Lab Exercise: Forecasting for the Pacific Northwest
Originally developed for the COMET® Satellite Meteorology Course
Don Moore, Ron Miller, Tony Mostek
Computer and data support from Tim Alberta, Dolores Kiessling, Liz Page, Kevin
Schrab, Patrick O'Reilly
 |
This laboratory exercise consists of 2 parts which are designed to simulate
an operational forecast process on the day shift. However due to time constraints,
not all aspects of an operational forecast process will be simulated. GARP
data will be used to analyze observational and model data. Many questions
asked in this laboratory exercise can be answered using little to no model
data, including questions regarding the forecast. In order to complete this
laboratory exercise in the time allowed, model forecast analysis should
be limited to 10 fields or less.
The objectives are to: 1. Use GOES data in combination with surface observations to analyze conditions over the Eastern Pacific Ocean, including the identification of fronts and areas of precipitation. 2. Use GOES data to identify numerical model errors and improve upon forecasts provided by numerical models. 3. Create an accurate precipitation forecast for the Pacific Northwest using a combination of observations and numerical model data, with an emphasis on GOES data. |
Introduction
Precipitation forecasts for the Pacific Northwest are complicated by limited surface observations over the Pacific Ocean. Numerical model initializations and forecasts are created using both GOES and POES data, minimal ship and buoy observations, and previous numerical model forecasts. Because of the limited number of observations available over the Pacific, numerical models rely more heavily on previous numerical model runs over the oceans compared to land. As a result, similar model errors may occur from run to run for a particular event until the storm finally makes landfall, where observation density is greater.
The following laboratory exercise will require a precipitation forecast to be made during strong zonal flow into the Pacific Northwest. Three main factors should be considered when forecasting for the Pacific Northwest in zonal flow.
1) In zonal flow, plumes of high precipitable water, extending past the dateline will often be "grabbed" by areas of low pressure approaching the Canadian and U.S. coasts. Precipitation, both stratiform and convective in nature, is closely tied to the location of the plume of highest precipitable water.
2) Temperature advection patterns from the surface to 700 mb are important for the generation of or lack of precipitation during zonal flow, especially east of the Washington and Oregon Cascades. Typically neutral to low level cold-air advection will inhibit precipitation across much of Washington and Oregon east of the Cascades due to strong down sloping effects. In addition, warm-air advection will not always produce precipitation in Washington and Oregon east of the Cascades. Conversely, along the coast and western slopes of the Cascades as well as Northern Idaho, up slope conditions will allow for precipitation to occur in both warm and cold air advection in the low levels. The intensity and coverage of precipitation will be greatest during warm air advection episodes.
3) Another significant factor for precipitation in zonal flow is short wave forcing. Typically, numerous weak short waves imbedded in strong flow will produce periods of precipitation across Washington, Oregon, and North Idaho. Precipitation in up slope regions may be more persistent, but periods of heavier precipitation will still occur.
In zonal flow, the combination of short wave forcing and warm air advection in areas of high precipitable water are the main factors for precipitation in Washington, Oregon and North Idaho. Without both positive temperature advection and short wave forcing, precipitation will likely be light with low areal coverage outside of the up slope regions.
Due to limited observations, numerical models may not be able to accurately simulate temperature advection fields until the storm makes landfall. Frequently, the models do not accurately predict the depth of surface lows approaching the coast. Numerical models are also often unable to accurately simulate the passage of short waves. Since numerical models poorly forecast short waves during strong zonal flow and misrepresent actual temperature advection fields, close scrutiny of observational data over the Pacific is required to make accurate precipitation forecasts.
When creating a forecast for this laboratory exercise, you should maximize the use of satellite to 1)identify the location and movement of greatest precipitable water, 2)recognize short waves and accurately time their passage through the Pacific Northwest, 3)determine the true strength of surface lows moving inland, and 4)identify the strength of actual temperature advection fields over the Pacific and time their progression inland.
To assist with the laboratory exercise, a high resolution topography image is provided along with average yearly precipitation maps for Washington, Oregon and Northern Idaho.
Part I - approximately 1 hour
Section A - 5 to 10 Min + time for discussion
Limit data fields to GOES-10 along with fixed buoy observations...
On the North American scale, at 16Z 11/23/99, load into 3 different GARP
windows:
Water Vapor
IR
Surface Observations: fixed buoy observations, and station plots
Overlay Lat/Lon grid to all panels.
For the location from the West Coast to 170W and from 35N to 60N:
1) Identify the location of any synoptic scale surface features. Draw the fronts on the map provided.
2) Make a guess to the location/orientation of maximum precipitable water (PW) at 11/23/99 18Z. Outline the area of greatest PW on the map provided.
3a) Identify any features in water vapor that would favor the development of surface low pressure or upward motion.
3b) Identify the areas of precipitation.
Group Discussion (5 Min) - How difficult was it to identify surface features and maximum PW? What channel proved most useful to identify these features? What did you see in water vapor to favor precipitation and surface low development?
Section B - 5 to 10 Min + time for discussion
Limit data fields to GOES-10, along with fixed buoy observations...
Add Eta PW to a new GARP window.
4) How does your initial guess to the location/orientation of maximum PW compare to ETA PW?
5) Using simple extrapolation, when and where do you expect precipitation to initially begin on the US West coast? When will the precipitation begin at UIL and AST? If the precipitation makes it across the Cascades, when would begin at GEG?
Group Discussion (5 Min) - How did your initial guess to location of precipitation and maximum PW compare to ETA PW? What is your timing of precipitation onset? Were you able to get the PW structure seen via ETA PW?
Section C - 15 to 20 Min + time for discussion
Use all data fields available...
Create the following products to answer the questions below:
Upper level RH, vorticity, and winds on water vapor
mean sea level surface pressure on visible
AVN PW on ETA PW.
Note: the AVN is at 3 hour intervals and will not always match up with the ETA
6) Perform a satellite and model comparison using GOES and other observations. Based on your comparison, which model do you feel is and will be most accurate? Be sure to consider your initial guess of surface features. Explain your reasoning.
7) Suggested fields: 850 temp advection and qpf.
Do you feel your model of choice will accurately forecast the
timing and location of precipitation through 36 hours? If so, why?
If there are any inaccuracies, what are they and why? Be sure to consider
your initial guess to precipitation onset.
8a) Make a QPF forecast for 11/24/99 00Z to 11/24/99 00Z and for 11/24/99 12Z to 11/25/99 12Z for UIL, AST, GEG. What 6 hour period(s) will experience the heaviest precipitation for UIL, AST, GEG?
9) Time permitting, outline on the map provided the area you expect to receive 0.25, 1" and 2" or more of liquid equivalent precipitation from 11/24/99 12Z to 11/25/99 12Z.
Group Discussion (10 Min) - What model did you choose and why? What is your QPF forecast? Did you deviate from the "best" model's QPF forecast, why or why not?
Part II - approximate 1.25 hours
Overview of data up to 18Z 11/24/99. When did precipitation onset occur? How much precipitation has fallen? - 5 min discussion.
Section A - 5 to 10 Min + time for discussion
Limit data fields to GOES-10 along with fixed buoy observations...
On the North American scale,using 20Z 11/24/99 data, load into 3 different
GARP windows:
Water Vapor
IR
Visible with fixed buoy observations
Overlay Lat/Lon grid.
For the location from the West Coast to 170W and from 35N to 60N:
1) Identify the location of any synoptic scale surface features. Draw the fronts on the map provided.
2) Make a guess to the location/orientation of maximum precipitable water (PW) at 11/24/99 18Z. Outline the area of greatest PW on the map provided.
3a) Identify any features in water vapor that would favor the development of surface low pressure or upward motion.
3b) Identify the areas of precipitation.
Group Discussion (5 Min) - How confident are you in your location of surface features and maximum PW? What channel did you find most helpful? What did you identify in water vapor imagery that would favor upward motion and precipitation?
Section B - 10 Min + time for discussion
Limit data fields to GOES, along with fixed buoy observations...
Using the fields created in section A, add ETA PW to a new GARP window.
4) How does your initial guess to the location/orientation of maximum PW compare to ETA PW?
5) Can you estimate where the max PW plume will be located at 11/25/99 06Z and/or 18Z with confidence?
Group Discussion (5 Min) - Did your initial guess to location of precipitation and maximum PW compare to Eta PW improve from section A? Were you able to estimate where the plume of highest PW be located for the specified times?
Section C - 15 to 20 Min + time for discussion
Use same times and data as Section B
Create, on the North American scale, at 18Z 11/24/99, 2 GARP windows of:
visible
ETA PW.
7a) How do the associated cloud features identified in the previous question move through 11/24/22Z? Does this movement give any indication to the movement of the plume in a north of south direction?
7b Can you estimate where the plume of highest PW will be located 11/25/99 06Z and 11/25/99 18Z with more confidence? If so, draw the 1" PW contour on the map provided.
8) The precipitation has now tapered off in Western Washington and
is in the process of ending in Eastern Washington. Do you expect more precipitation
to spread into western and/or eastern Washington overnight. How about on the
25th?
If so, when will that occur at UIL and GEG?
9a) Do you expect greater than 0.05" of liquid equivalent precipitation
to occur at UIL, AST, or GEG between 11/25/99 00Z 11/25/99 12Z and/or from 11/25/99
12Z to 11/26/99 00Z? What 6 hour period(s) will experience the heaviest precipitation
at UIL, AST, GEG?
9b) Do you expect the precipitation to end at UIL, GEG, and/or AST by 11/26/99 00Z?
Group Discussion (10 Min) - Were you able to find any correlation between the clouds seen in visible imagery and highest PW seen via ETA? Which direction were they moving?
