R. Goldberg and F. Herrero
NASA/Goddard Space Flight Center, Greenbelt, MD
Visualization is a powerful method to study the space-time variability of energy deposition due to precipitating particles. Such methods are also useful in the study and comparison of existing models relevant to these phenomena. However, new techniques for the visual display of relevant data sets are required. Specifically, the study of effects caused by highly relativistic electron events (HRE), which have recently received wide attention, will be considered. These events provide the most energetic electron fluxes observed to date with energy deposition occurring mainly in the mesosphere. Furthermore, HRE electrons can significantly modulate the concentrations of OH and O3 in the mesosphere, which in turn can affect heating and transport processes there. The study of this phenomena may involve the analysis of remotely sensed and in situ data from sounding rockets or spacecraft including particles and fields, soundings of minor atmospheric constituents and temperature as well as comparisons with models. The diversity and complexity of these data require care in their presentation so that artifacts due to the visualization process are not introduced and erroneously interpreted as features in the data. The present format of these data is ill-suited for the study of phenomena that occur over a large continuous and non-rectilinear region (i.e., it tears the data). In addition, they are not uniformly available for the regions being studied. The data examined are generally not geographically or geomagnetically co-registered, defined on differing geometric structures, irregularly sampled and often noisy.
To be effective, visual correlation of the aforementioned data must be done in a coordinate system suitable for three-dimensional, dynamic presentation. Hence, the notion of a natural coordinate system to provide a context for three-dimensional analysis, viewing and interaction is introduced. They provided representations of the data, which are registered with relevant terrain and political boundary maps. This approach is derived from the concept of correlative visualization, where each data set to be examined is processed independently and merging takes places only at render time.
Such a methodology has applications in the analysis of archived observations as well as operational forecasting. To demonstrate the validity of this methodology for such application, the techniques being developed will be applied to several candidate data sets.
To begin to address these issues, candidate data sets acquired during a sounding rocket experiment were utilized. This effort, NASA 33.059, was launched at 21:29 UT from Poker Flat, AK (65.1 N, 147.5 W, L = 5.5) on May 13, 1990. The rocket trajectory and the geographic domain are illustrated in Figure 1. This image also shows the framework described above for interaction and visualization. Data from 13 to 258 seconds after launch were available. This experiment was designed to observe the interaction of highly relativistic electrons with the earth's upper atmosphere (mesosphere). In particular, a high level of deposition of energy from such events is believed to affect the composition (e.g., OH and ozone) as well as heating and transport in the mesosphere.
The rocket carried a number of instruments. Of primary interest here are the measurements made by two solid-state electron detectors, which measure electrons in the energy range of .09 - 3.7 MeV. The low-energy detector (.09 - 1.0 MeV) has 12 channels while the high-energy detector (.43 - 3.7 MeV) has 8. Derived integral electron flux data from those sensors were examined. In addition, there are attitude and trajectory data for the instruments and rocket, as well as pitch angle samplings. These characteristics of the experiment are summarized in Figures 2, 3 and 4. Two periods of interest are from about 125-155 and 210-215 seconds after launch, during which particles seemed to be injected at higher flux rates.
The techniques described herein have been developed with IBM Visualization Data Explorer, a general-purpose software package for scientific data visualization and analysis. It employs a client-server architecture with an extended data-flow execution model and is available on Unix workstations (e.g., Sun, Silicon Graphics, Hewlett-Packard, IBM, DEC and Data General) and Intel-based personal computers running Windows NT. An end-user, portable application to study these data was implemented to provide direct interaction in flight coordinate system of NASA 33.059 with the ability to create a number of ancillary displays (spectrograms, time history plots). Motif widgets for indirect interaction are also available to select operation modes, display options, and data of interest.
The application built to support this research provides a main viewing enviroment in the "natural" geographic coordinate system described earlier. A topographic map with the launch point highlighted is shown with the parabolic trajectory of the rocket. Thus, direct interaction is provided in the flight coordinate system. The measurements made by the solid-state electron detector are the key data from this program. The user may select either the low- or the high-energy detector to examine. This basic level of interaction is illustrated in Figure 5. The user may select one channel to display along the rocket trajectory in the main viewing window. That same channel may optionally be displayed as a time plot, if that option from the Display Controls panel is selected. The path of the rocket is colored by the integral electron flux from a specific channel from one of the two solid-state detectors. The Solid-State Detector panel is used to select which channel and detector.
The selected channel may be interrogated, by picking a point of interest along the rocket trajectory by using Pick mode in the main viewing window, which is illustrated in Figure 6. The data value and the time since launch at that point will be displayed and will also be marked in ancillary displays.
A number of ancillary displays may be requested by the user such as a time-energy integral electron flux spectrogram of the selected solid-state detector or a time-altitude spectrogram of the selected energy deposition data. The spectrograms may be displayed as a pseudo-color image or deformed surface. The user may choose either a perceptually appropriate continuous or segmented color map. These are illustrated in the interactive application in Figures 5, 6, 7, 8 and 9, and in more detail in Figures 12, 13, 14, 15, 16 and 17. Figure 5, 6, and 7 show two-dimensional time-energy spectrograms of the integral electron flux from the high-energy detector, with one channel being shown on the rocket trajectory, and on a time plot in Figure 7.
Figure 8 now displays the spectogram as a deformed surface, but shows the correlation with the aformentioned pitch angle variation.
/
For the time-altitude spectrogram, one of four energy deposition data sets may be selected: atmospheric density, energy deposition from energetic electrons, energy deposition from bremmstrahlung x-rays, or total energy deposition. The source of the energy deposition derivation can be either the low- or high-energy solid-state detector. This capability is illustrated in Figure 9, where both the high-energy detector spectrum and the energy deposition spectrum from that detector have been chosen.
The user may choose one parameter each from two data sets from various electron flux, trajectory, attitude and pitch angle data to plot as a function of time during the rocket flight. Each data set has from one to 19 parameters. In addition, the ability to create a linear or semi-log plot is available. Examples of the interactive selection of such plots are shown in Figures 7, 8, 9 and 10.
The application also has an animation mode under the control of the Sequencer, through which the motion of the rocket may be shown along the flight path, which is illustrated in Figure 10 with data from the low-energy detector.
The Animation panel may be used to control the characteristics of the animation. The point on the path is shown by a pair of arrows. A cyan arrow shows the direction that the solid-state detector is looking. A red arrow shows the velocity of the rocket. A snapshot from such animation for the high-energy detector is show in Figure 11.
Figures 12, 13, 14, 15, 16 and 17 show detailed spectrograms of the integral electron flux observations. Figure 12, 13 and 14 are data from the low-energy detector, while Figures 15, 16 and 17 are from the high-energy detector. The former set shows that channel 6 (out of 12) from the low-energy detector is missing due to instrument failure. For each spectrum, the results are shown with three techniques: segmented colormap with perceptual ordering with a two-dimensional spectogram (Figures 12 and 15), isomorphic or perceptually continuous colormap with a two-dimensional spectogram (Figures 13 and 16), and an isomorphic colormap with a "three-dimensional" spectogram, where flux value is redundantly encoded as a linear deformation in height to yield a deformed, shaded surface (Figure 14 and 17). The segmented colormaps used in Figures 12 and 15 easily highlight the bursts of higher flux in red, observed about 125-155 and 210-215 seconds seconds after launch. While the other techniques do impart more of a continuum in representation, they only illustrate qualitatively the coarse structure of the spectra, and not the features of interest.
With easy-to-use tools to access, reorganize, realize and render yet preserve the salient characteristics of multiple data sets, a scientist can readily and appropriately scrutinize such data at many different levels through disparate techniques. The support of a plethora of visualization strategies properly coupled with powerful manipulation functions promotes the (visual) exploration and correlation of diverse data sets and thus, enables a scientist to extract knowledge from complex data. Specifically, the introduction of flight path coordinates as a metaphor for direct manipulation shows promise as a way of examining observations from sounding rocket experiments by providing a context for analysis.
Therefore, the next step is to apply these techniques to other rocket measurements of similar events as well as relevant spacecraft observations such as soundings (both constituents and dynamics) of the mesosphere. Once that is established, the integration of both classes of data as well as model results can proceed. Currently, the three-dimensional coordinate system for interaction is geographic. An obvious enhancement to consider is the ability to interactively transform the data and interaction to geomagnetic coordinates.
R. A. Goldberg, C. H. Jackman, D. N. Baker, F. A. Herrero. "Changes in Concentration of Mesospheric O3 and OH During a Highly Relativistic Electron Precipitation Event". The Upper Mesosphere and Lower Thermosphere: A Review of Experiment and Theory Geophysical Monograph 87, pp. 215-223. American Geophysical Union, 1995.
R. A. Goldberg, D. N. Baker, F. A. Herrero, C. H. Jackman, S. Kanekal and P. A. Twigg. "Mesospheric Heating during Highly Relativistic Electron Precipitation Events". J. Geomag. Geoelectr., 47, pp. 1237-1247, 1995.