Labrador Sea Deep Convection Experiment (LabSea) 1997

During January-March 1997 a unique multiagency complement of instruments for observing ocean surface emission and scattering signatures was installed and deployed on the NASA/Wallops Flight Facility (WFF) Orion P-3B aircraft (N426NA). The primary goal of the Ocean Winds Imaging (OWI) complement was to collect data to verify the utility of passive ocean wind vector sensing in high seas, with secondary goals being to better characterize the thermal emission and scattering signatures of the wind-driven ocean surface. The complement was flown under a variety of meteorological conditions in coordinated patterns over both ocean buoys along the eastern U.S. coast and an instrumented research vessel - the R.V. Knorr - within the Labrador Sea. The conditions represented a wide range of surface wind speeds, cloud and water vapor states, and fetch lengths.

During the Labrador Sea deployment the OWI complement observed the ocean across sections of several convective meteorological systems driven primarily by arctic cold-air outbreaks moving eastward from Labrador, Canada over the open water. Several coordinated underflights of low-orbitting satellites sensors, including the NASA Scatterometer (NSCAT) and the Defense Meteorological Satellite Program (DMSP) SSM/I sensors were also performed. The flights resulted in the acquisition by the OWI science team of the first high-resolution polarimetric passive microwave conically-scanned imagery of the ocean in a broad set of bands, and the first combined joint high-resolution passive and active imagery of the ocean surface. Total flying time for the effort (including integration and engineering check flights) amounted to ~48 flight hours.

Figure 1. Azimuthal wind direction harmonics observed during the Laborador Sea Deep Convection Experiment on March 4, 1997. From Piepmeier and Gasiewski [2000].

Participating institutions in the integration and observation phases of the OWI/Labrador Sea experiment included the Georgia Institute of Technology School of Electrical and Computer Engineering (GIT/ECE), the NOAA Environmental Technology Laboratory Ocean Remote Sensing Division (NOAA/ETL), the University of Massachusetts at Amherst Microwave Remote Sensing Laboratory (U-Mass/MIRSL), the NASA/WFF Observational Science Branch (OSB) and Aircraft Programs Branch (APB), the National Center for Atmospheric Research (NCAR), the Woods Hole Oceanographic Institution, and the Aerospace Corporation. In addition, assistance in the design and fabrication of critical aircraft equipment was obtained from the Raytheon Corporation, the NASA/Ames Research Center (ARC), and the Georgia Tech Research Institute (GTRI).

The OWI complement included two active radar scatterometers for measuring both Bragg and specular return (the U-Mass C-band Scatterometer - C-Scatt and the NASA/WFF/OSB Radar Ocean Wave Spectrometer - ROWS), two passive polarimetric scanning radiometers for imaging the upwelling thermal emission from the ocean surface (the Polarimetric Scanning Radiometer - PSR/D and the U-Mass Ka-band Scanning Polarimetric Radiometer - KaSPR), and precision fixed-beam radiometers for measuring the sub-track upwelling polarimetric emission from the ocean (the ETL Ka-band polarimetric radiometer - Kapol) and the above-track thermal emission from clouds and water vapor (the ETL Cloud and Water Vapor Radiometer - CWVR). The complement also included a nadir infrared pyrometer for measuring sea surface temperature, a GPS dropsonde package (provided by NCAR) for measuring subtrack pressure, temperature, humidity, and wind vector, and several video cameras (provided by GIT) for recording ocean foam and cloud conditions.

The OWI experiment provided data to develop an extended geophysical model function (GMF) for ocean surface thermal emission for T_v, T_h, and T_U at 10.7 and 37.0 GHz, and for T_v and T_h at 18.7 GHz (Figure 1). When the raw data were averaged over a large number of scans to reduce both instrument and geophysical noise, excellent agreement between the PSR and SSM/I 37 GHz azimuthal harmonic amplitudes was obtained for the range of wind speeds from near calm to ~16 m/sec over large footprint areas and for an ensemble of ocean-atmosphere states. The GMF also exhibits excellent consistency between amplitudes and phases of the azimuthal harmonics for the 10.7, 18.7, and 37.0 GHz channels, with harmonic amplitudes monotonically increasing by ~50% from 10.7 to 37.0 GHz.

On smaller spatial scales, the T_v and T_h data often showed local brightness variations of up to 10-15 K that are related to the presence of convection and/or unstable air-sea conditions. Atmospheric conditions suggest that these brightness perturbations are caused by surface wave spectrum variations. Such conditions did not influence the T_U imagery nearly as much, and in many instances virtually no impact is seen in T_U. The degree to which T_U rejects such convection-related perturbations indicates that the polarization signature will provide valuable information on surface wave, and hence wind, direction.

The PSR GMF was used to demonstrate the first aircraft-based passive microwave maps of ocean surface wind fields using a maximum likelihood (ML) estimator with adaptive channel weights. One-dimensional wind line plots were developed using full 360° azimuthal scans, and two-dimensional wind field maps over a region of ~14 x 100 km were developed using a two-look technique. Both one- and two-dimensional techniques show that ocean wind direction signatures over mesoscale-sized regions of ~15 km or greater size are consistent enough for satellite mapping using a two-look polarimetric technique that includes T_U. Moreover, adaptive channel weighting allows use of all radiometric channels during conditions where the azimuthal harmonics have high signal-to-noise ratio (SNR), but emphasizes the lower frequency channels during conditions when the higher frequency channels become perturbed by either clouds or surface roughness.

Satellite simulations using the PSR data downsampled to spot sizes similar to that of the SSM/I, but including both tri-polarimetric channels (T_v, T_h, and T_U) and two-looks show retrieval accuracies of ~8.4° and ~0.8 m/sec without directional ambiguities. These results strongly support the concept of satellite-based passive ocean surface wind vector mapping. When the third Stokes parameter is removed the simulated retrieval accuracy was reduced to ~12.6° and ~1.0 m/sec, with a significant increase in ambiguity rate (11.1%). When a tri-polarimetric system is used but with only one look, the accuracy is ~14° and ~1.6 m/sec, although the ambiguity rate is slightly lower (7.7%). It is noted that these simulation accuracies are consistent with error values predicted using the Cramer-Rao error bound for retrievals of multichannel sinusoidal signal phases in the presence of satellite radiometer measurement noise. The simulations also show that wind direction information significantly benefits the retrieval of wind speed by resolving small azimuthal brightness perturbations that would otherwise produce wind speed errors.

The OWI experiment also resulted in several innovations useful for spaceborne passive microwave wind direction sensing. The use of a 1 GS/sec three-level digital polarization correlator in the PSR showed for the first time that accurate calibrated polarimetric measurements can be made using only standard unpolarized ambient and cold calibration targets. Such calibration standards are readily available in a space environment. The correlator demonstration also required the development of new expression for the response of a digital correlator to the input correlation coefficient, correlator A/D converter offset, and A/D converter hysteresis. The utility of a polarized calibration standard for ground-based pre-flight calibration of the three first three Stokes parameters was demonstrated. Such a calibration standard should be useful for pre-launch calibration of future polarimetric sensors such as WindSat and CMIS.