NOAA Center for Satellite Applications and Research banner
National Oceanic & Atmospheric Administration website NOAA Center for Satellite Applications and Research website

STAR Satellite Rainfall Estimates

Validation Documentation

At appoximately 2000 UTC each day, archive files for the 24 hours ending the previous 1200 UTC are created and stored in the archive. The archive from the past 3 days are stored online; older data are on disk and must be ordered via e-mail at

File specifications are as follows:

  • Data format:

    The files are stored as 16-bit little Endian-ordered integers (I*2), representing the rainfall amount for the hour indicated by the YYYYMMDD.HH file name (e.g. all.20030401.03 contains data for 0200-0300 UTC on 1 April 2003) in hundredths of millimeters. Negative numbers indicate missing values. NOTE: If you are using a big Endian machine (e.g. AIX, HP-UX, IBM mainframe, Macintosh, SGI, Solaris) you will need to use the code here to convert these little Endian files to big Endian.

  • Grid projection / size:

    The data are on the HRAP grid, covering the region from X=1 to X=1075 and Y=26 to Y=825 (i.e., 800 rows by 1075 elements). The file is read by rows and the first point in the data file is (X=1, y=26); i.e., the file begins reading from the lower left. The data can be converted to lat/lon using the following code:






  • Data Fields:

    The file contains four fields in the following order:

    • Stage IV

    • QMORPH

    • Operational Hydro-Estimator (east&west combined)

    • SCaMPR

Date Range 1300 UTC 11 April 2003 – 1300 UTC 7 July 2003 1400 UTC 7 July 2003 – 1700 UTC 5 Nov. 2003 1800 UTC 5 Nov. 2003 – 1200 UTC 28 July 2004 1300 UTC 28 July 2004 – 1200 UTC 10 Nov. 2004 1200 UTC 10 Nov. 2004 – 1500 UTC 1 March 2005 1600 UTC 1 March 2005 – 1200 UTC 5 Feb. 2013 1300 UTC 5 Feb. 2013 – 1200 UTC 10 June 2013 1300 UTC 10 June 2013 – present
File size 1046 x 769 1075 x 800 1075 x 800 1075 x 800 1075 x 800 1075 x 800 1075 x 800 1075 x 800
Record 1 Stage IV Stage IV Stage IV Stage IV Stage IV Stage IV Stage IV Stage IV
Record 2 A-E A-E A-E A-E A-E A-E A-E QMORPH
Record 3 H-E (operational) H-E (operational) H-E (operational) H-E (operational) H-E (operational) H-E (operational) H-E (operational) H-E (operational)
Record 4 H-E (radar screen) H-E (radar screen) H-E (radar screen) H-E (radar screen) H-E (radar screen) SCaMPR SCaMPR SCaMPR
Record 5 GMSRA v. 1 GMSRA v. 1 GMSRA v. 2 GMSRA (radar calibration) SCaMPR GMSRA    
Record 6 GMSRA v. 2 GMSRA v. 2 GMSRA (radar calibration) NRL-Turk GMSRA (radar calibration) NRL-Turk    
Record 7 NRL-Turk NRL-Turk NRL-Turk   NRL-Turk      

  • Converting Code:

    The code is designed to convert the all* files from I*2 binary to ASCII and to add latitude and longitude for each point. convert2arch.f


  • Starting 16Z on 1 Mar. 2005, EAST and WEST combined HE will replace HE with Radar and HE without Radar (Both of them are GOES-EAST only).
  • Starting 10 Nov. 2004, SCaMPR has been added to the validation package.
  • Starting 28 Jul. 2004, GMSRAf (for "fixed calibration") has been taken off from the validation package. The current validation program only includes five algorithms.
  • As of 1800 UTC on 05 Nov. 2003, previous GMSRA version 2 has been switched to version 1 (now called GMSRAf (for "fixed calibration")). The current version 2 (named as GMSRAr (for "radar calibration")) is a real- time radar calibrated rain algorithm. Due to programming error, GMSRA version 2 data is not usable between Nov. 5 and Nov. 18 2003.
  • As of 1900 UTC on 07 July 2003, the size of the validation grid files has been changed from 1046 by 769 to 1075 by 800. This was done because the original grid size did not cover northern Maine and southeastern Florida.
  • Due to a programming error, all of the operational Hydro-Estimator (HE) estimates (i.e. without the radar correction) from 15Z Monday, 23 June 2003 through 15Z Wednesday, 25 June 2003 were multiplied by 0.15. The problem can be corrected in part by dividing all of the estimates by 0.15; however, since the original HE has a reporting precision of approximately 0.3 mm, the corrected HE estimates will have a precision of 2 mm. The files on the archive have been left unadjusted and it is the user's discretion as to how to proceed.