Resolving the role of groundwater in multi-scale land-atmosphere dynamics using simulation, sensor networks and satellites: Juniata River Basin

Prediction of flood and drought has proven particularly challenging in small upland river basins (1 to 10,000 km2 and 1st to 4th order channel networks) which represent the major water generating areas to downstream, higher order rivers. Current prediction schemes for the most part, rely upon statistical methods, not physics-based prognostic models at these scales. Further, while advances in weather prediction have come from improved representation of soil moisture and vegetation fluxes, existing land surface schemes using in NWP models are limited to vertical moisture transport in the soil column and largely ignore deeper soil moisture processes and ground water. This proposal investigates the value of integrating a physics-based, highly dataconstrained model of ground water hydrology for a river basin in central Pennsylvania into flood/drought prediction, and into NWP models.

We will address two primary hypotheses: 1) A bedrock-to surface layer hydrologic modeling system, driven by satellite observations of the land surface and meteorological reananalyses, will improve simulation of flood and drought conditions in the Juniata River basin at 1-10,000 km2 spatial scales; and 2) Reanalysis of the ground water hydrology of the Juniata river basin using PIHM will significantly improve the accuracy of predictions of the basin-wide, daily surface energy balance at time scales where groundwater hydrology is predictable (days to months), relative to a prediction that does not include explicit modeling of ground water hydrology. The project will bring together resources including the Penn State Integrated Hydrologic Model (PIHM), the National Science Foundation sponsored Critical Zone Observatory (CZO) at Shale Hills, PA, and the Penn State ensemble Kalman filter (EnKF) data assimilation system. PIHM will be implemented across the Juniata River basin, initialized with high quality, static land surface characteristics, driven by a combination of North American Regional Reanalysis (NARR) meteorological and Moderate Resolution Imaging Spectroradiometer (MODIS) vegetation inputs, and optimized using basin-wide stream flow and ground water table data for the period from 2000 through 2010. The optimized model and basin-wide hydrologic reanalyses will then be used to evaluate the skill of the model-data assimilation system in predicting flood/drought conditions in the basin as well as sensible and latent heat fluxes in the basin relative to current operational models. In the short-term, we anticipate that this reanalysis of basin hydrology could be used to improve flood and drought forecasting in this and other river basins. Longer-term goals of the research are to describe the influence of seasonal, inter-annual and decadal climate variability and change on extreme events (floods and droughts), and to progress towards a fully-coupled, multiscale hydrologic and atmospheric modeling system that could yield important benefits in long-term weather forecasting. 

In terms of the FY2010 Climate Prediction Program for the Americas request for proposals, this proposal squarely addresses the request to “improve hydrologic predictions at regional scales at intraseasonal to interannual time scales,” and to, “improve understanding and process modeling of land surface physics including soil moisture, vegetation, snowpack, groundwater and processes in complex terrain,” two of the three elements called for in the request. This project has relevance to the first element of the call in that it has the potential to contribute to “climate predictability at intraseasonal to interannual time scales focusing on land memory effects.” The research also addresses CPPA priority 3. “Climate-based hydrologic and water management applications at regional scales”.

We will address two primary hypotheses: 1) A bedrock-to surface layer hydrologic modeling system, driven by satellite observations of the land surface and meteorological reananalyses, will improve simulation of flood and drought conditions in the Juniata River basin at 1-10,000 km2 spatial scales; and 2) Reanalysis of the ground water hydrology of the Juniata river basin using PIHM will significantly improve the accuracy of predictions of the basin-wide, daily surface energy balance at time scales where groundwater hydrology is predictable (days to months), relative to a prediction that does not include explicit modeling of ground water hydrology. The project will bring together resources including the Penn State Integrated Hydrologic Model (PIHM), the National Science Foundation sponsored Critical Zone Observatory (CZO) at Shale Hills, PA, and the Penn State ensemble Kalman filter (EnKF) data assimilation system. PIHM will be implemented across the Juniata River basin, initialized with high quality, static land surface characteristics, driven by a combination of North American Regional Reanalysis (NARR) meteorological and Moderate Resolution Imaging Spectroradiometer (MODIS) vegetation inputs, and optimized using basin-wide stream flow and ground water table data for the period from 2000 through 2010. The optimized model and basin-wide hydrologic reanalyses will then be used to evaluate the skill of the model-data assimilation system in predicting flood/drought conditions in the basin as well as sensible and latent heat fluxes in the basin relative to current operational models. In the short-term, we anticipate that this reanalysis of basin hydrology could be used to improve flood and drought forecasting in this and other river basins. Longer-term goals of the research are to describe the influence of seasonal, inter-annual and decadal climate variability and change on extreme events (floods and droughts), and to progress towards a fully-coupled, multiscale hydrologic and atmospheric modeling system that could yield important benefits in long-term weather forecasting.