ITR: Advanced Imaging and Information Technology for Assessing the Ecological and Economic Impact of Brazilian Free-Tailed Bats on Agroecosystems

在美国西南部，每年夏夜，数以百万计的巴西无尾蝙蝠（Tadarida brasiliensis）都会贪婪地吃掉大量昆虫。这些蝙蝠提供了一种鲜为人知的农业害虫控制服务，科学界和决策者对此知之甚少。拟议的工作将评估该物种对全国自然和农业生态系统的生态和经济影响。该项目在信息技术发展方面具有创新性，在复杂性和规模上具有独特性。它需要计算机科学家、应用数学家、气象学家、生态学家和生态经济学家的共同努力。拟开展的活动包括：传感技术。-设计、开发、部署和评估用于数百万蝙蝠和害虫的热、超声波和雷达传感的算法和系统。这样的算法和系统目前还不存在，但对于提供全国范围内无尾蝙蝠种群和过程的可靠普查至关重要。计算机视觉技术将用于分析蝙蝠个体和群体夜间出现、飞行路径和觅食行为。计算建模。-设计、开发、解决和验证农业-昆虫-蝙蝠系统跨时间和空间尺度的计算模型。在较低层次的组织过程，如蝙蝠个体及其生理功能，将被分析，以解决从种群到景观的较高生态层次的组织所带来的问题。特定洞穴或桥梁的当地人口模型将被推广到德克萨斯州的空间明确区域模型，然后再推广到描述巴西无尾蝙蝠对农业生态系统的全国影响的空间明确景观模型。尽管蝙蝠在世界上无处不在，并且被认为是哺乳动物中的第二大目，但目前还没有完整的无尾蝙蝠个体或种群生活史模型，事实上也没有任何一种蝙蝠的生活史模型。系统集成。-提出的模型将使用概念和空间层次进行整合。这些传感技术提供的结果将结合起来代表觅食和迁徙的过程。这些结果以及分子工具（如粪便DNA）、昆虫学和农业信息、能量学、气象和毒理学数据库的整合，将为蝙蝠种群的计算模型提供输入参数和验证数据。这些模型将综合考虑饮食因素、能量消耗和分配、天气模式、毒物影响等对蝙蝠及其猎物的出生率和死亡率以及种群规模的相互依赖性。拟议活动产生的更广泛影响自然虫害防治服务的重要性往往只有在它们被人类活动削弱或消除时才显现出来。拟议中的研究将对巴西无尾蝙蝠的经济影响提供现实的估计。所提出的技术可以推广到其他物种，因此在生物学、生态学、生态经济学和农业领域具有广泛的影响。本文提出的图像分析方法可以应用于其他大规模视频跟踪应用，例如其他蝙蝠物种、昆虫、群居哺乳动物（如海豹、驯鹿）、殖民地海鸟的群体行为分析、人类群体行为分析、人类运动视频数据挖掘、国土安全视频监控等。拟议的研究对社会的一个重要影响将是联邦、州或地方当局根据对自然虫害控制的基本生物学和经济原则的根本改进的了解制定适当的政策反应。拟议的努力以多种方式解决了ITR计划的目标。它本质上是多学科的，在计算科学和生态学领域之间提供了一个新的桥梁。这可能会在这两个领域产生新的、意想不到的见解和技术。提议的努力将训练生物学和计算机科学的学生充分整合信息技术和科学。这两个学科的研究生和本科生将学习进行联合实地实验、分析数据和使用计算模型。受生物学问题启发的计算机科学项目对女性有特殊的吸引力，我们希望我们的项目能鼓励更多的女性探索生物学的计算方面。

Intellectual Merit of the Proposed ActivityMillions of Brazilian free-tailed bats (Tadarida brasiliensis) voraciously consume enormous quantities of insects each summer night throughout the southwestern United States. These bats provide an agricultural pest control service little understood by the scientific community and policy makers. The proposed effort will evaluate the nationwide ecological and economic impact of this species on both natural and agricultural ecosystems. The project is innovative in its development of information technology and unique in its complexity and scale. It requires the collaborative efforts of computer scientists, applied mathematicians, meteorologists, ecologists, and ecological economists. Proposed activities involve:1. Sensing Technologies.-Design, develop, deploy, and evaluate algorithms and systems for thermal, ultrasonic, and radar sensing of millions of bats and insect pests. Such algorithms and systems currently do not exist but are crucial for providing a reliable census of the nationwide free-tailed bat populations and processes. Computer vision techniques will be developed to analyze nightly emergence, flight paths, and foraging behaviors of individuals and groups of bats.2. Computational Modeling.-Design, develop, solve, and validate computational models of the agricultural-insects-bats system across temporal and spatial scales. Processes at lower levels of organization, such as the individual bat and its physiological functions, will be analyzed to solve problems posed at higher ecological levels of organization from the population to the landscape. Local population models for particular caves or bridges will be generalized to a spatially explicit regional model for Texas, and then to a spatially explicit landscape model that describes the nationwide impact of Brazilian free-tailed bats on agricultural ecosystems. Currently, no complete individual or population life-history models exist for free-tailed bats nor, indeed, for any bat species, in spite of the fact bats are ubiquitous throughout the world and are distinguished as the second largest order of mammals.3. System Integration.-The proposed models will be integrated using both conceptual and spatial hierarchies. Results provided by the sensing technologies will be combined to represent the processes of foraging and migration. The integration of these outcomes as well as molecular tools (such as fecal DNA), entomological and agricultural information, energetics, meteorological and toxicological databases, will provide input parameters and validation data to computational models of bat populations. The models will integrate the interdependencies of dietary factors, energy consumption and allocation, weather patterns, effects of toxicants, etc. on birth and mortality rates and population sizes of bats and their prey.Broader Impacts Resulting from the Proposed ActivityThe importance of natural pest-control services becomes evident often only when they are degraded or eliminated by human activity. The proposed research will provide realistic estimates of the economic impact of Brazilian free-tailed bats. The proposed techniques may generalize to other species and thus have a broad impact in the fields of biology, ecology, ecological economics, and agriculture. The proposed methods of image analysis may apply to other large-scale video tracking applications, for instance, the analysis of group behavior of other bat species, insects, herding mammals (e.g., seals, caribou), colonial seabirds, the analysis of human crowd behavior, data mining of video of human motion, and video surveillance for homeland security. An important impact of the proposed research on society will be the development appropriate policy responses by federal, state, or local authorities based on the fundamentally improved understanding of the underlying biological and economic principles of natural pest control. The proposed effort addresses the goals of the ITR Program in a number of ways. It is multidisciplinary in nature, providing a new bridge between the fields of computational sciences and ecology. This may lead to novel, unanticipated insights and technologies in both fields. The proposed effort will train students in biology and computer science to fully integrate information technology and science. Graduate and undergraduate students from both disciplines will learn to conduct joint field experiments, analyze data, and work with computational models. The projects in computer science that are inspired by questions in biology promise to have a special appeal to women, and we expect that our project will encourage more women to explore the computational aspects of biology.