Introduction: Buildings that collapse or are damaged by earthquakes are responsible for the majority of earthquake-related casualties. High-precision building data are the key to improving the accuracy of risk assessments of earthquake disaster loss. Many countries and regions have also proposed varying regional building exposure models, but most of these models are still based on administrative-level (city or county) statistical data; furthermore, they cannot accurately reflect the differences among buildings in different towns or villages.Methods: Although field investigation-based “township to township” methods can obtain more accurate building inventory data, considering costs and timeliness, remote sensing and other diverse data should be combined to acquire building data. Based on the field survey data of three cities in shanxi Province, combined with Global Human Settlement Layer (GHSL) data, this study is conducted on building inventory data. Data regarding the proportion of each building type and corresponding lethality level in each township are obtained based on the classification of building height, and the overall lethality level at the building level and township level is calculated on this basis.Results: The fitting results between the calculated results and the field survey results are good, the error is within 0.15, and the fitting R² values of Xian, Baoji and Ankang are 0.6552, 0.5788 and 0.5937, respectively. Therefore, an earthquake disaster loss risk assessment is conducted based on the building level.Discussion: The findings indicate that the risk of casualties caused by the same building type can vary by city. Generally, the areas with high disaster loss risk in the three cities are distributed mainly in urban areas; the disaster loss risk in the newly built areas of each city is relatively low. According to the quantitative assessment results for each city, Xi’an has the highest loss risk, while Baoji and Ankang have the same loss risk. Based on the method constructed in this paper, we can realize the quantitative assessment of earthquake disaster loss risk at the building level to better target pre-earthquake emergency preparation and post-earthquake auxiliary decision-making.

Population exposure is an important element of disaster loss assessment. High–resolution spatiotemporal distribution assessment of population exposure could improve disaster preparation and reduce the risk. This study proposed a model for assessing the spatiotemporal distribution of indoor people at the scale of single buildings by distinguishing the differences in people densities for various functional buildings. The empirical study results in the Dongcheng Subdistrict, Xichang City, China, showed that this method could determine the indoor population change in different single buildings at different times of day and map fine–scale spatiotemporal distribution of the regional indoor population. Due to the difference in the use function of buildings and human activities, the distribution of indoor populations in different functional buildings changes significantly during the day. Compared with the spatiotemporal changes in the indoor population on non–weekdays, the changes were more obvious on weekdays. The change in indoor population exposure during the daytime is significantly stronger than that at night. The results could provide an important reference for regional disaster preparedness and post–disaster emergency rescue.

Based on the collected mobile phone location data, this paper analyzes changes in four mobile location-based indicators and their spatiotemporal distribution characteristics before and after the earthquake, summarizing crowd movement patterns and communication behaviors after the M_S6.8 Dingri earthquake. By comparing natural neighbor interpolation and Thiessen polygon interpolation methods, we explore novel rapid assessment approaches for earthquake perception ranges, combined with actual seismic intensity maps. The results indicate an uneven distribution of population and differing dynamics in mobile phone signal activity. This reflects different behavioral patterns and the potential perceived extent of the earthquake. Within 50 km of the epicenter, all four indicators showed varying degrees of decline post-earthquake, while areas beyond 100 km exhibited short-term surges, reflecting differentiated behavioral responses based on seismic impact severity. In areas experiencing strong shaking, risk avoidance behavior predominated, while in areas where shaking was noticeable but less severe, communication behavior was more prominent. Mobile data decline zones showed high spatial correlation with intensity VIII+ regions, proving their effectiveness as rapid indicators for identifying strongly affected areas. Notably, mobile location data enabled accurate identification of strongly affected zones within 30 min post-earthquake.