Empirical modeling and mapping of urban air temperatures and human comfort in Baltimore, MD

Impacts of climate change on human health and well-being depend in part upon the added effect of local- and micro-scale climatic influences in the human environment.  The small-scale climatic deviations from regional climates can be especially important in urban ecosystems.  As a contribution to the Baltimore Ecosystem Study (BES), a Long Term Ecological Research site in the National Science Foundation LTER program, we began continuous measurements of air temperature and humidity at the 1.5-m height and wind speed at 2 m at five sites near the city of Baltimore, MD in June 2003.  These sites are in a large open pasture, a woodlot near the pasture, a residential area with heavy tree cover but few buildings, a lawn area with nearby trees between two large apartment complexes, and a residential area with some trees and large lawn areas.  An object of the measurements is to empirically model and map below-canopy air temperature and wind speed expressed as differences in temperature, DT, and wind speed, DU, from a reference site.  We anticipate that such a model will lead to the ability to map an index of urban structural and vegetation influences on outdoor human thermal comfort.  Other potential applications include use of modeled temperature for studies related to carbon cycling, soil and stream temperatures, ozone formation, effects of UV radiation on human health, and energy use for building space conditioning.  In addition to our measurements, data are available from two National Weather Service Automated Surface Observing System (ASOS) sites—in downtown Baltimore and at the Baltimore/Washington International Airport (BWI).  We used the downtown ASOS site as the reference point for DT on an hourly basis between each site and the reference.  The values of DT were related by regression analysis to predictor variables derived from differences in upwind tree, impervious, and water cover from the National Land Cover Database (NLCD) for 2001.  Additional predictor variables for DT were atmospheric stability (Turner Class, derived from BWI cloud and wind speed data), vapor pressure deficit, antecedent precipitation, and descriptors of topography.  Interactions between stability and the other predictor variables were included to account for the predictor influences varying with stability.  In an initial analysis for one summer, atmospheric stability was even more significant than land cover in influencing DT, and topographic relationships were also important.  We used wedge-shaped focal analysis functions in ESRI ArcGIS for cover analysis to prepare GIS maps of predicted DT for every cell in 30-m resolution vegetation and DEM images of the Baltimore region.  Despite the numerous sources of variability in the regression modeling, the method produced reasonable map patterns of DT.  Continuing research will model DU by similar methods.  Predictions of human comfort will be based on the OUTCOMES model (http://www.fs.fed.us/ne/syracuse/Tools/tools.htm) with input from the modeled DT and DU, and average sky view by land use categories as estimated from hemispherical photographs at 200 sample field points in the city of Baltimore.