Several general algorithms have been developed to predict mortality changes during heat waves. Buechley et al. (1972) developed the following algorithm for heat-related mortality at temperatures above 90deg.F:

TMR = cycle + 0.10e[0.2(F[1] - 90)] (1)

where TMR is the temperature-specific mortality ratio (the predicted mortality for the day divided by the average annual daily mortality), cycle is the expected mortality ratio for that day of the year (an attempt to account for the impact of seasonality on mortality), and F[1] is yesterday's temperature. Cycle is computed from several years of mortality data and varies in a sinusoidal fashion, peaking in the winter and reaching a minimum at the end of the summer. Each day has a distinctive cycle value de pending upon the mean mortality rate for that time of year. The following example represents a hypothetical calculation of TMR. Assume that the maximum temperature on a given day is 100deg.F, and the cycle is 0.95. TMR = 0.95 + 0.1e[0.2(100 - 90)], which equals 1.70. Thus the equation predicts that mortality on the day following the 100deg. maximum temperature will equal 170% of the annual mean daily mortality. Oechsli and Buechley (1970) had previously developed a related algorithm, the age- and temperat ure-specific mortality ratio model (ATMR):

ATMR = 98.806 + e[(-15.23 + .0385 Age + .1655 F)] (2) where F is the present day's maximum temperature.

In a more recent study, Marmor (1975) attempted to develop a model that accounted for acclimatization effects. This led to his sensitivity index, which decreased as the population was exposed to more hot days during the season. Sensitivity (S[d]) equals:< p> 1 / (1 + e[(Ad - 6) / 0.46)] (3)  where Ad is the total number of previous days with temperatures over 90deg.F.

This sensitivity value was added to a newer version of the TMR algorithm, producing the following: