Puerto Rico Testsite for Exploring Contamination Threats (PROTECT)

波多黎各探索污染威胁的试验场（PROTECT）

• 批准号: 10382038
• 负责人: Akram N Alshawabkeh
• 金额: $ 1.5万
• 依托单位: NORTHEASTERN UNIVERSITY
• 依托单位国家: 美国
• 财政年份: 2021
• 资助国家: 美国
• 起止时间: 2021-07-21 至 2022-01-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/10382038
• 关键词: Affect; Area; Artificial Intelligence; Award; Big Data Methods; Catchment Area; Characteristics; Classification; Computational Technique; Consumption; Data; Detection; Environment; Fresh Water; Human Resources; Imagery; Modeling; Output; Pattern; Precipitation; Public Health; Puerto Rico; Research; Research Personnel; Source; Surface; Surface Properties; System; Training; Water; externship; feature extraction; ground water; insight; machine learning algorithm; mathematical model; remote sensing; response; transmission process

Project Abstract Karst terrains show distinctive surface and subsurface features associated with sinkholes, springs, and caves. Many of the hydrological and hydraulic characteristics associated with these features make groundwater systems in karst areas a significant freshwater resource for human consumption and ecological integrity. Unfortunately, these characteristics also make karst aquifers highly vulnerable to contamination. Karst springs, specifically, provide researchers insight to the interaction between input precipitation, karst aquifers as the transmission medium, and discharge as the output response. A karst catchment area or spring catchment is a mapping unit defined by the surface area and the subsurface drain system that contributes to spring discharge. Therefore, being able to determine this catchment area, and understand how hydrologic conditions in karst environments affect contaminant fate and transport in springs, is crucial to establishing effective source water protection areas (SWPAs) to safeguard groundwater and public health. However, due to the inherent complexity of karst systems, it has been challenging for the PROTECT team to (1) forecast spring response, and (2) infer subsurface properties from surface features in karst environments, since both require the application of advanced computational techniques, such as big data analytics and machine learning (ML) algorithms. Hence, the proposed project for the KC Donnelly Externship award will allow me to train along with Mr. Sean Griffin and his team on the core concepts and models behind geospatial artificial intelligence to automate object detection, object classification, and feature extraction from remotely sensed imagery. Mastering the application of ML algorithms will allow me to discover patterns in my research data, and to construct predictive mathematical models using these discoveries.

项目摘要 喀斯特地形显示出与天坑相关的独特的地表和地下特征， 泉水、洞穴。许多与这些相关的水文和水力特征 喀斯特地区地下水系统的特征使之成为人类重要的淡水资源 消费和生态完整性。不幸的是，这些特征也使得喀斯特 含水层极易受到污染。具体来说，喀斯特泉水为研究人员提供了 洞察输入降水、作为传输介质的岩溶含水层之间的相互作用， 并放电作为输出响应。喀斯特流域或泉水流域是一个映射 由表面积和有助于弹簧的地下排水系统定义的单位 释放。因此，能够确定这个集水区，并了解如何 喀斯特环境中的水文条件影响污染物的归宿和泉水的迁移， 对于建立有效的水源保护区（SWPA）以保护水源至关重要 地下水和公共卫生。但由于喀斯特系统固有的复杂性， PROTECT 团队面临的挑战是 (1) 预测弹簧响应，以及 (2) 推断地下 喀斯特环境中表面特征的特性，因为两者都需要应用 先进的计算技术，例如大数据分析和机器学习 (ML) 算法。因此，拟议的 KC Donnelly 实习奖项目将使我能够 与 Sean Griffin 先生及其团队一起培训背后的核心概念和模型 地理空间人工智能可自动执行对象检测、对象分类和特征分析 从遥感图像中提取。掌握机器学习算法的应用将有助于 我发现研究数据中的模式，并构建预测数学模型 利用这些发现。