Estimation of long term post mortem intervals by ATR-FTIR spectroscopy of weathered blow fly puparia

Fourier transform infrared (FTIR) spectroscopy is often used in forensic science due to its ability to obtain a (bio-)chemical fingerprint of the sample. This method could also be used in forensic entomology to obtain additional information about empty blow fly puparia, which can still be found at crime scenes even several months or years after the event. Accurately estimating the age of these samples would significantly improve the determination of the minimum post-mortem interval. This is because the current method only provides day-specific information for the first few weeks after death - i.e. for as long as the first wave of colonization has not yet fully developed. Blow fly puparia of four different forensically relevant species were stored for up to 9 months outdoors (in leaf litter or in soil), or under room conditions. Puparia were collected monthly, cleaned and infrared spectra were recorded. To assess the weathering time of the puparia, partial least squares regression and support vector regression (SVR) were applied. SVR-based models were favoured as they resulted in higher R² values and lower prediction errors (1-1.5 months) when presented with previously unseen data. Training a model with a combination of all weathering conditions and species resulted in a prediction error of 1.2 months. This shows that it is possible to assess weathering time without using a model specific to the weathering condition or blowfly species. Being able to infer the age of an insect sample without a previous species identification is a novelty in forensic entomology.