Skip to main content
. Author manuscript; available in PMC: 2021 Dec 1.
Published in final edited form as: Environ Res. 2020 Aug 20;191:110065. doi: 10.1016/j.envres.2020.110065

Table 2.

Variables used in the various generalized additive models for Charlottesville (CHO) and Roanoke (ROA). The variable that served as the cross-basis (cb) in the dlnm is indicated for each model; other values for the temperature variables indicate the number of degrees of freedom used in the spline term (8×10 in the “trend” term is 10 degrees of freedom for each of 8 years). T=1 p.m. temperature; AT=1 p.m. apparent temperature; Td=1 p.m. dew point temperature, DTR=diurnal temperature range; HW= heat wave; CW=cold wave, both using the “medium” thresholds (see Table 1); “Holi”=holidays, “DOW”=day of week; and “Snow”=snow events with precipitation over up to two days of at least 5 inches. The last column shows the adjusted R2 of the gam.

T AT Td DTR HW CW Holi DOW Snow Trend Adj. R2
CHO
Total cb x x x 8×10 0.275
White cb 10 x x x 8×10 0.212
Non-white cb 10 x x x 8×10 0.274
Male cb 10 x x x 8×10 0.199
Female cb 10 x x 8×10 0.176
Elderly cb 10 x x x 8×10 0.159
 
ROA
Total cb 10 10 x x x x x 8×10 0.473
White cb 10 x x x 8×10 0.376
Non-white cb 10 x x x 8×10 0.273
Male x x x x 8×10 0.363
Female x x x x 8×10 0.292
Elderly cb 10 x x x x x 8×10 0.405