By Tracy Hansen
The Smart Tools concept comprises a framework within the Graphical Forecast Editor (GFE) that allows NWS forecasters to create and modify their own edit tools for incorporating forecasting techniques specific to their local forecast area.
The initial GFE design provided forecasters more or less a "paint" program. It contained a framework and user interface for editing forecast grids, but the tools available were unable to recognize meteorological aspects. For example, if forecasters wanted to define snow amounts in areas above 7,000 feet, they would first have to find the areas over 7,000 feet and then manually paint in the various snow amounts for several forecast time periods. With approximately 10 different weather elements to define, this methodology quickly became too labor-intensive. For a typical 3-day forecast period covering an area the size of a typical forecast office domain, there are roughly 3–5 million data points that must be defined. Clearly, the GFE needed tools that would remove the tedium and make this daunting task possible.
Emergence of Smart Tools
Forecasters work more efficiently when they are allowed to think in meteorological terms. In the situation described in the above example, we need a tool that would automatically identify the areas over 7,000 ft. and set the snow amount according to meteorologically related weather elements in the system, such as temperature and precipitation amount. Forecasters need to be able to think in concepts rather than in "mouse movements and edit operations."
Another factor shaping the emergence of Smart Tools is the wide diversity of weather over the United States in relation to terrain and climate. No single set of tools will work effectively in the many regions that are as diverse as the frozen tundra in Alaska to tropical beaches in Puerto Rico. These tools must be tailorable to accommodate the weather features of each local site.
Finally, we needed to enforce meteorological consistency across these massive datasets, ensuring that the physical laws of meteorology are not violated in the forecast. As a simple example, we need a tool that would ensure that the dewpoint never exceeds the temperature.
The Smart Tools Solution
To solve these problems, we developed the concept of Smart Tools, which encompasses small bits of Python code that "plug-in" to the GFE. This is similar in concept to macros used in other forecaster packages such as PC-grids and the General Meteorological Package (GEMPAC). These bits of code implement meteorological equations and concepts. The framework contains a library of low-level functions that free the forecaster to think in higher-level concepts. For example, the sounding functions provide access to a vertical sounding of all elements at any point. For the purpose of developing tools that operate in the vertical dimension, this abstraction is ideal for the developers of Smart Tools.
Figure 1 shows how Smart Tools work through the Smart Tools framework. The forecast grids are stored and managed by the GFE server. These are not model grids, but represent surface parameters such as maximum temperature (MaxT), quantitative precipitation forecast (QPF), and probability of precipitation (PoP). The system also stores D2D model grids, topography information, and surface model grids. Through the Smart Tools framework, written in Python and C++, Smart Tools have access to all of the various grid data, soundings, and unit conversion functions. With this information, Smart Tools calculate new values, modify the forecast grids, and enforce consistency. Forecasters can use these tools to quickly and easily express meteorological concepts and then execute them to make the gridded data conform to a particular meteorological concept.
The Smart Tools framework is not limited to simply editing gridded data in a meteorological manner. It can be deployed to enforce consistency between particular weather elements, used to compare the forecast against observations in real time (monitoring), or at a later time (verification). We suspect that such a flexible framework will allow Smart Tools to be used for future tasks and functions not yet conceived.
Future of Smart Tools
This evolving technology is in its infancy and highly experimental. We are very excited to work with the NWS field to refine these concepts, find out what will be useful to the forecasters, and, for the first time, to make gridded forecasting a real possibility.
(Tracy Hansen can be reached at email@example.com, or phone (303)497-6569.)