Research

Fuelberg Lab

Go Back a Page
Current Research :: Past Research :: Dr. Henry E. Fuelberg :: Fuelberg Lab Home


Hydrologic Modeling Using WAM and High Resolution WSR-88 Precipitation Data

A Research Project for the Florida Department of Environmental Protection

Modelers attempting to estimate rainfall-driven runoff volumes and nonpoint source pollutant loadings traditionally have had to rely on rainfall estimates from surface gages. The horizontal spacing of the gage data is relatively coarse, and many only provide information at 24-hour intervals.

To estimate runoff volumes over large areas, various statistical techniques have been applied to "distribute" the rainfall from several gages (e.g., Thiessen polygons). Unfortunately, these estimates may have large errors, and that error is propagated directly into the hydrologic model. It is clear that gage data alone are inadequate for estimating runoff volumes.

The Florida Department of Environmental Protection (FDEP) wishes to develop a high resolution historical database of precipitation to support surface and ground water pollution modeling requirements of the Total Maximum Daily Loading (TMDL) program. Approximately ten years of data are needed to calibrate the rainfall driven models. The goal of this project is to develop that database for the FDEP. Specifically, we are merging rainfall estimates from precision weather radars with those from traditional rain gages into a single statewide set of archived (historical) hourly rainfall estimates on a 4 x 4 km grid. This database will be available to all analysts that require the use of hourly (or daily historical) rainfall volumes over a given area.

Theoretical Considerations

Although radars provide excellent spatial and temporal coverage, their precipitation amounts may be erroneous. Conversely, gages provide excellent precipitation values, but, as noted above, their spatial and temporal coverage is inadequate. The National Weather Service (NWS) River Forecast Centers (RFCs) seek to mitigate the limitations of radar-derived precipitation by combining these data with gage-derived values.

The National Weather Service (NWS) has implemented software that integrates gage- and radar-derived precipitation values. Called the RFC-wide Multisensor Precipitation Estimator (denoted RFC-wide or MPE), it provides real time gridded precipitation estimates covering a River Forecast Center's entire area of responsibility. The procedure uses hourly digital precipitation data at a 4 km resolution from the radars within the computational area. These hourly data are called the Digital Precipitation Array (DPA). Since most grid points within the computational area are covered by more than one radar, a precipitation climatology is prepared for each radar to determine which radar provides the best coverage of each individual grid point. This climatology then is used by RFC-wide for all subsequent calculations over the watershed. The result is that numerous radars are used to calculate precipitation over the total area, with the radar providing the best coverage at each individual 4 km grid point being used there.

Once the 4 km grid of radar information is obtained, the MPE software calculates hourly bias correction factors for each radar to improve the remotely sensed precipitation values. Then, the bias corrected radar-derived precipitation data are merged with hourly rain gage observations using optimal interpolation. RFC-wide is considered by many to be a major advancement compared to older procedures. The MPE software is being used in the current research described below.

Current Research

During the current contract period January - December 2006, we will perform the following tasks for FDEP:

  1. Extend FSU's historical MPE database to 10 years, making the complete database period 1996-2005. Years 1996-2004 were completed as part of earlier contracts. The domain of our data set is shown in Fig. 1. An example of the MPE product is shown in Fig.2. One should note the enhanced spatial detail that is provided by the radar and the agreements between maximum/minimum values of both products.

  2. Figure 1. Northern extent of domain (bold red line) for which FSU is preparing the MPE historical dataset. Figure 2. Rain gauge only product, and (Right) MPE final product for the period 0800-1600 UTC 6 August 2001. Note the spiral bands associated with Tropical Storm Barry in the MPE product that are not detected by the rain gauges.

     

  3. Quantify differences in rain volumes between point source rain values (gauges), Thiessen polygons, and the MPE product over various sized basins within Florida,

  4. Learn to run the Watershed Assessment Model (WAM) through training from its developers at Soil and Water Engineering Technology, Inc. (SWET) of Gainesville, FL,

  5. WAM currently is configured for rain gauge data. FSU will configure WAM for the optimal use of MPE input data,

  6. Disseminate our findings at conferences and in the literature so that other agencies will know of FDEP's state-of-the-art efforts,

  7. Train FDEP personnel to run the MPE procedure, and

  8. Collaborate with FDEP personnel to place the MPE dataset online at FDEP so that users will have ready access.

Extended plans for 2007 include running the WAM model on various watersheds and during various rainfall scenarios to document the ability of fully distributed hydrologic models with high resolution input (MPE) data to correctly simulate streamflow.

Graduate Students on this Project

John Sullivan

Click here to view Dennis VanCleve's Master's thesis entitled An Intercomparison of Mean Areal Precipitation from Gauges and a Multisensor Procedure.

Click here to view Joe Marzen's Master's thesis entitled Development of a Florida High-Resolution Multisensor Precipitation Dataset for 1996-2001 -- Quality Control and Verification.

Click here to view Greg Quina's Master's thesis entitled Statistical and Hydrological Evaluations of Rain Gauge- and Radar-Derived Precipitation for the Florida Peninsula.

Click here to view related research by Bryan Mroczka entitled Convective Patterns Across the Florida Peninsula Driven By The Position of The Subtropical Ridge (and other research).

Click here to reach the web site of the Florida Department of Environmental Protection.

Click here to reach the web site of Soil and Water Engineering Technology, Inc. (SWET).

Click here to reach the web site of the Southeast River Forecast Center.

last updated February 1, 2007