Abstract. We proposed earthquake forecasting models for Albania, one of the most seismogenic regions in Europe, to give an overview of seismic activity by implementing area source and smoothing approaches. The earthquake catalog was firstly declustered to remove foreshocks and aftershocks when they are within the derived distance- and time-windows of mainshocks. Considering catalog completeness, the events with M≥4.0 during the period of 1960–2006 were implemented for the forecast model learning. The forecasting is implemented into an area source model that includes 20 sub-regions and a smoothing model with a cell size of 0.2° x 0.2° to forecast the seismicity in Albania. Both models show high seismic rate along the western coastline and at the southern part of the study area, consistent with previous studies which discussed seismicity in the area and currently active regions. To further validate the forecast performance from the two models, we introduced the Molchan diagram to quantify the correlation between models and observations. The Molchan diagram suggests that both models are significantly better than a random distribution, confirming their forecasting abilities. Our results provide crucial information for subsequent research on the seismic activity, such as probabilistic seismic hazard assessment.

<p>Abstract<br>We proposed an earthquake forecasting model for Albania, one of the most seismic<br>regions in Europe, to give an overview of seismic activity by implementing area<br>source and smoothing approaches. The earthquake catalogue was firstly declustered<br>to evaluate the completeness time window and magnitude of completeness for shallow<br>events. Considering catalogue completeness, the events with M≥4.0 during the period<br>of time 1960 – 2006 were implemented for forecasting seismicity in 20 area sources<br>covering the region of study and each grid cell with a size of 0.2 x 0.2 degrees. Our<br>results from both models show a high seismic rate along the western coastline and<br>south part of the study area, consistent with previous studies and currently active<br>regions. To further evaluate the seismicity results from the models, we introduced a<br>Molchan diagram to investigate the correlation between a model and observations of<br>earthquake events. The catalogue from 1960 to 2006 is regarded as the “learning<br>period” for model construction, and the catalogue data covering the period of time<br>2018-2020 is the “testing period” for comparing and validating the results. The<br>Molchan diagram suggests that both models are significantly better than random<br>distributed, confirming their forecasting abilities. Our results could provide crucial<br>information for subsequent probabilistic seismic hazard assessment.</p><p><br>Keywords: area sources, declustering, earthquake catalogue, Molchan diagram,<br>probabilistic seismic hazard assessment, smoothing model,.</p>