This readme.txt file outlines in detail the correction and calibration processes applied to the GPM Ground 
Validation Doppler on Wheels (DOW) OLYMPEX V2 dataset

General Information: Cfradial files were produced from the dorade (sweep) files (using NCAR's
RadxConvert) that were produced from the raw I&Q time series collected during IOPs. Upon request, the
I&Q data can be made available to the PIs. DOW6 uses two independent transmitters with “high” (9.55
GHz) and “low” (9.40 GHz) frequencies; files are segregated/named accordingly. Either frequency can be
used for analysis. Following the IOP on 12 November 2015, only the “low” frequency is available.

Volume Descriptions: RHI volumes contain scans at 22 azimuths – one every degree starting at 50.4 and
ending at 71.4 degrees - with elevations ranging from 0 to approximately 71 degrees. Sector volumes
contain scans at 6 elevations – approximately 2.8, 3.0, 5.0, 7.0, 9.0, and 11.0 degrees – with azimuths
ranging from approximately 39.2 to 83.6 degrees.

Navigation: Latitude and longitude values were obtained from GPS measurements recorded at the
beginning of the project, and are included in the header values of the translated cfradial files. Altitude
was retrieved using Google Earth, with 3 meters added to account for the ground-to-antenna height.
Truck heading was obtained from a solar alignment conducted on 3 November 2015. On 8 December
2015, the truck was repositioned, and a new heading obtained. The translated dorade files were oriented
to the north.
	Latitude: 47.488456 N Obtained from GPS
	Longitude: -123.869193 E Obtained from GPS
	Altitude: 64 m Approximated from Google Earth
	Heading (start to 12/7): 251.4 deg Solar Alignment
	Heading (12/8 – end) 251.1 deg Solar Alignment
	
Clutter-Filtered Fields: To help mitigate ground clutter, several clutter-filtered fields (in addition to the
non-clutter-filtered fields) are provided. Based on the recommendations of Mike Dixon (NCAR), we chose
a special adaptive filter.

ZDR Calibration: During the OLYMPEX IOPs, near-vertical, full 360-degree scans were collected every 10
minutes. In light to moderate rain, the peak of a ZDRM distribution of such a scan should come to 0 dB,
since falling drops on average look circular to a vertically pointing radar. In reality, there is often an offset
from 0 dB due to minute inconsistencies in the transmitter. To help mitigate this issue, the negative of the
observed offset is added to ZDRM, and placed in a new variable called ZDRC. This is regarded by us as the
final value of differential reflectivity and is the preferred product for analysis. We accept an error of +/-
0.2 dB in this calculation. To find the observed offset, histograms of ZDRM were generated with the
following conditions: 1.) no ZDRM values with DBZHC lower than 10 dBZ were included, 2.) no ZDRM
values with RHOHV less than 0.97 or greater than 1 were included, 3.) no ZDRM values were included with
range less than some value from the radar, typically between 1.2 – 2.2 km. For this third condition, the
value was determined from the observation, in the range where ZDR values are often observed to be
abnormally high due to radar signal over-saturation. The peak of the approximately normal distribution of
remaining values was taken to be the ZDR offset. Peaks with unusually high standard deviation, or a small
sample set were not used. In case where offsets could not be determined from available data (if any),
RHIs could be used with data from elevation angles above 70 degrees. In the case where no offset data
could be determined or an extended period of time, an offset of 0 dB was accepted. Offsets were
collected, and ZDRC was modified as needed for every ten minutes of scans. A spreadsheet of applied
offsets is provided in the folder ZDR_Offsets. The first worksheet, called “Data”, is a list and plot of our
initial observations of offset from the raw swp files. These were initially measured in the field following
each IOP, and utilize both vertical scans and RHIs. The second worksheet, called “FreqCharts” is a plot of
transmit frequency vs offset. The third worksheet, called “Applied Offsets” is a list and plot of the offsets
actually applied to the final product.

Reflectivity calibration offset: Based on a discrepancy discovered between DOW reflectivity data and that
from NPOL/NEXRAD, a comparison between DOW6 and WSR-88Ds using OLYMPEX and SNOWIE
data, as well as data collected locally near Denver with matched scans between radars, revealed an offset
estimate of 3.2 (+/- 1) dB, with the DOW being too low. This offset of +3.2 dB has been added to V2 of the
DOW dataset. In the error estimate, variations between receiver calibrations were also considered (varying
up to 1 dB, but typically less). The caveat to this offset is that it doesn’t necessarily hold in melting levels
(where Mie scattering becomes an issue) or values over ~35 dBZ. The offset was still applied in these
scenarios, but these caveats should be kept in mind when analyzing the data.

3 December 2015 issue: Data from 21:52:10 UTC on 2 December 2015 through 17:21:06 UTC on 3
December 2015 were reprocessed to correct for a 2-gate offset. ZDR offsets were recalculated during this time period, which are reflected in the variable ZDRC. These changes are reflected in updated versions of
the ZDR offsets spreadsheets.