Determining the frequency of open windows in residences: a pilot study in Durham, North Carolina during varying temperature conditions

Abstract

Listen

Translate article

Air pollution exposures in the residential microenvironment can be significantly affected by air exchange rate (AER). A number of studies have shown that AER in residences is significantly affected by the number and location of open windows and doors. A pilot study was conducted in Durham, North Carolina, to determine whether useful data on open windows and doors could be acquired through a visual survey. The study consisted of 72 2-h survey sessions conducted between October 24, 2001 and March 13, 2003. During the first hour of each session, a technician selected a set of corner residences in one of 48 census tracts and completed a survey form and meteorological measurements for each residence. During the second hour, the technician revisited each residence surveyed during the first hour. The resulting database included data on 2200 "residential visits" (1100 residences times two visits per residence). The technician observed one or more open windows during 20.0 percent of the residential visits. One or more open doors were observed during 13.4 percent of the residential visits; 28.2 percent of the residential visits were associated with at least one open window or door. A series of stepwise linear regression analyses were performed on the data to identify factors associated with open windows and doors. Results of these analyses indicated that the likelihood of one or more windows being opened tended to increase under the following conditions: occupancy at time of visit; session during April, May, or June; high population or housing density; window air conditioning (AC) units; absence of AC; large number of doors; and wind speed above 2 mph. The likelihood of open doors tended to increase under the following conditions: occupancy at time of visit; residence within city limits; session during April, May, or June; detached one-story residence; large number of doors; high housing density; school out; and residence within 10 m of road. Transition probabilities (closed to open and open to closed) were determined for windows and doors by time of day.