Estimating the Post-Mortem Interval Under Extreme Heat Environments: A Climate-Adaptive Case Series Based on...

Area of Science:

• Forensic Science
• Artificial Intelligence
• Environmental Science

Background:

• Accurate post-mortem interval (PMI) estimation is challenging in extreme heat due to accelerated decomposition and altered indicators.
• Hyperthermal environments can lead to mummification and significant overestimation of PMI using classical methods.

Purpose of the Study:

• To analyze forensic cases impacted by climate-driven decomposition anomalies.
• To present and evaluate a climate-adaptive, AI-assisted diagnostic framework for improved PMI interpretation in extreme heat.

Main Methods:

• Retrospective case series analysis of three individuals recovered during heatwaves.
• Integration of crime scene data, PMCT, autopsy, genetics, and meteorological data.
• Application of a multimodal AI model (Random Forest-LSTM) with engineered climate-stress indices.

Main Results:

• Classical morphological assessments overestimated PMI (1-6 months vs. actual ~20 days or ~42 hours).
• The AI model provided estimates consistent with verified PMIs, including prediction intervals.
• Explainability analysis highlighted thermal load and desiccation as key drivers of decomposition anomalies.

Conclusions:

• Extreme heat significantly alters decomposition, rendering traditional PMI methods unreliable.
• A climate-aware AI framework enhances interpretability and provides uncertainty-aware PMI estimates.
• The AI framework shows promise for next-generation PMI diagnostics in hyperthermal forensic settings.