This article aims to increase accuracy in estimating the postmortem submersion interval (PMSI) for bodies recovered from rivers in the United Kingdom. Data were collected from closed case files, crime scene reports, and autopsy files concerning bodies recovered over a 15-year period from the River Clyde, Scotland, and the River Mersey and canals in northwest England. One hundred and eighty-seven cases met the study criteria and were scored by quantifying the overall amount of decomposition observed in each case. Statistical analysis showed that the duration of a body's submergence in water and the temperatures to which it was exposed, as measured in accumulated degree days (ADD), had a significant effect on the decay process. Further analysis indicated that there were no significant differences in decomposition between the waterways. By combining the data from all study samples, it was possible to produce a single linear regression model for predicting ADD from observed decomposition.

Comparison of data from a variety of environments and ambient temperatures has previously been difficult as few studies used standardized measures of time/temperature and decomposition. In this paper, data from previous studies and recent experiments are compared using simple conversions. These conversions allow comparison across multiple environments and experiments for the first time. Plotting decomposition score against logADD allows the exponential progression of decomposition to be expressed as a simple linear equation. Data comparison from many environments and temperatures shows no difference in decomposition progression when measured using Accumulated Degree Days. The major effector of change in rate was insect presence, regardless of depositional environment, species, or season. Body size is significant when carcasses are accessed by insects; when insects are excluded, while bodies are indoors, submerged, or buried, then decomposition progresses at the same rate regardless of body size.

Megyesi et al.'s (J Forensic Sci, 2005, 50, 618) paper was important to forensic anthropology as it introduced a quantitative framework for estimating time since death in human cadavers, based upon physical appearance by way of scoring on a novel scale. However, errors concerning rounding, temperature scale, and incorrect use of a statistical regression model render their predictive formula unusable. Based upon only their more reliable data, a more appropriate regression model to predict accumulated degree days (ADD) from total body score (TBS) is presented. The new model is also a superior fit (r(2)  = 0.91) and produces markedly narrower confidence intervals than the original, which also allowed impossible, negative ADD values. Explanations of the shortcomings in the original analysis and calculations are presented, which it is hoped will help forensic scientists avoid making similar mistakes.