Statistical and machine learning-based development of future typical meteorological years using city-specific GCMs for major cities in Indonesia

Abstract

The present study establishes a novel city-specific future typical meteorological year dataset by incorporating future climate projections across 26 Indonesian cities. The projected monthly mean changes in four climatic variables (surface air temperature, precipitation, wind speed, and relative humidity) were obtained from city-specific global climate models for four future decades (2021–2030, 2031–2040, 2041–2050, 2051–2060) under the Shared Socioeconomic Pathway 5–8.5 scenario. These projections were morphed into the 2011–2020 TMY to create four future TMY (FTMY) datasets. The results indicate that, compared with 2011–2020, the annual mean surface air temperature across all cities may increase by 0.15°C to 2.90°C from 2021 to 2060. By 2051–2060, 19 cities may experience a precipitation increase of more than 20%, with minimal changes in relative humidity and wind speed. FTMY data generated from the morphing method were fed into a CatBoost algorithm to predict future solar irradiance. Projections indicate that by 2051–2060, shifts in solar irradiance components across most cities will be imperceptible. Analysis of future thermal discomfort indicated a significant increase in heat stress exposure across all cities over time. The FTMY datasets developed through this research provide crucial inputs for climate-responsive applications, including energy-efficient building design, energy planning, and supporting urban sustainability under future climate scenarios.