Weather and Climate Prediction

1
Weather and Climate Prediction

• Cliff Mass
• University of Washington

2
Outline

• Evolution of numerical weather prediction
• Application to regional climate and seasonal
  forecasting
• Some ideas for climate.com

3
The Resolution Revolution in Numerical Modeling
4
NGM, 80 km,1995
5
1995
6
2007-2008
4-km UW MM5 System
7
2013 WRF Model at 1.3 km
8
Last Week
9
WRF, 1.3 km
10
(No Transcript)
11
1.33 km resolution temperature
12
(No Transcript)
13
(No Transcript)
14
(No Transcript)
15
But just as important as the computer revolution
has been the weather data revolution, with
satellites giving us three dimensional data over
the entire planet
16
Example The Pacific Data Void No Longer Exists
17
(No Transcript)
18
Cloud Track Winds
19
Better than Star Trek!
20
(No Transcript)
21
NOAA Polar Orbiter Weather Satellite
22
Satellite Sensors Provide Thousands of High
Quality Vertical Soundings Daily over the Pacific
23
Cosmic GPS Satellites Provide More Soundings!
24
(No Transcript)
25
(No Transcript)
26
Impacts

• The addition of massive amounts of new
  observations is causing a steady improvement in
  weather prediction
• We are now starting to see frequent examples of
  forecast skill past one week
• Hurricane Sandy is only one example

27
Superstorm Sandy well predicted over a week
ahead of time
ECMWF Forecast of Sea Level Pressure
28
Observed 180 hr (7.5 days)
29
Skill Improvements (ECMWF)
Major improvements, mainly due to satellite data
and improved models
30
A Fundamental Problem

• The way we have been forecasting has been
  essentially flawed.
• The atmosphere is a chaotic system, in which
  small differences in the initializationwell
  within observational error can have large
  impacts on the forecasts, particularly for longer
  forecasts.
• Not unlike a pinball game.

31
A Fundamental Problem

• Similarly, uncertainty in our model physics
  (e.g., clouds and precipitation processes) also
  produces uncertainty in forecasts.
• Thus, all forecasts have some uncertainty.
• The uncertainty generally increases in time.

32
This is Ridiculous!
33
Forecast Probabilistically

• We should be using probabilities for all our
  forecasts or at least providing the range of
  possibilities.
• There is an approach to handling this issue that
  is being explored by the forecasting
  communityensemble forecasts

34
Ensemble Prediction

• Instead of making one forecastmake manyeach
  with a slightly different initialization or
  different model physics.
• Possible to do this now with the vastly greater
  computation resources that are available.

35
(No Transcript)
36
Ensemble Prediction

• Can use ensembles to give the probabilities that
  some weather feature will occur.
• Ensemble mean is more accurate than any
  individual member.
• Can also predict forecast skill!
• When forecasts are similar, forecast skill is
  generally higher.
• When forecasts differ greatly, forecast skill is
  less.

37
The Transition

• Numerical Weather Prediction is progressively
  transitioning to ensemble prediction and ensemble
  data assimilation

38
The Data Assimilation Revolution

• The combing of observations and model output to
  provide a three-dimensional description of the
  atmosphere is called data assimilation.
• Until recently the leading technology was 4DVAR,
  4D Variational Data Assimilation. NWS has lagged
  in using this.
• Ensemble-based data assimilation has many
  advantages and is increasingly being used.
• Future convergence between ensemble prediction
  and data assimilation is probable.

39
The Technology of Regional NWP Can Be Used for
Seasonal or Climate Prediction
40
Regional Dynamical Downscaling

• For regional numerical weather prediction we can
  embed high resolution models within a coarse
  resolution global forecasts.
• Can do the same thing for climate/seasonal
  prediction by simply replacing global weather
  forecasting models with global climate models
  (GCMs) or seasonal global prediction models
  (e.g., NOAAs Climate Forecast System-CFS)
• Just need the computer resources.

41
UW Regional Dynamical Downscaling

• Have completed a number of 100-year regional
  climate simulations using the WRF model at 12-km
  grid spacing.
• Driven by a half-dozen different climate models
  and emission scenarios.

42
Change in Winter Surface Air Temperatures (F)
43
(No Transcript)
44
(No Transcript)
45
(No Transcript)
46
Change in Snowpack from 1990 to 2090
-40
0
40
47
Climate Simulations

• Will be running with many more climate model
  driven simulations.
• Now evaluating the use for monthly and seasonal
  prediction at high resolution using output from
  the NOAA CFS model, a coupled atmosphere/ocean
  modeling system.
• Is there useful predictive skill at 1-9 months
  for mean quantities?

48
(No Transcript)
49
(No Transcript)
50
(No Transcript)
51
Some Ideas
52
Providing Useful Climate InformationBased on
Historical Records

• One of the greatest deficiencies of the climate
  community and the NOAA/NWS.
• There is a huge amount historical climate data
  available (station data, reanalysis information)
  but it is difficult or impossible for folks to
  get the actionable information they need.

53
Some Climate Questions

• When is the best time for wedding in Seattle?
• What is the windiest time of the day in Tucson?
• When are the high temperatures in Rome between 60
  and 70?
• I want to take a vacation the second week in
  March. Where will temperatures between 70 and 80
  with less than a 30 chance of rain, within a 7
  hr flight?
• What is the climatological last day of freezing
  temperatures at my house?

54
Climate Information Today

• Pre-generated tables and graphics.

55
National Climatic Data Center (NCDC)
56
God Help You if You are a Layman Looking for
Climate Information at the NCDC Site
57
Low Hanging Fruit

• Secure U.S. and International Climate/Historical
  Weather Data (available from NCAR, NCDC, and
  others for minimal costs). I assume climate.com
  already has it.
• Put into a relational data base.
• Build an interface/inquiry engine using natural
  language queries if possible.

58
Climate Apps The Surface Has Been Barely
Scratched
59
GardenKeeper

• Using calibrated radar-based precipitation data,
  tells you when watering is necessary at your
  location (considering water demands of your
  plants and evapotranspiration based on recent
  weather)
• Warns when freezing conditions are imminent
  during the winter.
• Tells you when you can plant seeds and young
  plants in the spring

GardenKeeper
60
Custom Automated Pinpoint Forecasts
61
The Idea

• The owner of a vineyard wants accurate forecasts
  that considers the microclimate of his property.
• An owner of a private airport wants forecasts
  tailored exactly to his airfield.
• The harbormaster of a yacht club wants accurate
  forecasts at his location.

62
There is a way.

• They contact climate.com for pinpoint forecasting
  service.
• Working with the client, weather instrumentation
  is installed at the exact locations of interest,
  with the data retrieved via wifi, cell phone, or
  wired connection.
• As soon as several weeks of data are available,
  statistical postprocessing is applied to
  operational models (e.g., GFS, NAM) to provide an
  optimal forecast at the observation location.

63
(No Transcript)
64
There is a way

• The longer the observations are in place the
  better the statistical postprocessing.
• Could use linear regression, extended logistical
  regression, or other approaches.
• Forecast biases could be radically reduced at
  such sites.
• Could use ensembles or analog methods to give
  probabilistic predictions.

65
The End