A high-resolution daily temperature reconstruction for Zhejiang Province during 1961–2020 using statistical residual interpolation

High-resolution, long-term near-surface air temperature ( T a ) data are fundamental for regional climate monitoring and extreme temperature hazard assessment, particularly in topographically complex regions with sparse observations. However, achieving both temporal continuity and spatial detail in daily T a fields remains challenging under heterogeneous landscapes. This study develops a spatial-background residual framework to generate a 1 km resolution daily T a dataset (ZJ-DAT) for Zhejiang Province, China, spanning 1961–2020, covering daily minimum, mean and maximum temperatures. The framework integrates a high-resolution climatological baseline derived from hourly 1 km gridded temperatures for 2008–2018 with spatially interpolated daily anomalies from meteorological stations. Three interpolation schemes are evaluated for comparison: inverse distance weighting (IDW), lapse-rate–adjusted IDW (IDW + LR), and multiple linear regression (MLR); cross-validation results indicate that inverse distance weighting without lapse-rate correction provides the most robust overall performance. Further analysis reveals strong spatiotemporal non-stationarity in near-surface temperature lapse rates, which approach zero or become negative in winter lowlands but exceeding 7.0 °C km⁻¹ in summer mountainous regions, indicating that uniform lapse rate can introduce systematic elevation-related biases over complex terrain. ZJ-DAT shows high agreement with station observations and reanalysis products, demonstrating its reliability for climate applications. Based on this dataset, extreme temperature characteristics and associated hazards are further examined. Results show that ZJ-DAT provides spatially coherent and temporally consistent estimates of temperature extremes, enabling improved assessment of extreme temperature hazards across different regions and return periods. The proposed framework is transferable to other topographically complex, data-limited observations and provides a reliable basis for high-resolution climate monitoring, hazard assessment, and adaptation planning.