Dear Sandra,

I computed the drop size distributions. They are at
ftp.pa.op.dlr.de in the directory /pub/martin

There are 2 files dsd_dlr.tar.gz and dsd_wash.tar.gz. The *d files contain
the N(D) for the 20 classes. Note that there are 2 times, one in hh:mm and
the other one in decimal hours. The *t files contain a tabel with the
countet drops (n_tot) and the deat-time corrected drops (nktot), rainrate,
reflectivity (D^6), lwc, N0 and D0 according to Waldvogel (1974), the
accumulated sum and for the DLR files 2 empty colums for rain gauge data
(here at Oberpfaffenhofen I run the disdrometer together with an rain gauge).
Every 1-minute spectra was filtered to have more than 20 drops.
Then the 10 and 60 minutes DSD were generated. There must at least 8 or 48
dsd's in that time interval, and the rainrate of the averaged dsd must
pass 0.2 mm/h.

I attached a word file with a table containing the various a and b for
both disdrometers and different time interval. I also looked how the
minimum rain rate affects the a and b.
I only used the averaged dsd to compute the rainrate and to filter. I
alternatively could only use those dsd which pass the filter for
averaging. This might give different results.

Best regards
         Martin