Pacts of Free- Roaming Dogs on Spationtemporal Niches of Native Canivores in Taiwan
Abstract
Among animals of the same trophic level, niche partitioning often occurs to facilitate coexistence. However, when a new common predator appears, they may adjust their activity temporally and spatially to avoid the predator. Such changes may increase their niche overlap with each other. Dogs (Canis familiaris) are invasive carnivores that are distributed widely, and they pose serious threats to native mammals in Taiwan. We hypothesize that free-roaming dogs dominate spatial and temporal resources, which forces native mesocarnivores to adjust their realized niches to avoid dogs. We predict that threats from free-roaming dogs result in (1) the activity levels of native carnivores being correlated negatively with those of dogs, (2) native carnivores undergoing spatial or temporal avoidance in response to the presence of dogs, and (3) increased spatial or temporal niche overlap among native carnivores. This study covers four mesocarnivores in Taiwan: the masked palm civet (Paguma larvata), ferret badger (Melogale moschata), small Indian civet (Viverricula indica), and crab-eating mongoose (Herpestes urva). With data from 1270 camera traps, we used relative activity index (RAI), occupancy models, and kernel density estimation to analyze changes in abundance and spatiotemporal niche overlaps. A significant negative correlation was found between dog RAI and those of ferret badgers (p < 0.001) and crab-eating mongooses (p < 0.05). The ferret badgers also avoided dogs spatially (Species Interaction Factor, SIF < 1). However, crab-eating mongooses co-occurred with dogs spatially (SIF > 1), possibly due to their strong dependence on specific environments (i.e., forests around freshwater environments). We did not detect changes in spatiotemporal niche overlap among native carnivores. Variations in niche overlap may need to be observed at finer scales or across other dimensions. This study demonstrates the impact of free-roaming dogs on ferret badgers and crab-eating mongoose, which highlights the importance of conserving riparian environments from the threat of dogs. Future research that integrates other niche dimensions is necessary to better understand the impact of dogs.
Methodology and Implementation Steps
我們整合了來10份自動相機資料集，來源包含政府機關、私人公司以及學術單位。2010年至2021年間，共蒐集1,270部自動相機，海拔範圍1至2540公尺。我們隨後依據不同模型的需求篩選資料。
我們也納入多項環境因子為控制變因，包括道路密度、海拔、坡度、降雨量與森林覆蓋率。此外，考量到食蟹獴對河岸環境的依賴性，我們在食蟹獴的模型中納入與水體距離為因子。
為了檢驗原生食肉動物的活動程度是否與遊蕩犬呈負相關（預測1），我們以這些物種的相對活躍指數（RAI，每100天的獨立照片數）建立對數常態分佈的廣義線性模型。為了檢驗原生食肉動物是否在空間或時間上與犬隻進行棲位分化（預測2），以及犬隻是否導致原生食肉動物之間的空間或時間棲位重疊增加（預測3），我們分別運用條件式兩物種佔據模型(Conditional Two-species Occupancy Model)與核密度估計(Kernel Density Estimation)來探討空間交互作用與活動模式。前者可直接納入環境因子計算，並得出物種交互因子(Species Interaction Factor, SIF)。該因子若大於一，代表兩物種共域；小於一，代表劣勢物種迴避優勢物種；等於一，則代表兩物種獨立。核密度估計則可運算出活動模式重疊度(Δ)，我們再以負二項式分佈的廣義線性模型判斷各因子對於Δ之影響。
Innovation and Cross-Disciplinary Collaboration
過去研究多假設，若原生物種未迴避入侵物種，則表示其未受顯著影響。然而，本研究除應用生態棲位相關模型，亦進一步納入相對活躍指數分析，補足單一類別模型的解釋限制。我們發現，即使食蟹獴與遊蕩犬的生態棲位高度重疊，其活躍程度仍因犬隻的存在而顯著下降。該結果顯示，共域並非物種相安無事的結果，而是因原生物種受限於棲地，無法完全迴避犬隻。本研究提供更細緻的影響評估方式，並可作為未來入侵種影響研究的標竿。
TOP