Iranian Journal of Veterinary Medicine

Iranian Journal of Veterinary Medicine

Spatiotemporal Clustering and Hotspot Detection of Reported African Swine Fever Cases in the Cordillera Administrative Region, Philippines, January 2023–February 2025

Document Type : Original Articles

Author
Mathematics Department, Abra State Institute of Sciences and Technology, Philippines
10.22034/ijvm.20.5.1006131
Abstract
Background: African Swine Fever (ASF) remains a major threat to swine production in the Philippines. In the Cordillera Administrative Region (CAR), understanding the spatial and temporal distribution of reported ASF cases is important for improving local disease surveillance and control.
Objectives: This study examined the spatiotemporal clustering and hotspots of laboratory-confirmed ASF cases recorded and validated by the DA—CAR Animal Health Unit in CAR from January 2023 to February 2025.
Methods: The ASF cases were summarized by year, month, and province. Standard Deviational Ellipse analysis was applied to examine changes in geographic center and directional distribution. Global Moran’s I, Local Indicators of Spatial Association, Getis-Ord Gi*, and SaTScan space-time cluster analysis were used to assess spatial and space-time clustering. The Prophet time-series model was applied to weekly cases to examine temporal trends.
Results: Most reported ASF cases occurred in 2024 and were temporally concentrated from August to October. At the provincial level, Apayao recorded the highest number of cases, followed by Abra, Ifugao, and Kalinga. Standard deviational ellipse results indicated a shift in the geographic center and orientation of reported cases between 2023 and 2024. Global Moran’s I showed weak but statistically significant positive spatial autocorrelation in 2023 (I=0.0593, P=0.0005), 2024 (I=0.0942, P=0.0001), and the complete study period (I=0.0898, P=0.0001). Local Moran’s I identified localized clusters and spatial outliers. Meanwhile, Getis–Ord Gi* detected significant hotspots and cold spots. SaTScan identified significant high- and low-rate space–time clusters. The Prophet model showed an upward trend and suggested a possible yearly pattern with higher values around September to October.
Conclusions: Reported ASF patterns in CAR varied across locations and periods. The findings supported flexible surveillance, prioritized field monitoring, and targeted biosecurity activities in repeatedly affected and emerging areas. Coordination among neighboring local government units may strengthen outbreak detection and response.
Keywords