基于航拍数据的南极维多利亚地企鹅源温室气体排放量估算

Aerial photography based estimation of greenhouse gas emissions from penguins in Victoria Land,Antarctica

海洋动物是南极气候变化的＂生物指示剂＂,其排泄物中丰富的碳（C）和氮（N）等营养物质为土壤中温室气体的产生与排放提供了有利条件,企鹅作为一种重要的海洋动物,因此其聚居区成为甲烷（CH4）和氧化亚氮（N2O）等温室气体排放的潜在＂热点＂区域.然而,受企鹅数量遥感资料的限制,区域尺度上企鹅源温室气体排放总量尚缺乏精确估算.以南极维多利亚地难言岛企鹅聚集区为研究对象,基于0.1 m分辨率航拍照片发展了面向像元的RGB颜色模型法（pixel-oriented RGB color model）识别企鹅数量,通过企鹅粪便CH4和N2O排放通量、企鹅排便量等数据建立了企鹅源温室气体估算模型.结果显示,航拍照片中企鹅像元在RGB彩色空间模型中的R值（17~104）与其他背景像元（〉110）存在显著差异,该差异可以作为将企鹅与背景像元有效分离的理论依据;南极维多利亚地难言岛企鹅总数为19150只,企鹅源CH4和N2O排放总量分别约为275和2.99 kg.

Marine animals are ＂bio-indicators＂ of climate change in the Antarctic. The abundance of nutrients in their excreta, such as carbon（C） and nitrogen（N）, promotes the emission of greenhouse gases（GHGs）, including methane（CH4） and nitrous oxide（N2O）. Adélie penguins are important marine animals, and their colonies have therefore become potential ＂hotspots＂ of GHG emissions. Field observations have been conducted to study penguin excreta as a source of CH4 and N2O emissions in the Antarctic Peninsula. However, due to the lack of penguin population data, the total GHG emissions have not been estimated at the regional scale. With the development of remote sensing monitoring techniques in recent years, high-resolution satellite imagery and aerial photographs have made it possible to monitor penguin populations in the Antarctic, with the results used as a foundation to estimate GHG emissions. Inexpressible Island, a small, rocky island on the shore of the Ross Ice Shelf in Terra Nova Bay, Victoria Land, Antarctica, houses an important colony of Adélie penguins. This study extracted information about the penguins from aerial photographs using a pixel-oriented RGB（red, green, and blue） color space model of Victoria Land. Adélie penguin populations on Inexpressible Island could be estimated via a penguin shadow analysis in the RGB model. To validate the accuracy of the RGB model, we randomly selected several images and used the RGB method alongside an artificial visual interpretation method, with the results from both methods compared using the relative deviation index. In addition, a GHG model was developed to estimate CH4 and N2O emissions from Adélie penguins. The model included a variety of parameters, such as the CH4 and N2O fluxes from penguin excreta, the population number, the fresh weight of excreta produced by a penguin each day, and the duration of CH4 and N2O emissions from the excreta. The results revealed significant differences in the R values between the penguin（R=17–104） and non-penguin（R110） pixels in the aerial photographs assessed by the RGB model. We effectively separated the penguin pixels from background pixels based on the R values. The total number of penguins on Inexpressible Island was estimated to be 19150 and the average percentage deviation of the estimation was 11.04%. The total emissions of CH4 and N2O from these penguins in summer was 275 and 2.99 kg, respectively. Two factors might have led to GHG estimation uncertainties：（1） the GHG emission factors were obtained from an incubation experiment conducted during a previous study that was conducted much earlier than 2013（i.e., the date of this study）;（2） The parameters used in the GHG estimation model, such as the penguin population, fresh weight of penguin excreta, and the duration of the study period, would affect the total estimation of CH4 and N2O emissions. This study established a pixel-oriented method to extract information regarding penguin populations from aerial photographs, and estimated the total GHG emissions from penguins at the regional scale. The results will provide important reference values for the estimation of the GHG budget from marine animals.