AFPS Quarterly Report (95Q3)

Table of Contents

1. Introduction
2. Accomplishments
3. Presentations/Visitors/Travel
4. TDL activities
5. Plans for the next quarter
The FSL AFPS Team
Using the AWIPS Forecast Preparation System (AFPS)
Using Numerical Model Output to Provide Initial Forecasts of Surface Weather for the AFPS

AFPS Quarterly Report FY95 Q3: April - June 1995

1. Introduction

Most of this report covers FSL work; "we" here refers to FSL. The TDL Activities section is based on information provided by Matt Peroutka of TDL.

We reached a watershed of sorts this quarter. During the initial years of AFPS development, we (both FSL and TDL) were working toward a rather ephemeral AWIPS implementation goal, and deadlines tended to be far-off, and often receding. Late last year, with the creation of WFO-Advanced in FSL, a much more immediate deadline appeared, and we set out a list of requirements to meet the scheduled August and October/November WFO-Advanced exercises.

It became clear by early spring that we were not keeping to the schedules required to meet these exercise deadlines, and we tried to adopt an accelerated development schedule, to complete Level 1 (milestone 1d) by June. It soon became apparent that taking design shortcuts, while addressing immediate needs, is detrimental to our progress in the long run. The first minor adjustment we made in our plans was in early April, when we decided to concentrate on completing all spatial depictions and editors before moving on to temporal work.

As noted in previous reports, we are committed to preparing high-quality software by doing a thorough design, conducting design reviews, and submitting all code for peer review. Experience has verified that this process allows us to produce high-quality code, and saves time in the long run.

In late May, we decided that it was unrealistic to continue to advertise that we could produce a system by October that can support generation of terminal forecasts (TAFs). We believe, however, that we can complete a usable public/agriculture system, including initialization, visualization and editing, interpolation, derivation, and text generation. At that time, we revised our schedule as follows:

• Level 1d (7/95): spatial visualization and editing for all data types, interpolation, map backgrounds;
• Level 2a (10/95): temporal editor, derivation server, initialization, formatters, user and training documentation;
• Level 2b (2/96): point database; TAF worksheet, editors, initialization, and formatter.

2. Accomplishments

On the initialization front, three separate lines of work are being pursued:

• We are initializing AFPS grids from the RUC every cycle (3 hours); these grids are available for viewing in FSL's Forecast Center (Wier, based on algorithms developed by Jim Ramer of FSL's Forecast Research Division). The figure below shows the spatial editor with RUC weather, QPF, and wind elements loaded.
• TDL is routinely initializing AFPS weather elements from MOS, and storing grids in the AFPS database (Mark Oberfield of TDL, with assistance from Mathewson). A necessary prerequisite for this was installing the AFPS gatekeeper (manager of databases) and database documentation at TDL (Mathewson)
• LAPS (FSL's Local Analysis and Prediction System) analysis grids have been stored in the AFPS database on a test basis (Frank Merrem of FRD, in consultation with Mathewson and LeFebvre).

Other activities:

• Legends were added to the system (Longstaff and Romberg). These optional legends occupy their own floating windows, so users can position them where they like.
• A preliminary version of undo is included in the Level 1d system. This is a significant help to users (LeFebvre).
• Additional feedback is provided to users regarding the effect of edit actions on the data. For example, when using the Paint, Spray, or Bulldozer tools, a matrix of dots shows the user exactly which grid boxes will be affected. In addition, help/information strings are now included at the bottom of the editor window (LeFebvre). The text provides a guide to mouse button actions, plus shows the value or offset to be applied by the edit tool, while dynamically indicating the current value of the grids. An example of this information for the Paint tool is shown here.
• Detailed plans have been prepared for testing the worksheet, spatial, and temporal editors. Many of these tests have been performed. We continue to make extensive use of our bug-reporting and -tracking system (Howard).
• Worksheet time scales have been modified as requested by AFWG members, to use pre-set time spans (Romberg).
• Interpolation code has been completed and integrated into the prototype (Wier, Romberg). An interpolation server accepts user requests, performs the necessary work, and returns grids for display and editing (Bacco). Testing of this system is under way.
• Several sections of our code have been made more efficient. Part of this was facilitated through the use of Quantify, a commercial system which identifies which parts of the software are most used. In one instance, this helped isolate the source of the bulk of contouring time to searching for contour-level limits, which led to a several-times improvement in contouring performance. Other optimizations included improvements in the performance of the weather display by over 50 times. (Mathewson, LeFebvre)
• Along these lines, code was developed to throw away intermediate edit points where unnecessary (e.g., when moving a block of data, it's not necessary to show every intermediate point) and to add points when necessary (e.g., when painting, one wants to hit every intermediate point, even if the mouse drag is very rapid). This improves both performance and quality of editing (Romberg, Mathewson, LeFebvre).
• A prototype version of a temporal editor window was written, testing our ability to manage and scroll multiple time lines (Mayer).
• We have prepared an initial user guide for Level 1d (Wakefield).
• Two abstracts were submitted for the 12th IIPS conference (Atlanta). Copies of these are included at the end of this report.

3. Presentations/Visitors/Travel

• a group of congressional staffers, accompanied by Lou Boezi, 11 April
• some staff of the Space Environment Laboratory, 18 April, interested in tools for managing and viewing solar forecasts
• Frank DiGialleonardo, Director of the NOAA Information Systems and Finance Office, 20 April
• Dale Shelton, NWS Office of Hydrology, 25 April
• five people from the Korean Meteorological Administration, 1 May
• Tim Barker of the Missoula NWSO, 2 May and 17 May
• attendees at a planning meeting for the FAA's WARP (Weather and Radar Processor) system, 5 May
• Barry Hanstrum, Perth, Australia, 10 May
• members of COMET's COMAP class, plus Linda Butler of PRC's HCI staff, 12 May
• seven Air Weather Service visitors, 18 May
• Jens Daabeck, ECMWF, 19 May
• Brad Colman, Seattle WSFO, 2 June
• the National Research Council NWS Modernization committee, accompanied by Mary Glackin and Doug Sargeant, 5 June
• Denny Walts and a GSA official, 7 June
• the Department of Energy Meteorological Coordinating Council, 7 June
• a joint PRC/NWS AWIPS User Interface design team, 7 June

• Mark Mathewson and Joe Wakefield went to NWS headquarters the 17th and 18th of April for two WFO-Advanced presentations, including AFPS.
• We provided some slides to Andy Edman (NWS Western Region Scientific Services Division), for a presentation at Lawrence Livermore Laboratory in mid-May.

4. TDL activities

TDL's grid initialization programs were enhanced to allow offsets to selected weather elements; for example, the temperature at a grid point can be initialized to the MOS forecast for a site minus 5 degrees. This technique should improve grid initialization in areas where MOS stations are lacking. Its creation opened the way for TDL to produce MOS-based initialization fields for AFPS. Check-out is underway, and these data are now flowing into AFPS' reference worksheets.

Forecasters at Charleston, Norman, and Peachtree City are also evaluating a number of product generation programs which will be useful in AFPS. The product list includes most public and closely-related products (Zone Forecast, Agricultural Forecast, Coded Cities Forecast, and a variety of coded and tabular products). AFPS products will be stored in the WFO-Advanced text database, which is taking shape.

Back to Table of Contents

5. Plans for the next quarter

Before that meeting, we plan to distribute the Level 1d software and user documentation to those AFWG members able to use it. Level 1d will run on a SAC (SOO Application Computer), as currently outfitted. The only question is one of distribution medium. We hope to send the software over the Internet.

We will begin work on the TAF database and editor concepts, in anticipation of significant work in this area toward the end of 1995.

Back to Table of Contents

The FSL AFPS Team

Dave Howard Quality Assurance Specialist (testing,
303-938-2088 bug tracking, configuration management)
dhoward@fsl.noaa.gov

Tom LeFebvre Meteorologist/Programmer (design,
303-938-2086 graphic editors, database)
lefebvre@fsl.noaa.gov

Jennifer Longstaff Programmer (graphics, user interface)
303-938-2069
longstaff@fsl.noaa.gov

Mark Mathewson Technical Manager -- Meteorologist/
303-938-2061 Programmer/Lead Designer
mathewson@fsl.noaa.gov

Bob Mayer Programmer (user interface, design,
303-938-2075 graphics)
rmayer@fsl.noaa.gov

Mike Romberg Programmer (graphics, user interface,
303-938-2084 network)
romberg@fsl.noaa.gov

Joe Wakefield Project Manager -- Meteorologist
303-938-2089
wakefield@fsl.noaa.gov

Stuart Wier Programmer (initialization,
303-938-2078 interpolation, graphics displays)
wier@fsl.noaa.gov

Our fax number is 303-497-3096.

Information about AFPS and the EFT Branch is available on the World-Wide Web via URL http://www-md.fsl.noaa.gov/eft/EFTHome.html.

Please direct comments on or questions about this report to Joe Wakefield.

Back to Table of Contents

Using the AWIPS Forecast Preparation System (AFPS)

The modernization program of the National Weather Service (NWS) is bringing new technology to Weather Forecast Offices (WFOs) to aid forecasters in interpreting the weather. Systems like doppler radar, new satellites, and the Advanced Weather Interactive Processing System (AWIPS) provide much more insight into the atmosphere than current capabilities allow. At the same time, the NWS is also looking to improve the forecaster's efficiency in many areas. One particular area where improvements can be made is in routine forecast product generation.

The NWS has specified that forecasters will maintain a database of gridded digital forecasts. Routine forecast products will be automatically generated from these forecast grids. The system for preparing and managing the forecast grids is known as the AWIPS Forecast Preparation System (AFPS).

Forecasters currently spend a significant amount of time composing forecasts as text. With AFPS, the forecaster no longer writes text products, but instead "draws" the forecast graphically using a series of highly interactive displays and tools. The forecast is stored as forecast grids. Many different types of products can quickly be generated from the forecast grids, including new product formats such as graphics and gridded products. Forecast amendments can be "drawn" once and the entire suite of products generated with reduced effort from the forecaster.

This paper describes the major components of AFPS and how forecasters will use AFPS to generate forecasts.

Back to Table of Contents

Using Numerical Model Output to Provide Initial Forecasts of Surface Weather for the AFPS

The AWIPS Forecast Preparation System (AFPS), being developed at the Forecast Systems Laboratory (FSL) in Boulder, Colorado, and the National Weather Service (NWS) Techniques Development Laboratory (TDL) in Silver Spring, Maryland, will support preparation of most routine forecasts at NWS Weather Forecast Offices (WFOs) when it is deployed in the late 1990s.

Using AFPS, the forecaster prepares graphical depictions of forecast weather. These graphics fix the contents of a forecast database, which is then used by automated forecast product generators to generate text forecast products. The forecaster begins by selecting a complete suite of weather elements from initial forecast conditions derived from numerical weather models, or alternately, other sources. The initial forecasts are edited, where necessary, to compose the desired depiction of the atmosphere. The process of using AFPS is designed to be similar to the mental process that forecasters presently use to select and revise features from guidance to include in the forecast.

Providing the best possible initial conditions is essential to the success of AFPS. In the final form, AFPS will offer the forecaster a choice of initial conditions based on MOS, LAMP (Local AWIPS MOS Program), NCEP manually-prepared grids, Eta, NGM, AVN, MRF, RUC, and, where available, local numerical models. Algorithms or local models for special weather conditions or specific elements, such as lake-effect snow or wave heights, can also be used. Other available data will include the current forecast database, recent observations, and climate.

Using grids from NCEP's Eta and RUC model runs, AFPS now provides initial forecast values of many surface weather elements. The process of deriving these elements from the model output and an assessment of the results for correctly forecasting the weather will be presented.

Since the Eta and RUC grids are available at NWS offices nationwide, the programs described here to generate surface weather conditions might be useful at NWS offices today.

Back to Table of Contents