Civilization as we know it is threatened by environmental degradation on a global scale which, if continued unchecked, will soon reach a critical mass that will devastate the Earth's ecological life support systems just as surely as could a full scale nuclear war. We summarize first-hand observations of the current ecological situation in Eastern Europe and around the world. We report on critical situations such as new diseases, the water supply and its contamination that have come to light only in recent years. We find that our greatest threat today is not from other nations trying to conquer us, but from our own capacity to destroy our environment and with it ourselves along with our planetary flora and fauna. Indeed, this is the greatest enemy of our times, one that we need to recognize and fight before we lose a battle that we did not foresee. To counter this imminent threat, we propose a new post-war role for the US military in a global environmental defense initiative in which our military capabilities would be restructured into a planetary environmental defense force.

Polarimetry covers the entire electromagnetic spectrum and 'Low Frequency Polarimetry' applies to the polarimetric vector field nature of the magnetic and electric fields within the Ultra-Low Frequency (ULF: below 3 Hz), the Extremely Low Frequency (ELF: 3 Hz to 3 kHz) and the Very-Low Frequency (VLF: 3 kHz to 30 kHz) spectral regions. The acoustic complementary spectral regions are identified as Infrasonic (.01 - to 20 Hz) and Audiosonic (20 Hz - to - 20 kHz). These low frequency waves play an important role in coupled lithospheric-ionospheric wave interaction especially in the upper ULF (above .01 Hz) to lower VLF (below 10 kHz) for which the radiowaves display pronounced polarimetric effects, i.e., strong field component dependence. In this paper, an introduction to this exciting re-emerging field is presented, paying major emphasis on recent findings regarding the observation of elevated electromagnetic low frequency noise, strongly peaking in the 0.1 to 5 Hz spectral region before an earthquake seismic distress event. Also, polarimetric signatures resulting from the interaction of large space vehicles with the ionosphere will be discussed.

The fields of radar and optical polarimetry are compared. Both fields have defined scattering matrices that describe the coherent scattering of monochromatic electric fields. This paper compares the mathematics of the two disciplines by redeveloping Kennaugh's optimum polarization theory in the context of optical polarimetry. In addition, a novel definition of 'eigenpolarization' is given.

A standard Ku-band FM-CW scatterometer which can transmit in the horizontal (H) and vertical (V) states was converted to a full 'Stokes Parameters' polarimeter. The system was used to evaluate the polarization response of three cereal crops (spring wheat, durum wheat, and barley), canola and fallow fields. It was found that all the targets caused partial unpolarization (a decrease in the polarization ratio) of the transmitted state but no depolarization in that the polarized portion of the received signal had the same linear polarization as the transmitted state (H or V). As a consequence, one can recover all the polarization information about these targets from the standard scatterometer measurements. This can be stated as the 'Scatterometer Polarization Theorem', in which the co-polar scatterometer components (HH or VV) contain all the polarized power plus half the unpolarized power, while the cross-polar components (HV or VH) contain exactly half of the unpolarized power. There was more attenuation by wheat of V than H transmissions. This indicates that, for the V transmit state, the polarized component of the signal is anti-correlated with the plant water content. For canola, a randomly structured target medium, the unpolarized component of the signal correlates positively with plant moisture.

Some far-field radar imaging techniques (like spotlight-mode synthetic aperture radar and holography) are shown to have a tomographic formulation. For short radar-target distance (much shorter than the distance required in the far-field condition) the far-field computer aided tomography algorithm (CAT) may blur the reconstructed image. After studying first the near- field theory of some microwave imaging techniques, this paper shows that a unified near-field microwave tomography algorithm is possible. The received signal in SAR, which is a convolution between the spherical projection of the target and the emitted signal, can be interpreted as a Fourier transform of the spherical projection of the image as in near-field holography and tomography. Reconstruction of an SAR image may then be accomplished using a back projection algorithm. This study is an interchange of ideas between these different disciplines and it is a step towards a unified near-field algorithm. An optimal method can be used for combining the four polarimetric images (hv basis) to reduce speckle compared with the span image and single channel image.

Among the engineering scientists who have most decisively contributed toward forefront advances for the 'Development of POLARIMETRIC Radar Theory, Techniques and Target Phenomenology', Dr. Jean Richard HUYNEN stands out as one of the towering giants. This paper is dedicated to him, and it is a great honor for the author to present here a summary on some of the main Dr. J.R. HUYNEN's contributions in the development of Polarimetric Radar Techniques, including the 'Mueller Matrix Decomposition' approaches which this Senior Radar Polarimetrist considers as his 'life's main contribution'.

This leads to three world-views. One is the world of basic symmetry we live in, while the other is applicable to an exotic world where there is a preference for helices with right sense. The other exotic world caters to a world composed of helices with left sense. Hence we have arrived at a unique N-target decomposition which caters to the world of basic symmetry we live in. In this world it is still possible to have helices of any kind, but predominantly the symmetry is preferred, i.e., the N-target which we separate out of the data has no symmetry (A^N_o equals 0) and is completely non-symmetric (hence the name N-target). Other types of decomposition are based on eigenvalues and eigentargets. These are not transparent in general as to their physical significance. All these physically based arguments lead us to conclude that the N-target decomposition is unique and physically realizable in all cases.

The paper gives an overview about the activities at DLR in the field of ground based polarimetric radar imaging. Some features of the measurement equipment are given. The basic principle of the generation of complex target microwave images from scattered fields is outlined, and the derivation of target characteristic parameters as well as target characteristic polarizations is described. Special emphasis is given to the analysis of localized scattering centers. Two methods for studying bistatic scattering effects on a monostatic range are introduced. The physics of radar backscatter from randomly distributed targets is explained especially with respect to the measurement of scattering matrix time series. Further, the determination of signal degradations on Earth-satellite paths by means of the DLR weather radar is dealt with.

A succint overview of basic events in the development of polarimetry--low frequency to optical regions--is followed by a succinct summary of the basic underlying principles of 'Direct and Inverse Methods in Polarimetry', including some pertinent examples. A critical assessment of solved and unresolved problems is followed with a listing of recent workshops, providing the incentives for planning and executing a series of coordinated international meetings such as the SPIE, Radar Polarimetry Conference, San Diego, CA, 1992 July 23-24 the Second Polarimetric Radar Workshop, JIPR-2, Nantes, France, 1992 Sept 7-10, and the 'Wideband Imaging and Sensing Polarimetry: WISP', Adelaide, South Australia, 1992 August 14.

Geometrical concepts which have long been central to the Poincare sphere and Stokes vector representations are extended to include vector representations of target scattering matrices via a new derivation based on spinor algebra techniques. New results include the geometric significance of an absolute target phase, an interpretation of Graves' power scattering matrix as a four-vector, and its relationship to the complex target vector. The mathematical and physical significances of the interpretation of Stokes vector space as having a Minkowskian geometry are discussed. The consequences of Lorentz invariance are explored in an application to incoherent backscatter optimal polarimetry.

Basic principles of radar polarimetry are introduced and various optimization procedures for the propagation (scattering) range operator equation and the received power expressions are presented and compared. Based on a complete description of isolated and distributed scatterers, target classification, target-versus-clutter discrimination, and optimal contrast enhancement algorithms are derived and shown to be of great utility in the proper interpretation of POL-RAD/SAR microwave signatures in terrestrial and planetary remote sensing.

The characteristic polarization theory for the monostatic reciprocal coherent case in the cross- polarized radar channel is developed using a Stokes vector formalism. Power representation in terms of the Stokes vector and the Mueller matrix is utilized to determine the characteristic polarization states for optimal reception of the coherent wave. The optimization procedure for the channel power leads to an eigenvalue equation which explains the characteristic polarization state properties mathematically and physically. A total of six theoretical characteristic polarization states for a single target can be derived by this method. Several target examples are considered to show a perfect agreement with the results obtained by an alternative method of using the polarization transformation ratio formulation for the coherent Sinclair matrix.

In this paper we applied Cloude's decomposition to imaging radar polarimetry. We show in detail how the decomposition results can guide the interpretation of scattering from vegetated area. For multi-frequency polarimetric radar measurements of a clearcut area, the decomposition leads us to conclude that the vegetation is probably thin compared to even the C-band radar wavelength of 6 cm. For a forested area, we notice an increased amount of even number of reflection scattering at P-band and L-band, probably the result of penetration through the coniferous canopy resulting in trunk-ground double reflection scattering. The scattering for the forested area is still dominated by scattering from randomly oriented cylinders, however. It is found that these cylinders are thicker than in the case of clearcut areas, leading us to conclude that scattering from the branches probably dominate in this case.

The nearfield reflectance of a buried, dielectric object of circular shape was measured with an antenna operating in its and the object's nearfield region for two orthogonal linear polarizations. Reflectance depended on polarization. Mutual coupling between individual antennas was also measured and found to depend on polarization.

A fully polarimetric radar system consists of orthogonal dual polarized transmission mode and a dual channel receive mode that are typically set to be copolar and cross polar to the transmit state of polarization. The transmit polarization state can be switched alternately between the two states. This paper presents techniques to construct instantaneous fully polarimetric response from switched systems. Three different algorithms are presented for the reconstruction of instantaneous full polarimetric response in the context of observation of atmospheric targets at microwave frequencies. The performance of the suggested algorithms is studied using data from switched dual channel polarimetric radar at C-band.

Classification, decomposition and modeling of polarimetric SAR data has received a great deal of attention in the recent literature. The objective behind these efforts is to better understand the scattering mechanisms which give rise to the polarimetric signatures seen in SAR image data. In this paper, a new approach is described, which involves the fit of a combination of three simple scattering mechanisms to the polarimetric SAR observations. The mechanisms are volume scatter from a cloud of randomly oriented dipoles, even-bounce scatter from a pair of orthogonal surfaces with different dielectric constants and Bragg scatter from a moderately rough surface. This composite scattering model is used to describe the polarimetric backscatter from naturally occurring scatterers. Results are presented of application of this new algorithm to different types of scene, including multi-frequency polarimetric SAR images of a tropical rain forest, a boreal forest, a pine forest, geologic targets, urban areas and agricultural fields. Fitting the model to polarimetric SAR data of the tropical rain forest for example, allows clear discrimination between flooded and non-flooded forest. The model can be used to estimate the overall contribution from each of the three basic scattering mechanisms for each SAR image pixel.

In this paper, an adaptive polarimetric system is proposed, which allows for the minimization of the contribution of a stationary clutter by adapting the polarization state of the receiving antenna to the polarimetric signature of the environment by using the minimum power criterion. From the knowledge of the polarimetric signature of the environment and by applying the least mean square criterion, we can detect the presence of targets and estimate their polarimetric parameters. An algorithm of CFAR (Constant False Alarm Rate) is developed to detect the presence of a radar target and of its disappearance.

When determining the nature of speckle in microwave POL-SAR image analysis, various known methods of speckle reduction are reviewed and compared for the purpose of enhancing image quality in polarimetric matched image filtering. Basic principles of the Polarimetric Matched Image Filter (PMIF) are presented and by utilizing the now widely used NASA-JPL C-Band POL-SAR data sets, collected over the San Francisco Bay area, the PMIF concept, together with various speckle reduction methods are verified and interpreted. Specific emphasis is placed on demonstrating the efficiency and usefulness of applying the PMIF method to the optimization of image discriminants in POL-SAR image analysis.

During normal SAR processing, a flat earth is assumed when performing radiometric corrections such as antenna pattern and scattering element size removal. Here we examine the effects of topographic variations on these corrections. Local slopes will cause the actual scattering element size to be different from that calculated using the flat earth assumption. It is shown that this effect, which is present for both airborne and spaceborne SAR data, may easily cause calibration errors on the order of a dB. In the case of airborne systems, the errors introduced by assuming a flat earth during antenna pattern removal are also significant; errors of several dB can easily result. The effect of topography on antenna pattern removal is expected to be negligible for spaceborne SARs.

In this paper, a systematic theoretical analysis of polarization filtering for a bistatic system is developed for two applications: maximization of signal to noise ratio and discrimination between two target types. The described method finds the optimum receive antenna polarization analytically but relies on a numerical approach to find the optimum transmit antenna polarization. The method uses the Stokes matrix representation, and therefore applies to analyze the partially polarized scattered field from extended targets. Examples of the technique are presented for the NASA CV990 polarimetric L-band radar. Image enhancement filters maximizing the signal-to-noise ratio are developed for different noise characteristics and different target types (urban and forest). A filter is also developed to maximize the power ratio between urban and natural targets. Results show that the filter maximizing the contrast between urban and forest area is essentially the same as the one maximizing the contrast between urban and ocean areas.

Two supervised classification procedures are presented and applied to relative polarimetric SAR images in order to identify the different Earth terrain components. The first one is the classical Bayes classifier, and the second is an original polarimetric method based on a neural network modelization. The subject of this paper is to show that it is possible to classify polarimetric data by using neural network techniques, without knowing the a-priori statistic distributions of the different classes. The purpose being to show that POLARIMETRY and IA theories can become complementary sciences.

Polarimetric measurements made by Synthetic Aperture Radar (SAR) may be in some cases, depending on the polarimetric response of distributed targets to be imaged, severely limited in their accuracy due to the joint effect of range ambiguities and weak crosspolarized signal response. Due to the utilization of alternate transmission of pulses at orthogonal polarizations, each ambiguous swath gives rise to one different kind of interference, depending whether its order is even or odd. Interference arising from even-order ambiguous swaths, differently from that arising from odd-order swaths, is generated by pulses transmitted on the same polarization channel of the pulse soliciting the desired echo signal, that they corrupt. Evidently, interference arising from odd-order swaths and affecting crosspolar measurements is most harmful, together with that arising from zones at low incidence angle, which carries a strong reflectivity contribution to the total interference on the desired signal. The paper discusses the utility of appropriate interpulse phase coding strategies, depending on the SAR geometry, than can be devised and utilized in the polarimetric interleaved-pulse measurement technique, with the task to reduce the interference generated by range ambiguities and affecting those target scattering matrix elements, whose measurement is expected to be most critical.

A brief summary of early polarimetric radar development pertinent to radar meteorology in North America over the past forty years (1950 - 1991) is presented. Rather than providing an in-depth assessment, only important events leading to major advances are succinctly reviewed by providing the pertinent references.

The necessity for using a three parameter raindrop size distribution in weather radar applications is quantitatively established. Using detailed theory for the microphysical interpretation of averaged complex correlations, an algorithm for extending the dual polar method for radar determination of a 3 parameter raindrop-size distribution is detailed with an illustrative example.

Addressed is the transformation of the polarization state of an electromagnetic wave as it propagates through a rain field. A random medium formulation reflecting the well known anisotropicities of a rain fall is used. Presented are solutions for the mean polarization state and its variance, along with a preliminary example for linear and circular incidence.

Mueller matrices are important in radar polarimetry as they characterize a target or a random medium at the incident frequency. Many random media which exist in nature have random rough surfaces and randomly distributed spherical particles. Two models for a slab of such random media are presented. Both models contain randomly distributed spherical particles. Planar surfaces are considered for the first model and Gaussian statistical rough surfaces are considered for the second one. Kirchhoff rough surface scattering theory is used to calculate the scattering of optical waves from the rough surface. Vector radiative transfer theory is used to calculate the volume scattering due to the particles. The scattered wave is computed for an arbitrarily polarized incident wave for optical thickness (tau) equals 0.1 to 5, size parameters ka equals 0.4 to 1 and various surface roughnesses. The scattered wave is expressed in terms of the modified Stokes vectors and then used to construct the Mueller matrices. Polarization signatures are constructed from the Mueller matrices of the reflection side at the backscattering direction. The Mueller matrices are found to have some symmetrical properties and there are eight nonvanishing elements.

A formulation for internal emission from a thermal scattering layered medium with irregular boundaries is developed. The matrix doubling method is utilized to obtain the upward and downward thermal intensities inside the scattering layer. The scattering layer is assumed to be a slab imbedded with randomly-positioned thermal scattering particles, and the layer top and bottom boundaries are assumed to be random rough surfaces. The developed emission model may be used for the understanding of the internal emission problem where a radiometer is located inside an emitting medium with irregular rough boundaries.

This essay documents an informal talk about the central theme in the author's research career. That has mainly related to the visual physiology and orientation of aquatic animals, particularly with regard to underwater polarized light. This required pioneer measurements of underwater polarized light patterns, proof that oriented behavior could be determined by e- vector direction independently of intensity patterns or other secondary clues and a demonstration of the retinal dichroic mechanism involved, at least in crustacean compound eyes. The relevant visual data processing by two orthogonal channels was also analyzed with regard to oriented swimming behavior. Some current research by others and major unsolved problems are mentioned and the relevant part of the author's bibliography is appended.