非洲人類猴痘的地理分布：運用生態棲位模型理解環境變數與自然宿主物種

Abstract

猴痘(mpox) 是一種再現的人畜共通傳染病，目前更在非洲中部和西部地區引發公共健康關注。本研究採用生態棲位模型(ecological niche modeling,ENM）以預測人類猴痘的潛在地理分布，並識別高風險外溢區域。本研究分析2010年至2022年間相關報告確認的人類猴痘病例之550地點，並結合從最初33個環境變數經相關矩陣分析後，篩選出的16個變數進行 最大熵模型( maximum entropy model, MaxEnt)建模分析。這些變數包括氣候資料、土地覆蓋類型、海拔高度、人口數量、人類足跡，以及三種已知囓齒類自然宿主物種（Cricetomys、Funisciurus 和 Graphiurus 屬）的分布情形。最大熵模型建模的表現良好（平均AUC值為0.867），並指出海拔高度、人口數量、最冷季節的降水量（BIO19）、葉面積指數以及Graphiurus屬為最具影響力的預測因子。高風險區主要集中在剛果民主共和國、奈及利亞南部、喀麥隆、賴比瑞亞以及獅子山共和國。整合人口資料後發現，剛果民主共和國有超過2,900萬人、奈及利亞有1,300萬人居住於高度適合猴痘傳播的環境區域內。本研究強調將宿主物種納入風險模型的重要性，因為其存在對空間預測結果具有顯著影響。研究結果顯示，結合生態、環境與人口統計數據，有助於識別優先監測區域並制定公共衛生預防措施。

Mpox is re-emerging zoonotic disease with increasing public health concern across Central and West Africa. This study used ecological niche modeling (ENM) to predict the potential geographic distribution of human mpox and identify areas at high risk of spillover. A total of 550 confirmed mpox case locations reported between 2010 and 2022 were analyzed using maximum entropy model (MaxEnt), along with 16 environmental variables selected from an initial set of 33 following a correlation matrix analysis. These variables included climate data, land cover types, elevation, population count, human footprint, and distributions of three known rodent reservoir species (Cricetomys, Funisciurus, and Graphiurus spp.). Maxent model performed well (mean AUC = 0.867) and identified elevation, population count, precipitation of the coldest quarter (BIO19), leaf area index and Graphiurus spp. as the most influential predictors. High-risk zones were concentrated in the Democratic Republic of the Congo (DRC), southern Nigeria, Cameroon, Liberia, and Sierra Leone. Integrating population data revealed that over 29 million individuals in the DRC and 13 million in Nigeria reside in zones of high environmental suitability for mpox transmission. This study emphasizes the importance of including reservoir host species in risk models, as their presence significantly shaped the spatial predictions. These findings demonstrate the value of combining ecological, environmental, and demographic data to identify priority areas for surveillance and public health intervention.