Abstract
The spatial variability of the magnitude–frequency distribution is important to improve earthquake forecasting capabilities at different time scales. Here, we develop a novel approach, based on the weighted maximum‐likelihood estimation, to build a spatial model for the ‐value parameter of the Gutenberg–Richter law and its uncertainty, also for earthquake catalogs with a time‐varying completeness magnitude. Then, we also provide a guideline based on the Bayes factor to measure the importance of the ‐value spatial variability with respect to a model having a spatially uniform ‐value. Finally, we apply the procedure to a new Italian instrumental earthquake catalog from 1960 to 2019 to investigate the ‐value spatial variability over the Italian territory.