Work

Spatial interpolation alters heat-wave detection despite small temperature errors in data-sparse regions

Year: 2026
Type: article
Abstract: Climate assessments frequently rely on temperature fields generated by spatial interpolation, yet the influence of interpolation methods on derived climate indicators remains insufficiently understood. This study investigates how commonly used distance-based interpolation methods affect temperature representation and heat-wave detection in the state of Säo Paulo, Brazil. Hourly air-temperature observations from meteorological stations (2021–2025) were evaluated using a leave-one-out cross-validation framework. The Nearest Neighbor (NN) method and multiple configurations of Inverse Distance Weighting (IDW) were compared, and interpolated temperature series were generated for the centroids of 645 municipalities to derive daily temperature statistics and heat-wave indicators. The best-performing configuration (IDW with a 250 km search radius and distance-decay exponent of 2) achieved an RMSE of 1.97 $$^\circ $$ C and an $$R^2$$ of 0.877. Although IDW adequately preserved the central tendency of maximum and mean temperatures, it systematically reduced temperature variability, particularly for minimum temperatures, attenuating cold extremes and modifying temperature distributions. These differences became substantially more pronounced for threshold-based climate indicators. Compared with NN, IDW consistently identified fewer and shorter heat-wave events, with cumulative heat-wave duration differing by up to 199 days in some municipalities over the study period. The results demonstrate that the simple distance-based interpolation methods evaluated here can show satisfactory predictive accuracy while still altering the representation of climate extremes and derived heat-wave indicators. Consequently, evaluating interpolation methods solely using conventional error metrics may be insufficient for applications involving climate-risk assessment and urban climate analysis.
Language: English
Open Access status: hybrid
Estimated APC paid: $3,090
OpenAlex ID: W7203760431
FWCI:
Cites: 42
Cited by:
Related to: 10