Machine learning–GIS integration for spatiotemporal analysis of regional biogas potential: a case study of Eastern Anatolia, Türkiye

Biogas derived from organic waste is a dispatchable renewable energy source that can be stored and converted into electricity when needed, supporting both energy security and circular economy practices. Accurate estimation of regional biogas potential under current and future conditions is therefore essential for energy planning. Regions with intensive livestock production are particularly suitable for manure-based biogas systems. In this context, Eastern Anatolia, one of Türkiye’s major livestock regions, provides a representative case for assessing cattle manure-based biogas potential (BMBP) at the subnational scale. This study presents a provincial-level assessment of BMBP across all 14 provinces of Eastern Anatolia and analyses its spatiotemporal dynamics by integrating machine learning and geographic information systems (GIS). Artificial Neural Networks (ANN) and an Autoregressive Integrated Moving Average (ARIMA) model were used to forecast BMBP up to 2035 based on data from 2004 to 2021. Regional BMBP was estimated at 2868 GWh in 2021, with 2035 projections ranging from 3347 GWh (ANN) to 3709 GWh (ARIMA). Erzurum showed the highest provincial BMBP, indicating substantial potential to offset local electricity demand. Although these estimates represent theoretical planning-level potential, the proposed machine learning–GIS framework is transferable to other livestock-intensive regions.