Abstract
A newly developed numerical model has been developed that enables case-specific calculations of the rate of cooling of a cadaver following death. The finite element method is used; it is an advanced computational method that accounts for body mass, thermal environment, presence or absence of clothing, etc. The method was trained against 20 experiments using data collected from 20 recently deceased persons. Comparisons between measured and calculated rectal temperatures show excellent agreement for a subset of the cases, while other cases had lesser agreement. On average, the mean signed error of rectal temperatures was 0.21 °C (0.38 °F) with a 95 percent CI ranging from −0.41 °C to 0.83 °C for the experimental training values, indicating no evidence of systematic bias in the temperature predictions. The mean absolute error was 1.02 °C (1.84 °F) with a 95 percent CI ranging from 0.63 °C to 1.41 °C. Next, the method was applied to a new set of blinded studies in which the time of death was unknown, using the numerical method. After the calculations were completed, the results were compared to the recorded time of death. It was found that the numerical method was able to accurately predict the times of death, with an average error of 47 min or 8.2 percent. These results indicate that the method can reasonably estimate the time of death based on cadaver core temperatures, even when the cadaver’s ambient conditions change over time.