Climate Classification of Nigeria Using K-MeansClustering and MODIS Data

Understanding spatiotemporal climate patterns is essential for agricultural planning, water resource management, and biodiversity conservation, particularly in climatically diverse regions like Nigeria. Traditional climate classification systems, such as Köppen-Geiger, rely on long-‎term monthly averages and stationary weather station data, which often suffer from spatial discontinuities and latency. This study proposes ‎an alternative, data-driven approach to climate zoning by applying unsupervised machine learning (K-means clustering) to high-resolution ‎remote sensing data. Utilizing the Google Earth Engine (GEE) cloud computing platform, we processed Terra and Aqua MODIS time-‎series data for the year 2020, specifically the Land Surface Temperature (LST) and Normalized Difference Vegetation Index (NDVI) products. By extracting seasonal phenological signatures and thermal profiles across the Nigerian landscape, the study generated a spatially continuous climate classification map. The analysis revealed distinct clusters corresponding to Nigeria's major ecological zones: Man-‎grove/Swamp Forest (South), representing 25.0% of the total, Rainforest (South-West/South-East) with a percentage of 13.4%, Guinea ‎Savannah accounting for 28.8%, Sudan Savannah at 15.9%, and Sahel Savannah at 17.0%. The results demonstrate that K-means clustering ‎of MODIS data effectively captures the hydro-thermal gradients driving Nigeria’s climate variability, offering a robust, scalable methodology for dynamic climate monitoring at a high spatial resolution (1km)‎.