The collection and management of vast quantities of meteorological data, including satellite-based as well as ground- based measurements, is presenting great challenges in the optimal usage of this information. To address these issues, the Army Laboratory has developed neural networks for combining for combining multi-sensor meteorological data for Army battlefield weather forecasting models. As a demonstration of this data fusion methodology, multi-sensor data was taken from the Meteorological Measurement Set Profiler (MMSP-POC) system and from satellites with orbits coinciding with the geographical locations of interest. The MMS Profiler-POC comprises a suite of remote sensing instrumentation and surface measuring devices. Neural network techniques were used to retrieve temperature and wind information from a combination of polar orbiter and/ or geostationary satellite observations and ground-based measurements. Back-propagation neural networks were constructed which use satellite radiances, simulated microwave radiometer measurements, and other ground-based measurements as inputs and produced temperature and wind profiles as outputs. The network was trained with Rawinsonde measurements used as truth-values. The final outcome will be an integrated, merged temperature/wind profile from the surface up to the upper troposphere.
Our approach in addressing complex terrain scenarios is to utilize a high resolution wind (HRW) model that will provide a high resolution microscale analysis of the surface layer horizontal wind and temperature fields. The wind model uses Gauss' Principle of Least Constraints for a variational adjustment of an initial estimated wind field in a single surface layer to conform with terrain structure, mass conservation, and buoyancy forces. From meteorological measurements taken at a single location, HRW can calculate wind vectors over a 5 x 5 km area with grid points spaced 100 meters apart. These micrometeorological data as well as canopy information can then be used as inputs to the Rachele-Tunick model to calculate optical turbulence parameters at each grid point within the microscale area. The output will be 2D optical turbulence contours from which integrated line of sight turbulence calculations can be made.
It is well known that the presence of aerosols in the atmospheric boundary layer can have a significant impact on electromagnetic propagation, and the underlying physical processes involving extinction, multiple scattering, and thermal emission are reasonably well understood. In this paper we examine a related, but less well understood, aspect which we term aerosol-induced `radiative damping' that can alter the local atmospheric stability and the vertical profiles of temperature and humidity which, in turn, can alter the vertical profiles of optical turbulence and hence image propagation.
Microwave radiation emitted by the atmosphere contains information about the composition and temperature structure of the atmosphere. Given the vertical temperate profile and humidity structure of the atmosphere, it is possible to compute the microwave radiance received by an upward-looking surface-based radiometer operating in the 20 - 60 GHz range. The inverse problem is ill-posed and requires additional information for its solution. We have found a method for solving this problem using the techniques of neural networks. In our calculation, the network has as inputs the set of microwave brightness temperature measurements made at the surface, surface temperature, and surface pressure, or surface altitude, or both. The output is the atmospheric temperature as a function of height. The neural network computes the outputs from the inputs using its internal weights which have been established by a training process. The training consists of presenting the system with input brightness temperatures calculated from radiosonde observations, and comparison of the resulting outputs with the radiosonde temperature profiles using a backpropagation training rule. The trained system was then tested on a separate set of brightness temperatures. Similar techniques should work with many types of inverse problems.
It is desirable to determine the internal structure of clouds during smoke tests. One method for doing this involves using lidar returns from a cloud and inversion of the lidar equation. Conventionally, the latter requires a knowledge of the attenuation at the end of the path of interest. This is seldom known or measurable with sufficient accuracy. In this presentation, a new iterative inversion algorithm is described. The new algorithm uses the total attenuation through the desired path length as input, rather than the attenuation of the last range cell. The total attenuation is easy to determine accurately. For example it may he measured using the lidar and a backstop of known reflectivity placed at the end of the desired path.
A simple technique for extracting the extinction and backscatter coefficients from lidar returns signals is presented. The technique uses lidar signals from two different wavelengths taken with the same lidar system, and is based on the assumptions that the extinction coefficient at one wavelength is related to the extinction at the other wavelength, likewise for the backscatter coefficients. The technique is tested on data generated by the Nd:YAG laser 2nd (532 nm) and 3rd (355 nm) harmonics. The relationship between extinction coefficients was obtained by a power law fitting to data generated by the UVTRAN model. For the backscatter coefficient a simple proportionality law is assumed.
A prototype expert system program has been developed to assist in the identification of fluorescence materials from UV lidar returns. Features such as emission peak locations, peak broadness at half height, number of peaks, and decay lifetimes comprise the knowledge base for a particular category of substances. The expert system will ask the user for features from an unknown substance and then will compare these features with the knowledge base of known substances. Preliminary results show that the expert system is capable of discriminating single substances from a small data set of fluorescence substances with their associated features. To handle multicomponents, it is suggested that preprocessing of the data be performed in which only the peak locations of single substances and combinations of peak locations of multiple substances be searched.