Michael H. Bettenhausen

A nonlinear optimization algorithm for WindSat wind vector retrievals

WindSat is a space-based polarimetric microwave radiometer designed to demonstrate the capability to measure the ocean surface wind vector using a radiometer. We describe a nonlinear iterative algorithm for simultaneous retrieval of sea surface temperature, columnar water vapor, columnar cloud liquid water, and the ocean surface wind vector from WindSat measurements. The algorithm uses a physically based forward model function for the WindSat brightness temperatures. Empirical corrections to the physically based model are discussed. We present evaluations of initial retrieval performance using a six-month dataset of WindSat measurements and collocated data from other satellites and a numerical weather model. We focus primarily on the application to wind vector retrievals.

WindSat radio-frequency interference signature and its identification over land and ocean

Radio-frequency interference (RFI) in the spaceborne multichannel radiometer data of WindSat and the Advanced Microwave Scanning Radiometer-EOS is currently being detected using a spectral difference technique. Such a technique does not explicitly utilize multichannel correlations of radiometer data, which are key information in separating RFI from natural radiations. Furthermore, it is not optimal for radiometer data observed over ocean regions due to the inherent large natural variability of spectral difference over ocean. In this paper, we first analyzed multivariate WindSat and Scanning Multichannel Microwave Radiometer (SMMR) data in terms of channel correlation, information content, and principal components of WindSat and SMMR data. Then two methods based on channel correlation were developed for RFI detection over land and ocean. Over land, we extended the spectral difference technique using principal component analysis (PCA) of RFI indices, which integrates statistics of target emission/scattering characteristics (through RFI indices) and multivariate correlation of radiometer data into a single statistical framework of PCA. Over ocean, channel regression of X-band can account for nearly all of the natural variations in the WindSat data. Therefore, we use a channel regression-based model difference technique to directly predict RFI-free brightness temperature, and therefore RFI intensity. Although model difference technique is most desirable, it is more difficult to apply over land due to heterogeneity of land surfaces. Both methods improve our knowledge of RFI signatures in terms of channel correlations and explore potential RFI mitigation, and thus provide risk reductions for future satellite passive microwave missions such as the NPOESS Conical Scanning Microwave Imager/Sounder. The new RFI algorithms are effective in detecting RFI in the C- and X-band Windsat radiometer channels over land and ocean.

A credit assignment approach to fusing classifiers of multiseason hyperspectral imagery

A credit assignment approach to decision-based classifier fusion is developed and applied to the problem of land-cover classification from multiseason airborne hyperspectral imagery. For each input sample, the new method uses a smoothed estimated reliability measure (SERM) in the output domain of the classifiers. SERM requires no additional training beyond that needed to optimize the constituent classifiers in the pool, and its generalization (test) accuracy exceeds that of a number of other extant methods for classifier fusion. Hyperspectral imagery from HyMAP and PROBE2 acquired at three points in the growing season over Smith Island, VA, a barrier island in the Nature Conservancys Virginia Coast Reserve, serves as the basis for comparing SERM with other approaches.

A statistical approach to WindSat ocean surface wind vector retrieval

WindSat is the first space-based polarimetric microwave radiometer. It is designed to evaluate the capability of polarimetric microwave radiometry to measure ocean surface wind vectors from space. The sensor provides risk reduction for the National Polar-orbiting Operational Environmental Satellite System Conical Scanning Microwave Imager/Sounder, which is planned to provide wind vector data operationally starting in 2010. The channel set also enables retrieval of sea surface temperature, and columnar water vapor and cloud liquid water over the oceans. We describe statistical algorithms for retrieval of these parameters, and a combined statistical/maximum-likelihood estimator algorithm for retrieval of wind vectors. We present a quantitative analysis of the initial wind vector retrievals relative to QuikSCAT wind vectors.

Passive Remote Sensing of Sea Foam using Physically-Based Models

Demonstrating the high variability of sea foam fraction (whitecap coverage) on the ocean surface, we justify the need for estimating foam coverage with space-based passive remote sensing as an alternative to conventional photographic measurements. We outline the concept and prove the feasibility of a method for deriving foam coverage from satellite measurements. The encouraging results of an initial implementation motivate further work addressing its drawbacks. We describe data and modeling improvements to the initial algorithm and report results on satellite-derived whitecap coverage using physically based models.

Identification of Ocean-Reflected Radio-Frequency Interference Using WindSat Retrieval Chi-Square Probability

Ocean retrievals using passive microwave radiometers are sensitive to small fluctuations in ocean brightness temperatures. As such, the signals emanating from geostationary satellites that reflect off the ocean surface can result in large errors in ocean retrievals. Since geostationary communication satellites maintain fixed positions above the Earth and constantly transmit to predetermined regions while most other error sources, e.g., precipitation, are transient, time-averaged retrieval error statistics can be used to identify regions of measurements contaminated with radio-frequency interference (RFI). This letter describes a new method of identifying regions of ocean where ocean retrievals are affected by geostationary communication (television) satellites by using geophysical retrieval chi-square probability (goodness-of-fit) estimates. A three-month time-averaged collection of retrieval chi-square estimates is used to identify regions of the ocean where RFI may be present. This information is combined with information on geostationary satellite bandwidths, locations, and antenna contours to identify the source of the RFI. A mask derived from the analysis is used, in conjunction with satellite geometry calculations, to flag individual channels for RFI. These channels can then be ignored in the geophysical retrieval processing in order to produce uncontaminated ocean retrievals.

Automatic land-cover classification of a barrier island in the Virginia Coast Reserve using HYMAP imagery: an intercomparison of methods

Automatic land-cover maps were developed from HYMAP hyperspectral imagery acquired May 8, 2000 over Smith Island, VA in the Virginia Coast Reserve. Both unsupervised and supervised classification approaches were used to create these products. Ground surveys made by us in late October and early December, 2000 provided ground truth data for various land-cover types. We used GPS data from these surveys to extract spectral end-members used in supervised land-cover classification models. Both approaches to the classification problem produced consistent results for some categories such as Spartina alterniflora, although there were differences for other categories.

The Impact of Radio-Frequency Interference on WindSat Ocean Surface Observations

To study the effects of radio-frequency interference (RFI) on remote sensing data, we examined five years of WindSat ocean observations from regions affected by reflected X-band emissions from geostationary communication satellites. We compared measured brightness temperatures to modeled brightness temperatures obtained using mitigated retrievals as input to the WindSat parameterized radiative transfer model, demonstrating a considerable contribution to the radiometric measurements from interference. Comparisons of potentially contaminated retrievals of sea surface temperature (SST), wind speed, and wind direction with surface reanalyses confirmed that the presence of RFI can bias retrievals while also increasing retrieval uncertainty. Mitigation removed most of the biases. The quality of mitigated SST and wind speed retrievals was comparable to uncontaminated retrievals; however, the performance of first-rank wind directions was noticeably degraded. Recommendations are given on the use of contaminated and mitigated data, from radiance assimilation to climate studies.

Automatic detection of an invasive plant species on a barrier island in the Virginia

Invasive plant species such as Phragmites australis pose a threat to coastal habitats. This study compares a number of methods for automatically detecting Phragmites using a HYMAP scene of Smith Island, Virginia, acquired on May 8, 2000, and an IKONOS scene of the same region acquired on June 6, 2000. The best model for the phragmites distributions used both spectral and spatial-spectral input windows from HYMAP and combined Projection Pursuit (PP) for feature extraction and dimensionality reduction with the traditional ISODATA clustering technique. Although not perfect, this hybrid, unsupervised approach produced the lowest false alarm rate when compared with supervised learning models. Supervised algorithms found phragmites in the open and along the swale edges, but had inordinately high false alarm rates when compared with the PP-ISODATA hybrid.

The WindSat polarimetric radiometer and ocean wind measurements

The wind vector affects a broad range of naval missions, including strategic ship movement and positioning, aircraft carrier operations, aircraft deployment, effective weapons use, underway replenishment, and littoral operations. Furthermore, accurate wind vector data aids in short-term weather forecasting, the issuing of timely weather warnings, and the gathering of general climatological data. WindSat is a satellite-based multi-frequency polarimetric microwave radiometer developed by the Naval Research Laboratory for the U.S. Navy and the National Polar-orbiting Operational Environmental Satellite System (NPOESS) Integrated Program Office (IPO). It is designed to demonstrate the capability of polarimetric microwave radiometry to measure the ocean surface wind vector from space. The sensor provides risk reduction for the development of the Conical Microwave Imager Sounder (CMIS), which is planned to provide wind vector data operationally starting in 2010