Work

Meteorological and Biometeorological Evaluation of a Population-Aware Heat Warning Framework for Mauritius

Year: 2026
Type: preprint
Abstract: Human biometeorological heat assessment requires metrics that distinguish local climatic rarity, multi-day persistence, nocturnal heat, combined thermal stress and spatial population exposure. We developed and evaluated a population-aware heat-warning framework for Mauritius using reference station observations, ERA5-Land reanalysis, calendar-day temperature percentiles, Excess Heat Factor (EHF), the Universal Thermal Climate Index (UTCI), nocturnal heat diagnostics and fixed-standard 2022 census weights. ERA5-Land reproduced day-to-day maximum-temperature variability but showed station-dependent cool bias ($-0.28$ to $-2.53\degC$) and weak same-day 99th-percentile event detection. At Plaisance, derived wet-bulb temperature closely reproduced observed November--April measurements (bias $-0.076\degC$; root-mean-square error $0.191\degC$; $r=0.992$). A hybrid UTCI series combining observed air temperature, humidity and wind with ERA5-Land mean radiant temperature correlated strongly with ERA5-only UTCI ($r=0.962$) but was $2.18\degC$ warmer. Historical heat episodes expressed contrasting biometeorological signatures: December 2013 combined thermal stress, EHF, hot-day and warm-night burden; February 2025 produced island-wide UTCI $\ge38\degC$ and 99th-percentile hot-day/warm-night burden without EHF severity $\ge3$; and March 2019 exposed a marked reanalysis-observation disagreement. The 1991--2025 archive contains 12,784 daily records with six-zone, island-maximum and population-fraction diagnostics. The approach provides a transferable biometeorological validation framework for data-constrained tropical islands; epidemiological calibration remains necessary before health-based alert thresholds are defined.
Language: English
Open Access status: green
OpenAlex ID: W7225998890
Cites: 11
Cited by:
Related to: 10