Quinn P. Remund

Sea ice extent mapping using Ku band scatterometer data

Although spaceborne scatterometers such as the NASA scatterometer have inherently low spatial resolution, resolution enhancement techniques can be used to increase the utility of scatterometer data in monitoring sea-ice extent in the polar regions, a key parameter in the global climate. The resolution enhancement algorithm produces images of A and B, where A is the normalized radar backscatter coefficient σO at 40° incidence and B is the incidence angle dependence of σO. Dual-polarization A and B parameters are used to identify sea ice and ocean pixels in composite images. The A copolarization ratio and vertically polarized B are used as primary classification parameters to discriminate between sea ice and open ocean. Estimates of the sea-ice extent are obtained using linear and quadratic (Mahalanobis distance) discriminant boundaries. The distribution parameters needed for the quadratic estimate are taken from the linear estimate. The σO error variance is used to reduce errors in the linear and Mahalanobis ice/ocean classifications. Noise reduction is performed through binary image region growing and erosion/dilation techniques. The resulting edge closely matches the NASA Team algorithm special sensor microwave imager derived 30% ice concentration edge. A 9-month data set of global sea-ice extent maps is produced with one 6-day average map every 3 days.

An iterative approach to multisensor sea ice classification

Characterizing the variability in sea ice in the polar regions is fundamental to an understanding of global climate and the geophysical processes governing climate changes. Sea ice can be grouped into a number of general classes with different characteristics. Multisensor data from NSCAT, ERS-2, and SSM/I are reconstructed into enhanced resolution imagery for use in ice-type classification. The resulting twelve-dimensional data set is linearly transformed through principal component analysis to reduce data dimensionality and noise levels. An iterative statistical data segmentation algorithm is developed using maximum likelihood (ML) and maximum a posteriori (MAP) techniques. For a given ice type, the conditional probability distributions of observed vectors are assumed to be Gaussian. The cluster centroids, covariance matrices, and a priori distributions are estimated from the classification of a previous temporal image set. An initial classification is produced using centroid training data and a weighted nearest-neighbor classifier. Though validation is limited, the algorithm results in an ice classification that is judged to be superior to a conventional k-means approach.

A cloud-removal algorithm for SSM/I data

Microwave radiometers, while traditionally utilized in atmospheric and oceanic studies, can also be used in land surface applications. However, the problem of undesirable atmospheric effects caused by clouds and precipitation must be addressed. In this paper, temporal composite surface brightness images are generated from special sensor microwave/imager (SSM/I) data with the aid of new algorithms to eliminate small-scale distortion caused by clouds or precipitation. Mean, second-highest value, modified maximum average (MMA), and hybrid compositing algorithms are compared. The effectiveness of each algorithm is illustrated through simulation and real data distribution analysis. The results show that the second-highest value algorithm is biased high. MMA provides a more accurate brightness temperature estimate in areas of atmospheric distortion, while the mean is superior in regions with little or no distortion. A hybrid algorithm is developed that is a combination of MMA and mean. It utilizes the strengths of both to create a superior algorithm for regions with varying levels of distortion. Uses of composite images produced by these algorithms include studies of vegetation change, land cover classification, and surface parameter extraction.

Large-scale inverse Ku-band backscatter modeling of sea ice

Polar sea ice characteristics provide important inputs to models of several geophysical processes. Microwave scatterometers are ideal for monitoring these regions due to their sensitivity to ice properties and insensitivity to atmospheric distortions. Many forward electromagnetic scattering models have been proposed to predict the normalized radar cross section (/spl sigma//spl deg/) from sea ice characteristics. These models are based on very small scale ice features and generally assume that the region of interest is spatially homogeneous. Unfortunately, spaceborne scatterometer footprints are very large (5-50 km) and usually contain very heterogeneous mixtures of sea ice surface parameters. In this paper, we use scatterometer data in a large-scale inverse modeling experiment. Given the limited data resolution, we adopt a simple geometric optics forward-scattering model to analyze surface and volume scattering contributions to observed Ku-band signatures. A model inversion technique based on recursive optimization of an objective function is developed. The result is a least squares estimate of three surface parameters: the power reflection coefficient at nadir, the rms surface slope, and the volume scattering albedo. Simulations demonstrate the performance of the method in the presence of noise. The inverse model is implemented using Ku-band image reconstructed data collected by the National Aeronautics and Space Administration scatterometer. The results are used to analyze and interpret /spl sigma//spl deg/ phenomena occurring in the Antarctic and the Arctic.

Sea ice mapping algorithm for QuikSCAT and Seawinds

Polar sea ice extent is an important input to global climate models and is considered to be a sensitive indicator of global climate change. Studies have shown that Ku-band scatterometer data are sensitive to the presence of sea ice. An algorithm has been developed for sea ice extent detection using data from the NASA scatterometer (NSCAT). This paper discusses the extension of that algorithm to data from future scatterometers, QuikSCAT and Seawinds. Simulated Seawinds data are generated from NSCAT data. Experiments are conducted using Seawinds data as inputs to the NSCAT algorithm. The results show that these data can be used to estimate the ice edge although with a lower degree of accuracy than when NSCAT /spl sigma//sup o/ data are used. While NSCAT requires 6 days of data to effectively implement the algorithm, Seawinds will only require 1-2 days of data due to the wider swath, lack of a nadir gap, and better single pass cell overlap.

Polar sea-ice classification using enhanced resolution NSCAT data

The NASA scatterometer (NSCAT) collected Ku-band scatterometer measurements from September 1996 to June 1997. These data are converted high resolution six day images of the polar regions through the use of the scatterometer image reconstruction with filter (SIRF) algorithm. SIRF produces images of A and B where A is /spl sigma//sup 0/ at 40/spl deg/ incidence and B is the incidence angle dependence of /spl sigma//sup 0/. A simple four-dimensional classification technique is proposed which uses the dual polarization parameters A/sub v/, A/sub h/, B/sub v/, and B/sub h/. A k-means clustering classification can be used to separate pixels of the images with differing scattering mechanisms. This method also adapts to the seasonal characteristics of cluster migration by converging to the locally optimal cluster centroids. While validation data was not available at the time of this writing, the method is shown to have high correlation with the NSIDC SSM/I derived multiyear ice maps.

The ozone mapping and profiler suite (OMPS): on-orbit calibration design

The Ozone Mapping and Profiler Suite (OMPS) will collect total column and vertical profile ozone data and continue the daily global data produced by the current operational satellite monitoring systems, the Solar Backscatter Ultraviolet radiometer (SBUV/2) and the Total Ozone Mapping Spectrometer (TOMS), but with higher fidelity. The collection of this data will contribute to fulfilling US treaty obligations to monitor ozone depletion for the Montreal Protocol. OMPS has been selected to fly on the National Polar-Orbiting Operational Satellite System (NPOESS) spacecraft - the next generation of polar orbiting environmental satellites. The first OMPS flight unit will fly on the NPOESS Preparatory Project (NPP) spacecraft. On-orbit calibration of the OMPS instruments is critical to maintaining quality data products. A number of signal corrections and calibrations are applied on-board the sensor and in ground processing to account for instrument non-idealities and to convert measured digital signals to calibrated radiances and irradiances. Three fundamental on-orbit calibration measurements are made to provide the required data to perform the radiometric calibration and trending.

Discrimination of Africa's vegetation using reconstructed ERS-1 imagery

Because of persistent cloud cover typical of equatorial regions, active microwave imagery such as ERS-1 SAR is preferred over high-resolution visible and near infrared spaceborne sensors. However, no active microwave instrument designed to provide continental coverage at resolutions similar to the popular AVHRR has ever flown. Satellite scatterometers are calibrated active microwave radar instruments originally designed to measure the radar backscatter of the oceans surface under all-weather conditions. The European Space Agency successfully launched its ERS-1 satellite into a quasi-polar mission-adjustable orbit in the summer of 1991. The instrument payload included the Active Microwave Instrument (AMI) which is capable of operating in a wind scatterometer mode (5.3 GHz) for the production of wind vector products. However, an image reconstruction technique developed by D.G. Long et al. (1993) has improved the AMI scatterometer resolution to 14 km (from 50 km), making it a candidate for coarse monitoring of cloudy global areas such as the Arctic, Antarctic, and continental equatorial forests. The goal of the research reported in this paper is to evaluate reconstructed ERS-1 scatterometry for discriminating between vegetation classes of continental Africa.

Iterative estimation of Antarctic sea ice extent using SeaWinds data

Polar sea ice extent is an important input to global climate models and is considered to be a sensitive indicator of climate change. An algorithm has previously been developed for sea ice extent detection using data from the NASA scatterometer (NSCAT). This paper discusses the extension of that algorithm to data from SeaWinds on QuikSCAT. Due to differences in the two instruments, several modifications to the algorithm are needed for adaptation to SeaWinds data. Enhanced resolution, reconstructed SeaWinds imagery is generated using the scatterometer image reconstruction (SIR) algorithm. Images of the modified polarization ratio, h-pol /spl sigma//sup 0/, and the hand v-pol /spl sigma//sup 0/ estimate error standard deviations are used in the discrimination. An iterative maximum likelihood classifier is developed to segment sea ice and open ocean pixels. Residual classification errors are reduced through binary image processing techniques. The resulting ice extent estimates are highly correlated with the NASA Team algorithm 30% ice concentration edges derived from SSM/I data.

Validation of the SIRF resolution enhancement algorithm for scatterometer data using SAR imagery

The inherently low resolution of spaceborne scatterometer measurements, limits the utility of this data in studies of land and ice surfaces. The scatterometer image reconstruction (SIR) algorithm was developed to increase the resolution of reconstructed scatterometer imagery by using multiple passes of the satellite. SIRF imagery created from data collected by the NASA scatterometer (NSCAT) is generated and compared to ERS-2 and RADARSAT SAR imagery. The comparisons show that SIRF images have higher resolution than nonenhanced gridded images. Low-pass filtered strips from the SAR image as well as small point targets are used to show that the effective resolution of the NSCAT imagery is on the order of 10 km.