The purpose of Algorithm 3B-31 is to compare and accumulate the instantaneous rainfall products from the TMI (2A12) and Combined (2B31) algorithms.  This is done because the combined algorithm should, in principle, produces better rainfall estimates than the TMI alone.  Its swath, however, is only 220 km compared to the 760 km swath of the TMI.  Thus, for monthly accumulations, the best possible solution was thought to be consists of the TMI product (for sampling), but normalized by its ratio to 2B31.  Product 3B31 contains these accumulations and ratio.   The algorithm also provides the ratio rain water contents for 14 vertical layers as well.  References for Products 2A12 and 2B31 are available as part of their descriptions.

    Output data of 2A-12 and 2B-31 are input to the 3B-31 and then accumulated for each month at  0.5º grid elements. Monthly surface rainfall totals are derived by multiplying the mean rainfall rate by the total number of hours in each month.  Since the rain water content profile of 2A12 and rain rate profile of 2B31 are output at different vertical resolutions, averaging the outputs of 2B-31 over several radar range gates to match the 14 layers of 2A12 is necessary.  Also, the rain rate (mm/hr) at each radar gate in 2B-31 is converted into rain water content (g/m^3) to match that in 2A-12.  The relationship between rain rate and rain water content can be found in Haddad et al., (1996).

File Format

    The basic product consists of the 2A12 monthly accumulations and its ratio to 2B31 (from which the monthly accumulation of 2B31 can be computed).  The ratio is also available for the 14 levels defined by the TMI profiling algorithm.  The file content description can be obtained from the Vol. 4 ­ Level 2 and 3 File Specifications provided by the TRMM Data and Information System (TSDIS).  It is available at:  http://tsdis02.nascom.nasa.gov/Documents/ICSVol4.pdf

Notes

1) Current version of 3B31 only checks various quality flags in the 2A12, but not the 2B31.  Recent version of 2B31 has provided an uncertainty index.  The quality check of 2B31 will be added in the next version to remove any input data with poor quality.

2)  Pixels for which the 2A12 is either zero, or unreliable, currently do not allow the user to reconstruct the 2B31 original sollution.  This will be addressed in the next reprocessing.  It may require a file format modification.

References

Haddad, Z. S., E. A. Smith, C. D. Kummerow, T. Iguchi, M. R. Farrar, S. L. Durden, M. Alves and W. S. Olson, 1996: "The TRMM ‘Day-1’ Radar/Radiometer Combined Rain-Profiling Algorithm",  http://pequod.jpl.nasa.gov/paper.html