Weather radars produce large amounts of data and this has important im
plications for the archiving and analysis of data. The need for method
s to deal with weather radar data sets will only increase as the Unite
d States National Weather Service (NWS) continues deployment of its 13
7 WSR-88D radars as part of the NEXRAD program. In this article we des
cribe a compression and archiving strategy for weather radar data and
present results for 62 days of reflectivity data from a radar operated
by the NWS, as well as results for 60 days of reflectivity data for a
radar operated by the Bureau of Meteorology in Australia. In their or
iginal format, these two sets require 60 GB of storage. In the format
we describe, they require 4.8 GB and the data is portable across many
platforms. The software for manipulating the converted data is simple,
efficient, and easy to implement in C or Fortran. The savings in disk
space and reduction in reading time compare favorably with what is at
tainable with deflation, the algorithm used in the popular gzip compre
ssion program. (C) 1997 John Wiley & Sons, Ltd.