Modeling acute respiratory illness during the 2007 San Diego wildland fires using a coupled emissions-transport system and generalized additive modeling

Environmental Health

Research

Tatiana Loboda¹  Shiliang Wu²  Michael Billmire³  Nancy HF French³  Brian Thelen³  Benjamin W Koziol⁴  Robert Chris Owen⁵  Jeffrey Johnson⁶  Michele Ginsberg⁶ 

[1] Department of Geographical Sciences, University of Maryland, 2181 LeFrak Hall, 20742, College Park, MD, USA;Department of Geological and Mining Engineering and Sciences and Department of Civil and Environmental Engineering, Michigan Technological University, 1400 Townsend Drive, 49931, Houghton, Michigan, USA;Michigan Tech Research Institute, Michigan Technological University, 3600 Green Court, Suite 100, 48105, Ann Arbor, MI, USA;Michigan Tech Research Institute, Michigan Technological University, 3600 Green Court, Suite 100, 48105, Ann Arbor, MI, USA;Currently at NESII/CIRES/NOAA Earth System Research Laboratory, 325 Broadway, 80305, Boulder, CO, USA;Michigan Tech Research Institute, Michigan Technological University, 3600 Green Court, Suite 100, 48105, Ann Arbor, MI, USA;Currently at U. S. Environmental Protection Agency, Office of Air Quality Planning and Standards, Research Triangle Park, 27711, Durham, NC, USA;San Diego County Health & Human Services Agency, 3851 Rosecrans Street, 92110, San Diego, CA, USA;

关键词: Wildland fire;    Particulate matter emissions;    Syndromic surveillance;    Generalized additive modeling;    Air quality;    Respiratory health;    San Diego County;    California;   

DOI  :  10.1186/1476-069X-12-94

 received in 2013-02-25, accepted in 2013-10-28,  发布年份 2013

来源: Springer

【 摘 要 】

BackgroundA study of the impacts on respiratory health of the 2007 wildland fires in and around San Diego County, California is presented. This study helps to address the impact of fire emissions on human health by modeling the exposure potential of proximate populations to atmospheric particulate matter (PM) from vegetation fires. Currently, there is no standard methodology to model and forecast the potential respiratory health effects of PM plumes from wildland fires, and in part this is due to a lack of methodology for rigorously relating the two. The contribution in this research specifically targets that absence by modeling explicitly the emission, transmission, and distribution of PM following a wildland fire in both space and time.MethodsCoupled empirical and deterministic models describing particulate matter (PM) emissions and atmospheric dispersion were linked to spatially explicit syndromic surveillance health data records collected through the San Diego Aberration Detection and Incident Characterization (SDADIC) system using a Generalized Additive Modeling (GAM) statistical approach. Two levels of geographic aggregation were modeled, a county-wide regional level and division of the county into six sub regions. Selected health syndromes within SDADIC from 16 emergency departments within San Diego County relevant for respiratory health were identified for inclusion in the model.ResultsThe model captured the variability in emergency department visits due to several factors by including nine ancillary variables in addition to wildfire PM concentration. The model coefficients and nonlinear function plots indicate that at peak fire PM concentrations the odds of a person seeking emergency care is increased by approximately 50% compared to non-fire conditions (40% for the regional case, 70% for a geographically specific case). The sub-regional analyses show that demographic variables also influence respiratory health outcomes from smoke.ConclusionsThe model developed in this study allows a quantitative assessment and prediction of respiratory health outcomes as it relates to the location and timing of wildland fire emissions relevant for application to future wildfire scenarios. An important aspect of the resulting model is its generality thus allowing its ready use for geospatial assessments of respiratory health impacts under possible future wildfire conditions in the San Diego region. The coupled statistical and process-based modeling demonstrates an end-to-end methodology for generating reasonable estimates of wildland fire PM concentrations and health effects at resolutions compatible with syndromic surveillance data.

【 授权许可】

CC BY   
© Thelen et al.; licensee BioMed Central Ltd. 2013