Song Shu

This project focuses on developing and implementing cutting-edge algorithms, software tools, and corresponding data products for monitoring key reservoir variables across the contiguous United States. Our goal is to provide high-temporal-resolution estimates of reservoir water levels, surface areas, storage volumes, inflows, and outflows in a near-real-time framework. These data products will substantially enhance the calibration and validation processes of NOAA's national reservoir operation models. Crucially, this project will enable the expansion of these operation models to a vast number of ungauged reservoirs throughout the country, thus broadening their applicability and utility. To realize this goal, we will initially compile a consistent, georeferenced reservoir inventory using Sentinel-1 SAR data at 10 m spatial resolution. Leveraging the SWOT satellite’s unique capability, we will construct reservoir surface area-water level rating curves and river width-water level rating curves for the reservoir’s inflow and outflow rivers. With these established rating curves, we will be able to convert reservoir surface area measurements and river width measurements obtained from regular SAR (Sentinel-1A, Sentinel-1C, and NISAR) and optical (Landsat-8/9 and Sentinel-2A/B) satellite image observations to corresponding water level estimations, thereby greatly improving the temporal resolution of water level measurements for both reservoirs and their inflow/outflow rivers. We will also use SWOT wide-swath altimetry observations, nadir-looking satellite altimetry satellites (Jason-3, Sentinel-3A/B, Sentinel-6, and ICESAT-2), and indirect water level measurements from SAR and optical satellite image data to develop an innovative hybrid CNN-LSTM deep learning model. This model will be capable of estimating, hindcasting and forecasting the water level, surface area, and volume change of reservoirs and the water levels and discharges of their inflow and outflow rivers. The accuracy and reliability of these estimates and forecasts will be rigorously evaluated.

This project will provide lake water quality and storage monitoring services for management decisions in the Great Lakes region of Eastern and Southern Africa by integrating multi-sensor Earth observations, machine-learning models, and cloud computing technology. Specific objectives include: 1) Establish well-calibrated and validated algorithms/models and associated software tools for monitoring and mapping lake water quality parameters (chlorophyll-a, toxic algal blooms, turbidity, suspended sediment concentration, colored dissolved organic matter, water temperature), invasive water hyacinth, lake water level and volumetric change for three of the largest African Great Lakes in this stage of the project, with the potential to extend such services by applying the calibrated models to other major lakes in Eastern and Southern Africa in the future. 2) Conduct joint field surveys to collect to collect on-lake measurements on various water parameters, and co-develop remote sensing algorithms with the personnel of the regional hub at RCMRD, and other regional professionals. 3) Implement water quality and quantity algorithms/models into opensource software tools, and transfer the source codes and software tools to the regional hub-RCMRD through technical workshops and co-development activities. 4) In partnership with RCMRD, develop training materials and hold training workshops for regional professionals and end users to build their capability in using multispectral, thermal, radar altimetry Earth observations and geospatial information technologies for water availability and water quality monitoring, assessment, and prediction. 5) Develop online maps and a visualization website with RCMRD, provide user-tailored services, and organize user conferences for stakeholders, policy makers and resource managers to use Earth observations and geospatial tools for water resource management and decision-making activities. We will greatly expand previous SERVIR’s “Lake Victoria Water Quality and Ecosystem Management” project to include the three largest lakes (Lake Victoria, Lake Tanganyika, Lake Malawi) of the African Great Lakes as focus lakes for field and in situ measurements for satellite model calibration and validation. Besides lake water quality parameters, lake water quantity variables (lake water level, water storage volume) will be derived and monitored by applying innovative remote sensing algorithms/models to multiple Earth observations. Our team has many years of on-the-ground-and-water field experience in different countries/continents. The three Points of Contact at the regional hub RCMRD in Nairobi, Kenya will serve as co-Is of this project. This project will address SERVIR’s priority thematic topic “Water Resources and Hydroclimatic Disasters”. Data products, calibrated models, software tools, and web-based maps from this project will help the regional hub personnel, professionals, and users to document lake water level and storage volume change, harmful algal blooms, sediment plumes, and other threats to their freshwater resources. Lake water level and storage information will meet the regional needs for drought and flooding early warning as well as for water availability and supply assessment for human drinking, agriculture, aquaculture and food security. The satellite-based lake water quality information will benefit Eastern and Southern Africa by providing the early warning and monitoring of toxic algal blooms and invasive water hyacinth outbreaks to protect approximately 25% of Earth’s unfrozen freshwater resources and the health of millions of people in the African Great Lakes region. Our project will build and improve the capacity of regional professionals, government policy makers and resource managers in using new generation of Earth observations and geospatial technologies to support water resource management, lake aquatic ecosystem protection, natural hazard mitigation, and decision-making activities.

Dams across the entire U.S. have been taken down in increasing numbers in the past decade as they have filled with sediment, become unsafe or inefficient (Bellmore et al. 2017; Connor et al. 2015). After a dam is removed, there is usually an acute release of water and sediments from the previous impoundment to downstream, which could rapidly alter the downstream channel, due to the redistribution of the sediment and the damage of the shoreline (Doyle et al. 2002). In addition, long-term soil erosion and sediment scour could occur on the riverbed and on the two banks of the upstream area, which was previously submerged by the high-level water in the impoundment and is then revealed to air after dam removal. Both processes above will significantly change the bathymetry of the up- and down-stream river channel. Traditionally, the monitoring and assessing of dam removal impacts on stream geomorphology, bathymetry, and bank soil erosion requires temporally frequent in-situ surveys (e.g., cross-sectional and longitudinal surveys) on the up- and down- stream before and after dam removal. These temporally frequent surveys require a lot of manpower and cost a great amount of time. The advent of Unmanned Aerial Vehicles (UAVs) offers a new approach that can be used to efficiently monitor soil erosion and bathymetric changes caused by dam removal and save significant time and money. However, few dam removal efforts have incorporated this newly emerged technology, and significant information and technology gaps still exist in planning, mapping, and assessment. Ward’s Mill Dam, a 130-foot-long, 20-foot-high rock and concrete dam, was built in 1890 and impounded the Watauga River downstream in Valle Crucis, NC (about six miles downstream). The dam was used for hydroelectricity to power a sawmill and to provide electricity for local homes in the valley, but it had been inactive since 2016 and became a significant blockage on the Watauga River for aquatic species migration and for public recreational activities, such as kayaking and canoeing. Watauga County, therefore, decided to remove the dam on May 12 - 14, 2021. This project first aims to take advantage of the recent removal of Ward’s Mill Dam (Figure 1), located on the Watauga River in Western North Carolina. We would like to compare UAV-based pre/post-removal data on the erosion/sediment transport regime in order to develop a better understanding of how southern Appalachian river systems like the Watauga will respond to the removal of dams of similar type and size, which are common throughout southern Appalachia. In particular, we are interested in the river's bathymetric response to dam removal. Bathymetric Light Detection and Ranging (LiDAR) has emerged with significant capabilities to improve data collection and quality capabilities for monitoring and inspections. The accuracy of UAV-based data, including LiDAR data and optical data will be first assessed. Then, the collected data will be used to detect the changes occurred to the banks and riverbed after dam removal. In addition, the possibility of combining different UAV-based sensors (e.g., LiDAR and optical camera) for the survey of topography and bathymetry of different waterbodies (e.g., ocean coasts, lakes, and rivers) will be evaluated. We are particularly interested in evaluating the applicability of UAV-based LiDARs for a broader range of areas (e.g., waterbodies with different levels of water clarity).

Dams across the whole U.S. have been taken down in increasing numbers in the past decade as they have filled with sediment, become unsafe or inefficient (Bellmore et al. 2017; Connor et al. 2015). After a dam is removed, there is usually an acute release of water and sediments from the previous impoundment to downstream, which could rapidly alter the downstream channel, due to the redistribution of the sediment and the damage of the shoreline (Doyle et al. 2002). In addition to the quick and short-term impacts on the downstream channel, a long-term soil erosion could occur to the two banks of the upstream area that is previously submerged by the high-level water in the impoundment and is then revealed to air after the removal of the dam. Ward’s Mill Dam is a large structure (approximately 40 m long and 6 m high) that impounds the Watauga River downstream of Valle Crucis, NC (about 10 km downstream of Valle Crucis). The dam was built in 1890 and has been damaged by floods and repaired several times. It was used for hydroelectricity to power a sawmill and to provide electricity for local homes in the valley. The dam has been inactive since 2016 and the power generating facilities have been decommissioned. Currently, it becomes a significant blockage on the Watauga River for aquatic species migration and for public recreational activities, such as kayaking and canoeing. Watauga County, therefore, has determined to remove the dam in late February or early March 2021. Traditionally, the monitoring and assessing of dam removal impacts on stream geomorphology and bank soil erosion requires temporally frequent in-situ surveys on the up- and down- stream before and after the dam removal, for example cross-sectional and longitudinal surveys. These temporally frequent surveys require a lot of human work and cost a great amount of time. The advent of Unmanned Aerial Vehicle (UAV) offers us a new approach that can be used to monitor efficiently the dam removal induced soil erosion and to save significantly human work and time. However, few dam removal efforts have incorporated this newly emerged technology for that purpose. This research project aims to take advantage of the planned removal of Ward’s Mill Dam to evaluate the applicability of incorporating UAV technology in the dam removal efforts and to develop a better understanding of how dam removal of similar type and size will impact the soil erosion on rivers like Watauga which are common in southern Appalachia.