A194-06
A novel high-resolution long-term gridded meteorology dataset including urban heat islands over the contiguous United States for health studies

Tuesday, 15 December 2020: 08:50
Virtual
Andrew James Newman, National Center for Atmospheric Research, Boulder, CO, United States, Andrew Monaghan, University of Colorado at Boulder, Boulder, United States, Heather Holmes, University of Utah, Department of Chemical Engineering, Salt Lake City, UT, United States, Howard H. Chang, Emory University, Department of Biostatistics and Bioinformatics, Atlanta, GA, United States, Lyndsey Darrow, University of Nevada Reno, Environmental Sciences, Reno, NV, United States, Joshua Warren, Yale University, School of Public Health, New Haven, CT, United States and Matthew Strickland, University of Nevada Reno, Department of Environmental Health, Reno, NV, United States
Recorded Presentation
Abstract:
It is well-documented that preterm infants, born before 37 completed weeks of gestation, are at greater risk of a wide range of acute and chronic health conditions, including developmental disabilities and mortality in the neonatal period and longer term. Recent epidemiologic studies suggest that higher temperatures increase the risk of preterm birth but results are inconsistent. To understand the currently differing conclusions, a large, multi-site framework is necessary to systematically assess the impacts of extreme heat events on pregnancy duration.

A fundamental component of a study covering a large area is the meteorological data used to determine heat events. Previous investigations have often used point observation station data located at airports that may not be representative of the population’s exposure in a given area. Over the contiguous United States (CONUS), several high resolution (1-4 km) gridded meteorological datasets exist and have recently been used in a number of studies related to health. However, these datasets may be inadequate for quantifying heat exposure because they do not fully represent the impacts of land surface heterogeneity, particularly the magnitude of the so-called ‘urban heat island’ which can magnify extreme heat events in cities and disproportionately impact urban populations.

We are using the High Resolution Land Data Assimilation System to create a 1 km dataset of near-surface (2 m) meteorology spanning 1980-2018 over the CONUS to account for heterogeneous land surface types (e.g., rural versus urban) in our estimates of near surface sensible weather. This dataset will be made publicly available and should become a valuable resource for epidemiological research due it its ability to resolve fine-scale temperature and humidity gradients. Initial results show the product is able to represent the strength of the urban heat island over large cities, with subsequent increases in summer temperatures over many population centers. We also show that explicit representation of the spatial variability of temperature modifies exposure estimates when compared to assuming an essentially spatially homogeneous temperature based off of one or only a few observation stations.

<< Previous Abstract | Next Abstract >>