AGU Fall Meeting 2023 is coming. I will serve as a chair of the GeoHealth (GH23): Surviving the Dry Spell: Understanding the Connection Between Drought Health, which will be held from 14:10 to 15:40 on December 12. Here is the abstract of the session.
Drought is a natural disaster that can have a profound impact on the environment, economy, and human health. Droughts can lead to food and water scarcity, malnutrition, water-borne diseases, and increased environmental hazards, such as smoke, dust, and heat. Moreover, droughts can exacerbate existing health conditions such as respiratory diseases, infectious diseases, and mental health problems. As droughts are becoming more frequent and severe due to climate change, it is important to understand the relationships with health for improved public health mitigation strategies. This session will bring together researchers, public health officials, and policymakers to share their latest research findings and insights on the relationship between drought and human health. The session will cover topics such as the direct and indirect health impacts of drought, the mechanisms linking drought and health outcomes, and effective strategies for reducing the negative effects of drought on human health.
I will also present in this session for my recent work on investigating the relationships between the droughts in the hydrological process and cardiovascular and respiratory mortality in California. This is the abstract and the poster.
Previous studies provide evidence that droughts have impacts on human health-related aspects, such as infectious diseases and mortality. Under climate change, drought impacts on human health will likely increase because drought events are often projected to become more intense and severe. However, drought monitoring and impact assessment are primarily focused on other sectors, such as the agricultural and ecological sectors. Hence, drought indices are mainly developed based on hydro-meteorological processes and vegetation characteristics with limited attention to human health. Most studies on drought and health are inclined to choose commonly-used drought indices, such as the Standardized Precipitation-Evapotranspiration Index (SPEI) and the Evaporative Demond Drought Index (EDDI), as their default options. However, the relationships between various existing drought indices and health-related observations have rarely been explored. Therefore, in this proposed study, we developed an exploratory analysis of investigating the spatiotemporal relationships between drought indices and respiratory- and cardiovascular-related mortality in selected case studies in the United States, considering social vulnerability and crucial environmental factors, such as air quality, air temperature, and relative humidity. We also investigated the associations between mortality and various drought characteristics, including severity, duration, and onset. The outcomes of this study are expected to help instruct the application of different drought indices to reflect drought impacts related to human health and well-being and provide suggestions for future work on building compound indices to monitor drought impacts on health risks.
I am also the co-first auther of an exploratory study applying reinforcement learning to investigate human-flood interactions. Here is the abstract and the poster. The work will be presented at NH41B: Applications of AI/ML Using Remote Sensing, Social Sensing, and Model Data to Study Hazard from 8:30 to 12:50 on December 14.
Floods can severely affect human and natural environments, causing extensive damage to property, infrastructure, and agricultural land. The updated USGS water cycle graph highlights that human activities can significantly affect the hydrological environment. However, only limited studies have focused on investigating the interactions between anthropogenic impacts, extreme precipitation, and floods. Reinforcement learning (RL) can be a powerful tool in flood risk mitigation as it can learn an optimal reward function and decisions that adapt to dynamic natural environments and human activities. In this study, we aim to couple a process-based hydrological model and socio-economic information, such as the social vulnerability index (SVI), to understand the interactions between natural hydrological processes and human activities at the watershed level. The interaction between agent and environment will be studied using the RL algorithm: Deep Deterministic Policy Gradient (DDPG), which combines policy-based and value-based methods to search for the optimal policy that maximizes the expected cumulative long-term reward. This study can improve our understanding of how societies and individuals respond to floods (through implementing or improving flood preventive structures, taking adaptive measures, and improving land use scenarios) and provides valuable insights into how these actions can significantly change the impact of flooding.
Moreover, I participate in a study of drought risk on specialty crops in the United States lead by my colleague at the National Drought Mitigation Center (NDMC), which will be presented in H41L: Drought Risk Assessment: Monitoring, Modeling, and Prognosis from 8:30 to 12:50 on December 14. Here is the abstract:
Measuring drought impacts on agricultural crops can be difficult due to the complex interactions between the atmosphere, land surface and vegetation. While ample impact data exists for commodity crops in the form of news articles, images, and field data, less data regarding impacts has been recorded for specialty crops (e.g. pumpkins, sweet corn, potatoes, sugar beets, etc.). This study aims to provide a methodology and understanding of drought risk to specialty crops. Using a mixed-methods approach, we investigate the feasibility of using United States Department of Agriculture (USDA) Risk Management Agency (RMA) crop insurance data as a proxy for drought impacts on specialty crops in the U.S. Midwest. Preliminary results show that, by combining quantitative drought indicators (e.g. SPI, SPEI, EDDI, Soil Moisture) and qualitative data from focus groups and surveys, specific months and indicators are critical for predicting if there will be loss, measured in the form of an insurance claim. Quantifying the likelihood of a claim being filed for drought based on the change of drought indicators can help farmers be more prepared for future loss and provide timely information regarding possible impacts. This study provides a basis for more detailed studies and the possibility for ML integration in efforts to better understand drought risk to specialty crops. Furthermore, this study also demonstrates how a mixed methods approach to investigating climate-related impacts can provide context-specific information to produce more meaningful results.
If you are interested in any of them, please feel free to stop by and chat with me on these days. Also, please email me if you have any additional suggestions or comments. I look forward to seeing you in San Francisco!