Climate projection datasets pose major challenges for traditional GIS workflows due to their size and dimensionality. This study uses the NASA NEX-GDDP-CMIP6 dataset, comprising 65 years of historical simulations and four 85-year future scenarios across multiple models, totaling approximately 7.4 TB of data. To enable scalable analysis, we employ Dask and Xarray to construct spatiotemporal data cubes with lazy loading, chunking, and pixelwise computation. This workflow enables mesoscale GIS-ready analysis of changes in the thermodynamic drivers of humid-heat across the Arabian Peninsula.
