It was agreed to priortise these indicators as they can provide regular, reliable, and comparable data to monitor climate-related health impacts using state-of-the-art statistical methods.

Further justification for the focus on this indicator is explained in the Topic introduction , which also outlines supplementary and additional indicators that are not in scope for the current SOSCHI framework but are recommended for future development.

Each indicator has four component parts:

For guidance on the process for calculating these indicator metrics, see the Methodology section below.

These indicators can give insights into how changes in temperature and extreme rainfall events influence their risk of waterborne diseases, and how these impacts vary geographically and among the most vulnerable groups in society, informing local adaptation needs.

While diarrheal disease incidence does not capture all pathways through which climate change affects health, it is a well established and sensitive indicator of waterborne health impacts, particularly in setting with limited access to safe water, sanitation and hygiene (WASH) services.

This indicator can therefore:

• Monitor and quantify the health burden associated with extreme rainfall and extreme temperature through waterborne disease outcomes.
• Inform national adaptation plans (NAPs) and resilience strategies
• Guide environmental and public health policy to reduce risks and improve outcomes

The indicator uses epidemiological methods to examine short-term links between climate exposure and waterborne disease outcomes. The analysis uses a spatiotemporal Bayesian model combined with a Distributed Lag Nonlinear Model (DLNM). The DLNM captures both nonlinear and delayed effects of climate on monthly diarrhea counts, while the Bayesian framework accounts for spatial and temporal correlations.

The following data are needed for this indicator production:

• Monthly count of diarrhea reported cases from national administrative records.
• Climate data from the National Meteorological Agency.
• Population estimates from national statistical agencies.

Where this data is unavailable or observations are incomplete, proxy datasets may be used. Further information on these potential sources as well as required variable names and formats are outlined in section 2 of the Waterborne disease and water-related illnesses: methodology

Diarrheal diseases are a key indicator of climate-sensitive waterborne health outcomes. Exposure to extreme temperatures and rainfall can influence diarrhea risk through multiple pathways, including pathogen survival and replication, water contamination, overload of sanitation systems, and changes in hygiene behaviour.

These indicators estimate the burden of diarrhea associated with climate conditions that increase or decrease risk at a monthly time scale, supporting climate adaptation planning, WASH interventions, and public health decision-making. Diarrhea risk is influenced by multiple interacting factors, and extreme temperatures and rainfall represent one component of this broader system.

In the analysis, extreme temperature refers to the monthly maximum temperature, while extreme rainfall refers to the cumulative monthly rainfall. Extreme conditions are identified using relative risk (RR) rather than percentile-based thresholds. Months where RR > 1 are considered harmful, meaning the temperature or rainfall conditions are associated with increased diarrhea risk, while RR < 1 indicates protective conditions with lower risk.

The relative risk is estimated from a statistical model linking monthly diarrhea incidence to monthly maximum temperature and cumulative rainfall. Extremes, therefore, reflect health-relevant risk conditions, rather than rare values in the climate distribution. See Topic introduction for more details.

No. These are model-based attribution estimates, not direct observations of climate-caused diarrhea cases. The analysis models the association between monthly diarrhea counts and monthly maximum temperature and cumulative rainfall, while accounting for delayed and cumulative effects, and spatiotemporal variability at the monthly scale.

“Attributable” refers to the estimated number, rate, or fraction of diarrhea cases associated with months in which climate conditions increase risk. In this analysis, attribution is calculated only for months with a relative risk (RR) greater than 1, indicating harmful temperature or rainfall conditions. It represents the portion of the diarrhea burden that would not have occurred in the absence of these risk-increasing climate conditions, as estimated from the modeled exposure–response relationship.

The analysis uses a spatiotemporal Bayesian model combined with a Distributed Lag Nonlinear Model (DLNM). The DLNM captures both nonlinear and delayed effects of climate on monthly diarrhea counts, while the Bayesian framework accounts for spatial and temporal correlations.

This combination enables robust estimates of relative risk and attributable burden across multiple locations over time using the R-INLA package. Depending on the dataset size and spatial coverage, running the model can take 3 to 5 minutes.

The reliability of these estimates depends on the quality of monthly diarrhea case data, the representativeness of monthly maximum temperature and cumulative rainfall, and the strength of the modeled climate–diarrhea relationship.

Reliability is strengthened because the model can:

• Borrow strength across space and time, producing more stable estimates in areas or months with low case counts.
• Capture delayed and nonlinear effects of temperature and rainfall on diarrhea risk, reflecting cumulative or lagged impacts rather than only immediate associations.
• Account for spatial and temporal correlations, recognizing that diarrhea risk is related across neighboring regions and adjacent months.
• Quantify uncertainty rigorously through credible intervals, which incorporate variability from the data, model structure, and parameter estimation.

While this approach improves robustness compared with raw counts alone, users should still interpret results cautiously, particularly at sub-national scales or for smaller populations. Detailed methodological limitations are described in waterborne disease and water-related illnesses: methodology.

Yes, waterborne diarrhea indicators are produced at multiple spatial levels, including national, regional, and district scales, and for key sub-populations such as children under five years of age.

Estimates for the under-five population are considered reliable because diarrhea incidence is higher in this group, providing sufficient data for stable modeling. As with all analyses, uncertainty intervals (credible intervals) should still be considered when interpreting results, particularly at finer spatial scales.