. This is consistent to an extent with our expectation that climate hazards, such as floods and droughts, disproportionately affect those in rural areas due lower access to improved water sources (15, 16) and infrastructure in general. While we had not anticipated that storms would be associated with higher water insecurity in urban but not rural settings, this may be due to built-up urban areas being more vulnerable to these hazards due to their drainage systems that can become overwhelmed (34). This pattern further indicates the need for differentiated investments based on specific geographic factors, including urbanicity as well as climate zone. For example, investing in greater flood and drought resilient infrastructure may be more important in rural areas while storm-proofing is more important in urban areas.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWhilst economic factors, specifically subnational wealth and individual income, were associated with lower relative risk of water insecurity, these factors did not seem to influence the vulnerability to climate hazards (H3). Although there were some exceptions; storms were associated with higher water insecurity in areas with low subnational wealth, and floods were associated with higher water insecurity among people with low individual incomes. The lack of differential associations observed for the other hazards makes it difficult to draw strong conclusions about how these economic factors affect vulnerability. Access to resilient infrastructure, better management of water systems and access to humanitarian assistance in disasters may play an important role in buffering the effect of climate hazards (19, 20, 35). \u0026nbsp;Further research is required to understand how these infrastructural investments and social programs may influence vulnerability to climate hazards beyond economic status.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eExperiencing water insecurity has substantial implications for people’s overall health and well-being. For example, it is closely associated with higher food insecurity and higher likelihood of experiencing poor physical and mental health outcomes (7). Given the close associations found between climate hazards and water insecurity, and that they differed by geographic vulnerabilities, this study highlights the importance of investing in climate adaptation measures that are context-specific. Monitoring using experiential measures of water security can allow for a more nuanced understanding of disparities, and how these may fluctuate because of these climate hazards. This in turn can help guide strategic investments within countries to mitigate the risks brought about by these hazards.\u003c/p\u003e\n\u003cp\u003eIn addition to these policy implications, this study identifies future research needs. One is greater investment in climate data. Many have highlighted the challenge of defining droughts as these events do not manifest as visibly as flood events, which may lead to underreporting (36-38). Whilst alternative metrics such as the Standardized Precipitation Evapotranspiration Index (SPEI) may potentially better capture meteorological drought than reported disaster or SPI data (39), geographic coverage of this data remains limited. Droughts are also typically slow onset, so a longer timeseries of water insecurity data than one year may be required to better understand the potential effects of this hazard. Future analyses might also explore how climate hazards shape each of the four domains of water security (7).\u003c/p\u003e\n\u003cp\u003eAlthough this study provides novel insights into the risks associated with climate hazards, using large nationally representative datasets, there are some limitations. Whilst climate hazards are exogenous and reverse causality is not plausible, we cannot infer causality as these data are cross-sectional; there may be other factors that increase vulnerability that were not captured. Furthermore, there are some temporal limitations in our ability to interpret the associations of these hazards with water insecurity, as the recall period of water insecurity was over the last 12 months, but people were classified as exposed to a climate event if they had been exposed to at least one event (or for SPI in at least one month) over that year. Collecting water insecurity experiences more frequently in longitudinal datasets using shorter recall periods, such as one month, would enable a more accurate understanding of the effect of climate events. A further limitation is that there may be inaccuracies in the reporting of climate events captured in the EM-DAT database. Individual water insecurity data could only be matched with climate event data at the administrative level, so we cannot be certain if all the respondents living within that unit were equally affected, i.e., there might be some exposure misclassification. Finally, the international wealth index was used as proxy variable, but it covers large administrative areas within which there might be variations in access to infrastructure that may influence the associations between climate hazards and water insecurity (13).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn conclusion, climate hazards were associated with higher water insecurity across 29 LMICs. Data on water insecurity experiences made evident the heterogenous toll of climate hazards and can guide appropriate climate adaptation strategies. Given the anticipated increase in frequency and intensity of climate events due to climate change, efforts should be made to address vulnerabilities and disparities that may occur due to both climate hazards and geographic and economic factors. More robust (higher resolution, higher frequency) measures of both climate events and water insecurity can help us to better understand the risks of climate events and support the development of strategies to mitigate the harms of climate hazards to both water security and health in places where risks are the greatest. \u0026nbsp;\u003c/p\u003e"},{"header":"4. Methods ","content":"\u003cp\u003e\u003cem\u003e4.1 Survey design of the Gallup World Poll\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eNationally representative data on individuals’ experiences with water insecurity were collected in 29 countries between September 2020 and February 2021 by the Gallup World Poll (GWP). These countries included 20 from Sub-Saharan Africa (\u003cem\u003eBenin, Burkina Faso, Cameroon, Congo Brazzaville, Cote d’Ivoire, Ethiopia, Gabon, Ghana, Guinea, Kenya, Mali, Mauritius, Namibia, Nigeria, Senegal, South Africa, Tanzania, Togo, Uganda, Zambia\u003c/em\u003e), 4 from North Africa (\u003cem\u003eAlgeria, Egypt, Morocco and Tunisia\u003c/em\u003e), 3 from Latin America (\u003cem\u003eBrazil, Guatemala and Honduras\u003c/em\u003e) and 2 from Asia (\u003cem\u003eIndia and Bangladesh\u003c/em\u003e), see Figure S1. GWP data collection procedures are described in detail elsewhere (25). Briefly, non-institutionalized individuals aged 15 or older were eligible and were randomly sampled using stratified sampling methods depending on sampling frame (telephone vs face-to-face). Surveys were conducted predominantly via telephone due to COVID-19 restrictions, except for Mali, Senegal and two out of three of the survey waves in India that were conducted face-to-face.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eRespondent geolocations were determined by recording the GPS location of the primary sampling units in face-to-face surveys and via a series of respondents’ answers to location-related questions in telephone surveys. If the location of a respondent could not be determined from the telephone surveys, then the centroid of their administrative unit (level 1) was assigned (which is typically a large subdivision within a country such as a province, region, or state). In these cases, the respondent was flagged as not having an accurate geolocation and excluded from the analysis. Geolocations were dislocated by 2km in urban and 5km in rural in any direction, and an additional random 1% of rural by 10km, from the original location to ensure anonymity; all dislocation remained with the original level 1 administrative unit.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e4.2 Outcome: water insecurity\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe primary dependent variable was experiences with water insecurity measured at the individual level using the Individual Water Insecurity Experiences (IWISE) Scale, which has been established as reliable, valid, and equivalent for making cross-country comparisons (25). The 12 items in the scale query about universal experiences with water, including limitations to water-related behaviours (e.g., unable to wash hands), psychosocial effects (e.g., worrying about water), and interruptions in water supply (Supplementary Table S12). Respondents were asked how frequently they had experienced these issues over the last 12 months, with answer options “never (score 0)”, “rarely (1-2 months, score 1)”, “sometimes (in some but not every month, score 2)” or “often/always (in almost every month, score 3)”. Responses were summed to create a score with a range of 0 to 36, with higher scores indicating higher water insecurity. Respondents missing one or more IWISE responses were excluded (n=1,483, 3.6%).\u003c/p\u003e\n\u003cp\u003eFour ordinal categories of water insecurity were made using established cut-points: “no-to-marginal” (0-2); “low” (3-11), “moderate” (12-23); or “high” water insecurity (scores of 24-36)\u0026nbsp;(29). \u0026nbsp;For the analyses exploring how geographic and economic vulnerabilities modified the association between climate hazards and water insecurity, we operationalised a binary outcome representing moderate-to-high water insecurity\u0026nbsp;(IWISE score \u003cu\u003e\u0026gt;\u003c/u\u003e12), referred to as “water insecurity” in the results (cf. 2.3 \u0026amp; 2.4).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e4.3 Exposure variables: climate events and precipitation anomalies\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe primary exposure variables were reported climate disaster events (drought, flood and storms) and precipitation anomalies based on the Standard Precipitation Index (SPI).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe extracted reported climate disaster events from 2019-2021 that occurred in the countries corresponding to the GWP dataset from The International Disaster Database (EM-DAT)(26). The EM-DAT data includes events that have reported at least ten deaths, at least 100 affected people and/or a response that involves either a call for international assistance or a declaration of an emergency. We extracted events classified as a “drought”, “flood” and “storm” (26). There was one glacial lake outburst that we reclassified as a flood. We linked each event to the lowest possible administrative unit using the Global Administrative Areas (GADM) administrative unit boundaries (40), based on the location names and administrative unit provided in the EM-DAT database. We assumed that an event impacted all areas of an administrative unit equally given the geographical information available in the data. We considered a respondent to be exposed to an event category if their administrative unit had experienced at least one event in the last 12 months prior to the interview.\u003c/p\u003e\n\u003cp\u003eWe extracted Standardised Precipitation Index (SPI) data from the Climate Engine\u0026nbsp;(27, 28), which uses satellite-derived rainfall data from The Climate Hazards Group InfraRed Precipitation with Station (CHIRPS)(41)) \u0026nbsp;to classify precipitation anomalies from the long-term average. SPI values are calculated using the number of standard deviations from the mean of accumulative precipitation for a given location and period. We classified areas as being exposed to high precipitation if the SPI values were \u0026gt; 1.5 (corresponding to extremely and very wet conditions) and low precipitation if SPI values were \u0026lt; -1.5 (corresponding to extremely dry or severely dry conditions (42, 43)). We derived frequencies for each respondent in the GWP by collecting monthly SPI values which give the deviation, for a given month, from the monthly mean between 1991 and 2020. These values were calculated for each CHIRPS grid cell (5x 5km resolution) corresponding to the countries within the GWP dataset. We extracted SPI data for each respondent based on dislocated geolocations.\u0026nbsp;We then used these monthly classifications to count the frequency each respondents’ location had been exposed to high or low precipitation (as defined previously) in the 12 months prior to survey. For our analysis, we considered ‘exposed’ to be at least 1 month exposure to high or low SPI.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e4.4 Vulnerability variables: geographic and economic factors\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eWe investigated how geographic (climate zone and urbanicity) and economic (sub-national wealth and individual income) factors might modify the risk of water insecurity associated with climatic events, specifically, cf. Figure 2.\u003c/p\u003e\n\u003cp\u003eWe classified the climate zone of each respondent in the GWP using the Köppen-Geiger climate classification maps of 10km spatial resolution, version 1 (30). Five groups are included in this classification: tropical, arid, temperate, cold (continental), and polar. Due to the small sample size of respondents in cold and polar regions (n=44), we re-assigned them to the most common climate zone in their country.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe level of urbanicity for each respondent was provided by the GWP according to their DEGURBA classification (31). Based on these DEGURBA classifications we generated a dichotomous variable representing whether lived in an urban/ peri-urban areas (“1”) or rural areas (“0”).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSubnational wealth was established using the International Wealth Index (IWI) at subnational level (32), from Global Data Lab (44, 45). This index is based on assets derived from household survey data that are comparable between countries (32). The subnational area names were matched to GADM administrative units, and IWI values were then linked to respondents based on their administrative unit. Tertile IWI brackets were generated per country and assigned to respondents based on their subnational region. These tertiles were combined to form a dichotomous variable representing low (lowest two tertiles) and high (highest tertile) subnational wealth.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIndividual income was operationalised based on the income quintile for each respondent in GWP within the income distribution for that country. We created a dichotomous variable for individual income, pooling the quintiles of the lowest three brackets to represent low individual income, and the highest two brackets to represent high individual income.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e4.5 Other covariates\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eWe also extracted other individual covariates from the GWP dataset that have been shown to be relevant for water insecurity elsewhere, as these may be related with individual vulnerability: perceived adequacy of household (HH) income (a self-reported measure of wealth beyond what is captured using income; \u003cem\u003eliving comfortably or getting by on HH income “0” or finding it difficult or very difficult on HH income “1”\u003c/em\u003e), \u0026nbsp;gender (\u003cem\u003ewomen “1” or men “0\u003c/em\u003e”), age (c\u003cem\u003eontinuous variable\u003c/em\u003e), household size \u0026nbsp;(c\u003cem\u003eontinuous variable\u003c/em\u003e), education level \u0026nbsp;(\u003cem\u003eprimary “0”, secondary “1\u003c/em\u003e”, tertiary “2”), and COVID-19 pandemic-related life disruptions (a self-reported measure of how the pandemic has disrupted their lives; \u003cem\u003enot at all “0”, some “1\u003c/em\u003e”, a lot “2”) (24, 46).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e4.6 Statistical analysis\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eWe linked GWP respondents to the exposure (climate hazards) and vulnerability variables (climate zone, sub-regional wealth) using the geolocation data. All analyses were conducted using R\u0026nbsp;(47).\u003c/p\u003e\n\u003cp\u003eOf a total 41,048 respondents, we excluded those with missing IWISE items (n = 1,483), country-coordinate mismatches (n = 2), flagged inaccurate geolocations (n = 1,876), or missing covariates (n = 1,345), yielding 36,342 respondents across 29 countries. Survey weights were normalised and rescaled within country for pooled analysis. To assess any exclusion bias from geolocation flags, we compared WISE category distributions between all vs non-flagged respondents per country (chi-squared; Table S13); no evidence of bias was found.\u003c/p\u003e\n\u003cp\u003eBecause SPI measures rainfall anomalies, with low SPI indicative of abnormally dry conditions and high SPI indicative of abnormally wet conditions, exposure to low SPI might be associated with exposure to reported droughts and exposure to high SPI might be associated with exposure to reported floods. We tested these associations using a contingency table analysis; these results are reported in the participants characteristics section in the results section (cf. 2.1, Table S1).\u003c/p\u003e\n\u003cp\u003eWe estimated associations of water insecurity categories with (i) reported climate events and (ii) SPI using survey-weighted multinomial logistic regression with country fixed effects, adjusting for geographic (climate zone, urbanicity) and economic (individual income and subnational wealth) vulnerabilities, and other key covariates: perceived adequacy of household income, gender, age, household size, education level and COVID-19 pandemic-related life disruptions. \u0026nbsp;Because multiple hazards can occur over a 12-month period, models for reported events included all three climate event types simultaneously; SPI models included high and low SPI simultaneously.\u003c/p\u003e\n\u003cp\u003eTo understand if geographic factors (H2), specifically climate zone and urbanicity, modified the association between water insecurity and climate hazards (reported events and SPI), we tested for an interaction using logistic regression models. In these models we classified the outcome as a binary variable representing moderate-to-high water insecurity (IWISE score \u003cu\u003e\u0026gt;\u003c/u\u003e12) to reduce the complexity of the models and aid interpretability of the results. Because evidence of an interaction was found with some of these climate events, we ran stratified models to examine the differences in these effects, to further aid interpretability. Models were stratified by climate zone: arid, temperate and tropical. Models were also stratified by urbanicity: rural and urban (including peri-urban). Sensitivity analysis was conducted using a self-reported definition of urbanicity, which yielded similar results (Supplementary Text 1, Figure S3).\u003c/p\u003e\n\u003cp\u003eTo understand if economic factors (H3), specifically subnational wealth and individual income, modified the association between water insecurity and climate hazards, we ran logistic regression models stratified by subnational wealth (low and high) and individual wealth (low and high).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAll models included country-level fixed effects, were survey-weighted and adjusted for perceived adequacy of household income, gender, age, household size, education level and COVID-19 pandemic-related life disruptions. \u0026nbsp;Models were also adjusted for geographic (urbanicity, climate zone) and economic (individual income and subnational wealth) factors, as appropriate. We report \u0026nbsp; report adjusted relative risk ratios (aRRR) and adjusted odds ratios (aOR) with 95% Cis.\u003c/p\u003e\n\u003cp\u003e4.7 Ethics\u003c/p\u003e\n\u003cp\u003eGallup World Poll sought approval for their survey from theh\u0026nbsp;governing bodies required in each country and obtained informed consent from all participants. For this analysis we received deidentified data from GWP. \u003cstrong\u003e\u003c/strong\u003e\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003eThis study was based on deidentified data made available by Gallup. Gallup World Poll followed their standard protocol for obtaining consent from participants. The authors of this paper were not involved with the consent or data collection process. No participants were involved in study design, implementation, or dissemination, including the writing of this manuscript. Informed consent of all survey participants was obtained, and survey protocols were approved by Gallup\u0026rsquo;s Internal Review Board and by the governing bodies in countries where approval is required\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eIPCC. Climate Change 2022: Impacts, Adaptation, and Vulnerability. Contribution of Working Group II to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change Cambridge University Press. Cambridge University Press, Cambridge, UK and New York, NY, USA,; 2022.\u003c/li\u003e\n \u003cli\u003eCaretta AMMA, Arfanuzzaman RBM, Morgan SMR, Kumar M. Water. In: Climate Change 2022: Impacts, Adaptation, and Vulnerability. Contribution of Working Group II to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change. 2022.\u003c/li\u003e\n \u003cli\u003eKuzma S, Saccoccia L, Chertock M. Water Resources Institute. 2023. [23/01/2025]. Available from: https://www.wri.org/insights/highest-water-stressed-countries.\u003c/li\u003e\n \u003cli\u003eJepson WE, Wutich A, Colllins SM, Boateng GO, Young SL. Progress in household water insecurity metrics: a cross-disciplinary approach. WIREs Water. 2017;4(3):e1214.\u003c/li\u003e\n \u003cli\u003eYoung S, Miller J, Bose I. Measuring human experiences to advance safe water for all. Evanston: Institute for Policy Research, Northwestern University. 2024.\u003c/li\u003e\n \u003cli\u003eO\u0026apos;Neill BC, van Aalst M, Zaiton Ibrahim Z, Berrang-Ford L, Bhadwal S, Buhaug H, et al. Key risks across sectors and regions. Cambridge University Press, Cambridge, UK and New York, NY, USA,; 2022.\u003c/li\u003e\n \u003cli\u003eRosinger AY, Young SL. The toll of household water insecurity on health and human biology: Current understandings and future directions. WIREs Water. 2020;7(6):e1468.\u003c/li\u003e\n \u003cli\u003eMellor JE, Levy K, Zimmerman J, Elliott M, Bartram J, Carlton E, et al. Planning for climate change: The need for mechanistic systems-based approaches to study climate change impacts on diarrheal diseases. Science of The Total Environment. 2016;548-549:82-90.\u003c/li\u003e\n \u003cli\u003eLevy K, Woster AP, Goldstein RS, Carlton EJ. Untangling the Impacts of Climate Change on Waterborne Diseases: a Systematic Review of Relationships between Diarrheal Diseases and Temperature, Rainfall, Flooding, and Drought. Environ Sci Technol. 2016;50(10):4905-22.\u003c/li\u003e\n \u003cli\u003eBroyles LM, Pakhtigian EL, Rosinger AY, Mejia A. Climate and hydrological seasonal effects on household water insecurity: A systematic review. Wiley Interdisciplinary Reviews: Water. 2022;9(3):e1593.\u003c/li\u003e\n \u003cli\u003eRosinger AY. Household water insecurity after a historic flood: Diarrhea and dehydration in the Bolivian Amazon. Social Science \u0026amp; Medicine. 2018;197:192-202.\u003c/li\u003e\n \u003cli\u003eBose I, Dreibelbis R, Green R, Murray KA, Ceesay O, Kovats S. Coping strategies for household water insecurity in rural Gambia, mediating factors in the relationship between weather, water and health. BMC Public Health. 2024;24(1):3150.\u003c/li\u003e\n \u003cli\u003eBose I, Dreibelbis R, Green R, Murray KA, Ceesay O, Kovats S. Climate change, seasonality and household water security in rural Gambia: A qualitative exploration of the complex relationship between weather and water. PLOS Water. 2024;3(6):e0000239.\u003c/li\u003e\n \u003cli\u003eV\u0026ouml;r\u0026ouml;smarty CJ, Green P, Salisbury J, Lammers RB. Global Water Resources: Vulnerability from Climate Change and Population Growth. Science. 2000;289(5477):284-8.\u003c/li\u003e\n \u003cli\u003eBain RES, Wright JA, Christenson E, Bartram JK. Rural: urban inequalities in post 2015 targets and indicators for drinking-water. Science of The Total Environment. 2014;490:509-13.\u003c/li\u003e\n \u003cli\u003eUNICEF, WHO. Joint Monitoring Programme for Water Supply, Sanitation and Hygiene (JMP) https://washdata.org/ [\u003c/li\u003e\n \u003cli\u003eHoward G, Bartram J, World Health O, United Kingdom. Dept. for International D, United States of America. University of North Carolina at Chapel H. Vision 2030 : the resilience of water supply and sanitation in the face of climate change: technical report / Guy Howard, Jamie Bartram. Geneva: World Health Organization; 2010.\u003c/li\u003e\n \u003cli\u003eMacAllister DJ, MacDonald AM, Kebede S, Godfrey S, Calow R. Comparative performance of rural water supplies during drought. Nat Commun. 2020;11(1):1099.\u003c/li\u003e\n \u003cli\u003eLeal Filho W, Totin E, Franke JA, Andrew SM, Abubakar IR, Azadi H, et al. Understanding responses to climate-related water scarcity in Africa. Science of The Total Environment. 2022;806:150420.\u003c/li\u003e\n \u003cli\u003eCharles KJ, Howard G, Villalobos Prats E, Gruber J, Alam S, Alamgir ASM, et al. Infrastructure alone cannot ensure resilience to weather events in drinking water supplies. Science of The Total Environment. 2022;813:151876.\u003c/li\u003e\n \u003cli\u003eGeremew G, Cumming O, Haddis A, Freeman MC, Ambelu A. Rainfall and Temperature Influences on Childhood Diarrhea and the Effect Modification Role of Water and Sanitation Conditions: A Systematic Review and Meta-Analysis. International Journal of Environmental Research and Public Health. 2024;21(7):823.\u003c/li\u003e\n \u003cli\u003eMeehan K, Jepson W, Harris LM, Wutich A, Beresford M, Fencl A, et al. Exposing the myths of household water insecurity in the global north: A critical review. Wiley Interdisciplinary Reviews: Water. 2020;7(6):e1486.\u003c/li\u003e\n \u003cli\u003eMeehan K, Jurjevich JR, Chun NM, Sherrill J. Geographies of insecure water access and the housing\u0026ndash;water nexus in US cities. Proceedings of the National Academy of Sciences. 2020;117(46):28700-7.\u003c/li\u003e\n \u003cli\u003eYoung SL, Bethancourt HJ, Ritter ZR, Frongillo EA. Estimating national, demographic, and socioeconomic disparities in water insecurity experiences in low-income and middle-income countries in 2020\u0026ndash;21: a cross-sectional, observational study using nationally representative survey data. The Lancet Planetary Health. 2022;6(11):e880-e91.\u003c/li\u003e\n \u003cli\u003eYoung SL, Bethancourt HJ, Ritter ZR, Frongillo EA. The Individual Water Insecurity Experiences (IWISE) Scale: reliability, equivalence and validity of an individual-level measure of water security. BMJ Global Health. 2021;6(10):e006460.\u003c/li\u003e\n \u003cli\u003eEM-DAT [Internet]. 2023 [cited February 2023]. Available from: www.emdat.be.\u003c/li\u003e\n \u003cli\u003eClimate Engine, version 2.1. [Internet]. 2023 [cited 14th November 2023]. Available from: http://climateengine.org.\u003c/li\u003e\n \u003cli\u003eHuntington JL, Hegewisch KC, Daudert B, Morton CG, Abatzoglou JT, McEvoy DJ, et al. Climate Engine: Cloud Computing and Visualization of Climate and Remote Sensing Data for Advanced Natural Resource Monitoring and Process Understanding. Bulletin of the American Meteorological Society. 2017;98(11):2397-410.\u003c/li\u003e\n \u003cli\u003eFrongillo EA, Bethancourt HJ, Miller JD, Young SL. Identifying ordinal categories for the Water Insecurity Experiences Scales. Journal of Water, Sanitation and Hygiene for Development. 2024:washdev2024042.\u003c/li\u003e\n \u003cli\u003eBeck HE, Zimmermann NE, McVicar TR, Vergopolan N, Berg A, Wood EF. Present and future K\u0026ouml;ppen-Geiger climate classification maps at 1-km resolution. Scientific data. 2018;5(1):1-12.\u003c/li\u003e\n \u003cli\u003eDijkstra L, Florczyk AJ, Freire S, Kemper T, Melchiorri M, Pesaresi M, et al. Applying the degree of urbanisation to the globe: A new harmonised definition reveals a different picture of global urbanisation. Journal of Urban Economics. 2021;125:103312.\u003c/li\u003e\n \u003cli\u003eSmits J, Steendijk R. The International Wealth Index (IWI). Social Indicators Research. 2015;122(1):65-85.\u003c/li\u003e\n \u003cli\u003eColston JM, Ahmed T, Mahopo C, Kang G, Kosek M, de Sousa Junior F, et al. Evaluating meteorological data from weather stations, and from satellites and global models for a multi-site epidemiological study. Environmental Research. 2018;165:91-109.\u003c/li\u003e\n \u003cli\u003eZhou Q. A review of sustainable urban drainage systems considering the climate change and urbanization impacts. Water. 2014;6(4):976-92.\u003c/li\u003e\n \u003cli\u003eStringer LC, Mirzabaev A, Benjaminsen TA, Harris RMB, Jafari M, Lissner TK, et al. Climate change impacts on water security in global drylands. One Earth. 2021;4(6):851-64.\u003c/li\u003e\n \u003cli\u003eLloyd-Hughes B. The impracticality of a universal drought definition. Theoretical and applied climatology. 2014;117:607-11.\u003c/li\u003e\n \u003cli\u003eWilhite DA, Glantz MH. Understanding: the drought phenomenon: the role of definitions. Water international. 1985;10(3):111-20.\u003c/li\u003e\n \u003cli\u003eStanke C, Kerac M, Prudhomme C, Medlock J, Murray V. Health effects of drought: a systematic review of the evidence. PLoS Curr. 2013;5.\u003c/li\u003e\n \u003cli\u003eTirivarombo S, Osupile D, Eliasson P. Drought monitoring and analysis: Standardised Precipitation Evapotranspiration Index (SPEI) and Standardised Precipitation Index (SPI). Physics and Chemistry of the Earth, Parts A/B/C. 2018;106:1-10.\u003c/li\u003e\n \u003cli\u003eGADM database of Global Administrative Areas, version 2.0 [Internet]. 2012 [cited July 2022]. Available from: https://gadm.org.\u003c/li\u003e\n \u003cli\u003eFunk CC, Peterson PJ, Landsfeld MF, Pedreros DH, Verdin JP, Rowland JD, et al. A quasi-global precipitation time series for drought monitoring. Report. Reston, VA; 2014. Report No.: 832.\u003c/li\u003e\n \u003cli\u003eGuerreiro MJ, Lajinha T, Abreu I. Flood analysis with the standardized precipitation index (SPI). 2007.\u003c/li\u003e\n \u003cli\u003eOlanrewaju CC, Reddy M. Assessment and prediction of flood hazards using standardized precipitation index\u0026mdash;A case study of eThekwini metropolitan area. Journal of Flood Risk Management. 2022;15(2):e12788.\u003c/li\u003e\n \u003cli\u003eSmits J. GDL Area Database. Nijmegen, The Netherlands. 2016:16-101.\u003c/li\u003e\n \u003cli\u003eGlobal Data Lab International Wealth Index, version v1.0. [Internet]. Available from: https://globaldatalab.org/wealth/.\u003c/li\u003e\n \u003cli\u003eYoung SL, Bethancourt HJ, Frongillo EA, Viviani S, Cafiero C. Concurrence of water and food insecurities, 25 low-and middle-income countries. Bulletin of the World Health Organization. 2023;101(2):90-101.\u003c/li\u003e\n \u003cli\u003eTeam RC. _R: A Language and Environment for Statistical Computing_. R Foundation for Statistical Computing, Vienna, Austria; 2023.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-7959507/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7959507/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Climate hazards (floods, storms, droughts, and rainfall anomalies) are intensifying, threatening national water security. Individual water security, however, likely varies by exposure to climate hazards and people’s geographic and economic vulnerabilities. We spatiotemporally linked climate hazards with nationally representative survey data on individual water insecurity experiences from 29 low- and middle-income countries (n=36,342). We estimated associations between each hazard and the probability of experiencing water insecurity, and examined heterogeneity by vulnerability (climate zone, urbanicity, subnational wealth and individual income). Floods, droughts and storms were associated with higher water insecurity, rainfall anomalies were not. Geographic and economic vulnerabilities modified these associations, but the directionality varied by hazard type. For example, floods and droughts were associated with higher water insecurity in rural areas, but storms were associated with higher water insecurity in urban areas. Water insecurity experiences make evident the heterogenous toll of climate hazards, and can guide appropriate climate adaptation strategies.","manuscriptTitle":"Climate hazards and water insecurity, by geographic and economic vulnerabilities","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-11-10 09:08:54","doi":"10.21203/rs.3.rs-7959507/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"7e726a39-5777-402c-8fae-6104a857be7f","owner":[],"postedDate":"November 10th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":56946981,"name":"Health sciences/Risk factors"},{"id":56946982,"name":"Earth and environmental sciences/Environmental social sciences/Climate-change impacts/Environmental health"},{"id":56946983,"name":"Earth and environmental sciences/Natural hazards"},{"id":56946984,"name":"Earth and environmental sciences/Hydrology"}],"tags":[],"updatedAt":"2025-11-27T14:41:41+00:00","versionOfRecord":[],"versionCreatedAt":"2025-11-10 09:08:54","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7959507","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7959507","identity":"rs-7959507","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}