Global warming is increasing extreme heat events, directly impacting labour health and economic productivity. High-resolution global reanalysis datasets, such as ERA5-Land, are enabling worldwide and regional assessments of heat-related labour impacts. This study systematically examines the reliability and bias of ERA5-Land-derived labour loss estimates across diverse Köppen-Geiger climate zones (A, C, D) and land cover types (urban and natural land cover) at 37 sites (including 36 WMO stations) in advanced-economy regions of Pacific Asia (Japan, Korea, Taiwan, and Singapore) from 2014 to 2023. Using the physics-based Liljegren wet-bulb globe temperature (WBGT) model and established exposure-response functions, we quantified annual cumulative labour loss and productivity loss while accounting for spatial labour distribution. A comparison of WBGT‑derived labour‑loss estimates from 37 ground‑station datasets with ERA5‑Land-driven estimates indicates strong concordance in temperate zones (R2 ≈ 0.85) and urban areas (R2 ≈ 0.94), contrasted by a systematic underestimation in tropical regions (up to ~ 18%, R2 ≈ 0.27). Incorporating a weighting approach that integrates population density and labour force participation rate (LFPR) significantly improves the assessment of the effective labour force scale, with ERA5-Land’s R2 in tropical regions increasing from 0.27 to 0.70. Based on our findings, we recommend (1) application of ERA5-Land in temperate and urban local-climate-zone (LCZ)-dominant dominant areas with standard caution, (2) for macro-level, climate- or land-cover-aggregated assessments, a 20% safety margin for tropical climate class (Köppen-Geiger Type A) and a 15% adjustment for natural environments to address underestimation uncertainty; and (3) for fine-scale, coupled climate × land-cover applications, context-specific correction (e.g., 31.55% for tropical natural areas) where localised land-cover data is available. These guidelines refine the operational use of reanalysis data for precise heat-risk assessments, and inclusion of population-density weighting in standard assessments, which improves index performance by 15–50%. While these guidelines refine the use of reanalysis data and offset policy under-protection risks, future applications must still account for microclimatic heterogeneity and the inherent assumptions of uniform sector-specific labour distributions. Ultimately, integrating these safety margins and weighting protocols provides a practical framework for informed public health interventions.
Lai et al. (Sun,) studied this question.