Ahrens Ch. 14 - Weather Forecasting

• World Meteorological Organization (WMO): located in Geneva, Switzerland; includes over 130 nations; sets the standards for weather observing and the international exchange of weather data.

• WMO Data Centers: NCEP, Washington DC, USA, Melbourne, Australia and Moscow, Russia.

• Over 10,000 land-based weather stations report hourly weather conditions, including "specials" in rapidly changing weather conditions.

• Automated Surface Observation System (ASOS): a fully computerized and automated weather array that continuously reports weather conditions.

• Hundreds of ships, buoys and C-mans report weather conditions every 6 hours over the water areas (large lakes, seas and oceans).

• Weather balloons (radiosondes) are launched every 12 hours around the world at 00Z (8 PM EDT) and 12Z (8 AM EDT); atmospheric soundings of temperature, humidity and winds are created for each launch site.

• SATELLITE and aircraft observations are available throughout the day.

• Weather RADAR (conventional and Doppler) in operation all day.

National Weather Service

• Until the mid-1950s, forecasters used hand plotted and analyzed maps.

• Numerical Weather Prediction (NWP): weather data is "fed" into the models to create an "initial analysis" ... data is retrieved for each "grid point" in the particular model ... each point is forecast into the future for all the various weather parameters over small time intervals ... weather forecasts (progs) are generated.

• Advanced Weather Interactive Processing System (AWIPS): latest hi-tech system for data communications, storage, processing and display capabilities to better help the forecaster assimilate all the available information ... including ASOS, weather satellites, Doppler RADAR and wind profilers.

• Weather Forecast Office (WFO): uses all the above in conjunction with the latest computer forecast models such as NGM, Eta, MRF here in the USA and those of other countries as well (GEM-Canada, ECMWF- Europe, etc.).

• Model Output Statistics (MOS): a model-based, computer-generated forecast of weather specifics, such as Max & Min, Precip Prob, Wind, etc.

• Nowcasting: the WFO makes a short range forecast, usually for less than 6 hours into the future; uses many current weather information, RADAR and satellite. (Computer models are helpful in making the longer range predictions.)

• "Rules of Thumb" ... some as simple as "looking out the window!"

Weather Satellites

• TIROS I: the first weather satellite, launched on April 1, 1960; used television cameras to photograph clouds.

• Polar Orbiting Satellite: these satellites have an orbit that passes directly over the north and south poles, closely paralleling the earth's meridian lines; the earth rotates under the satellite, eventually allowing for a composite photograph of the entire earth ... in great detail.

• Geostationary Satellite: they orbit the equator at the same rate the earth spins and as a result, remain above a fixed point on the earth's surface.

• Geostationary Operational Environmental Satellite (GOES): uses radiometers to "see" the clouds, such as an imager (which provides better resolution ... down to 1 nm) and a sounder (which gives a more accurate profile of temperature and moisture at different levels in the atmosphere).

• VISIBLE: the radiometer is set to "see" reflected sunlight ... so that clouds appear to be WHITE; object with a low "albedo" appear dark, such as the oceans and forests; SNOWCOVER will give the appearance of clouds.

• INFRARED: the radiometer is set to "see" outgoing infrared radiation ... so that warm clouds (land & oceans) will appear DARK and the colder, cloud tops will appear WHITE, resulting in a "gray scale" for in-between temperatures.

• ENHANCED INFRARED: the "gray scale" is color enhanced to increase the contrast between features and their backgrounds.

• WATER VAPOR: the radiometer is set to "see" just 6.7 micro-meters, which is a unique absorptivity wavelength of water vapor; the areas of high moisture will appear WHITE and the dry areas will be DARK; due to re-radiation in all directions ... this method of sampling only "sees" the middle and upper troposphere ... it cannot "see" to the ground; sometimes, the water vapor pictures are color-enhanced to bring out details.

Weather RADAR

• RADAR (Radio Detection And Ranging): emits pulses of microwave radiation are sent out and measures those that are back-scattered to the RADAR dish; is affected by GROUND CLUTTER (i.e., buildings, trees, mountains, etc.).

• Conventional RADAR: developed in World War II; used by the NWS until the 1990s; measures base reflectivity and can determine cloud tops; has six (6) levels of reflectivity; cannot detect drizzle or light snow; can be "fooled" by large hail and sleet.

• Doppler RADAR (WSR-88D): can measure base reflectivity and clouds tops, as well as wind movement (radial velocity and vertical wind profiles); has 16 levels of reflectivity; can detect drizzle and light snow (can actually detect insects!), as well as strong wind gusts and meso-cyclones (tornadoes); in the radial velocity mode, wind coming toward the RADAR site is, by convention, a "cool color" (greens & blues), whereas wind going away from the RADAR is, by convention, a "warm color" (reds & oranges); a tornado would appear as a small area, that rapidly changes from green/blue to red/orange; can "integrate" the base reflectivity to estimate how much precipitation (rain or snow) has fallen over a given areas; ground clutter can be "removed" from the RADAR display; can still be "fooled" by large hail and sleet.

All This Technology - What Happened?

Why Did The Forecast Bust?

• Computer models make assumptions ... which can be way off!

• There are boundary errors on regional models. (The global model is thus preferred, but at the sake of a large grid spacing for computational speed.)

• Errors in the data.

• Lack of data ... especially over oceans and in high latitudes. (GOES satellite does help in those areas.)

• Meso-scale features (thunderstorms, tornadoes, and sometimes small tropical storms) too small and "slip through the cracks" in the grid.

• Cannot accurately model such things as interactions of water, ice and local terrain (i.e., hills, lakes, mountains and even ... the oceans).

• CHAOS !! Countless, small, unpredictable atmospheric fluctuations that can become dominant (i.e., amplify) as the forecast is extended into time. (Denser observing network and a "perfect" model cannot overcome chaos!)

Other Forecasting Methods

• PERSISTENCE: a prediction that the current weather will continue ... unchanged. If it's "sunny and unseasonably warm" today ... then, the forecast would call for "sunny and unseasonably warm" tomorrow; this is the worst kind of forecasting method.

• STEADY-STATE or TREND METHOD: assumption is made that weather systems will tend to move in the same direction and approximately at the same speed as they have been moving, provided no evidence exists to indicate otherwise (e.g., a blocking weather pattern upstream, a strengthening storm, etc.); accuracy falls off greatly after about 12 hours into the future; used for "nowcasting."

• ANALOGUE METHOD (Pattern Recognition): when existing weather features on today's weather map "look familiar" and fit a recurring weather pattern (e.g., Bermuda High, chinook winds, heat wave, etc.).

• WEATHER TYPES: employs the analogue method, but categorizes weather patterns into "groups", that behave in a similar fashion for a 3-5 day period and sometimes longer; this method was used in the 1800s and early 1900s.

• CLIMATOLOGICAL: based solely on long term averages for the region in question; this forecast can be made for today, for next month, for next year, for the year 2020.

• TELECONNECTIONS: when the weather in one part of the world has a direct effect (a link) to another part of the world; El Nino has very strong teleconnections worldwide.

• ENSEMBLE FORECASTING: several forecast models (or different versions of the same model) are run, using slightly different initial conditions; the resulting forecasts are compared for similarity; a strong agreement would indicate a higher level of confidence in the forecast.

• PROBABILITY FORECASTING: most weather forecasters (especially the NWS), use probabilities in making precipitation forecasts.

• FORECAST ACCURACY: 12-24 hours is very good ... 1-3 days is fairly good ... 4-6 days, temperatures are good, precip forecasts fair to poor ... 7 days or greater ... ???