Start Date | Start Time | End Date | End Time | Data Quality Metric |
---|---|---|---|---|
03/01/1994 | 0005 | 04/04/1996 | 2359 | Suspect |
Subject: | SGP/MWR/B1/B4/B5 - Reprocess: Error in MWR calibration |
DataStreams: | sgp5mwravgB4.c1, sgp5mwravgB5.c1, sgp5mwravgB1.c1 |
Description: | The effect of this error is small. At most, it results in a bias of about -0.015 cm in precipitable water vapor and -0.015 mm in liquid water path during clear sky conditions. The error is largest when the brightness temperatures are small (i.e. clear skies and low PWV). The error results from failing to correctly account for the effect of the Teflon window covering the radiometer mirror. Although the contribution of the window is subtracted when the tip curve data are reduced to determine the true zenith brightness temperature, it is not added back in when the zenith brightness temperature is used to calibrate the noise diode. This would still not be a problem if the contribution of the window where not subtracted (again) during zenith line-of-sight (LOS) operations. But it is and the net effect is to subtract the contribution of the window twice. The calibrations ('Noise Injection Temperatures') are off by a factor of 1.00164 and 1.00217 for the 23.8 and 31.4 GHz frequencies, respectively. The magnitude of the error is equal to the emissivity of the window multiplied by the difference between the brightness temperature and the temperature of the window. The latter is taken to be equal to the temperature of the internal blackbody target (which is about 10 deg C above ambient.) The emissivity of the window is 0.00164 at 23.8 GHz and 0.00217 at 31.4 GHz. For a reference temperature of 292.6 K and brightness temperatures of 32.3 and 20.8 K at 23.8 and 31.4 GHz respectively, this amounts to errors of -0.43 and -0.59 K at the respective frequencies. The average PWV for this date (5 April 1995) was 1.4 cm. At higher levels of PWV and for cloudy conditions, the brightness temperatures are higher and so the error is smaller. I will adjust the calibrations of all SGP radiometers to account for this problem by the end of tomorrow (4 April 1996). |
Suggestions: | |
Measurements: | sgp5mwravgB4.c1: more
sgp5mwravgB5.c1: more sgp5mwravgB1.c1: more |