This website houses data from "version 1" of the Iowa Daily Erosion Project. This website is now deprecated. Please check out our "version 2" Daily Erosion Project.

Rainfall Estimates Methodology

Introduction:

The most important variable in our erosion modeling effort is the estimation of rainfall. This page describes the process by which these estimates are made. The most important point is to understand that our rainfall product is an estimate. There are simply not enough observing platforms in the state to create a high resolution depiction of rainfall that is needed for this project.

Rainfall Grid


Rainfall cells over Polk County

Our rainfall estimates are produced on the HRAP (Hydrologic Rainfall Analysis Project) grid. The HRAP grid has square cells on polar stereographic map projection with vertical longitude ("standard longitude") 105° W and "standard latitude" 60° N. This grid was choosen, since it is the grid on which the Stage 4 precipitation analysis is produced on.


Stage 4 Precipitation:

The stage 4 (stage4 hereafter) precipitation product is produced at the National Center for Environmental Prediction (NCEP). The stage4 product is a nationwide composite of regional analyses made at the 12 River Forecast Centers (RFC). The advantage of this product is that some manual quality control is done at the RFCs. The disadvantage is that it can sometimes take many hours for this product to be posted at NCEP.

The stage4 product is derived from multiple data sources including the National Weather Service (NWS) NEXRAD system. This data is analyzed in hourly intervals. For the purposes of this project, sub hourly data is needed, so we needed a way to temporally downscale the stage4 rainfall product.

Downscaling the Hourly Stage4 data:

The stage4 information is produced at hourly intervals. To downscale this to 15 minute intervals, we derived a method to use NEXRAD derived rainfall rates at 15 minute intervals to give us a template for downscaling the hourly stage4 data. We now present the procedure of this downscaling.

Ingesting NEXRAD information:

Information from these six NWS WSR-88D radars are used to distribute the hourly stage4 data into 15 minute bins. Their site identifiers and names are:

  • KDVN: Davenport, Iowa
  • KDMX: Des Moines, Iowa
  • KARX: LaCrosse, Wisconsin
  • KMPX: Minneapolis, Minnesota
  • KOAX: Omaha, Nebraska
  • KFSD: Sioux Falls, South Dakota

The NIDS Composite Reflectivity product is used from the radars to generate a rainfall rate via the standard Z-R relationship. This rainfall rate is assumed to be an accumulation over 15 minute periods. So for a given hour, four values of rainfall are produced for each individual cell in the HRAP grid.

Comparison with Stage4 Product

After the NEXRAD 15 minute estimates are made, a program compares the stage4 precipitation estimates with the 15 minute NEXRAD estimates to distribute the hourly stage4 data into 15 minute bins. Here is the logic for each cell in the HRAP grid:

First, lets define the 4 NEXRAD 15 minute estimates as N0, N1, N2, & N3, the sum of these 15 minute NEXRAD estimates as NT, the hourly stage4 data as ST and the resulting estimates as E0, E1, E2, & E3.

If ST equals zero, then E0, E1, E2, & E3 are set to zero. Justification: The process of distributing the stage4 data into 15 minute bins does not produce more rainfall than what orginally existed in the hourly data. Even if the NEXRAD estimates were greater than zero, the stage4 zero sum is maintained in the estimates.

If ST is greater than zero and ET is zero, then E0, E1, E2, & E3 are simply set to ST/4. Justification: Since the NEXRAD is not providing us with any insight into the sub-hourly distribution, we simply assume an even distribution.

If ET and ST are greater than zero, E0 is set to (ST/NT)*N0, E1 is set to (ST/NT)*N1, E2 is set to (ST/NT)*N2, and E3 is set to (ST/NT)*N3. Justification: If both the stage4 and NEXRAD data is indicating rainfall, then we want to distribute the hourly stage4 data into 15 minute bins using the same temporal distribution as is found in the NEXRAD information.

Result:

The result of this methodology is hourly stage4 data distributed into 15 minute bins. Accumulated stage4 rainfall is maintained, while we also try to get a better depiction of the high intensity rainfalls that produce erosion.

Revised: 5 Jun 2004.