Developing statistical downscaling to improve water quality understanding and management in the Ramganga sub-basin

通过“改善所有人获得清洁水和卫生设施的机会”来增强社会包容性，需要深入了解与水相关的生态系统及其可以为社会提供的好处。然而，“目前通过可持续发展目标进程收集的全球数据并不能反映淡水生态系统的总体状况或已知趋势”。新的数学科学研究对于融合地球观测和地面数据源以填补知识空白、提高对水质的认识和应对发展中国家面临的水管理挑战至关重要。该提案将提供世界领先的统计研究，支持恒河流域Ramganga次流域水质监测和建模框架的开发。传统的水采样基于少量站点，非常劳动密集且昂贵，而我们的提案汇集了来自新的原位传感器的数据，以高时间频率提供数据，再加上密集耦合的高端现场水上和水下的Ramganga的光学特性的特征与测量校准的高光谱成像辐射计部署从无人机和数据从新的卫星任务（哨兵2）。总之，这些提供了一个有效的和前所未有的手段，收集重要的数据，在一系列的环境和污染排放的情况下，光学水类型在Ramganga盆地。将这些数据与水质量的传统测量方法相结合，将为以迄今无法实现的空间和时间分辨率外推水质量数据提供非常理想的框架。恒河流域覆盖了印度总陆地面积的26%，其水质和水资源对世界上最大和最密集的人口之一的福祉至关重要然而，由于快速的工业化和城市化等活动，它们正在受到损害，并且由于缺乏历史和当代排放和质量数据，减缓工作受到阻碍。该项目将专门为Ramganga次流域制定和实施新的统计方法，将新的和现有的水质数据与遥感卫星数据（多个传感器的历史数据和最近哨兵任务的新检索数据）相结合。为了应对Ramganga子流域的水质挑战，并充分利用新的数据流，将开发新的统计降尺度和数据融合方法，通过变系数，分层贝叶斯建模框架，将河流网络结构和流量分位数模型结合起来。这些方法支持整合不同的数据来源，以便能够预测水资源状况和相关的不确定性，为一系列社会经济和气候变化情景下的风险模型提供信息，并为今后的监测设计提供信息。全流域水质量估计的输出将提供给政策制定者和未来的研究人员，以指导政策和设计未来的采样点和时间频率。

Empowering social inclusion through 'improving access to clean water and sanitation' for all requires arobust understanding of water-related ecosystems and the benefits that they can provide to society.However, 'the global data currently collected through the SDG process do not reflect the general state or trends known about freshwater ecosystems'. Novel mathematical sciences research is essential to enable fusion of Earth observation and on-the-ground data sources to fill the knowledge gaps, provide improved understanding of water quality and address the water management challenges faced by developing countries. This proposal will deliver world leading statistical research supporting the development of a water quality monitoring and modelling framework for the Ramganga sub-basin of the Ganges river basin.Traditional water sampling is based on a small number of sites, and is very labour intensive and expensive, and our proposal brings together data from new in-situ sensors, delivering data at high temporal frequency, coupled with intensive coupling high-end in-situ above and below water characterisation of the biogeo-optical properties of the Ramganga with measurements from calibrated miniaturised hyperspectral imaging radiometers deployed from drones, and data from new satellite missions (Sentinel 2). Together, these provide an efficient and unprecedented means of collecting significant data across a range of environments and pollution discharge scenarios of optical water types in the Ramganga basin. Coupling these data with conventional measures of WQ will provide the much desired framework for extrapolating WQ data at hitherto unachievable spatial and temporal resolutions.Covering 26% of India's total landmass, water quality and water resources in the Ganges basin are vitalfor the wellbeing of one of the largest and densest global populations (43% of India's population).However, they are being compromised due to activities such as rapid industrialization andurbanization, and mitigation efforts are hampered by lack of historical and contemporary discharge and quality data. This project will develop and implement new statistical methodology specifically for the Ramganga sub-basin to integrate the new and existing water quality data with remote sensing satellite data (both historical data from multiple sensors and new retrievals from recent Sentinel missions). To address the water quality challenges in the Ramganga sub-basin and to fully utilise the new data streams, novel statistical downscaling and data fusion methodologies through a varying coefficient, hierarchical Bayesian modelling framework will be developed to incorporate river network structure and model quantiles of flow. These approaches support integration of disparate data sources to enable prediction of water resource condition and associated uncertainties to inform risk-based modelling under a range of socio-economic and climate change scenarios, and provide tools to inform future monitoring design. The output of catchment-wide WQ estimates will be made available to policy makers and future researchers to guide policy and design future sampling sites and temporal frequency.