Integrating Observations and MM5/CMAQ Predictions to Study the Link between Climate Variability, Air Quality, and Health in New York State – Project Description and Initial Results

This paper presents initial results from a project aimed at developing and applying new methods for analyzing the potential link between climate variability, air quality, and health in New York State. The specific aims of the project are to: (1) develop fine-scale gridded maps of hourly surface weather, ozone, and particulate matter (PM) over NY State over the 15-year period 1988 to 2002 using station observations integrated with simulations from the MM5/CMAQ modeling system, (2) analyze the relationship between climate variability and episodes of extreme PM, ozone, and heat, (3) measure the independent and joint effects of air quality and weather on acute mortality and hospitalization risks at the county level across NY State from 1988-2002. Here, we focus on the initial stages of the first objective, i.e. the development of concentration maps using a rich data base of observations augmented by regional-scale weather and air quality models. In this stage, spatially continuous meteorological and air quality fields will be derived from observed values using geostatistical interpolation techniques such as kriging. We will discuss possible shortcomings of this method such as oversmoothing and the lack of spatially and temporally homogeneous correlation structures utilized during the kriging process. We will then outline and illustrate the use of model output from MM5/CMAQ simulations at 12 km to explore possible ways to generate enhanced pollutant maps by integrating model-predicted spatial gradients into interpolated observations while accounting for possible model bias. Finally, we describe how the daily gridded estimates of weather, ozone, and PM will be matched with daily mortality and hospital admissions data at the county level in future stages of the research project. The health outcomes of interest will be county level data for all of New York State on daily mortality (obtained from the National Center for Health Statistics (NCHS)) and daily hospital admissions (obtained from the New York Statewide Planning and Research Cooperative System (SPARCS) database). These data will be processed to compute daily, county-specific counts for total internal (i.e., excluding accidents, homicides, and suicides), total respiratory, and total cardiovascular causes.