Greenhouse-gas emission trends and decarbonization outlook for the Rabat-Salé-Kénitra region (Morocco): an exploratory inventory and STIRPAT-machine learning assessment

Abstract

Morocco faces increasing climate pressures, while regional greenhouse gas (GHG) emissions remain poorly documented. This study reconstructs the GHG inventory of the Rabat-Salé-Kénitra (RSK) region, Morocco’s second-largest economic hub, for 1989–2021 using the IPCC 2006 Tier-1 method. It then explores socio-economic drivers and decarbonisation pathways to 2050 within a framework adapted to a short annual series (n = 33). Total emissions increased from about 5.5 to 10.4 Mt CO₂e (+88%, ~2% per year), with the energy sector rising from 61% to 75% of total emissions. To avoid circularity, only four non-circular structural drivers (population, GDP per capita, urbanised area and vehicle fleet) were retained in an extended STIRPAT model estimated by Ridge regression, while tree ensembles and a naïve baseline were used only for comparison. Under leave-one-out cross-validation, STIRPAT-Ridge achieved R² = 0.965 and MAPE = 3.4%, without overfitting. GDP per capita was the only statistically robust driver. Compared with a BAU pathway reaching ~19.9 Mt CO₂e in 2050, moderate and ambitious scenarios reduce emissions by 20% and 40%, respectively. Beyond RSK, the study provides a transferable protocol offering circularity control, regularised STIRPAT and small-sample validation for regions with limited data in emerging economies.