Skip to content

ISSE Seed Project: TVA Database Mining

Team: Liem Tran, Geography, and John Schwartz, Civil and Environmental Engineering. Additional team members: Carol Harden and Kelsey Ellis, Geography; Richard Strange and Ben Keck, Forestry, Wildlife and Fisheries; Ungtae Kim and Jon Hathaway, Civil and Environmental Engineering

2014-2015 Report

picture of raingages and streamgages
Raingages and streamgages on a street basemap with time series plot and histogram of selected station shown on the left panel.

Efforts in 2014-2015 were to continue to develop and maintain the GIS-based, integrated database initially set up in 2013-2014, and to increase the number of layers and their metadata in the geodatabase (87 layers by the end of 2014-2015). More common GIS tools for data display, classification, attribute and spatial queries have been added to the GIS web interface. In addition, an all-new web application for TVA using the latest and most popular open-source solution has been developed during 2014-2015. This application is used to query and visualize and spatial and temporal data in the TVA region. Compared with the previous implementation based on ESRI products, this version is mainly based on open source software but much faster and more efficient. For example, the new web application is able to provide time series plots and histograms of precipitation and stream gauge data.

The TVA geodatabase has laid the groundwork for collaboration between the watershed group and the Landscape Architecture Program in developing a specialized database for a water quality project in southeast Tennessee (e.g., adding more GIS data, e.g., sewer system, greenways, mining operations, etc., and developing tools for spatial analysis serving water planning purposes). In addition, continued effort with the TVA River Operations group includes interfacing with their new initiative to use Deltare database software for hydrology data.

2013-2014 Report

This project, which completed its second year in 2014, is jointly funded with UT’s Office of Research and Engagement. The goal of the project is to create a an integrated database that includes hydrologic, water quality, biodiversity, weather, climate, and land use data, to enable interdisciplinary systems-level investigations of water resource issues within the TVA system. The recent focus is on data entry into a GIS-based database with a web interface. More than 80 datasets/layers and associated metadata have been processed and organized into an ESRI geodatabase. In addition more than two dozen GIS tools for data display, classification and attribute, and spatial queries have been implanted in the web interface.

Our current effort is to develop a data portal—the Tennessee Valley Authority (TVA) Data Portal—based on the already collected data and GIS layers as well as new data (e.g., hydrologic and hydraulic data, population censuses, land cover data, elevation data, climate data, soil data, and more). The main goals of the TVA data portal are to (1) collect, preserve, integrate, and merge data sets to create custom outputs that measure hydrologic, environmental, and population changes in the TVA region, (2) develop tools and procedures to manage and disseminate the data collections as well as tools to communicate with other TVA databases, (3) perform outreach to engage the scientific community and the public regarding the data products, and (4) establish a database structure to ensure the long-term use of the data portal.

Two specific projects were completed by Civil Engineering MS students who graduated Spring semester 2013; they were Joe Rungee and James Jones. Joe Rungee completed a reservoir routing optimization problem examining the effects of climate change specifically on Norris Dam operations. The title of his thesis is: Climate Change Impacts on the Tennessee Valley Authority Regulated River Operations: Long-Term Assessment of the Norris Reservoir. One manuscript for publication is in preparation. James Jones, utilizing TVA’s long-term climate data, explicitly examined trends over time and whether basin locations were co-related with any temporal climate changes. The title of his thesis is: Spatial and Temporal Variability in Precipitation and Streamflow in the Upper Tennessee Valley. Two manuscripts for publication were prepared—one manuscript was submitted to the Journal of Hydrology – Regional Studies, and has been accepted with minor revisions. Drs. Jon Hathaway and Kelsey Ellis were both involved with this work. The second manuscript will be submitted to Hydrological Processes.

A third project, involving Jennifer Starkey, is assisting TVA in the large-scale development of their river management system, specifically the development of a HEC-RAS unsteady flow model for the entire TVA system. Her research involves addressing and finding solutions to conditions that cause model instabilities. TVA’s goal to is have a fully integrated climate-hydrology-hydraulics modeling system for prediction of water surface elevations based on climate forecasts and potential dam operations and flow regulation. Data to run this model will be archived in a new database by Deltares.

UT’s development of a Data Portal will be linked with TVA’s Deltares Archive system. UT’s database development is more than the hydrological data that will be stored in the Deltares Archive, including biological, land use, water quality data. A broader array of database variables will enhance the capability of UT to pursue fundamental multidiscipinarily research grants. UT will continue its working relationship with TVA’s River Operations Division and long-term hydrological data management of the Tennessee Valley.

picture of map layers

2012-2013 Report

This seed project is a 50/50 partnership with UT’s Office of Research and Engagement and involves creation of an integrated database representing hydrologic, water quality, climate, weather, land use, macroinvertebrate, and fish data in the TVA system. The purpose of the database is to facilitate investigation of system level questions related to water resources, such as impacts of anthropogenic disturbances to watersheds that may take years to be observed, potentially highlighting subtle climate signal trends, and providing opportunities to examine responses of watersheds to Best Management Practice (BMP) implementations and restoration practices in a comparative way. The TVA watersheds also provide the unique opportunity to consider various river management constraints within the context of a complex energy system (hydro, coal, nuclear—all involve water resources). Project leaders include Drs. John Schwartz and Ungtae Kim, both from Civil and Environmental Engineering, Dr. Carol Harden from Geography, and Dr. Richard Strange from Forestry, Wildlife and Fisheries. This year, Dr. Liam Tran will join the team in order to help with GIS and database schema to enable the data to be stored, maintained, and queried in a robust manner.

Efforts in the first year were largely focused on gathering and understanding the available data in relation to the geography of the region (which sub-basins are the most data rich, thereby representing the best targets for intensive investigation). Preliminary analysis of the data has already begun, with discipline specific studies underway in assessing the hydrology, fish population dynamics, and reservoir management practices associated with the system. At the moment the data is segregated in a manner to address these discipline-specific questions. Activity in the coming year will focus on the creation of a GIS-based, integrated database to allow more comprehensive, interdisciplinary, systems-level questions to be addressed.

Initial work focused on mapping the locations of TVA rain gauges, TVA stream flow gauging stations, TVA aquatic biota sampling sites, TVA power generation facilities, USGS stream gauges, NWS climate stations, and TDEC biotic sampling sites in the context of geographical features such as terrain, road networks, stream/river networks, and political and watershed boundaries. These data sources were cross referenced with STORET, the EPA’s data archive of water quality data. Land use data for 1992 and 2006, acquired from the National Land-Cover-Database (Federal Multi-resolution Land Characteristics Consortiuum), were superimposed on the maps and changes over this period noted. These maps (seen on the following page) enabled the group to determine that the Emory and Little River watersheds had the greatest quantity of data available from TVA and other sources, thereby making them suitable candidates for future investigations.

Initial analysis of the hydrologic data is aimed at determining if precipitation and runoff relationships in the basin have been stationary over the last 30-year period. In addition the group is examining if land-use changes over the period 1992 to 2006 have affected the unit hydrographs within the basins. Additional analysis related to the adequacy of reservoir operational procedures to meet multi-objective criteria in the face of changing climate are also underway.

fisheries plot

Initial analysis of the fisheries data involves examining a dataset comprising 225 localities in the upper Tennessee River sampled between 2009 and 2011. Models were developed to explain variation in fish species and functional traits as a function of abiotic factors within the watershed. Initial results indicate that abiotic factors explained 10% of the variation in the fish species dataset and 15% of the variation in the functional traits dataset. The abiotic factors with the most explanatory power in both models were elevation, slope, and limestone (see figure on previous page). These results support a hypothesis that human-mediated biotic factors, e.g., land use and forest cover, may influence fish species occurrence and community assembly of fish functional traits more than abiotic factors. Currently, the researchers are seeking to increase the explanatory power of the model by adding abiotic factors such as stream substrate type and stream order, and biotic factors such as riparian vegetation and land use in the watershed. With this new data the researchers intend to make a comparison of the explanatory power of abiotic and biotic factors, and possibly discern if human mediated factors have more of an influence on fish communities than abiotic factors.

land cover map
land cover map