Leveraging social media and deep learning to detect rare megafauna in video surveys

Deep learning has become a key tool for the automated monitoring of animal populations with video surveys. However, obtaining large numbers of images to train such models is a major challenge for rare and elusive species because field video surveys provide few sightings. We designed a method that takes advantage of videos accumulated on social media for training deep-learning models to detect rare megafauna species in the field. We trained convolutional neural networks (CNNs) with social media images and tested them on images collected from field surveys. We applied our method to aerial video surveys of dugongs (Dugong dugon) in New Caledonia (southwestern Pacific). CNNs trained with 1303 social media images yielded 25% false positives and 38% false negatives when tested on independent field video surveys. Incorporating a small number of images from New Caledonia (equivalent to 12% of social media images) in the training data set resulted in a nearly 50% decrease in false negatives. Our results highlight how and the extent to which images collected on social media can offer a solid basis for training deep-learning models for rare megafauna detection and that the incorporation of a few images from the study site further boosts detection accuracy. Our method provides a new generation of deep-learning models that can be used to rapidly and accurately process field video surveys for the monitoring of rare megafauna.     

A comparison of fixed-site and non-fixed-site approaches for species protection

The efficacy of simultaneously advancing two distinct conceptual designs (referred to here as fixed-site and non-fixed-site) for species conservation and protection is addressed. In the literature, numerous models can be found that typically stem from a particular design, but rarely are comparisons made between approaches. This paper presents a more integrated optimization framework that models landowner behavior and species viabilities at a landscape scale. Regional demand for resource extraction is used as the economic driver, a variant of simulated annealing is used to solve the model under different species protection approaches, and a detailed species population simulator is utilized to measure biological responses. When directly comparing the outcomes of different species protection strategies from a case study in Oregon (USA), it was found that neither approach was universally superior in terms of financial value or degree of protection for two late seral forest dependent species.     

Invasive species change detection using artificial neural networks and CASI hyperspectral imagery

For monitoring and controlling the extent and intensity of an invasive species, a direct multi-date image classification method was applied in invasive species (salt cedar) change detection in the study area of Lovelock, Nevada. With multidate Compact Airborne Spectrographic Imager (CASI) hyperspectral data sets, two types of hyperspectral CASI input data and two classifiers have been examined and compared for mapping and monitoring the salt cedar change. The two types of input data are all two-date original CASI bands and 12 principal component images (PCs) derived from the two-date CASI images. The two classifiers are an artificial neural network (ANN) and linear discriminant analysis (LDA). The experimental results indicate that (1) the direct multitemporal image classification method applied in land cover change detection is feasible either with original CASI bands or PCs, but a better accuracy was obtained from the CASI PCA transformed data; (2) with the same inputs of 12 PCs, the ANN outperforms the LDA due to the ANN’s non-linear property and ability of handling data without a prerequisite of a certain distribution of the analysis data.     

Protecting endangered species under future climate change: From single-species preservation to an anticipatory policy approach

Anthropogenic climate climate change presents a unique challenge for endangered species policy and an opportunity for policy makers to develop a more predictive and robust approach to preserving the nation's biological resources. Biological and ecological reactions to shifting climate conditions and the potential feedbacks and synergistic effects of such changes may threaten the well-being of many species, particularly of those already in jeopardy of extinction. The United States Endangered Species Act of 1973 will fail to keep pace with increasing numbers of species needing protection as long as it remains focused on protecting species individually. The actmust not be abandoned, however; it holds tremendous promise for preserving biological diversity through a more proactive, anticipatory perspective. The current Endangered Species Act should be reinforced and improved by better integration of scientific expertise into habitat and community preservation listing decisions and recovery plan devlopment. Given the uncertainties surrounding long-term environmental consequences of human activities and resource use, a longer-term perspective must be integrated into all efforts to protect our biotic resources. Under appointment from the Graduate Fellowships for Global Change administered by the Oak Ridge Institute for Science and Ecducation for the US Department of Energy.     

北京1o个常绿树种颗粒物吸附能力研究

大气中颗粒物PM( particuIate matter)不仅造成环境污染,还对人体造成严重的危害.城市绿色植物作为大气过滤器,能够有效地提高城市空气质量,保护人体健康.因此了解不同树种对空气颗粒物的吸附滞纳作用是必要的.以北京植物园10种常绿植被为研究对象,应用空气气溶胶再发生器( QRJZFSQ_I)测定了北京市常见6种乔木和4种灌木叶片对空气总悬浮颗粒物( TSP)﹑PM10﹑PM2.5和PM1.0的吸附能力.结果表明①不同树种叶片表面附着颗粒物的能力差异明显,最高的是雪松( Cedrus deodara)和油松( pinus tabuliformis),吸附量分别是(18.95±0.71)μg.cm-2和(14.61±0.78)μg.cm-2,冷杉( Abies fabri)最小,为(8.02±0.4)μg.cm-2;②不同树种叶片单位面积对不同颗粒物的附着能力也存在差异,附着PM10能力最强的是油松和雪松,附着PM2.5能力最强的是雪松﹑铺地柏( Juniperus procumbens)﹑龙柏( Juniperus chinensis cv. kaizuka)和油松,附着PM1.0能力最强的是雪松﹑铺地柏﹑冷杉和油松;③不同月份叶片上附着的各粒级颗粒物( PM10﹑ PM2.5 )占TSP的比例不同.其中PM10在4~6月之间主要表现两种变化趋势,一是先上升后下降,主要为灌木树种;二是逐渐上升,主要的树种是乔木树种.而PM2.5则没有这种明显的变化趋势.     

安徽"两山一湖"区域生态不稳定因素辨识及其对策

通过对“两山一湖“区域生态系统的分析,认识到影响该区域生态不稳定的主要因素是人类活动对野生动植物栖息地的破坏、毁林垦种、外来物种入侵、森林火灾等;提出了保护森林资源、调整土地利用结构、实行退耕还林、推广水流域综合治理、加强物种保护、预防松材线虫病和森林火灾、建立以森林为主体的生态安全保障体系.     

聚合氯化铝中纳米Al13的层析法分离及其水解过程研究

采用聚丙烯酰胺凝胶层析法分离纯化聚合氯化铝(PAC)中的Al13形态,并采用Al-Ferron逐时络合比色法和27Al-NMR对Al13形态进行分析和表征;结果表明,在层析法分离中,随着洗脱时间的延长各种铝形态按分子的大小依次被洗脱下来,因此,截取中间组分即可得到最高含量为99%以上的纳米Al13形态;通过对水解过程的研究可以推断,各种铝形态最终水解产物的粒径多集中在20μm和100μm附近,不同的铝形态到达最终水解产物的速度各不相同,具有最高Alb含量的纳米Al13形态具有较高的水解稳定性.     

环境条件变化对太湖沉积物磷释放的影响

本文系统研究了环境条件变化对太湖沉积物磷释放的影响，结果表明：温度、ＰＨ、氧化还原条件对沉积物磷释放有很大影响，获得了在不同环境条件下磷释放的方程，结果还表明，厌氧下（ＤＯ〈１．０ｍｇ．ｌ＾－１）沉积物磷的最大释放量明显高于好氧条件（ＤＯ〉８．０ｍｇ．ｌ＾－１），其释放速率分别为３．１４μｇ．ｇ＾－１．ｄ＾－１和２．７５μｇ＾－１．ｄ＾－１，相应的动力学方程可表示为ＣＰ＝Ａｅ＾ｂ／ｔ，观察到藻与沉     

A Generic Impact-Scoring System Applied to Alien Mammals in Europe

Abstract: We present a generic scoring system that compares the impact of alien species among members of large taxonomic groups. This scoring can be used to identify the most harmful alien species so that conservation measures to ameliorate their negative effects can be prioritized. For all alien mammals in Europe, we assessed impact reports as completely as possible. Impact was classified as either environmental or economic. We subdivided each of these categories into five subcategories (environmental: impact through competition, predation, hybridization, transmission of disease, and herbivory; economic: impact on agriculture, livestock, forestry, human health, and infrastructure). We assigned all impact reports to one of these 10 categories. All categories had impact scores that ranged from zero (minimal) to five (maximal possible impact at a location). We summed all impact scores for a species to calculate "potential impact" scores. We obtained "actual impact" scores by multiplying potential impact scores by the percentage of area occupied by the respective species in Europe. Finally, we correlated species’ ecological traits with the derived impact scores. Alien mammals from the orders Rodentia, Artiodactyla, and Carnivora caused the highest impact. In particular, the brown rat (Rattus norvegicus), muskrat (Ondathra zibethicus), and sika deer (Cervus nippon) had the highest overall scores. Species with a high potential environmental impact also had a strong potential economic impact. Potential impact also correlated with the distribution of a species in Europe. Ecological flexibility (measured as number of different habitats a species occupies) was strongly related to impact. The scoring system was robust to uncertainty in knowledge of impact and could be adjusted with weight scores to account for specific value systems of particular stakeholder groups (e.g., agronomists or environmentalists). Finally, the scoring system is easily applicable and adaptable to other taxonomic groups.     

北方农牧交错带不同植被保护及恢复措施物种多样性研究

为研究不同生境类型及不同植被恢复方式植物群落结构特征及物种多样性,本文以位于我国西北农牧交错带的宁夏盐池为研究对象,根据不同植被恢复及保护措施特点,在宁夏盐池布设典型样地。样地类型分为:天然草地、退耕还草地、撂荒地、湿地、人工封育草地、流沙地等6种类型。植被调查采用样线和样方相结合的方法,内容包括:植物名称、株数、盖度、高度、生物量(鲜质量)等。运用生态统计学原理,结合BiodiversityPro、MicrocalOrigin等生态学统计软件对不同样地类型植物进行聚类、多样性指数计量等处理、分析。聚类分析表明:不同植被保护及恢复措施下,植物组成、群落结构、主要建群种等都表现出不同特点,不同样地具有各自不同的优势群落、主要建群种等;多样性指数计量结果表明:不同样地间由于生境类型、植被保护及恢复措施等不同造成样地间物种多样性的差异,湿地由于充足的土壤水分条件,无论是丰富度指数还是综合多样性指数都远高于其他样地类型,综合多样性指数D、H分别为12.783、2.915。此外,退耕还草地和天然草地多样性指数也较高,尤其是退耕还草地,D、H分别为11.514和2.696,是农牧交错带一种较为合理的植被恢复措施;多样性较低的是人工封育草地,D、H分别为2.599、1.576。造成人工封育草地物种多样性较低的主要因素一是随着封育时间的延长,优势种的优势地位不断加强、种间竞争等导致部分物种的消失等原因;二是随着封育时间的增加,土壤结皮加厚,水分入渗率和利用率降低从而导致植物群落衰退。