4 May 1999 revised: 10 Oct 1999 revised: 26 March 2001 revised: 10 Jan 2003 INTRODUCTION The ER-2 Doppler radar (EDOP) is an X-band (9.6 GHz) Doppler radar mounted in the nose of ER-2. The instrument has two fixed antennas, one pointing at nadir and the second pointing approximately 33 degree ahead of nadir. The beam width of the antenna is 3 degree in the vertical and horizontal directions which, for a 20 km altitude, yields a nadir footprint a the surface of 1 km. The ER-2 ground speed is nominally 210 m/s and the integration period used by the data system is 0.5 second. The transmit pulse is 0.5 micro-second and the gate spacing is over sampled at 37.5 meter interval. Minimum detectable reflectivity is about -10 dBZ at an altitude of 15 km and for a 0.375 meters range gate spacing. Additional details of the radar and processing are described by Heymsfield et al. (1996). EDOP DATA FORMAT EDOP data is in Universal Format (UF) as documented in the UF-IDL.guide. EDOP currently collects data at a rate of 2 Hz from 4 Doppler channels with 738 gates. IDL software is provided for reading the data. Other software packages such as developed by NCAR should be able to read the EDOP UF files although this has not been tested. The 4 EDOP channels are as follows: Nadir rain (dynamic range optimized for rain signal -10 to 70 dBZ) Nadir surface (dynamic range optimized for ocean surface 20 to 100 dBZ) Forward Co-polarization Forward Cross-polarization Each of these channels provides reflectivity, Doppler velocity, and spectral width estimated from the time series (autocovariance) approach. Four levels of UF format are implemented for EDOP data: Level 1: Conversion of EDOP uncalibated raw data format to UF (uncalibrated power, Doppler velocity). Level 2: Conversion of uncalibrated power to calibrated dBZ Level 3: Doppler unfolding and aircraft motion removal. Level 4: Attenuation correction of reflectivity. ALL DATA IS PROCESSED TO LEVEL 3 UNLESS NOTED. The distribution versions are as follows: .uf Version 0 is preliminary data which has been calibrated but not rigorously examined or used. Data is processed to Level 3 unless noted. .uf.X Version X higher numbers indicated a more recent, updated data set. The calibration may have been improved or some other aspect of the data was modified. The parameter names for the Level 2 files are as follows: PN Nadir_VV power PS Nadir_VV surface power DN Nadir_VV raw Doppler velocity [m/s] DS Nadir_VV raw surface Doppler velocity [m/s] ZN Nadir_VV reflectivity (dBZ) ZS Nadir_VV surface reflectivity (dBZ) Level 3 provides the following additional parameters: PF Forward_VV power PX Forward_VH (cross-pol) power DF Forward_VV raw Doppler velocity [m/s] DX Forward_VH raw Doppler velocity [m/s] VN Nadir_VV corrected Doppler velocity [m/s] VF Forward_VV corrected Doppler velocity [m/s] ZF Forward_VV reflectivity (dBZ) EDOP CALIBRATION The calibration of EDOP is described in Caylor et al. (1995) although this has been modified somewhat due to the change to the new digital-IF receiver and processing system in 1998. After this calibration is performed, EDOP surface measurements (sigma_0) from both the nadir and forward beams are compared with previously published values as well as with statistics from TRMM estimated values. In addition, the reflectivity measurements in rain are compared with simultaneous measurements from the TRMM Precipitation Radar and several ground-based radars. However, only the EDOP internal and external calibrations are used and NO adjustments are made to the EDOP reflectivites (unless noted below) based on these comparisons. EDOP has a linear receiver so all reflectivity integrations are performed linearly. Thus if any error in calibration exists, the bias will be constant over the full dynamic range of the measurements. The final calibration of EDOP should be accurate to approximately 1 dBZ. AIRCRAFT MOTION REMOVAL Aircraft motions must be removed from the Doppler measurements before they provide an estimate of the reflectivty-weighted Doppler hydrometeor motions. Aircraft motions are estimated using attitude and vertical acceleration data from the ER-2 high-speed (ARINC) navigation system. Vertical aircraft motions are estimated from vertical acceleration and GPS altitude using a third-order feedback loop similar to NCAR and NASA Ames approaches. Aircraft motions are batch processed in the distribution data sets and are hoped to be accurate to within 0.5 m/s. But in some cases such as strong convective towers where gravity waves are strong at the ER-2 altitude, the corrections can be in error by a much greater amount. These errors can be reduced by further analysis so users are encouraged to use Doppler measurements with caution. ARTIFACTS IN THE DATA 1) Second trip echo located at approximately 15 km altitude: This very narrow band of echo results when the transmit pulse reflects off the surface and illuminates the ER-2. The returned echo from the ER-2 appears in the next pulse. Since the ER-2 flies at 20 km altitude and the maxmimum unambiguous range is about 34 km, the second trip echo will be at about 14 km altitude. This can be edited out but it is tedious since we have not yet found an easy way to automate it. 2) Leakage between the forward and nadir beams: There is leakage between the forward and nadir receivers which are isolated by more than 80 db. The surface signal is quite strong (80-90 dBZ at nadir) so it leaks into the forward beam in a narrow band above the surface. This has been edited in the data. The forward beam surface return leaks into the nadir channel and it produces an artifical return below the nadir surface return. This leakage signal is very weak and has not been removed. Antenna side lobes and aircraft motions cause a wavy pattern below the nadir surface. Data regions (mirror image in rain) is relatively unaffected by this leakage. SPECIFIC ISSUES FOR FIELD CAMPAIGNS *************************************************************************** TEFLUN-A: During TEFLUN-A, the EDOP nadir front end amplifier degraded just after the pre-campaign external calibration and before the first data flights. The low-noise amplifier gain was reduced by about 10 dB for most of the experiment. The nadir channel was corrected based on the above calibration procedures and then the forward and nadir channels were compared in rainy regions to determine if any biases existed. A 1.85 dB adjustment (decrease) was incorporated in the current distribution of the nadir reflectivity, based on all avaliable knowledge. For the flight on 4 May, one digital receiver and processor boards were having problems. As a result, the nadir surface measurement uses a different board which required additional external calibrations after the field campaign. Again, the best estimate calibration data has been implemented on 4 May data. The forward cross polarzation is not available for this day. CALIBRATION CORRECTION AND OTHER KNOWN PROBLEMS (10/5/99): 1) A reflectivity calibration adjustment of -2.51 dB is required for all the Version 1 data sets. This adjustment is required because of the oversampling in range performed by EDOP. EDOP uses a digital bandpass filter and the filter in digital signal processing code was not being normalized by the pulsewidth. This resulted in a 3 dB calibration difference but filter losses were reduced (typically 1.8 dB) resulting in a net decrease in the calibration given above. This correction NOT included in Version 1 data. 2) The latitude and longitude values given in the UF files are incorrect in Version 1 data sets due to an incorrect scale factor used on the high-speed ARINC 429 navigation data. The scale factor used in Version 1 was 0.00017172 (S0), so it should have been 0.0001716614588 (SN). This results in an error to the west of about 0.02 degrees (~2 km) at 60deg West longitude (Rondonia Brazil), for instance. That is, Lon = (SN/S0) * Lon in current UF file. But it may be more accurate to use the navigation recorder data (in DAAC) until the data gets reprocessed. ************************************************************************** TEFLUN-A, TEFLUN-B, CAMEX-3, TRMM-LBA VERSION 2.0 - Released February, 2000 Changes from Version 1.0: 1) Problems 1 and 2 above have been corrected: - Calibration has been adjusted by 2.51 dB on all channels. - Latitude - longitude values have been corrected. ************************************************************************** TRMM-LBA See 1) and 2) above. Otherwise very clean data sets. ************************************************************************** CAMEX-4 No known problems. ************************************************************************** LIST OF FILES The flight tracks are provided in a compressed UNIX tar file. To extract the data files the following steps must be performed. 1) Uncompress: uncompress edop.tar.Z 2) Extract: tar -xf edop.tar The extracted files are datalog list of data file available README This file uf_software directory contains IDL softares to read the EDOP uf file REFERENCES: Caylor I.J., G.M. Heymsfield, S. Bidwell, and S. Ameen, 1994: NASA-ER-2 Doppler radar reflectivity calibration for the CAMEX project. NASA Tech Memo 104611, 15 pp. Heymsfield, G. M., S. Bidwell, I. J. Caylor, S. Ameen, S. Nicholson, W. Boncyk, L. Miller, D. Vandemark, P. E. Racette, and L. R. Dod, 1996: The EDOP radar system on the high-altitude NASA ER-2 aircraft. J. Atmos. Oceanic Tech., 13, 795-809. POINTS OF CONTACT: Gerald Heymsfield Laboratory for Atmospheres, Mesoscale Atmospheric Processes Branch Goddard Space Flight Center Code 912 Greenbelt, MD 20771 301-614-6369 heymsfield@agnes.gsfc.nasa.gov Lin Tian University Space Research Associates & Goddard Space Flight Center Code 912 Greenbelt, MD 20771 301-614-6318 tian@agnes.gsfc.nasa.gov =*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=