Transcript WX Analysis and Forecasting

Supplemental Topic
Weather Analysis and
Forecasting
Background
• Weather analysis
– Using maps, charts, tables, or any other data to determine
what is happening or what happened
– Presumably precise and accurate
• Weather forecasting
– Predicting what may occur using analyses, observations,
computer models, education, & experience
– Considered imperfect
• Must know composition of atmosphere
• Atmospheric conditions constantly changing
• Complicated further by complex terrain, limited understanding of
atmosphere, operating costs, disagreement among forecast models,
etc.
Use of Weather Maps in Analysis
• Two types of maps
– 1. Surface
– 2. Upper-air
• Surface and upper-levels of atmosphere
interact constantly at a wide range of scales
• Meteorologists/climatologists must be able
to keep both surface and upper-air in mind
when performing analyses and making
forecasts
Surface Maps
• General representation of
pressure patterns and
fronts (isobars, pressure
centers, precip, etc.)
• Can determine what
happened by looking at
consecutive maps
• Can infer what is
occurring using what
learned to this point
• Station models – symbols
that show conditions at a
location at observation
time
Station Model
Upper-air Maps
• Measurements for location of launch plotted
on thermodynamic chart for use in forecast
• Maps produced for “standard levels” –
1000mb, 850mb, 700mb, 500mb, 300mb,
250mb, 200mb
• Provide height of pressure level, temp,
dewpoint (or dewpoint depression), wind
speed and direction, and (at 300, 250,
200mb), location of jet stream
• Analyze for troughs, ridges, advection of
warm/cool air, divergence/convergence, etc.
Thermodynamic Chart
Examples of Upper-air Maps
850mb
500mb
700mb
300mb
Other Analysis Tools
• Satellite
– Generally use three types of
imagery: visible, infrared, water
vapor
– Also sounder data
• Radar
– Pulses of energy sent out by
radar, hit object (hopefully
precip.), energy returned to radar
reveals distance and number/size
of particles
– Useful for detecting
precipitation, fronts, tornadoes,
wind direction, and even masses
of insects or flocks of birds
Forecasting Procedures
• Analysis: just discussed
• Prediction
– Models use data to solve equations that estimate
behavior of atmosphere
– Equations solved at various intervals and use previous
results in subsequent forecast – why should this
concern a forecaster? (chaos)
• Post-processing
– Production of maps from model output
– Usefulness of model data varies by location, time of
year, conditions, and forecaster ability
– Model output statistics (MOS) try to take local factors,
such as topography, into account
Forecasting Methods
• Climatological
– Based on long-term averages
– Accuracy and precision vary by location
• Persistence
– Based on trends: current conditions will continue
– Usually reliable for short periods
• Analog
– Based on records: what happened in past under similar
conditions?
– Could lead to poor forecast – why?
• Numerical
– Based on output from computer models/programs
– Must rely on accuracy of model given initial condition
Three Types of Forecasts
• Quantitative
– Predicts specific value
– Examples: precipitation amount, temperature
• Qualitative
– Predicts class or category
– Examples: type of precipitation, partly cloudy
• Probability
– Chance of event occurring (in percent)
– Likelihood of rain
Three Types of Forecasts
TONIGHT...PARTLY CLOUDY... SCATTERED
SHOWERS AND THUNDERSTORMS
MAINLY IN THE EVENING. LOWS 68 TO 78.
WEST WIND 5 TO 10 MPH BECOMING
SOUTHEAST AROUND 5 MPH AFTER
MIDNIGHT. CHANCE OF PRECIPITATION
30 PERCENT.
Forecast Periods
• “Nowcasts”
– What’s happening now and what’s expected in next few hours
– Most detailed, usually most accurate
• Short-term
– Covers from 3 to 72 hours
– Fairly detailed, accuracy tends to decrease a bit
• Medium range
– 72 hrs to 7-10 days
– Utilize ensemble forecasts
– Detail and accuracy decrease, esp. after 7 days
• Long range
– Performed by Climate Prediction Center (in US)
– Ranges from end of medium range to weeks, months, or longer
– Made of period (not individual days), probability usually low, skill varies by
location, time of year, and variable forecasted
– Use climatology, numerical models, statistics, and subjective judgment
Assessing Forecasts/Post-processing
• Need to determine usefulness of data and
forecasts, justify expenditures, and predict
importance of changes
• Quality vs. value
– Quality: is forecast right?
: Example: did it rain in Tucson as predicted?
– Value: does forecast perform desired goals?
: Example – did hurricane warning work?
• Accuracy vs. skill
– Accuracy: how close are predicted and actual
conditions?
– Skill: is method used the best choice?
Limitations on Forecasting
•
•
•
•
•
•
•
•
Computing power
Understanding of atmosphere
Funding/expenses
Chaos
Model error
Stubborn and/or old-school forecasters
Human error
Coverage of instruments for
measurements/observations