Spatiotemporal pattern and suitable areas analysis of equine influenza in global scale (2005-2022)

Abstract

Equine influenza (EI) is a severe infectious disease that causes huge economic losses to the horse industry. Spatial epidemiology technology can explore the spatiotemporal distribution characteristics and occurrence risks of infectious diseases, it has played an important role in the prevention and control of major infectious diseases in humans and animals. For the first time, this study conducted a systematic analysis of the spatiotemporal distribution of EI using SaTScan software and investigated the important environmental variables and suitable areas for EI occurrence using the Maxent model. A total of 517 occurrences of EI from 2005 to 2022 were evaluated, and 14 significant spatiotemporal clusters were identified. Furthermore, a Maxent model was successfully established with high prediction accuracy (AUC = 0.920 ± 0.008). The results indicated that annual average ultraviolet radiation, horse density, and precipitation of the coldest quarter were the three most important environmental variables affecting EI occurrence. The suitable areas for EI occurrence are widely distributed across all continents, especially in Asia (India, Mongolia, and China) and the Americas (Brazil, Uruguay, USA, and Mexico). In the future, these suitable areas will expand and move eastward. The largest expansion is predicted under SSP126 scenarios, while the opposite trend will be observed under SSP585 scenarios. This study presents the spatial epidemiological characteristics of EI for the first time. The results could provide valuable scientific insights that can effectively inform prevention and control strategies in regions at risk of EI worldwide.

Keywords: Maxent model; equine influenza; spatial epidemiology; spatiotemporal pattern; suitable area.

Figure 1

The 517 global occurrence sites of equine influenza (EI) from 2005 to 2022.

Figure 2

The visualization matrix of correlation coefficients for the twenty-one environmental variables.

Figure 3

Global locations of the fourteen spatiotemporal clusters of EI occurrence.

Figure 4

The AUC of the Maxent model for EI occurrence and response curves of three important environmental variables affecting EI occurrence. (A) The AUC value of the Maxent model of EI occurrence; (B) Response curve of annual average ultraviolet radiation (uvb1) affecting EI occurrence; (C) Response curve of horse density (hd) affecting EI occurrence; (D) Response curve of precipitation of coldest quarter (bio_19) affecting EI occurrence.

Figure 5

The importance analysis of eight environmental variables for EI occurrence in the Maxent model by the Jackknife test.

Figure 6

Global suitable areas of EI occurrence under the current climate scenario.

Figure 7

Global suitable areas of EI occurrence under three future climate scenarios for the years 2050 and 2070. (A) 2050SSP126, (B) 2050SSP245, (C) 2050SSP585, (D) 2070SSP126, (E) 2070SSP245, (F) 2070SSP585.

Figure 8

Changes in global suitable areas of EI occurrence under three future climate scenarios for the years 2050 and 2070. (A) 2050SSP126, (B) 2050SSP245, (C) 2050SSP585, (D) 2070SSP126, (E) 2070SSP245, (F) 2070SSP585.

Figure 9

Centroid position of EI occurrence shifts under three climate scenarios from current to the years 2050 and 2070 (The blue, red and purple lines, respectively, represent the SSP126, SSP245, and SSP585 climate scenarios).