ARM/GCIP NESOB-97 DOE ARM SGP MWR Vertical Profiles Data 1.0 General Description The Department of Energy (DOE) Atmospheric Radiation Measurement (ARM) Southern Great Plains (SGP) Microwave Water Radiometer (MWR) Vertical Profiles is one of various data sets developed for the ARM/GCIP (Global Energy and Water Cycle Experiment [GEWEX] Continental-scale International Project) 1997 Near-Surface Observation (NESOB-97) Data Set. This data set contains vertical profiles of temperature and water vapor density every hour taken by the MWR instrument located at the Central Facility. This data set covers the period from 1 April 1997 through 31 March 1998. No additional quality control was performed by the University Corporation for Atmospheric Research/Joint Office for Science Support (JOSS). 2.0 Processing Done at UCAR/JOSS 2.1 Format Conversion The data arrived at UCAR/JOSS in netCDF formatted daily files. UCAR/JOSS converted the netCDF files to JOSS Cross Chain LORAN Atmospheric Sounding System (CLASS) Format (JCF). JCF is a version of the National Center for Atmospheric Research (NCAR) CLASS format and is an ASCII format consisting of 15 header records for each sounding followed by the data records with associated Quality Control (QC) information. Only selected parameters from the original data set are available within the JCF files (height, pressure, temperature, and water vapor density. Additional parameters within the orignial NetCDF files but not carried over into the JCF files include: Retrieved cloud liquid water content profile, Retrieved columnar water vapor, Retrieved columnar liquid water, Columnar water vapor, Columnar liquid water, and Average cloud base height used in the retrieval. All data at altitudes greater than 3000 m were removed from the data files. 2.2 Derivation of Relative Humidity The Relative Humidity (RH) was derived from the provided variables of temperature and water vapor density. The saturation vapor pressure was derived utilizing a form of the Clausius-Clapeyron (Wallace and Hobbs 1977). The vapor pressure was derived utlizing Wallace and Hobbs (1977) equation (2.12). The RH is 100 times the vapor pressure divided by the saturation vapor pressure. 2.3 Derivation of Dewpoint Temperature The Dewpoint Temperature was derived from the provided variable of temperature and the derived variable of RH utilizing the equations of Bolton (1980). 2.4 Detailed Format Description 2.4.1 Header Records The header records (15 total records) contain data type, project ID, site ID, site type/ID, profile location, profile time, and comments. The first five header lines contain information identifying the profile, and have a rigidly defined form. The following seven header lines are used for auxiliary information and comments about the profile, and vary from data set to data set. The last 3 header records contain header information for the data columns. Line 13 holds the field names, line 14 the field units, and line 15 contains dashes ('-' characters) delineating the extent of the field. The five standard header lines are as follows: Line Label (padded to 35 char) Contents 1 Data Type: Description of type and resolution of data. 2 Project ID: ID of weather project. 3 Profile Site Type/Site ID: Description of site. 4 Profile Location (lon,lat,alt): Position of site, in format described below. 5 UTC Profile Time (y,m,d,h,m,s): Time of profile, in format: yyyy, mm, dd, hh:mm:ss The release location is given as: lon (deg min), lat (deg min), lon (dec. deg), lat (dec. deg), alt (m) Longitude in deg min is in the format: ddd mm.mm'W where ddd is the number of degrees from True North (with leading zeros if necessary), mm.mm is the decimal number of minutes, and W represents W or E for west or east longitude, respectively. Latitude has the same format as longitude, except there are only two digits for degrees and N or S for north/south latitude. The decimal equivalent of longitude and latitude and station elevation follow. The seven non-standard header lines may contain any label and contents. The label is padded to 35 characters to match the standard header lines. 2.4.2 Data Records The data records each contain time (all missing for this data set), pressure, temperature, dewpoint, relative humidity, U and V wind components (all missing for this data set), wind speed and direction (all missing for this data set), ascent rate (all missing for this data set), balloon position data (all missing for this data set), altitude, and quality control flags (see the QC code description). Each data line contains 21 fields, separated by spaces, with a total width of 130 characters. The data are right-justified within the fields. All fields have one decimal place of precision, with the exception of latitude and longitude, which have three decimal places of precision. The contents and sizes of the 21 fields that appear in each data record are as follows: Field Format No. Width Parameter Units Missing Value ----------------------------------------------------------------------- 1 6 F6.1 Time Seconds 9999.0 2 6 F6.1 Pressure Millibars 9999.0 3 5 F5.1 Dry-bulb Temperature Degrees C 999.0 4 5 F5.1 Dew Point Temperature Degrees C 999.0 5 5 F5.1 Relative Humidity Percent 999.0 6 6 F6.1 U Wind Component Meters / Second 9999.0 7 6 F6.1 V Wind Component Meters / Second 9999.0 8 5 F5.1 Wind Speed Meters / Second 999.0 9 5 F5.1 Wind Direction Degrees 999.0 10 5 F5.1 Ascension Rate Meters / Second 999.0 11 8 F8.3 Longitude Degrees 9999.0 12 7 F7.3 Latitude Degrees 999.0 13 5 F5.1 Water Vapor Density g / m^3 999.0 14 5 F5.1 Azimuth Degrees 999.0 15 7 F7.1 Altitude Meters 99999.0 16 4 F4.1 QC for Pressure Code (see below) 99.0 17 4 F4.1 QC for Temperature Code (see below) 99.0 18 4 F4.1 QC for Humidity Code (see below) 99.0 19 4 F4.1 QC for U Component Code (see below) 99.0 20 4 F4.1 QC for V Component Code (see below) 99.0 21 4 F4.1 QC for Ascension Rate Code (see below) 99.0 Fields 16 through 21 contain the Quality Control information derived at the UCAR Joint Office for Science Support (UCAR/JOSS). Any QC information from the original data is replaced by the following JOSS codes: Code Description 99.0 Unchecked (QC information is "missing.") ("UNCHECKED") 1.0 Checked, datum seems physically reasonable. ("GOOD") 2.0 Checked, datum seems questionable on physical basis.("MAYBE") 3.0 Checked, datum seems to be in error. ("BAD") 4.0 Checked, datum is interpolated. ("ESTIMATED") 9.0 Checked, datum was missing in original file. ("MISSING") For this data set all QC flags are set to either unchecked (99.0) or missing (9.0). 2.4.3 Sample Data The following is a sample record of ARM/GCIP NESOB-97 MWR Vertical Profile data in JOSS CLASS format. The data portion is much longer than 80 characters and, therefore, wraps around to a second line. See section 2.2 for an exact format specification Data Type: Microwave Radiometer Profile Project ID: ARM/GCIP NESOB-97 Profile Site Type/Site ID: ARM Central Facility Profile Location (lon,lat,alt): 97 29.10'W, 36 36.36'N, -97.48, 36.61, 316.0 UTC Profile Time (y,m,d,h,m,s): 1997, 05, 21, 12:00:20 System Operator/Comments: RH derived from Rho and temp;DewPt derived from Rh and temp Additional comments: Field 13 contains Water Vapor Density (Rho) in g/m^3 Input File: sgpmwrprofC1.c1.970521.000020.cdf / / / Nominal Profile Time (y,m,d,h,m,s):1997, 05, 21, 12:00:20 Time Press Temp Dewpt RH Ucmp Vcmp spd dir Wcmp Lon Lat Rho Azim Alt Qp Qt Qrh Qu Qv QdZ sec mb C C % m/s m/s m/s deg m/s deg deg g/m^3 deg m code code code code code code ------ ------ ----- ----- ----- ------ ------ ----- ----- ----- -------- ------- ----- ----- ------- ---- ---- ---- ---- ---- ---- 9999.0 984.3 10.0 8.7 91.4 9999.0 9999.0 999.0 999.0 999.0 9999.000 999.000 8.6 999.0 316.0 99.0 99.0 99.0 9.0 9.0 9.0 9999.0 955.1 10.0 7.9 86.6 9999.0 9999.0 999.0 999.0 999.0 9999.000 999.000 8.1 999.0 566.0 99.0 99.0 99.0 9.0 9.0 9.0 9999.0 926.7 9.4 6.9 84.3 9999.0 9999.0 999.0 999.0 999.0 9999.000 999.000 7.6 999.0 816.0 99.0 99.0 99.0 9.0 9.0 9.0 2.4.4 References Bolton, D., 1980: The Computation of Equivalent Potential Temperature. Mon. Wea. Rev., 108, 171-180. Wallace, J. M. and P. V. Hobbs, 1977: Atmospheric Science: An Introductory Survey. Academic Press. 467 pp. 3.0 Note of this Documentation File The remainder of this document comes from the ARM program. An up-to-date version of this document (including figures mentioned within the text) is available on the WWW at: http://www.arm.gov/docs/instruments/static/mwr.htmll 4.0 General Purpose The Microwave Water Radiometer (MWR) provides time-series measurements of column-integrated amounts of water vapor and liquid water. The instrument itself is essentially a sensitive microwave receiver. That is, it is tuned to measure the microwave emissions of the vapor and liquid water molecules in the atmosphere at specific frequencies. 5.0 Primary Quantities Measured with System The MWR receives microwave radiation from the sky at 23.8 GHz and 31.4 GHz. These two frequencies allow simultaneous determination of water vapor and liquid water burdens along a selected path. Atmospheric water vapor observations are made at the "hinge point" of the emission line where the vapor emission does not change with altitude (pressure). Cloud liquid in the atmosphere emits in a continuum that increases with frequency, dominating the 31.4 GHz observation, whereas water vapor dominates the 23.8 GHz channel. The water vapor and liquid water signals can, therefore, be separated by observing at these two frequencies. 6.0 Detailed Description 6.1 List of Components Radiometrics WVR-1100 Radiometer Radiometrics Rain Sensor / Dew Blower Heater System Radiometrics instrument tripod and tribrach leveling base IBM ThinkPad or Consultronics LineBacker computer AC power cable serial communications cable 6.2 Description of System Configuration and Measurement Methods The water vapor radiometer receiver is composed of a gaussian optical antenna, a noise diode injection device, a dual junction isolator, a balanced mixer, an IF amplifier, a detector/video amplifier, and two Gunn diode oscillators. The receiver accepts input power from the antenna and supplies a voltage proportional to antenna temperature (plus antenna noise) via a square law detector to the radiometer voltage-to-frequency converter on the microprocessor (digital) board. Receiver frequency selection is accomplished by alternately powering the 23.8 and 31.4 GHz Gunn diode local oscillators. Brightness temperature calibration is provided by a noise source injected at the input (added to antenna temperature). The Gunn diode oscillators and noise source are powered by the radiometer analog board and controlled by the radiometer digital board. The MWR uses low noise, low power IF amplifiers. The receiver is linear with antenna power over a range of the sky and calibration observables. The receiver is thermally stabilized to ensure stability of the mixer and the noise diode and Gunn diode output and frequency. The sky brightness temperature is measured in the following manner. The small-angle receiving cone of the gaussian-lensed microwave antenna is steered with a rotating flat mirror. Both the 23.8 GHz and the 31.4 GHz waveband signals are transmitted by through a single waveguide into an isolator and then into the mixer section. Output from one of two Gunn diodes is injected into the local oscillator port of the mixer. The resultant IF signal is amplified, filtered to yield a 400 MHz wide dual sideband signal, detected, amplified again, and converted by a voltage-to-frequency converter. Zero crossings of this signal are counted, yielding the raw data in counts. Counts are then converted to brightness temperature through algorithms in the FORTRAN program. Water vapor, liquid water, and phase path delay are calculated using site-specific retrieval coefficients read from the configuration file. 6.3 Assessment of System Uncertainties for Primary Quantities Measured Measurement Uncertainty --------------------------- Sky 0.018 K Blackbody 0.12 K Blackbody+noise ~0.15 K Gain reference ~0.02 K Receiver gain ~0.09 K Receiver offset 0.035 K 6.4 Description of Observational Specifications Sample time user selectable -------------------------------------------------------------- Accuracy 0.3 K Resolution 0.25 K Radiometric range 0 to 700 K Operating range -20 to +50°C Power requirements 120W maximum Voltage requirements 90 to 130 or 180 to 260 VAC, 50 to 440 Hz Output ASCII data files to laptop computer via RS-232 Dimensions 50 x 28 x 76 cm Weight 17 kg Angular coverage all sky Pointing slew rate 3 deg/second, azimuth; >90 deg/second, elevation 7.0 Theory of Operations The instrument itself is essentially a sensitive microwave receiver. That is, it is tuned to measure the microwave emissions of the vapor and liquid water molecules in the atmosphere at specific frequencies. For a specific frequency, n, the amount of microwave radiation observed by a radiometer at the earth's surface looking directly upward can be expressed as I = Ic e-t(0,inf) + int(0,inf){B[T(z)] rke-t(0,z) dz}. (1) The first term represents the amount of cosmic (i.e. extraterrestrial) radiation entering at the top of the atmosphere Ic that reaches the radiometer. The exponential decay factor accounts for attenuation of the cosmic radiation by the intervening atmosphere; t is the optical thickness: t(0,z) = int(0,z){r(z) k(z) dz}, (2) where r is the density [mass per volume] or [number per volume] and k is the extinction coefficient [area per mass] or [area per number]. It is highly dependent on frequency. (Note that extinction is the sum of absorption plus scattering; however, because scattering is negligible in the microwave region of the electro-magnetic spectrum - except during heavy rain - k can be taken as the absorption coefficient alone.) The physical significance of t is that it represents an "effective thickness" of the atmosphere for a particular frequency: t will be large (and the attenuation e-t great) when either z, r or k is large. Put another way, if r and k are large enough that a very small value of z will still cause e-t =1, then the region is said to be "optically thick" - one cannot "see" very far into it. On the other hand, if r and k are sufficiently small, a very large value of z will be required to produce e-t =1 and the region is said to be "optically thin" - one can "see" a large distance at this frequency. The second term in Eq. (1) represents the sum of the contributions from the atmosphere along the line-of-sight (i.e., the path). B[T(z)] is the Planck function which describes the blackbody emission from the molecules at height z (which are at a temperature T(z)). The product rk is the amount of blackbody radiation that is emitted (i.e., not re-absorbed) by the molecules in the layer. The factor e-t accounts for the attenuation by the atmosphere between the source molecules and the microwave radiometer antenna. In the microwave region, the Planck function may be expressed as B(T) = 2KTc/l4 (3) where K is Boltzmann's constant, c is the speed of light and l is the wavelength of the radiation. We can rearrange this expression to define the equivalent blackbody brightness temperature: TB = I l4/2Kc. If Eq. (1) is divided through by 2Kc/l4 then TB = Tc e-t(0,inf) + int(0,inf){T(z) r(z) k(z) e-t(0,z) dz}, (4) where Tc = 2.75 kelvins. To actually calculate TB, the atmosphere is divided into a number of layers N which are considered isothermal: TB = Tc e-t(0,inf) + sum(i,N) Ti{int(zi,zi+1) r(z) k(z) e-t(0,z) dz}. (5) If N=1 (i.e., the entire atmosphere is taken to be isothermal), then TB = Tc e-t(0,inf) + TMR int(0,inf){r(z) k(z) e-t(0,z) dz} = Tc e-t + TMR[1 - e-t] (6) where t = t(0,inf), the total zenith absorption, and TMR is the mean radiating temperature of the atmosphere at this frequency. (In general, there is a different TMR for each frequency). Equation (6) is the "brightness temperature equation;" it is used to relate the observed emission TB to the absorption: t = ln (TMR - Tc/TMR - TB). V and L are then derived by relating them to the microwave absorption t: t = tdry + tvap + tliq = tdry + kV V + kL L. tdry is due to the (approximately constant) emission of O2 molecules; kV, and kL are calculated from climatology with sufficient accuracy to determine t. Observations of TB at a vapor sensitive frequency (23.8 GHz) and a liquid sensitive frequency (31.4 GHz) yield two linear equations which can be solved for the two unknowns, V and L. Note that the vapor sensitive frequency is chosen such that kV is not dependent on pressure - and thus not dependent on height. The unknowns V and L can be represented in terms of the optical depths, or directly in terms of the observed brightness temperatures: V = ao + a1(TB,v - TB,v) + a2(TB,l - TB,l) + a3(P - P) + a4(T - T) + a5(RH - RH) where the variables in italics are climatological mean values and the ai are determined through a linear regression (i.e., they are the "regression coefficients"). Note that ao = V. A similar expression can be written for the total liquid water, L. 8.0 ARM Data Quality Reports Nor reports issued. 9.0 Assessment of Instrument Calibration and Maintenance Procedures from work for site scientist Preventative Maintenance Procedure Summaries for the MWR at the Southern Great Plains (SGP) Site 10.0 Calibration Theory 10.1 Background The electrical output of the radiometer V in volts or digital counts is linearly related to the equivalent microwave brightness temperature TB: TB = Tref + (V - Vref)/G. Here Tref is the reference temperature of the internal blackbody target and Vref is the corresponding radiometer signal. G is the gain of the system in units of volts (or digital counts) per kelvin. One way to determine the gain G is through the use of tip-curves. The ARM radiometers use a reverse-biased diode which injects broadband microwave energy ("noise") directly into the waveguide when it is switched on, causing the signal output of the radiometer to increase by Vnd. The output of the noise diode is determined during the tip-curve procedure by pointing the elevation mirror at the reference blackbody target and measuring the radiometer output with the noise diode on and off: Vnd = Vref+nd - Vref. This "noise diode" output is calibrated from the tip-curve-derived gain to yield the noise injection temperature: Tnd = Vnd/G which is useful because it is nearly insensitive to changes in ambient temperature even though Vnd and G are strong functions of temperature. Because the instantaneous output of the noise diode represents a random process having a Gaussian distribution, the results of many tip-curves (>500) are used to compute G and Vnd and thus Tnd. Once calibrated, the system gain can be determined using the noise diode by viewing the blackbody target and measuring the change in the radiometer output due to switching on the noise diode, G = (Vref+nd - Vref)/Tnd. 10.2 Temperature Dependence of the Calibration The gain is very sensitive to the temperature of the radiometer components (i.e., the feed horn, waveguides, mixer, local oscillators, etc.). As a result, the stability of the gain is directly related to the thermal stability of the instrument. The microwave hardware in the ARM radiometers is mounted on a thick aluminum plate in an insulated enclosure and thermally stabilized to +/-0.25 K. Even so, the slight variations in the gain (which is the slope of the calibration curve) that arise from these slight temperature variations must be accounted for because the sky brightness temperatures typically range from 10-80 K whereas the blackbody reference temperature is at ambient (~300 K) so small errors in the gain will result in significant errors in the brightness temperature. Consequently, the tip-curve data are used to derive a linear relationship between the noise injection temperature and the ambient temperature (represented by the temperature of the blackbody target, Tref): Tnd = Tnd (290 K) + a(Tref - 290) where a is the temperature coefficient which typically ranges from -0.08 to 0.08 (K/K). 10.3 Tip Curves To perform a tip-curve calibration, measurements of optical thickness t are required along paths at various elevation angles, a. If the atmosphere can be assumed to be horizontally homogeneous, then the optical thickness along a path at an angle a above the horizon is directly proportional to the optical thickness at the zenith: t(m) = mt(1) where m = 1/sin(a) is the "air mass" (e.g., a line-of-sight path inclined at an angle of 30 degrees above the horizon or m = 2 traverses twice the mass of air as at 90 degrees or m = 1). The procedure is as follows: 1.Use the existing calibration as an approximation. 2.Use the approximate calibration and measure brightness temperatures at elevation angles corresponding to several different air masses. For ARM, the elevation angles are 19.5, 23.6, 30.0, 41.8, 90.0, 138.2, 150.0, 156.4, 160.5, and 90.0 which corresponds to m = 3.0, 2.5, 2.0, 1.5, 1.0, 1.5, 2.0, 2.5, 3.0, and 1.0. Angles on both sides of zenith are used to ensure horizontal homogeneity. 3.Relate these brightness temperatures to optical thickness: t(m) = ln [(TMR - Tc)/(TMR - TB(m))]. 4.Fit a straight line to the optical thickness as a function of air mass. (See Figure 1.) Since the absorption should be zero for m = 0 (no atmosphere), the intercept represents the error in the current calibration. If the correct brightness temperatures had been used the intercept would pass though the origin. Thus, the true zenith optical thickness t(1) is equal to the slope of the regression line. Note that the quality of the fit indicates the degree to which the atmosphere is horizontally homogeneous. In the presence of clouds, for example, the tip-curve calibration method is not valid because the absorption is not linearly related to air mass. For ARM, the regression must account for at least 99.8% of the observed variance (R2 = 0.998) to be considered valid. Angles on both sides of the zenith corresponding to the same air mass are used to assess horizontal homogeneity. 5.The true zenith optical thickness t1 is now used to compute the true zenith brightness temperature: TB(zenith) = Tc e-t1 + TMR (1 - e-t1). 6.Now the gain can be computed: G = (Vzenith - Vref)/(TB(zenith) - Tref). 11.0 Calibration History R is the minimum correlation coefficient of optical depth on airmass for a valid tip curve. RMS is the maximum root-mean-square of the brightness temperatures for a valid tip curve. N is the number of valid tip curves used to determine the calibration values. Tnd is the noise injection temperature adjusted to 290 K. tc is the temperature coefficient used to calculate Tnd at ambient temperature. dev is the mean deviation of the robust regression of Tnd on ambient temperature (the uncertainty in Tnd). S/N Tip Dates Date/Time R RMS N Tnd23 tc23 dev23 Tnd31 tc31 dev31 Implemented (K) (K/K) (K) (K) (K/K) (K) 004 06-17 May 94 0.998 650 163.13 +0.003 0.50 004 11-13 Apr 95 0.995 2456 162.75 -0.018 0.34 167.78 -0.013 0.53 004 09-11 May 95 0.995 1305 162.99 -0.003 0.35 167.13 +0.094 0.54 004 28-31 Jul 95 0.995 1824 162.65 +0.014 0.36 169.00 +0.013 0.57 004 22-25 Sep 95 0.995 1475 163.22 -0.016 0.21 168.71 -0.070 0.48 004 01-05 Apr 96 05 Apr 96/21 0.995 1581 340.20 -0.014 0.56 265.00 -0.028 0.28 004 09-13 Sep 96 13 Sep 96/01 0.990 5000 336.54 +0.020 0.86 257.10 -0.0006 0.27 004 09-30 Sep 96 0.990 9248 336.92 -0.005 0.92 256.78 +0.013 0.28 004 28-31 Jan 97 14 Mar 97/22 0.990 4357 335.67 -0.004 0.34 255.77 -0.005 0.23 004 06-09 Jun 97 0.990 2.0 2859 336.07 +0.022 0.73 256.53 -0.020 0.39 004 09-11 Sep 97 27 Oct 97/20 0.990 3.0 1725 334.29 +0.056 0.74 256.43 -0.035 0.42 004 04-07 Dec 97 17 Jan 98/23 0.990 2.0 3032 334.89 -0.032 0.31 256.19 -0.005 0.23 004 13-14 Feb 98 0.990 2.0 2508 335.14 -0.026 0.38 256.21 -0.003 0.29 004 21-24 Apr 98 19 May 98/00 0.990 2.5 3794 335.57 -0.026 0.54 256.66 -0.027 0.29 004 29-31 Jul 98 07 Aug 98/20 0.990 3.5 2968 334.39 +0.071 0.92 257.23 +0.003 0.48 010 30 Dec 93-3 Jan 94 0.995 1822 189.02 -0.048 0.20 183.43 +0.005 0.18 010 16-19 May 94 0.995 1492 190.26 -0.070 0.33 184.24 +0.016 0.18 010 28-29 Jul 94 0.995 678 191.40 -0.061 0.32 184.99 +0.025 0.16 010 22 Feb 95 0.995 51 190.51 -0.034 0.18 184.10 +0.134 0.10 010 28 Mar 95 0.995 641 190.78 -0.045 0.16 184.19 +0.024 0.14 010 09-11 May 95 0.995 641 191.32 -0.081 0.64 184.56 +0.037 0.57 010 28-31 Jul 95 0.995 79 191.00 -0.017 0.33 185.21 +0.027 0.16 010 21-25 Aug 95 0.995 1857 191.46 -0.052 0.30 185.26 +0.025 0.15 010 10-15 Feb 96 26 Feb 96/00 0.990 904 202.53 -0.056 0.30 190.39 +0.009 0.24 010 09-13 Sep 96 13 Sep 96/01 0.990 5000 202.75 -0.061 0.41 190.63 -0.019 0.21 010 09-30 Sep 96 0.990 8340 202.66 -0.056 0.36 190.52 -0.012 0.21 010 01-07 Oct 96 0.990 8238 202.94 -0.066 0.32 190.81 -0.016 0.26 010 28-31 Jan 97 14 Mar 97/20 0.990 2773 202.02 -0.042 0.23 189.73 -0.004 0.18 010 20-21 Mar 97 21 Mar 97/23 0.990 1088 205.13 -0.054 0.20 190.09 -0.004 0.16 010 06-09 Jun 97 0.990 3.0 492 205.04 -0.051 0.66 190.23 -0.004 0.41 010 23-26 Aug 97 27 Oct 97/20 0.991 2.6 68 205.78 -0.082 0.53 191.50 -0.052 0.29 010 26 Sep 97 0.990 3.0 33 191.84 +0.676 0.64 185.11 +0.294 0.33 010 26-30 Sep 97 03 Nov 97/22 0.990 3.0 375 205.38 -0.071 0.33 190.98 -0.026 0.19 010 04-07 Dec 97 17 Jan 98/23 0.990 2.0 287 204.71 -0.025 0.19 189.70 -0.008 0.16 010 13-14 Feb 98 20 Feb 98/20 0.990 2.0 196 205.45 +0.107 0.36 190.02 -0.099 0.30 010 21-24 Apr 98 0.990 2.5 60 204.94 -0.098 0.52 189.74 -0.024 0.31 010 01-04 May 98 05 May 98/? 0.995 2768 205.37 -0.074 0.39 189.41 -0.011 0.27 010 18-21 Jul 98 0.990 4.0 624 207.17 -0.141 0.67 190.99 -0.058 0.28 011 02-03 Apr 95 0.995 1219 199.66 +0.049 0.28 214.02 -0.003 0.19 011 09-11 May 95 0.995 424 200.38 +0.050 0.28 214.76 -0.031 0.23 011 28-31 Jul 95 0.995 2121 200.90 +0.036 0.44 216.01 -0.031 0.23 011 21-28 Aug 95 0.995 2259 201.27 +0.024 0.45 216.23 -0.027 0.19 011 22-31 Mar 96 05 Apr 96/00 0.995 2638 194.86 +0.063 0.42 215.67 -0.014 0.51 011 08-13 Aug 96 20 Aug 96/01 0.995 1450 197.34 +0.025 0.70 216.02 -0.042 0.32 011 09-13 Sep 96 13 Sep 96/01 0.990 1955 196.26 +0.057 0.51 215.03 +0.009 0.27 011 09-30 Sep 96 0.990 3174 196.01 +0.067 0.49 214.77 +0.019 0.24 011 28-31 Jan 97 14 Mar 97/21 0.990 788 195.82 +0.053 0.28 215.81 -0.001 0.25 011 06-09 Jun 97 06 Aug 97/01 0.990 3.0 2362 197.24 +0.034 0.60 217.97 -0.015 0.40 011 23-26 Aug 97 27 Oct 97/19 0.990 3.0 1336 197.57 +0.046 0.52 218.72 -0.012 0.30 011 04-07 Dec 97 0.990 2.0 1233 197.29 +0.042 0.32 218.34 -0.027 0.39 011 13-14 Feb 98 20 Feb 98/20 0.990 2.0 686 197.66 +0.064 0.33 219.08 -0.005 0.29 011 21-24 Apr 98 22 May 98/01 0.990 2.5 1315 193.12 +0.043 0.39 218.61 -0.009 0.24 011 18-20 Jul 98 06 Aug 98/15 0.990 4.0 1632 194.70 -0.014 0.91 220.31 -0.038 0.39 012 02-03 Apr 95 0.995 1164 180.35 -0.030 0.36 239.17 -0.003 0.29 012 01-30 Apr 95 0.998 1093 180.35 -0.029 0.33 239.16 0.000 0.29 012 09-11 May 95 0.995 900 180.77 -0.056 0.24 240.75 -0.036 0.23 012 01-31 May 95 0.998 738 180.72 -0.049 0.22 240.77 -0.038 0.20 012 28-31 Jul 95 0.995 614 180.09 +0.014 0.50 244.56 -0.099 0.42 012 01-31 Jul 95 0.998 362 181.48 -0.060 0.56 244.77 -0.110 0.48 012 21-28 Aug 95 0.995 1221 180.65 -0.032 0.31 243.48 -0.028 0.21 012 01-31 Aug 95 0.998 1719 180.64 -0.031 0.32 243.47 -0.026 0.25 012 25-26 Mar 96 0.995 748 178.76 -0.064 0.35 246.25 -0.038 0.36 012 25-26 Mar 96 05 Apr 96/00 0.995 1023 178.85 -0.055 0.22 246.36 -0.006 0.17 012 08-13 Aug 96 22 Aug 96/22 0.995 579 180.45 -0.036 0.41 244.63 +0.054 0.42 012 09-13 Sep 96 13 Sep 96/01 0.990 4929 179.89 -0.046 0.37 246.44 -0.037 0.28 012 28-30 Sep 96 0.990 3645 178.81 -0.040 0.25 244.53 -0.005 0.24 012 09-30 Sep 96 0.990 8764 178.77 +0.0045 0.46 244.54 +0.044 0.59 012 28-31 Jan 97 15 Feb 97/23 0.995 2202 168.28 -0.057 0.23 181.23 +0.025 0.20 012 14-15 Apr 97 18 Apr 97/19 0.995 428 168.45 -0.063 0.17 184.03 +0.022 0.14 012 06-09 Jun 97 06 Aug 97/01 0.990 3.0 2566 168.85 -0.071 0.33 186.15 -0.005 0.24 012 23-26 Aug 97 27 Oct 97/20 0.990 1.3 1229 169.97 -0.103 0.31 188.45 -0.029 0.19 012 01-03 Nov 97 17 Jan 98/23 0.990 3.0 3420 169.17 -0.082 0.24 188.76 +0.003 0.20 012 29-30 Jan 98 09 Feb 98/17 0.990 2.0 2340 218.64 -0.107 0.23 219.41 +0.007 0.18 015 15-31 Jul 95 0.995 11498 235.11 +0.009 0.57 242.35 +0.025 0.26 015 15 Jul-07 Aug 95 0.995 15206 235.07 +0.013 0.56 242.42 +0.027 0.31 015 09-13 Sep 96 13 Sep 96/01 0.990 1794 236.19 +0.025 0.42 243.62 +0.040 0.30 015 09-30 Sep 96 0.990 5369 236.50 +0.017 0.36 243.69 +0.035 0.24 015 12-13 Feb 98 25 Feb 98/18 0.990 1500 236.76 +0.000 243.50 +0.000 015 07-09 Nov 98 sent 12 Nov 98 0.990 1500 236.62 +0.010 244.10 +0.001 016 27 Mar-03 Apr 95 0.998 766 197.76 +0.006 0.64 174.99 +0.017 0.46 016 09-11 May 95 0.998 778 198.17 +0.017 0.32 175.16 0.000 0.22 016 15-16 Sep 96 sent 20 Sep 96 0.991 43 197.19 0.0 0.68 171.63 0.0 0.24 016 15-16 Sep 96 sent 11 Dec 96 0.995 6 199.00 0.0 0.20 172.27 0.0 0.09 016 15-16 Sep 96 sent 10 Jan 97 0.995 81 197.97 0.0 0.64 171.82 0.0 0.26 016 03-28 Mar 97 sent 23 Jun 97 0.990 5.0 14869 197.06 0.0 0.82 172.07 0.0 0.51 016 08-27 Jun 97 sent 20 Mar 98 0.990 6.0 15133 197.16 -0.038 0.55 172.69 -0.024 0.29 016 25-29 Apr 98 29 Apr 98/? 0.990 500 017 223.69 0.036 216.30 0.052 018 28 Jun-08 Jul 95 0.995 3603 268.83 -0.001 0.90 268.36 -0.041 0.52 018 01-31 Jul 95 0.998 3526 268.73 +0.008 0.73 268.18 -0.023 0.40 018 21-26 Aug 95 0.995 1385 270.23 -0.021 0.71 269.38 -0.030 0.39 018 12-18 Mar 96 05 Apr 96/01 0.996 3430 268.36 +0.049 0.54 269.42 -0.021 0.35 018 07-13 Aug 96 0.996 4862 271.00 -0.011 0.73 269.26 -0.077 0.57 018 09-13 Sep 96 13 Sep 96/01 0.990 4998 269.03 +0.050 0.62 266.54 -0.008 0.37 018 28-30 Sep 96 0.990 3837 268.42 +0.046 0.47 265.50 -0.016 0.36 018 09-30 Sep 96 0.990 8678 268.41 +0.072 0.57 265.45 +0.035 0.53 018 28-31 Jan 97 14 Mar 97/22 0.990 4449 267.85 +0.059 0.42 266.36 -0.001 0.37 018 06-09 Jun 97 06 Aug 97/01 0.990 2.0 2758 266.59 +0.035 0.71 264.42 -0.042 0.51 018 23-26 Aug 97 27 Oct 97/20 0.990 1.4 2032 267.38 +0.054 0.69 266.02 -0.067 0.46 018 04-06 Dec 97 0.990 1.5 2104 266.71 +0.052 0.45 266.58 -0.044 0.35 018 13-14 Feb 98 20 Feb 98/20 0.990 2.0 1955 266.57 +0.010 0.49 266.78 -0.078 0.41 018 01-04 May 98 07 May 98/15 0.990 1500 266.71 +0.031 266.64 -0.064 018 18-20 Jul 98 0.990 4.0 4301 265.95 +0.091 0.99 266.41 -0.027 0.45 020 09-13 Sep 96 13 Sep 96/01 0.990 4561 218.97 -0.012 0.37 285.55 -0.0013 0.31 020 28-30 Sep 96 0.990 3442 218.93 -0.011 0.24 286.44 +0.0007 0.25 020 09-30 Sep 96 0.990 8939 218.94 -0.011 0.32 286.47 -0.042 0.42 020 01-31 Oct 96 18 Dec 96/22 0.995 7815 218.68 0.0 0.26 288.91 0.0 0.33 020 sent 20 Nov 97 218.05 0.021 287.59 -0.003 020 08-13 Jul 98 15 Jul 98/21 0.990 1500 219.92 -0.050 281.18 -0.074 021 09-13 Sep 96 13 Sep 96/01 0.990 4941 351.74 -0.022 0.69 202.28 +0.016 0.21 021 28-30 Sep 96 0.990 3449 351.05 +0.007 0.49 201.56 +0.023 0.25 021 09-30 Sep 96 0.990 9406 351.19 +0.004 0.60 201.65 +0.044 0.30 021 01-15 May 97 17 May 97/01 0.990 0.5 7138 349.05 +0.030 0.39 199.50 +0.019 0.23 021 Oct 97 sent 6 Jan 98 349.49 0.000 200.25 0.000 12.0 Current Status and Locations Serial Property Installation Number Number Location Date Status ---------------------------------------------------------- 4 WD06605 SGP/BF4 92/06/12 operational 10 WD11023 SGP/CF 93/12/13 operational 11 WD14131 SGP/BF1 93/12/13 operational 12 WD14132 SGP/BF5 93/12/13 operational 15 WD14869 TWP/ARCS2 98/11/12 operational 16 WD13409 TWP/ARCS1 96/09/23 operational 17 WD13410 AIS/ARCS3 -- delivered 18 WD12906 SGP/BF6 95/06/27 operational 20 WD24769 NSA/Barrow 97/02/28 operational 21 WD24770 NSA/SHEBA 97/12/05 de-installed 13.0 Calculated Data 13.1 Value Added Procedures MWR PROF: retrievals of water vapor, liquid water, and temperature profiles from a suite of ground-based instruments. LSSONDE: radiosonde profiles, where the moisture profile is scaled to match the MWR's total precipitable water vapor. 13.2 Quality Measurements Experiments QME MWR PROF, comparisons of retrieved water vapor and temperature profiles from MWR PROF with radiosonde profiles. QME MWR/LBL, comparisons of observed versus calculated microwave radiance at two frequencies. QME MWR COL, comparisons of the MWR with an instrument performance model. 14.0 Examples of Data MWR Quick Plots, from the ARM Science Applications Group 15.0 Instrument Mentor notes on data quality control procedures Data quality control procedures for this system are mature. On a weekly basis, the instrument mentor produces and inspects plots of the precipitable water vapor (PWV) and liquid water path (LWP) versus time. The base level of LWP is evaluated for clear sky episodes and the PWV estimates are compared to those from the BBSS. DQRs are submitted when needed, and a summary report of data quality is sent periodically to the SGP site scientist team. 16.0 Explanation of Flags Applied During Data Ingest wet_window - Indicates when the moisture detector has activated the heater in the dew blower housing to accelerate evaporation of rain on the dielectric window. 0 = heater off 1 = heater on qcmin(SGP) - A bitwise summation of each field with a value below minimum threshold. "if (BITWISE_AND(qcfield, 2**(N - 1)) == 0) then the value of field N is above the stated threshold" qcmax(SGP) - A bitwise summation of each field with a value above maximum threshold. "if (BITWISE_AND(qcfield, 2**(N - 1)) == 0) then the value of field N is above the stated threshold" qc_info(TWP & NSA) - An information string of flags for each field. 0 = value of field is within stated thresholds 1 = value of field is outside stated thresholds The minimum and maximum thresholds are currently defined as follows: Field Field Index Name Units Min Max ------------------------------------ 1 tbsky23 K 2.73 100 2 tbsky31 K 2.73 100 3 vap cm 0.0 -- 4 liq cm -3*rms -- 5 ir_temp K 213 313 where rms is liquid_retrieval_rms_accuracy For any given field, if the field level attribute qcmethods does not have a min/max entry then no min/max test has been run for that field. Currently, vap and liq have no max entry. None of the fields have delta entries. 17.0 Frequently Asked Questions FAQs 1) What do the long runs of 'liq Max exceeded' alternating with 'liq OK' mean? Should we disregard the microwave data during these times? Why are these messages not completely coincident with the weather log reports of rain? The "maximum exceeded" alarm trips when the retrieved liquid water exceeds 1 cm. Such a value is not possible--it indicates a serious failure of the retrieval, most likely due to water standing on the instrument. There are two reasons why this may not be completely coincident with weather log reports of rain. First, the operators are only onsite from 8 am to 5 pm local time at the Central Facility; from 11:30 to 3:30 local time at the Boundary Facilities (except during Intensive Operating Periods or IOPs when 24-hour operations are required). If it rains when the operators aren't there, no log entry is made. Second, the problem can arise from standing water, not just rainfall; there is some time between the end of the rainfall and the evaporation of the water from the teflon window, which covers the mirror on the instrument. So, the operators could report that the rain has stopped but the microwave radiometer window is still wet and thus still reporting invalid data. 2) Can we use the microwave data when there are long runs of 'liq value below Min' (e.g., on 5, 7, 8 November 92)? Did you presumably make this problem go way on 21 Nov 92? Yes, you can use the data. I had originally set the lower limit to zero liquid, which is the physically plausible lower bound. I did not allow for the rms-accuracy of the retrieval (which is listed in the meta data portion of the data file). Actually, liquid values that are negative but within the rms-accuracy of the retrieval of zero may be considered equal to zero. On 21 November 1992 we changed the lower limit on the liquid to be "zero minus rms_liquid"--as indicated in the log--and we've not seen the nuisance alarms since. 3) How should we use those QC flags (qcmin, qcmax, qcdelta)? Just delete suspicious data? LWP and PWV that exceed the maximum should be eliminated - these usually indicate rain. PWV below zero is unphysical and arises during rain because of the opposite signs of the retrieval coefficients. Negative LWP is OK as long as it is within or close to the RMS uncertainty in the retrieval. The RMS uncertainties in the PWV and LWP are included in the meta data (which I hope you are examining): :vapor_retrieval_rms_accuracy[cm] = "0.057881" ; <--for December :liquid_retrieval_rms_accuracy[cm] = "0.003083" ; <--for December 4) When the weather log reports fog (more often than I would have guessed), is it possible that the microwave window gets covered with dew and leads to bad data even though no rain is reported? Dew has a fairly distinct signature - a smooth hump - in the retrieved vapor and liquid as well as in the underlying brightness temperatures. I have seen this quite a few times just at or before dawn; however, it usually goes away before the operators would have reported to the site. (This assessment is borne out by the surface station data which indicated that the dew point was within the sensor accuracy of the ambient temperature at about the time when the 'hump' first appeared.) For example, on one occasion the operators reported heavy ground fog at 1400 GMT (8 am local time) and the instrument showed 85-90 microns of liquid. I'd say the two were consistent. We have retrofitted all of the instruments with an "anti-dew" system. This is comprised of a continuous fan and a 500-750 W heater that is controlled by a resistive sensor mounted on top of the instrument. Normally, the fan blows ambient air over the teflon window to keep it clear of dust, etc. During condensing or precipitating conditions the heater turns on to prevent the formation of dew or the settling of fog on the window as well as to promote the evaporation of rain and snow. The condition of the heater (ON/OFF) is indicated in the netCDF files by the wet_window variable. This system seems to work quite well and the vendor has incorporated it into their design. The only problem that occurs now is maintaining the sensitivity of the sensor. Because it is a resistive element, it is somewhat temperature sensitive and so it periodically triggers unnecessarily on cold nights. Although this doesn't affect the PWV or LWP measurements, it does cause some confusion when using the wet_window data as an indication of rain or dew or fog. 5) When should we call the liquid water path zero (i.e. what is the noise level)? Why do we see significant (+/- 30 g/m3 = +/- 0.03 mm) positive/negative values of the liquid water path when the sky is clear according to the ceilometer? The noise level is very low: 0.003 mm = 0.0003 cm RMS. The problem is in the retrieval uncertainty. Statistical retrieval is essentially a multiple linear regression. Any regression will have a residual error. In the LWP retrieval the residual error or "theoretical accuracy" is 0.03 mm (RMS), 10 times the sensitivity or noise limit. So a value of LWP that is +/- 0.03 mm of zero *could* be clear sky. The real problem here is that the mean radiating temperature (T_mr) of the atmosphere, which is determined at the time the retrieval coefficients are computed, is assumed to only vary monthly with the retrieval coefficients. The truth is that T_mr varies diurnally - enough to cause the zero LWP to vary in a most annoying fashion within the uncertainty bounds of the retrieval. 6) What is a reasonable maximum liquid water path? Suppose the cloud averages 1 g/m3 of liquid water and is 1 km thick. Then the liquid water path would be 10**-6 g/cm3 x 10**5 cm = 0.1 g/cm2 = 0.1 cm = 1 mm. Thus, values above, say, 3 mm would be rare; such high values would probably be accompanied by rain and thus not measured anyway. 7) Why do we see occasional spikes way over 3 mm in the ARM microwave data? Two events cause the LWP to exceed 1 mm (or 3 mm). The first is rain or melting snow. The second is condensation (dew). A rule of thumb Ed Westwater (NOAA/ETL) uses is that brightness temperatures over 100 K aren't generally reliable (i.e. the optical depth can't get that large without precipitation or condensation on the Teflon window.) I have used this rule to set the upper limit for brightness temperatures; i.e. when the brightness temperatures exceed 100 K, a message is written to the Site Ops Log and a flag is set in the netCDF data file. 8) Are the data from the MWRs independent of the radiosondes? No, not entirely. Radiosondes from Oct 1992 - Dec 1993 were used to determine the "tuning functions" for the retrievals, and the retrievals are based on NWS radiosonde data from 1985-1990. 9) What are the "tuning functions"? The tuning functions linearly relate model-calculated microwave brightness temperatures (using radiosonde data) to brightness temperatures measured with a microwave radiometer. These are needed to account for imperfections in the microwave absorption model used to develop the retrievals which relate precipitable water vapor and liquid water path to the microwave brightness temperatures. The tuning functions should be independent of the instrument and of the location - they should depend only on the microwave absorption model used in the calculations. 10) How were the tuning functions determined? After each sonde launch, the model which computes the integrated vapor from the sonde as well as the microwave brightness temperatures is run automatically by the data system. I collected all of these modeled and measured brightness temperatures between Oct 92 and Dec 93, selected those for which the sky was clear (that is, for which the RMS variation in the liquid-sensing channel brightness temperature was less than 0.4 K) and calculated a regression for each channel. 11) Are the tuning functions still used? No. The so-called tuning functions were removed from the SGP CF MWR data that were collected between 950101 - 960409. Data collected before this date have already had these removed by Jim Liljegren, while data after this time window never had the tuning functions applied. At the time, it was commonly held that the sondes represented "ground truth" and that the "tuning functions" (i.e. regressions of model-calculated vs measured brightness temperatures) accounted for errors in the microwave absorption model upon which the retrievals were based. Jim Liljegren and Barry Lesht have since determined that the variation in the sonde calibration explains the differences between the model calculations and the microwave radiometric measurements. By removing the tuning functions, the PWV and LWP retrieved from the microwave radiometer are independent of the radiosondes. 12) What changes were made to the data ingest in October 1998? The MWR software was revised to provide additional functionality as described below. 1.Faster sampling rate -- Standard line-of-sight (LOS) observations can now be acquired at 15-second intervals vs. 20-second intervals previously. (The standard LOS cycle is comprised of one sky sample per blackbody sample and gain update.) 2.More flexible sampling strategy -- Multiple sky observations can be acquired during a LOS cycle, up to 1024 per gain update. This permits sky samples to be acquired at intervals of 2.67 seconds for improved temporal resolution of cloud liquid water variations and better coordination with the millimeter cloud radar during IOPs. 3.Separation of zenith LOS observations from TIP data -- When the radiometer is in TIP mode, the zenith LOS observations are now extracted, the PWV and LWP computed and reported separately in the output file. This eliminates the periods of missing LOS data during calibration checks/updates. 4.Automatic self-calibration -- The software now permits the calibration to be updated at specified intervals or continuously. To ingest the new format of the raw data, significant changes were made to the MWR Database Object Design. 18.0 Contacts 18.1 Instrument Mentor Victor Morris Pacific Northwest National Laboratory P.O. Box 999, MS K9-30 Richland, WA 99352 Phone: (509) 372-6144 Fax: (509) 372-6168 Email: vic.morris@arm.gov 18.2 Vendor/Instrument Developer Radiometrics Corporation 2840 Wilderness Place, Unit G Boulder, Colorado 80301-5414 Phone: (303) 449-9192 Fax: (303) 786-9343 Email: solheim@radiometrics.com 19.0 Glossary This information is currently unavailable 20.0 Acronyms AIS ARCS Integration Site ARCS Atmospheric Radiation and Cloud Station LWP Liquid Water Path NSA North Slope of Alaska PWV Precipitable Water Vapor SGP Southern Great Plains TWP Tropical Western Pacific WVR Water Vapor Radiometer 21.0 Citable References This information is currently unavailable