Integrating weather observations and local-climate-zone-based landscape patterns for regional hourly air temperature...

Area of Science:

• Urban Meteorology
• Climatology
• Geospatial Analysis

Background:

• Accurate hourly air temperature mapping is crucial for urban meteorology, urban climate, and climate change studies.
• Conventional methods face challenges in mapping fine-resolution surface air temperature due to limited observational data and complex urban environments.
• There is a need for advanced techniques to improve the spatial and temporal resolution of air temperature data in urban areas.

Purpose of the Study:

• To develop and validate machine learning (ML) techniques for high-resolution hourly air temperature mapping in Guangdong Province, China.
• To assess the accuracy and driving factors of ML-based air temperature mapping over a multi-year warm season period.
• To investigate the performance of the mapping technique during nighttime and its correlation with urban morphology.

Main Methods:

• Employed machine learning algorithms, specifically the Random Forest algorithm, for hourly air temperature mapping.
• Utilized meteorological and landscape data as input variables for the ML models.
• Conducted validation using R-squared, Root Mean Square Error (RMSE), and Mean Absolute Error (MAE) for the period 2008-2019.

Main Results:

• The ML-based hourly air temperature maps demonstrated high accuracy (mean R²=0.8001, RMSE=1.4821 °C, MAE=1.0872 °C) from 2008 to 2019.
• Meteorological factors, particularly relative humidity, were identified as the most significant drivers of air temperature, with landscape factors also playing a role.
• The maps maintained high accuracy during nighttime, revealing slower temperature decreases in metropolitan cores compared to urban fringes, correlating with local climate zones.

Conclusions:

• Machine learning reliably enhances the spatial refinement of hourly air temperature mapping in urban and surrounding areas.
• The study provides a novel, valuable, and reliable dataset for air temperature-related studies and implementations.
• The findings are essential for investigating nighttime urban climate conditions, including the urban heat island effect and its relationship with urban morphology.