動物調查努力量應隨棲地與季節調整

動物調查努力量大多固定，例如在所有棲地與季節以相同次數或頻度進行調查。但是群聚結構與個體偵測率可能隨空間或時間而有所變異；固定努力量可能使不同樣本之間的代表性或完整性不同，從而導致錯誤結論。本研究以6個物種—豐富度模型建構人工群聚，並以7個偵測率模擬抽樣，說明不同群聚結構及偵測率，需要的努力量不同。另以臺南大凍山低海拔闊葉林及七股濕地鳥類資料，實證不同棲地與季節需要不同的努力量；資料分析結果亦發現普遍採行的每月1次調查頻度及環保署環評技術規範的努力量大多嚴重不足。依據本研究結果，建議未來應避免在所有棲地與季節以相同努力量調查動物群聚，並在調查過程隨時評估樣本完整性。本研究也提出動物調查實施流程及停止調查之決策程序。 Most animal-survey projects use a fixed-effort sampling regime for every site during the entire survey period. However, the community structure and/or the individual's detectability might be different for various habitats and seasons. Therefore, fixed-effort surveys could yield samples with unequal completeness and cause misinterpretations of true natural state. To understand the influence of community structure and individual detectability on the required effort for animal surveys, I constructed artificial communities with six species-abundance models and randomly sampled from these using seven different individual detectabilities (i.e., sampling probabilities). I calculated the rarefaction curves and Chao2 estimated species richness for these simulated survey datasets. I also used two empirical datasets, one sampled from a woodland and one from coastal avifauna in order to determine their optimal sampling effort. Results suggest that both community structure and individual detectability, which might vary between habitats and/or seasons, significantly affect the required effort. Therefore, researchers should avoid to sample with fixed-effort when conducting biodiversity surveys. Species richness estimators (e.g., Chao2) and rarefaction curves can be used to assess the completeness of samples and therefore serve as a stopping rule for such surveys.