Meltpond 2000 PSR Processing - Final Report

Size: px

Start display at page:

Download "Meltpond 2000 PSR Processing - Final Report"

Brooke Crawford
5 years ago
Views:

1 Meltpond 2 PSR Processing Final Report Submitted to: Donald J. Cavalieri, Ph.D. NASA Goddard Space Flight Center Greenbelt, Maryland by: M. Klein A.J. Gasiewski NOAA/EL Boulder, Colorado 835 November 4, 22 As a part of the NASA Aqua AMSRE satellite validation program the NOAA Environmental Research Laboratory s Polarimetric Scanning Radiometer (PSR) was flown on a Navy P3 aircraft in JuneJuly of 2. Several flights over melting ice floes in Baffin Bay and Melville Sound were performed with the goal of imaging microwave brightness temperatures at high spatial resolution and under a variety of ice melt states. Conditions observed range from open water to completely frozen ice pack, including melting ponds. wo PSR scanheads were used during the experiment: (1) the PSR/C scanhead provided a Cband microwave radiometer with vertically and horizontally polarized channels at 6., 6.5, and GHz and a fully polarimetric channel at the AMSR band of 6.92 GHz, (2) the PSR/A scanhead provided fully polarimetric channels at 1.7 and 18.7 GHz, tripolarimetric channels at 37 and 89 GHz, and a dualpolarimetric channel at 21.5 GHz. Both scanheads provided an infrared radiometer operating from µm. Using the two scanheads all AMSRE frequencies were observed. In addition, L band data was obtained using the Scanning LowFrequency Microwave Radiometer (SLFMR). ABLE 1. MELPOND 2 FLIGH SUMMARY Date June 26 June 27 June 29 Region, Altitude (m) Baffin Bay, 58 Baffin Bay, 14 V Melville Sd., 15 Ice Camp, 3 Sensors PSR/A, SLFMR, IR PSR/A, SLFMR, IR PSR/A, SLFMR, IR Sensors Resolution PSR/A 1.7, 18.7, 21 GHz: 2. km; 37, 89 GHz:.6 km SLFMR: 3.3 km PSR/A 1.7, 18.7, 21 GHz:.5 km, 37, 89 GHz:.13 km SLFMR:.8 km PSR/A 1.7, 18.7, 21 GHz:.5 km 37, 89 GHz:.13 km SLFMR:.8 km

2 Date Region, Altitude (m) Sensors June V Melville Sd., 46 PSR/C, SLFMR, 3 Ice Camp, 3 IR July 5 Baffin Bay, 64 PSR/C, SLFMR, IR Sensors Resolution PSR/C: 2 km SLFMR: 2.6 km PSR/C: 2.7 km SLFMR: 3.6 km Meltpond 2 flight summary is shown in able 1, more comprehensive description of flights is in [1]. Maneuvers that have been processed to the level 2.3a are listed in Appendix A. his report provides a final assessment of the PSR Meltpond 2 data, as provided in the CD ROM entitled MELPOND, which includes PSR calibrated brightness temperatures over both ice and water. his data are also available on the NOAA EL PSR web site at hese results from the PSR and SLFMR show an excellent ability of highresolution microwave imagery to distinguish melting ice from water, and provide a means of determining the fractal structure of sea ice coverage. 1. PSR Data Reprocessing. An initial PSR Meltpond 2 data set was released to NASA sponsors at the end of August 2. his data set, while accurately calibrated, contained several shortcomings that have been addressed in the reprocessed CDROM. A discussion of each of these improvements is provided below. 1.1 Consolidation of Broken Flight Lines. During several Meltpond 2 flights, long flight lines were broken into several segments by the baseline PSR processing algorithm, as seen in Figure 1. he breaks occurred during parsing of the navigational data and were the result of unusual variations in the roll angle of the P3 aircraft. Reprocessing of the lines was required using a roll angle threshold variation of 9.

Figure 1. 1.7 GHz V brightness temperature maps from Meltpond 2 for the flight DF5 on June 27, 2 over Baffin Bay.

3 Figure GHz V brightness temperature maps from Meltpond 2 for the flight DF5 on June 27, 2 over Baffin Bay. he left map is the result of the initial processing, and resulted in several broken flight lines (see red circles). he reprocessed data are in the map on the right. 1.2 Level2 Data Processing for NADIR Observation Mode. In order to have the coincident angle of observation the PSR/A scanhead set to a constant incident angle using the nadir command of the PSR motion control system. he data from such observation mode, with constant incident angle, were processed to a three dimensional matrix with the second dimension (columns) is of size one. 1.3 Interference Removal from PSR/C Vertical Channels. he vertical channel PSR/C data on June 3 and July 5, 2 show high frequency spikelike noise (Figure 2) presumed to be interference. A nonlinear noise filter was developed (nlnf.m) to detect and remove this noise in the level 1.4 radiometric time series data. For comparison, Figure 3 shows the same B map after filtering. he filter is described in Appendix A.

Figure 2. Front look brightness temperature map of unfiltered PSR/C data at 6.

4 Figure 2. Front look brightness temperature map of unfiltered PSR/C data at GHz, vertical polarization during a flight over Baffin Bay on July 5, 2. he red dots are points exhibiting interference.

Figure 3. Front look brightness temperature map of filtered PSR/C data at 6.

Little interference remains in the image.

5 Figure 3. Front look brightness temperature map of filtered PSR/C data at GHz, vertical polarization during a flight over Baffin Bay on July 5, 2. Little interference remains in the image. he SLFMR data were processed for all data flights to level 2.3. Sample B map from the flight over Baffin Bay on June 27, 2 is on Figure 4.

6 Figure 4. SLFMR 1.42 GHz brightness temperature map from Meltpond 2 for the flight DF5 on June 27, 2 over Baffin Bay.

7 1.4 Outstanding Processing asks. Incident angle correction is not performed for Meltpond 2, since there is no method yet to identify open water points. 89GHz V channel is very noisy and the only explanation so far is some debris from iron epoxy. Radiometer data seems to be very noisy sometimes even with inverted B variations (colder spots on places where other channels show warmer signatures), but sometimes it looks that radiometer has reasonable B response. he option that was not tried yet is to handpick parts of the flights with reasonable B signatures. Reference: [1] D.J. Cavalieri: EOS Aqua AMSRE Sea Ice Validation Program: Meltpond2 Flight Report, NASA/M229972, NASA Goddard Space Flight Center, December 2.

8 APPENDIX A. Nonlinear interference filter description. An unbiased high pass filter was developed to identify and interpolate over radiometric image points that were corrupted by additive sporadic interference. he filter operates in the time domain, searching for points where the calibrated radiometric data exhibits nongeophysical spikelike behavior. he principal of the filter is illustrated in Figure A1. he filter assigns a weight w i to n neighboring points in on each side of the investigated point. he weight of the investigated point is assigned to be w =1. hus, the filter length is m=2*n+1. o insure that the filter is unbiased the weights are is unitary, viz: N i = w = w = 1 i N Acceptance or rejection of a point is determined based on the condition a timeseries of radiometric values, δ is a threshold value determined empirically, and * denotes convolution. Rejection of a point involves replacement by interpolation using its nearest healthy neighbors. v * w > δ where v is Figure A1. High pass filter impulse response. For the Meltpond 2 PSR/C vertically polarized data the following filter parameters were used: n=4 and δ=5. he Matlab mfile despiker.m for the filter is listed below: function [xout]=despiker(x,n,delta); %Nonlinear filter used to detect and replace sparse, sporadic high frequency %interference in uniformly sampled times series data. Noisy points are detected %using a highpass filter, then replaced using a interpolation over adjacent "healthy" points. % x is an input data vector or matrix with columns assumed to be different channels % n is the halfwidth of the filter (default: 4)

9 % delta is a noise threshold value that is used to select points to be replaced % (default: 5) %MK/AJG 12/1/1 %Set default values if these are not given as input parameters switch nargin case 1, n=4; delta=5; case 2, delta=5; end; debugg=1; %Set to zero after debbuging of filter parameters is done. %Construct highpass filter k=1/(n*(n+1)); w=[(1:n)*k,1,(n:1:1)*k]; nw=2*n; [npts,nchs]=size(x); %Perform highpass filtering operation, extending data at both ends of the vector, %using the first & last values xext=[repmat(x(1,:),n,1);x;repmat(x(end,:),n,1)]; xhpf=zeros(npts,nchs); for i=1:npts xhpf(i,:)=w*xext(i:i+nw,:); if mod(i,1e3)== fprintf('despiker is evaluating Point %d out of %d.\r',i,npts); end; end; %Identify interfering points and interpolate across xout=zeros(npts,nchs); for j=1:nchs indselect=find(xhpf(:,j)>delta); indorig=setdiff([(1:npts)],indselect); xout(indselect,j)=interp1(indorig,x(indorig,j),indselect); xout(indorig,j)=x(indorig,j); %Plot original, selected, and interpolated points for debbugging: if debugg figure; hold on; plot(x(:,j),'b.'); plot(indselect,x(indselect,j),'ro'); plot(indselect,xout(indselect,j),'g*'); legend('original vector','selected spikes','spike replacement values',); end end;

10 MELPOND Data Flight Catalog. List of maneuvers processed to the 2.3a level: Serial Numbe r Maneuver Identification 121 MP.DF4.M121.SLM. 123 MP.DF4.M123.SLM. 125 MP.DF4.M125.SLM. 126 MP.DF4.M126.SLM. 128 MP.DF4.M128.SLM. 129 MP.DF4.M129.SLM. Date (UC) 6/26/ 6/26/ 6/26/ 6/26/ 6/26/ 6/26/ 131 MP.DF5.M131.SLL. 6/27/ 133 MP.DF5.M133.SLL. 6/27/ 134 MP.DF5.M134.SLL. 6/27/ 135 MP.DF5.M135.SLL. 6/27/ 136 MP.DF5.M136.SLL. 6/27/ 138 MP.DF5.M138.SLL. 6/27/ imes (UC) 11:54:11 12:53:21 12:56:4 14:5:26 14:7:33 15:12:41 15:16:6 16:24:38 16:26:3 17:28:33 17:33:14 18:36:12 11:29:21 12:5:45 12:8:42 12:46:54 12:5:19 13:29: 13:33:35 14:17:15 14:19:16 14:54:59 14:58:24 15:34:2 Start Location (latitude, longitude) ø, ø ø, 65.49ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø 75.27ø, ø ø, ø Stop Location (latitude, longitude) ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, 69.62ø 14 MP.DF5.M14.SLL. 6/27/ 15:38: ø, ø,

11 Serial Numbe r Maneuver Identification Date (UC) imes (UC) 16:14: MP.DF5.M142.SLL. 6/27/ 16:17:35 16:56: MP.DF5.M143.SLL. 6/27/ 144 MP.DF5.M144.SLL. 6/27/ 145 MP.DF5.M145.SLL. 6/27/ 146 MP.DF5.M146.SLL. 6/27/ 148 MP.DF5.M148.SLL. 6/27/ 154 MP.DF6.M154.SLL. 6/29/ 155 MP.DF6.M155.SLL. 6/29/ 157 MP.DF6.M157.SLL. 6/29/ 158 MP.DF6.M158.SLL. 6/29/ 159 MP.DF6.M159.SLL. 6/29/ 161 MP.DF6.M161.SLL. 6/29/ 163 MP.DF6.M163.SLL. 6/29/ 17::45 17:35:25 17:38:22 18:15:9 18:17:57 18:21:41 18:21:56 18:24:29 18:31:33 18:37:58 13:31:8 14:37:43 14:4:54 15::16 15:4:29 15:22:43 15:26:25 15:49:7 15:49:42 16:2:1 16:9:24 16:27:2 16:34:59 16:57:11 Start Location (latitude, longitude) ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, 97.2ø Stop Location (latitude, longitude) ø ø, ø ø, ø ø, ø ø, 7.276ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø

12 Serial Numbe r Maneuver Identification Date (UC) 166 MP.DF6.M166.SLL. 6/29/ 167 MP.DF6.M167.SLM. 6/29/ 17 MP.DF7.M17.SLL. 6/3/ 172 MP.DF7.M172.SLL. 6/3/ 173 MP.DF7.M173.SLL. 6/3/ 175 MP.DF7.M175.SLL. 6/3/ 176 MP.DF7.M176.SLL. 6/3/ 177 MP.DF7.M177.SLL. 6/3/ 178 MP.DF7.M178.SLL. 6/3/ 18 MP.DF7.M18.SLL. 6/3/ 181 MP.DF7.M181.SLL. 6/3/ 182 MP.DF7.M182.SLL. 6/3/ 184 MP.DF7.M184.SLL. 6/3/ 186 MP.DF7.M186.SLL. 6/3/ imes (UC) 17:9:42 17:15:17 17:15:49 17:37:35 12:2:47 12:3:46 12:37:24 12:42:13 12:42:31 13::3 13:9:47 13:12:7 13:12:26 13:32:2 13:35:21 13:53:17 13:53:47 13:58:57 14:5:9 14:9:52 14:11:32 14:32:48 14:35:8 14:49:2 14:53:35 14:56:49 15:4:41 Start Location (latitude, longitude) ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø Stop Location (latitude, longitude) ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø ø, ø

13 Serial Numbe r Maneuver Identification Date (UC) imes (UC) Start Location (latitude, longitude) Stop Location (latitude, longitude) 15:14: MP.DF7.M188.SLL. 6/3/ 15:17:44 15:35: ø, ø ø, ø 19 MP.DF7.M19.SLL. 6/3/ 15:51: ø, ø, ø 15:58: ø 191 MP.DF7.M191.SLL. 6/3/ 16:1:12 16:26: ø, ø ø, ø 192 MP.DF7.M192.SLL. 6/3/ 16:28:51 16:44: ø, ø ø, ø 194 MP.DF7.M194.SLL. 6/3/ 16:47:26 17:6: ø, ø ø, ø 195 MP.DF7.M195.SLL. 6/3/ 17:9:38 17:23: ø, ø ø, ø 196 MP.DF7.M196.SLL. 6/3/ 17:23:32 17:26: ø, ø ø, ø 199 MP.DF7.M199.SLM. 6/3/ 17:34: ø, ø, ø 17:57: ø 2 MP.DF8.M2.SLM. 7/5/ 12:: ø, ø ø, ø 13:4:4 22 MP.DF8.M22.SLM. 7/5/ 13:8: ø, ø ø, ø 14:6:1 23 MP.DF8.M23.SLM. 7/5/ 14:9: ø, ø ø, ø 14:59:23 24 MP.DF8.M24.SLM. 7/5/ 15:3: ø, ø ø, 68.77ø 15:5:25 26 MP.DF8.M26.SLM. 7/5/ 15:54: ø, ø ø, ø 16:4:9 27 MP.DF8.M27.SLM. 7/5/ 16:4: ø, ø,

14 Serial Numbe r Maneuver Identification Date (UC) imes (UC) 16:37:54 7/5/ 28 MP.DF8.M28.SLM. 16:4:56 17:28:43 Start Location (latitude, longitude) ø ø, ø Stop Location (latitude, longitude) ø ø, ø

15 Level 2.3a data format description for PSR/A, PSR/C, and SLFMR instruments: First dimension index is the scan number. Second dimension index is the sample number within a scan. hird dimension index assignment if described for each instrument in tables below: PSR/A LEVEL2.3a DAA Format Description for MELPOND (L23ssss.mat) description firstday ssn Scanangle(:) scanhead scenel23a(:,:,1) scenel23a(:,:,2) scenel23a(:,:,3) scenel23a(:,:,4) scenel23a(:,:,5) scenel23a(:,:,6) scenel23a(:,:,7) scenel23a(:,:,8) scenel23a(:,:,9) scenel23a(:,:,1) scenel23a(:,:,11) scenel23a(:,:,12) scenel23a(:,:,13) scenel23a(:,:,14) scenel23a(:,:,15) scenel23a(:,:,16) scenel23a(:,:,17) scenel23a(:,:,18) scenel23a(:,:,19) scenel23a(:,:,2) scenel23a(:,:,21) scenel23a(:,:,22) scenel23a(:,:,23) Segment string description Julian day of the start of first flight Serial number (string) Nominal elevation angle per scan Scanhead identifier (string) 1.7v (B1Ch) 1.7h (B1Ch1) 18.7v (B1Ch4) 18.7h (B1Ch5) 21.5v (B1 Ch8) 21.5h (B1 Ch9) 37.vA (B2 Ch) 37.hA (B2 Ch1) 89.vA (B2 Ch6) 89.hA (B2 Ch7) 1umIR (B2 Ch12) (High rate) Pitch (degrees) (High rate) Roll (degrees) GPS latitude (dd.ddd) GPS longitude (dd.ddd) true heading (degrees) Inertial Altitude ime stamp. (Seconds from start of firstday) rue azimuth angle rue elevation angle Polarization angle Pixel latitude (calculated for sphere with the Earth radius) Pixel longitude (calculated for sphere with the Earth radius)

16 PSR/C LEVEL2.3a Data Format Description for MELPOND (L23ssss.mat) description firstday ssn scanangle(:) scanhead scenel23a(:,:,1) scenel23a(:,:,2) scenel23a(:,:,3) scenel23a(:,:,4) scenel23a(:,:,5) scenel23a(:,:,6) scenel23a(:,:,7) scenel23a(:,:,8) scenel23a(:,:,9) scenel23a(:,:,1) scenel23a(:,:,11) scenel23a(:,:,12) scenel23a(:,:,13) scenel23a(:,:,14) scenel23a(:,:,15) scenel23a(:,:,16) scenel23a(:,:,17) scenel23a(:,:,18) scenel23a(:,:,19) scenel23a(:,:,2) scenel23a(:,:,21) segment string description Julian day of the start of first flight Serial number (string) Nominal elevation angle per scan Scanhead identifier (string) 6.v (B1 Ch) 6.h (B1 Ch1) 6.5v (B1 Ch2) 6.5h (B1 Ch3) 6.92v (B1 Ch4) 6.92h (B1 Ch5) 7.32v (B1 Ch6) 7.32h (B1 Ch7) 1umIR (B1 Ch1) (High rate) Pitch (degrees) (High rate) Roll (degrees) GPS latitude (dd.ddd) GPS longitude (dd.ddd) true heading (degrees) Inertial Altitude ime stamp. (Seconds from start of firstday) rue Azimuth rue Elevation Polarization Pixel latitude (calculated for sphere with the Earth radius) Pixel longitude (calculated for sphere with the Earth radius)

17 SLFMR LEVEL2.3a Data Format Description for MELPOND (L23assss.mat) description firstday ssn scanhead scenel23a(:,:,1) scenel23a(:,:,2) scenel23a(:,:,3) scenel23a(:,:,4) scenel23a(:,:,5) scenel23a(:,:,6) scenel23a(:,:,7) scenel23a(:,:,8) scenel23a(:,:,9) scenel23a(:,:,1) scenel23a(:,:,11) scenel23a(:,:,12) scenel23a(:,:,13) scenel23a(:,:,14) scenel23a(:,:,15) Segment string description Julian day of the start of first flight Serial number (string) Scanhead identifier (string) Radiometric Channel SLFMR beam position azimuth (9 deg right from the airplane, 27 deg left from the airplane) SLFMR beam position (elevation) (deg from nadir) Aircraft (INS) latitude (deg) Aircraft (INS) longitude (deg) Pitch (deg) Roll (deg) rue Heading (deg) Inertial Altitude (feet) ime stamp in msec rue azimuth angle (deg) rue elevation angle (deg) rue polarization angle (deg) Pixel latitude (calculated for sphere with the Earth radius) Pixel longitude (calculated for sphere with the Earth radius)

18 IGARSS 21 paper: Sea Ice Brightness Imagery Observed During the Meltpond 2 Experiment M. Klein 1, A.J. Gasiewski 2, K. Schuler 3, D. Cavalieri 4,. Markus 4, A.Yevgrafov 1 1 CIRES, University of Colorado/NOAA Environmental echnology Laboratory, R/E/E1, 325 Broadway, Boulder, CO 835, Marian.Klein@noaa.gov, , , Aleksandr.Yevgrafov@noaa.gov, , , 2 NOAA Environmental echnology Laboratory, R/E/E1, 325 Broadway, Boulder, CO 835, Al.Gasiewski@noaa.gov, , University of Karlsruhe, Institute of High Frequency Electronics, Karlsruhe, Germany, karin.schuler@stud.unikarlsruhe.de 4 Laboratory for Hydrospheric Processes, NASA Goddard Space Flight Center, Mailstop 971., Greenbelt, MD 2771, don@cavalieri.gsfc.nasa.gov, , , thorsten@beaufort.gsfc.nasa.gov, , Abstract As a part of the NASA Aqua satellite validation program the NOAA Environmental Research Laboratory s Polarimetric Scanning Radiometer (PSR) was flown on a Navy P3 aircraft in JuneJuly of 2. Several flights over melting ice floes in Baffin Bay and Melville Sound were performed with the aim of imaging microwave brightness temperatures at high spatial resolution and under a variety of ice melt states. Conditions observed range from open water to completely frozen ice pack, including melting ponds. wo PSR scanheads were used during the experiment: (1) the PSR/C scanhead provided a Cband microwave radiometer with vertically and horizontally polarized channels at 6., 6.5, and GHz and a fully polarimetric channel at the AMSR band of 6.92 GHz, (2) the PSR/A scanhead provided fully polarimetric channels at 1.7 and 18.7 GHz, tripolarimetric channels at 37 and 89 GHz, and a dualpolarimetric channel at 21.5 GHz. Both scanheads provided an infrared radiometer operating from mm. Using the two scanheads all AMSRE frequencies were covered. In addition, Lband data was obtained using the Scanning Low Frequency Microwave Radiometer (SLFMR). Results from the calibration of the PSR brightness imagery are presented for both ice and water. hese results from PSR and SLFMR show an excellent ability of highresolution microwave imagery to distinguish melting ice from water, and provide a means of determining the fractal structure of sea ice coverage. I. MELPOND 2 EXPERIMEN he National Oceanic and Atmospheric Administration (NOAA) Environmental echnology Laboratory (EL) provided aircraft sensors for the National Aeronautic and Space Administration (NASA) Meltpond 2 experiment. Meltpond 2 [1] was the first in a series of Arctic and Antarctic aircraft campaigns planned as part of NASA s Earth Observing System Aqua sea ice validation program for the Advanced Microwave Scanning Radiometer (AMSRE). he Naval Air Warfare Center (NAWC) P3 aircraft was used to make five 8hour flights from hule Air Base over portions of Baffin Bay, Viscount Melville Sound, and an area immediately NW of Resolute, Canada between June 26 and July 5, 2. he flight summary is provided in able 1. he atmospheric and sea ice conditions varied during the duration of the experiment. he weather conditions experienced ranged from clear sky to precipitation within an extensive low pressure system. Sea ice conditions were also variable. Open water was present at the northern end of Baffin Bay with ice concentration increasing south west ward to 91%. ABLE 1. MELPOND 2 FLIGH SUMMARY Date Region, Altitude (m) Sensors Sensors Resolution June 26 Baffin Bay, 58 June 27 Baffin Bay, 14 June 29 V Melville Sd., 15 Ice Camp, 3 PSR/A, SLFMR, IR PSR/A, SLFMR, IR PSR/A, SLFMR, IR June 3 V Melville Sd., 46 Ice Camp, 3 PSR/C, SLFMR, IR July 5 Baffin Bay, 64 PSR/C, SLFMR, IR II. EXPERIMEN INSRUMENAION PSR/A 1.7, 18.7, 21 GHz: 2. km 37, 89 GHz:.6 km SLFMR: 3.3 km PSR/A 1.7, 18.7, 21 GHz:.5 km 37, 89 GHz:.13 km SLFMR:.8 km PSR/A 1.7, 18.7, 21 GHz:.5 km 37, 89 GHz:.13 km SLFMR:.8 km PSR/C: 2 km SLFMR: 2.6 km PSR/C: 2.7 km SLFMR: 3.6 km A comprehensive suite of microwave radiometers was operated by NOAA/EL during the Meltpond 2 experiment. he primary instrument was the Polarimetric Scanning Radiometer (PSR) [2] operating one of two available scanheads: PRS/A and PSR/C, see he PSR has three main components: a scanhead, a positioner, and aircraft control/acquisition electronics. he scanhead houses the PSR radiometers and is able to be arbitrarily moved in both azimuth and elevation. Several scanning modes are programmable, including conical, cross track, and along track. wo scanheads, PSR/A and PSR/C,

he PSR positioner includes mechanics to support and position a scanhead. he positioner also contains two external calibration targets, one at ambient temperature and one heated to ~ 7 C.

19 were used during Meltpond 2; the radiometric channels and polarizations, are listed in able 2. hese scanheads could be swapped during the deployment in as little as 2 hours, facilitating alternate flights using different bands. he PSR positioner includes mechanics to support and position a scanhead. he positioner also contains two external calibration targets, one at ambient temperature and one heated to ~ 7 C. he PSR data acquisition and control system is networked to the scanhead computers and data are available after a flight using a notebook computer. he preferred scanning mode of the PSR during Meltpond 2 was fullconical (36 in azimuth) with 55 incident angle. he Scanning Low Frequency Microwave Radiometer (SLFMR) is a cross track scanning radiometer operating at a center frequency of 1.41 GHz with bandwidth of 1 MHz [3]. he radiometer was built by Quadrant Engineering, Inc. primarily for ocean salinity mapping. he electronically steerable scan consists of eight vertically polarized beams (four on port and four on starboard) with incident angles varying from 6 to 47 from nadir. Unfortunately, during our experiment, only the four outermost beams were available due to an electronic failure, so only data with incident angles of 36 and 47 on both sides of the aircraft are available, otherwise the PSR and SLFMR worked reliably for the duration of the experiment and provided unique high spatial resolution brightness temperature ( B ) maps. integration. One exception was the 89 GHz vertical channel, which has not yet been calibrated due to spurious noise. he 37 GHz brightness temperature mosaic for horizontal polarization on June 27, 2 in Figure 1 shows breaking sea ice in Baffin Bay. he data were recorded during ~6 hours of flight time, from 11:3 to 17:3 UC. Excellent radiometric stability (typically better than 1 K) was observed from one flight line to another. Individual ice patches or patches of open water are seen to cleanly overlap across adjacent maneuvers even when observed nearly one hour apart. Figure 2 shows the same location observed using the SLFMR. he data show excellent correspondence in feature location. When comparing PSR and SLFMR B maps one must be aware of the different scanning geometries of the two instruments. Since the SLFMR is a crosstrack scanner the changing angle of incident and footprint size of the beam change the brightness and resolution significantly throughout the scan. Due to the lower spatial resolution of the SLFMR the ice boundaries in Figure 2 are thus not defined as sharply as on Figure 1, and striping between beam positions is present over water. ABLE 2. INSRUMENS AND CHANNELS FOR MELPOND 2. Channel Frequency Polarization Scanhead (GHz) Vertical SLFMR Vertical, Horizontal Vertical, Horizontal PSR/C Vertical, Horizontal Vertical, Horizontal 1.7 Vertical, Horizontal 18.7 Vertical, Horizontal 21.5 Vertical, Horizontal PSR/A 37 Vertical, Horizontal 89 Vertical*, Horizontal µm in both, PSR/A and PSR/C RGB Video Camera in both, PSR/A and PSR/C *Noisy III. OBSERVED MICROWAVE IMAGERY wo examples of observed imagery are shown in Figures 1and 2. he data in these B maps are not compensated for incident angle variations caused either by aircraft roll and pitch or, in the case of the SLFMR, by the crosstrack scanning mode. Even though the PSR scans at a fixed angles, aircraft pitch and roll variations influence the incident angle by as much as ±3 in angle, or roughly ±7 K for vertical and ±4 K for horizontal polarization over water [4]. Overall, the instrument suite covered the microwave spectrum from 1.4 to 89 GHz in seven distinct frequency bands (able 2). Absolute PSR calibration was accomplished using steep (6º) bank rolls to point the radiometers to cold space. All channels worked well providing ~.51 K rms noise for ~2 ms Figure 1. PSR/A 37 GHz horizontal polarization front look brightness temperature map observed over Baffin Bay on June 27, 2. he image is a composition of 1 flight lines over melting sea ice. he spatial resolution for the given flight level (14 m) and the 37 GHz channel beamwidth (2.3 ) is ~13 m. able 3 shows a comparison of theoretically calculated B values, using the Microwave Radiative ransfer (MR) model [4], and those observed by PSR/A for selected open water (648 points) and ice covered regions (29484 points) of Baffin Bay on June 26, 2. For the calculated B values

20 an atmospheric profile from a radiosonde launched from Resolute, Canada at 11:17UC was assumed, and ice emission coefficients were obtained from [6]. A calm sea water surface is assumed in the theoretical calculations along with a salinity of 3 ppt. he points over open water were used to adjust the absolute calibration curve (i.e. calibration load emissivity) of the PSR/A radiometers. he resulting standard deviation of B for the open water region after compensation for incident angle variations (as described in [4]) is shown in able 3. he level of stability for the PSR/A radiometric channels is documented in the values of σb for open water region and on Figure 3. he few outliers are the result of sun glint off the water surface. he values shown on Figure 3 were recorded during ~4.5 hours of flight. hus some of the B observed variation is likely of geophysical origin, probably due to surface roughness changes from wind speed variations. Most channels exhibited excellent stability over the sampling interval. ABLE 3. COMPARISON OF CALCULAED AND OBSERVED B FOR HE PSR/A FLIGH ON JUNE 26, 2 Parameter Channel Open water region Ice covered region (GHz) heory heory σ B First year ice Multiyear ice B Mean S 1 ( BV) (K) S 2 ( BH) (K) σ B Figure 3. he PSR/A 37 GHz vertical polarization observed B (blue points) over open water as function of incident angle. heoretically calculated dependence is shown as lines. he black line is for calm water and red line for 3 m/s wind. REFERENCES Figure 2. SLFMR brightness temperature map observed over Baffin Bay on June 27, 2. he image is a composition of 1 flight lines over melting sea ice. he spatial resolution for the given flight level (14 m) and SLFMR beamwidth is ~8 m. IV. CONLUSIONS he stability of the PSR radiometers during the Meltpond experiment was very good, and the flights yielded a data set useful for both ice imaging and theoretical radiative transfer model intercomparison. he values of observed B over ice are within the expected range of theoretically calculated values, and closer to those for first year ice, although meltponding is suspected of lowering the brightness values somewhat. [1] D.J. Cavalieri, EOS Aqua AMSRE Sea Ice Validation Program: Meltpond 2 Flight Report, NASA echnical Memorandum, NASA/M229972, December 2. [2] J.R. Piepmeier, A.J. Gasiewski, High Resolution Passive Polarimetric Microwave Mapping of Ocean Surface Wind Vector Fields, IEEE rans. Geosci. Remote Sensing, vol. 39, pp , March 21. [3] M.A. Goodberlet, C.. Swift, A Remote Sensing System for Measuring Estuarine and Coastal Ocean Surface Salinity, Quadrant Engineering, Nov. 2, [4] I. Corbella, A.J. Gasiewski, M. Klein, J.R. Piepmeier Compensation of Elevation Angle Variations in Polarimetric Brightness emperature Measurements from Airborne Microwave Radiometers, IEEE rans. Geosci. Remote Sensing, vol. 39, pp , January 21. [5] A.J. Gasiewski, D.H. Staelin: Numerical Modeling of Passive Microwave O 2 Observations over Precipitation, Radio Sci., vol. 25, pp , 199. [6] F.D. Carsey, Microwave Remote Sensing of Sea Ice,, Geophysical monograph 68, American Geophysical Union, Washington, USA, 1992.

Airborne Water Vapor Science, Radiometer Requirements, and Capabilities

Airborne Water Vapor Science, Radiometer Requirements, and Capabilities Professor Albin J. Gasiewski University of Colorado NOAA-CU Center for Environmental Technology (CET) al.gasiewski@colorado.edu 303-492-9688