As of 6 October 2015 15:18 UTC and continuing to present day, there has
been recurrence of the previously observed anomalous drop in the echo power
signal in the receiver of approximately 20 dB. Both noise and signal power
dropped but not by equal amounts. The Signal to Noise Ratio (SNR) dropped by
10 dB. Corresponding with this drop in power was an increase of approximately
0.20 mA in current in the Scatterometer Electronics Subsystem (SES). Due to
this anomaly the RapidScat Science Data Systems Team was initially unable to
process the telemetry data to backscatter (Level 1B, 2A) or wind vectors
(Level 2B) corresponding the following orbits: 5881, 5882, 5883, 5884, 5885,
5886, 5888, 5891, 5892, 5893, and 5894. Data from all other orbits beginning
with 5887 were processed and passed the standard quality assurance checks.
Throughout this time the operations team has continued to send commands to
the instrument to aid in diagnosing the anomaly and to aid in recovering data
production. The precise cause of the “reduced echo power” anomaly
has not yet been confirmed, but it has been determined to be an issue that is
outside of the control of ground operators.
As was the case with the previous “reduced echo power” anomaly,
the ground data processing is now able to process science quality data and
retrieve valid wind vector data with some degradation in performance at low
wind speeds.
There are several changes in the data going forward as a result of this
recurring anomaly, which are identical to the changes implemented in the
first anomaly. The following changes affect all orbits forward from 5887
until such time as the instrument switches back to its nominal high SNR
state:
- Wind direction (and to a lesser extent speed) accuracy are degraded for low winds. Winds above 6 m/s are not significantly affected but statistics for 3-6 m/s are noticeably worse especially in the middle of the swath.
- Rain flagging is now unavailable for the outer (single beam portion) of the swath. The IMUDH rain flag that the JPL wind retrieval has been using is corrupted because it uses the RapidScat Brightness temperature. With a 10 dB loss in SNR, brightness temperatures are no longer usable. So we are switching rain flags to use the Impact Quantity Rain flag which has been previously utilized to determine when to correct wind speeds for rain. The new rain flag does not use brightness temperature and is not available in the outer swath. The statistics of the rain impact quantity have not changed significantly as a result of this anomaly.
- The improved rain speed correction for tropical cyclones (version 1.1) has now been turned off because it relies on brightness temperature. We are going back to the original rain correction used in the Version 1.0 L2B data processing. It should be noted that the difference between the version 1 and version 1.1 rain corrections only impacts approximately 0.4% of the data values corresponding to the corrected wind speeds, thus the overall impact in that regard should be minimal. The chief impact is that rain corrected speeds near the center of tropical cyclones (or in other place with high rain rates and wind speeds) will be biased low. For users interested in this portion of the data the uncorrected wind speeds are preferable to the version 1.0 corrected speeds.
As a result of the anomaly, data flow of the L2B science data has not stopped
but has continued with the nominal 7-day latency.
The RapidScat project’s latest evaluation of the current RapidScat wind fields remains positive. The accuracy of the winds was similar to that of pre-anomaly data except for low wind speed regions. As a result, they have proceeded with a provisional release of the “reduced echo power anomaly” data to enable expedient access of new RapidScat to scientists with minimal gaps in the time series. The data during this anomaly are therefore provided for evaluation purposes and to aid in the calibration and validation of the data. However, they would like to stress that calibration activities are continuing, and they are still in the process of addressing concerns with the rain flagging and refining the calibration. They are also monitoring the data to make sure it is stable and that no anomalous trends are observed. Users should be aware, that the data will likely be reprocessed in the near future. The RapidScat project expects to complete a full reprocessing of the RapidScat wind data by 28 Feb 2016, but may reprocess earlier if more substantial impact to science data quality is observed in the current version.
Users should be mindful that this provisional data release also applies to
the data from the previous “reduced echo power anomaly” data as
referenced here:
http://podaac.jpl.nasa.gov/announcements/2015-09-02_RapidScat_Reduced_Echo_Power_Anomaly_and_L2B_Data_Disruption.
The data in which the SNR was temporarily restored to nominal levels is
also released provisionally but is processed in an identical manner to the
nominal high-SNR data before the occurrence of the first anomaly, as
referenced here:
http://podaac.jpl.nasa.gov/announcements/2015-09-24_RapidScat_Reduced_Echo_Power_Restored
Concerned data users who are not currently registered with the PO.DAAC email list are encouraged to register to receive timely data announcement updates via email by contacting podaac@podaac.jpl.nasa.gov.
PO.DAAC User Services