Europe has been identified as a heatwave hotspot, with numerous temperature records having been broken in recent summers and roughly 60,000 and 50,000 heat-related deaths occurring in the summers of 2022 and 2023, respectively. With recent summers, like that of 2022, projected to become the new norm, there is a pressing need to further develop heat-health warning systems to help society adapt to a warming climate. Here, we evaluate the skill of daily temperature related mortality forecasts, which can inform heat-warning systems, for the summers of 2022 and 2023. For most parts of Europe, enhanced temperature related mortality forecasts were associated with milder temperatures, close to the minimum mortality temperature. However, some of the hottest regions in Europe showed increased predictability associated with higher temperatures, suggesting that mortality forecasts can provide valuable information in regions also associated with high levels of temperature related mortality.

Diarrheal disease (DD) is a significant global public health issue and represents the leading cause of malnutrition and the second leading cause of death worldwide, resulting in the deaths of an estimated 829,000 people annually (Keddy, 2018). DD is a particularly urgent public health threat accounting for nearly 10% of global deaths of children under the age of five (Colombara, 2016).

Climate change and the natural environment are associated with DD risk. The World Health Organization (WHO) estimates that climate change will cause an additional 48,000 deaths per year from diarrhea between 2030 and 2050 (WHO, 2018). Much of the literature supports an association between high rainfall and increased risk of diarrhea (Mertens, 2019, Levy, 2016. How local variability in the natural environment impacts DD risk is understudied and requires remote, big data approaches to enhance community preparedness to reduce DD incidence among vulnerable populations.

This study applied Random Forest (RF) machine learning models to climatic, environmental, health, socio-economic, and geospatial data. The data used for this study originated from the Demographic and Health Surveys Program (DHS). Data was accumulated at the household (HH) level, with each HH belonging to a village that represents a collective of geographically proximal HHs.

Two RF classification algorithms were used to investigate feature importance at the HH and village scales. Model A was used for binary coding of DD occurrence at the HH level. Model B applied a prevalence threshold and coded households within a village according to DD prevalence.

The results of this study suggest that the contextual features most important for predicting DD are different at the household and village scales. At the household scale (Model A, Figure 1), the features identified as most important are related to the age and stunting status of the child, followed by geographic and weather-related variables. At the village scale (Model B, Figure 2), the features identified as most important were generally geographic and weather-related, including distance from the ocean and precipitation.

These results demonstrate the importance of scale when considering preparedness and emergency planning and demonstrate the utility of machine learning in preparedness, warning, and response efforts. Our findings suggest nuance is required in the analysis of DD across regions of the world where local variability in climate and natural hazard vulnerability may be an important consideration in devising appropriate preparedness and response strategies for reducing diarrheal disease burden.

Climate change has exposed communities to multiple hazards, making them vulnerable to more than one hazard. Scientific predictions indicate that global climate changes are likely to further increase exposure to multiple risks, affecting the magnitude, frequency, and spatial distribution of disastrous events. However, risk assessments often consider hazards as independent crises, ignoring their combined impacts. Consequently, there is a critical need to adopt a multi-hazard approach for risk assessment, enabling institutions and policymakers to devise more effective mitigation strategies. Against this backdrop firstly, the study aims to identify the hotspots in the Indian state of Rajasthan affected by the combined impacts of heat waves and droughts using a multi-hazard approach. The study seeks to understand the relationship between two extreme events in the region. Secondly, the study tries to examine the implementation of multi-risk governance approach at the institutional level, highlighting associated challenges and gaps. To better understand these gaps, the study includes a stakeholder analysis approach and grassroots community investigation through primary surveys in both rural and urban areas. A household survey of 150 farmers and three focus group discussions (FGDs) were conducted within the agricultural community in one of the identified rural hotspot region. To understand the urban scenarios, four FGDs were conducted in four slum regions of the capital city of Jaipur.

The study maps both extreme events as hazards in the region using the Standardized Precipitation Index and the India Meteorological Department classification, utilizing data from 165 meteorological stations. The study establishes a positive feedback mechanism between both events in the region. The finding of the study identifies the northeastern part of the state as the hotspot region, which is at greater risk from the combined impacts of heat waves and droughts. In the identified hotspot regions, the study establishes Stakeholder interviews at the institutional level, along with household surveys and focus group discussions (FGDs), revealed several governance challenges. These challenges span institutional, operational, social, economic, behavioral, communication, information and awareness, and political domains. To enhance community resilience to these events, it is essential to address these challenges through a bottom-up approach, ensuring the last-mile integration of policies and plans. A shift from recovery to preparedness is necessary, along with long-term planning, efficient communication, and collaboration among different stakeholders. Implementing Community-Based Disaster Management (CBDM) plans and conducting awareness drives are also crucial steps in fostering resilience. Thus, the finding of study opines that the drought and heat wave mitigation measures deployed in the state are inadequate, ineffective and mostly reactive in nature. As a result, the integration and bridging the gap between scientific risk assessments and the practices implemented at both institutional and community levels is crucial for enhancing disaster risk reduction (DRR) strategies and ensuring more effective and practical approaches to managing risks. The study further leaves the scope for a holistic framework for proactive multi hazard management which will focus on building a resilient and drought and heat wave proof society in the country by emphasizing on interdisciplinary research and collaboration for targeted interventions.