WMOD00477 - Laserfiche WebLink

Home Browse Search

MARCH 1987 MARK HEGGLI, ROBERT M. RAUBER AND J. B. SNIDER 205 1 The field evaluation was performed in two phases. The first, carried out in the SCPP, compared radio- metric measurements of water vapor with precipitable water vapor measurements made by a collocated ra- winsonde unit. The second phase was conducted during COSE. During this experiment two radiometers were brought together for an intercomparison test. A limited number of rawinsondes were also launched during routine radiometer calibrations to compare vapor measurements. A third phase of the experiment, not reported here, attempted to compare radiometric liquid measure- ments with in situ measurements by aircraft. To date, this later experiment has met with limited success, pri- marily because variations in cloud integrated liquid occur on time scales shorter than that required for an aircraft to descend through the cloud (e.g., Heggli and Reynolds, 1985; Rauber et al., 1986). 2. Instrumentation a. Dual channel microwave radiometer The dual channel microwave radiometer is a ground- based passive instrument which determines the bright- ness temperatures emitted by water vapor and cloud liquid water. The quantity of vapor and liquid is de- termined by the magnitude of the brightness temper- ature at specific microwave frequencies. The principles of microwave thermal emission from Rayleigh atten- utating clouds have been described by Westwater (1972). The design and operation of the dual-channel ground based microwave radiometer have been pre- sented by Hogg et aI., (1983) and Guiraud et aI., (1979). The SCPP radiometer was designed by NOAA and manufactured by the Hughes Corporation. The radiometer operates at two frequencies: 20.6 GHz, sensitive primarily to water vapor, and 31.65 GHz, sensitive primarily to liquid water. The antenna beam width is 2.50. Path-integrated amounts of water vapor and liquid water, expressed as depth in centi- meters and millimeters, respectively, are calculated from the radiometer measurements of brightness tem- perature using statistical retrieval algorithms (Hogg et aI., 1983). The coefficients in these algorithms are de- rived from a set of radiosonde data appropriate to the area in which observations are made. The coefficients include the effects of the small absorption by oxygen at each operative frequency. For nonprecipitating clouds or clouds with relatively low amounts ofliquid . water (i.e., < 1.5 mm zenith-equivalent), the retrieval coefficients are self-correcting and limit the algorithm induced measurement uncertainty to less than 5% (Westwater and Guiraud, 1980). This measurement uncertainty can be further reduced through the use of adaptive coefficients, which are a function of the initial estimates of the amounts of vapor and liquid present during the observation. However, due to the low amounts ofliquid water present in the clouds reported here, adaptive coefficients were not employed. Sources of error in radiometric measurements of liquid and vapor are associated with estimation of the mean radiating temperature of the emitter, the deter- mination of brightness temperature, uncertainties in dry attenuation, and uncertainty in the vapor and water attenuation coefficients (Westwater, 1978). Westwater estimated that with these uncertainties in the absorp- tion calculation, liquid retrieval accuracies better than 15% can be achieved for a wide range of liquid water contents. Rain or mixed phase precipitation also may obscure the measurement of cloud liquid. False signals can also be caused by melting hydrometers and/or a water coated reflector. Periods when melting occurred or the reflector was wet were removed from the dataset prior to this analysis. Ice crystals, on the other hand, have no discernible effect on the emissions by water vapor and liquid, whether it be falling or resting on the reflector. During SCPP, the USBR radiometer was located in the high alpine environment of the central Sierra Ne- vada at Kingvale, California, elevation 1857 m (all heights are relative to mean sea level). This radiometer was operated during the winter months of December 1984 through March 1985. Rawinsonde observations were taken concurrently during storms at Kingvale. The USBR and NOAA radiometers were collocated in COSE. The elevation of the radiometers was 2050 m. The two radiometer systems operated indepen- dently. However, the mean radiating temperatures and atmospheric attentuation coefficients used to retrieve vapor and liquid were identical. Retrieval coefficients were calculated from a sample of 352 soundings re- leased in previous COSE experiments from Craig, Col- orado. b. Rawinsondes Upper air observations were made during the SCPP and COSE. The Sierra Cooperative Pilot Project launches were made at Kingvale, where the radiometer was operated. The Colorado Orographic Seeding Ex- periment launched rawinsondes - 8 km from the col- located radiometers. TheVIZ Acculok rawinsonde was used at both field sites. The VIZ package is used in upper air measure- ments in more than 20 countries, and also by the Na- tional Weather Service. The rawinsondes were factory calibrated by VIZ at a temperature of 300C and 33% relative humidity, which is considered to be within the optimal calibration range. The humidity sensors were the most current sensor at the time. Relative humidity attained from these sensors at saturation was slightly lower than previous models. This is discussed by Shaffer (1982) and Nordahl (1982). Under these baseline con- ditions the relative humidity was expected to be deter- mined within 3 over the range from 50 to 97% relative