Known Problems

S-Pol MAP Italy, 1999

An updated, on-line version of this information is available

Azimuth Adjustment
Bad Cross-Pol Parameters ...
Time Glitches in Real-time Data
Polarization Feed Change-Out
Full and Partial Beam Blockage
Zdr Bias Adjustment
Inaccurate Clutter Filter Status

1.25° Azimuth Adjustment to Raw Data

Solar positioning checks during MAP showed consistent radar pointing. Hard target checks show that absolute azimuth was in error by 1.25°. An adjustment to the raw data is justified, and will be incorporated into all datasets generated after the end of the project. See the document: S-Pol Range and Azimuth Checks.

Bad Cross-Pol Parameters with Clutter Filter Use: AH, CH, AV, CV

From the beginning of September, through 11 October 1999, the full matrix cross-pol parameters AH, CH, AV, CV were improperly computed in regions where the clutter filter was switched on. While the problem was noted early in the experiment (by 6 September), it took several weeks for full diagnosis, correction, and installation of the fix.

Full Matrix variables for the early part of the experiment are uncorrectable and unrecoverable. At S-Pol during MAP, the clutter filter was invoked for ranges of 0 to 80 km, and at all azimuths and elevations.

Time Glitches in Real-time Data Stream

The date and time info on the datastream provided to the real-time systems in the field is occasionally flakey. There are two sources of time problems:

GPS clock rollover issues and the code implemented to correct these
a likely fault in the beam-by-beam FIFO buffer used to generate DORADE beams for real-time Zebra display (does not affect post-field processed data)

Specific symptoms include:

Bad date (but ok time) on transition between 12/13 Oct 1999 (a GPS correction code problem)
Time glitches, every n seconds, where n appears constant for a few hours, but varies over a day. n may vary from about 12 to 50 seconds (this is likely a FIFO buffer problem)
For data of early September, there was a single-beam time glitch every 12 seconds, where the time for that single beam was wrong by about 6.5 seconds.
This problem carries through to the DORADE tapes created in the field for Prof. Houze. These DORADE tapes were generated from the real-time Zebra DORADE files.

Many applications will not be greatly affected by the kind of time glitch reported here. For example, the ATD radar display software SOLO uses azimuth and elevation information to plot beam positions for PPIs and RHIs; it is only when SOLO displays data in vertical pointing mode (time series mode) that the time glitch becomes evident. Use of the distribution dataset should eliminate this problem.

Polarization Feed Change-Out

The S-Pol antenna feedhorn was swapped-out on 5-Oct-99, with first use in operations on 12-Oct-99. The new feed had been ordered after the Brazillian experiment (Jan/Feb 1999), out of a concern that the horizontal and vertical polarizations from the feed were not truely orthogonal.

The lack of transmit signal orthogonality resulted in loss of polarization isolation, limiting S-Pol's measurement of cross-polar quantities (this condition seems to have particularly affected LDR).

It was also noted in S-Pol antenna patterns that there was a "hot spot" about two degrees off-center and along one of the feedhorn support struts (at 45°). Such a hot spot is likely the cause of problems noted with negative Zdr values just below bright band layers. It was hoped that a new feed would also correct some of this type of Zdr artifact.

Evaluation of the new S-pol feed is ongoing. Initial results are encouraging.

Full and Partial Beam Blockage

An obvious problem, given the hills and mountains near the S-Pol site. This problem is worth a reminder, however, since use of the clutter filter at S-Pol removes some of the more visible effects of this blockage.

See the report section on blockage.

Zdr Bias Adjustment

NOTE: Results are preliminary; analysis is continuing (09-Feb-2000)

From analysis of vertical pointing data, it is known that there is a bias in Zdr for MAP. This bias varies over the course of the field project, and is affected by the feedhorn change-out, variations in system receiver response (a very small variation, but enough to affect Zdr, but not significant for reflectivity), and likely some unknown factors.

It is necessary to correct Zdr for the system bias. Bias corrections are as large as .4 dB. The varying bias is correctable, through use of periodic vertical pointing data and monitoring of system test pulse values. A full report is available. Corrections will be applied during creation of the S-Pol MAP distribution data set.

Inaccurate Clutter Filter Status

Compiled in the lists of radar scans is information on clutter filter status. This piece of information was considered to apply to a full scan volume (a series of elevation tilts, for example). However, the filter status word in the scan listings does not flag those cases where the clutter filter may have been manually altered during a sub-part of the volume, or a sub-part of a scan.

Manual switching of the clutter filter is apparent in image data for scans on 4-Nov-1999. See images in this summary near 01:25, and 20:46 to 24:00. Review the clutter near 280° and 70 km range. Other examples of clutter filter switching exist (for this day and some others). Clutter filter usage substantially affects several parameters, and a user must take the filter status into account in order to produce consistent analyses.

--- Credits --- / NCAR Atmospheric Technology Division
Created: 18-Dec-1999
Last modified: Sun 19 Dec 13:55:42 MST 1999