VISUALIZATION: Advanced Weather Data Visualization

可视化：高级天气数据可视化

• 批准号: 0500467
• 负责人: David Ebert
• 金额: --
• 依托单位: Purdue University
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2003
• 资助国家: 美国
• 起止时间: 2003-10-20 至 2007-03-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0500467&HistoricalAwards=false
• 关键词: VISUALIZATION; Advanced; Weather; Data; Visualization

The atmospheric science community requires visualization of observed, measured, and simulated data for accurate analysis of the atmosphere and improved weather prediction. Unlike many scientific communities, weather observers and atmospheric scientists rely heavily on important visual cues in the atmosphere to determine the potential severity of many storms.However, the current state-of-the-art in weather visualization from systems such as Vis5D, VisAD, or D3D, lack important visual information that is crucial for atmospheric scientists to fully understand the development and evolution of weather systems.Recognizing the importance of these visual cues, this project will significantly enhance the visualization of weather data through the development of innovative software techniques that will provide more accurate and effective visual representations of weather data. Simple visualization practices, such as depth cueing, isosurface texturing, volume shading, shadows, and correct natural color effects (such as sunlight) are absent in current weather data visualization software. While advanced computer graphics applications (e.g., movie production) have effectively used these techniques for some time, they have yet to be applied in a robust way to weather data. In this project, we will not only fill this gap to create improved, visually accurate weather data visualization, but also increase the quantity and clarity of the information conveyed from the resulting visualizations. Using mature numerical weather prediction software, the Advanced Regional Prediction System (ARPS), to generate numerically simulated severe weather events, new software techniques will be developed to enhance the visualization of this data and begin a new era in weather data visualization.Beyond current capabilities of standard isosurfaces, scalar volume renderings, and two-dimensional images lies important rendering capabilities for weather visualization, such as shaded volumes, shadows, light-transport, and simulated natural cloud modeling. In this project, we will develop, enhance, and apply these techniques to atmospheric data in ways which have yet to be attempted.The primary goal of our research is to produce visually accurate images of weather model data that will provide more accurate information than current methods and use the same cognitive model and analysis process as the forecasters already use, allowing them to increase their effectiveness. We will additionally develop techniques to effectively incorporate non-visual data and allow the selective visualization of the visual / non-visual weather data to enable better understanding of the relationships between these variables and quantities. Our goal is to develop these improved techniques, while also allowing interactive exploration of the observed, measured, and model data.Through the use of programmable graphics hardware with three-dimensional texture-mapping, we will implement techniques for interactive visually accurate weather visualization with low-albedo illumination, physics-based atmospheric scattering and attenuation, and volumetric shadowing. We will also implement slower high-albedo illumination models at coarser resolutions to give approximate multiple scattering effects and utilize this scattering information in the illumination calculation per-pixel fragment through three-dimensional texture mapping hardware. We will use perceptually motivated mapping of non-visual weather quantities (e.g., temperature, dewpoint, wind, atmospheric pressure, vorticity) to glyphs, particles, and isosurfaces to provide more information in an easily understandable manner, extending on our previous work in rceptually-motivated glyph rendering, fast isosurface rendering, and volume illustration. Given the capabilities of current graphics hardware, we won't be able to produce truly visually accurate images and animations of time-varying atmospheric data for at least the first half of the project, although we expect to be able to produce good approximations at interactive rates. We also plan to incorporate simple key-frame recording tools into the visualization system for off-line generation of atmospheric visualizations.The weather models produced contain multiple variables at each spatial location. By employing scientific-based combinations of these variables, it is possible to localize specific features contained in these models. We will extend our preliminary work in the development of multi-dimensional transfer function methods for multivariate data to effectively convey information from this complex model data. This improved interactive weather visualization system will increase the effectiveness of atmospheric analysis, improve severe storm forecasting, and enhance the formulation, parameterizations, and physics of numerical weather prediction models. Additionally, it will improve the training of weather observers and atmospheric science students (both undergraduate and graduate), and provide understandable animations to help in basic weather education at the K-12 level. The ultimate goal of this research is to produce a visually accurate, interactive rendering of a numerical severe thunderstorm simulation, thereby enhancing the ability of both the scientist and general user to discover and explore atmospheric processes in an unprecedented way.

大气科学界需要可视化观察到的，测量和模拟数据，以准确分析大气并改善天气预测。与许多科学社区不同，天气观察者和大气科学家在很大程度上依赖大气中的重要视觉提示来确定许多风暴的潜在严重性。但是，从系统中，Vis5D，Visad或D3D等系统的最新天气可视化的当前最新，缺乏重要的视觉信息，这对于大气科学的开发和进化的影响很大，这些信息是为了使这些视觉范围的重要性重要，这些信息actigentionss的发展范围很大程度上是在天气中的重要性。通过开发创新软件技术来可视化天气数据，这些技术将提供更准确，有效的天气数据的视觉表示。在当前的天气数据可视化软件中，不存在简单的可视化实践，例如深度提示，等距纹理，音量阴影，阴影和正确的自然颜色效果（例如阳光）。 虽然先进的计算机图形应用程序（例如，电影制作）有效地使用了这些技术，但它们尚未以可靠的方式应用于天气数据。在这个项目中，我们不仅会填补这一空白，以创建改进的，视觉上准确的天气数据可视化，而且还增加了从结果可视化中传达的信息的数量和清晰度。 Using mature numerical weather prediction software, the Advanced Regional Prediction System (ARPS), to generate numerically simulated severe weather events, new software techniques will be developed to enhance the visualization of this data and begin a new era in weather data visualization.Beyond current capabilities of standard isosurfaces, scalar volume renderings, and two-dimensional images lies important rendering capabilities for weather visualization, such as shaded volumes, shadows,轻型传输和模拟自然云建模。在这个项目中，我们将以尚未尝试的方式开发，增强和应用这些技术到大气数据中。我们的研究的主要目标是生成视觉上准确的天气模型数据图像，这些图像将比当前方法提供更准确的信息，并使用相同的认知模型和分析过程与预报员已经使用，从而提高了他们的有效性。我们还将开发有效合并非视觉数据的技术，并允许对视觉 /非视觉天气数据的选择性可视化，以更好地理解这些变量和数量之间的关系。我们的目标是开发这些改进的技术，同时还允许对观察到的，测量和模型数据进行交互式探索。通过使用三维纹理映射的可编程图形硬件的使用，我们将实施可视化的互动技术，以与低度艾尔比托微光进行互动准确的天气可视化，并具有低度艾尔比托，基于物理学的大气散射，基于物理学的大气散射和累计和量身定期，并构成了效率。我们还将在更粗的分辨率下实现较慢的高α照明模型，以通过三维纹理映射硬件在照明计算中提供近似的多个散射效果，并利用这些散射信息。 We will use perceptually motivated mapping of non-visual weather quantities (e.g., temperature, dewpoint, wind, atmospheric pressure, vorticity) to glyphs, particles, and isosurfaces to provide more information in an easily understandable manner, extending on our previous work in rceptually-motivated glyph rendering, fast isosurface rendering, and volume illustration.鉴于当前图形硬件的功能，我们将至少在项目的前半部分中产生真正的视觉精确图像和动画，尽管我们希望能够以交互速率产生良好的近似值。我们还计划将简单的钥匙框记录工具纳入可视化系统中，以进行离线生成大气可视化。产生的天气模型在每个空间位置都包含多个变量。通过采用这些变量的基于科学的组合，可以定位这些模型中包含的特定特征。我们将在开发多变量数据的多维传递函数方法中扩展我们的初步工作，以从此复杂模型数据中有效传达信息。这种改善的交互式天气可视化系统将提高大气分析的有效性，改善严重的风暴预测，并增强数值天气预测模型的配方，参数化和物理学。此外，它将改善对天气观察家和大气科学专业学生（本科生和研究生）的培训，并提供可理解的动画，以帮助K-12水平的基本天气教育。这项研究的最终目标是对数值严重的雷暴模拟产生视觉上准确的互动渲染，从而增强科学家和普通用户以前所未有的方式发现和探索大气过程的能力。