End-to-end cyberinfrastructure to support data-model integration in water connectivity studies

点击翻译按钮获取中文摘要

Hydrologic (or water) connectivity describes the movement of water between pairs of locations and has important implications for process understanding and watershed management. However, most hydrologic models do not explicitly depict connectivity, resulting in considerable knowledge gaps. Two emerging trends in the fields of hydrology and remote sensing are creating exciting, but largely untapped, opportunities in connectivity science. First, increased availability of data from optical and radar satellite sensors is transforming the way in which environmental dynamics are studied; most notably, dynamic surface water processes can now be observed at unprecedented spatial and temporal resolution. Second, agent-based models (ABMs) are commonly used in ecology and are powerful tools to simulate interactions between individual elements in an ecosystem. Some hydrologists have suggested that data-driven ABMs might be more efficient than equation-based models (EBMs) to simulate small-scale interactions leading to water connectivity, but that hypothesis has yet to be fully tested. The integration of remote sensing data into EBMs and ABMs has the potential to reveal formerly unknown connectivity dynamics, thus contributing to a more complete understanding of watershed processes and their management. However, multi-sensor satellite data processing and the application of EBMs and ABMs require significant computing power and data storage capacity, which is the motivation for this funding application. ******An end-to-end cyberinfrastructure will be developed to support data-model integration in water connectivity studies across agricultural and wetland-dominated watersheds. The cyberinfrastructure will comprise two multi-core and high memory computing servers, one network-attached data storage system, and one high-performance workstation. The servers will allow the processing of multi-sensor satellite images, the computation of connectivity metrics, and their integration into EBMs and ABMs for advanced water connectivity modelling. The network-attached data storage system will be configured to enable frequent data back-ups, with built-in redundancy in case of accidental data losses. Lastly, the high-performance workstation will serve as a link between the two computing servers, and as an interface for stakeholders to test the connectivity models and offer feedback based on specific watershed management objectives or targets. The research enabled by this cyberinfrastructure will be unprecedented in terms of volume of data processed and model complexity: it will not only address key questions relevant to connectivity science, but it will also help prioritize how, when and where corrective action should be taken within watersheds by providing spatially and temporally explicit information on water movements.