Utilizing Satellite Imagery to Help Analyze / Forecast Winter Storms Part 1

Utilizing satellite imagery to help analyze / forecast Winter Storms Part 1

1. Title. A prerequisite for this training session will be the VISIT teletraining session “Cyclogenesis: Analysis utilizing Geostationary Satellite Imagery”. An audio playback version is available for download to go through on your own time at the student guide web-page for that session: http://www.cira.colostate.edu/ramm/visit/cyclo.html

2. Objectives

3. Topics that will be covered.

4. Winter weather factors

5. Structure of winter storms

6. Conceptual model of conveyor belts. Note the cyclonically curved branch of the warm conveyor belt (WCB) which is referred to as the TROWAL by some meteorologists. Others refer to this feature as the secondary WCB (as referenced in VISIT Cyclogenesis training). The subtle difference is the timing, when occlusion begins it is generally referred to as the TROWAL.

7. Model Divergence. We are referring to different solutions between different models.

8. Let’s compare the differences between the Eta (green) and GFS (salmon) MSLP field through the forecast time. We can see that the Eta has the low further south. In terms of forecast implications the Eta would give northern Colorado a greater threat of precipitation with the low further south.

9. The GOES-10 IR loop shows that the low is following the more southerly track, meaning that the Eta should be the solution that would have a higher probability of verifying than the GFS.

10. Observed position of the low at 12:00 UTC 10 November based on HPC surface analysis. Notice that the position of the low at this time corresponds with the Eta. This increases confidence in choosing the Eta over the GFS during the forecast period.

11. Average snow to liquid water ratios between 1971-2000 by Marty Baxter of St. Louis University. Map shows that snow is generally drier in the west and in the lake-effect snow belts. These differences need to be considered when using automated snow algorithms (Does 10:1 usually work in your CWA?). The web-site (http://mnw.eas.slu.edu/CIPS/Research/snowliquidrat.html) provides the option of clicking on individual CWA’s (toggle) on the interactive map link. The example shown is the Boulder CWA. Note the variation across the small region, and the large standard deviation in the sample.

12. Introduction to satellite shear zone. Generally look for these ahead of the main trough. The stationary type of shear zone doesn’t move very much relative to the trough. The rotating type of shear zone rotates around the trough axis.

13. GOES-10 IR imagery from 01:00 UTC – 23:45 UTC 29 October, 2003. This is an example of the stationary type of shear zone. A long narrow dry slot combined with upper cloud deck moving slowly southward from Canada helps identify a shear zone. One vort max can be seen on the southwestern most portion of the deformation zone in southwest Canada, the other moving from southern Alberta into southern Saskatchewan. Sometimes we may see just one vort max but in this case we see multiple. Max PVA moving in a storm relative-sense just ahead of southwestern most vort max.

14. Denver/Boulder CWA for the March 17 – 20, 2003 blizzard. We will be forecasting for the 12-36 hour time period with a start time of the morning of 17 March 2003.

We will be presenting only enough model output to bring the class up to speed on the forecast problem for each of the cases that we present. The primary purpose will be to show how satellite data can be utilized to give value added to the forecast / nowcast situation. We simply don’t have the time to cover all aspects of each case you would normally go through.

15. Objectives for this case.

16. Eta 300 mb Heights and Isotachs from 12:00 UTC 17 March, 2003. Shows approaching trough. The trough becomes negatively tilted by 00:00 UTC 18 March and becomes cutoff 6 hours later. Note the jet max that develops around 06:00 UTC 18 March in northeast CO.

17. Eta 500 mb hgt and vorticity (shaded) from 12:00 UTC 17 March 2003 showing 12, 18, 24, 30 and 36 hour forecasts. 12 hr forecast shows vorticity lobe near Dalhart, TX. This vorticity center is forecast to move slowly northward over the forecast period. Note that the 500 mb low if forecast to consolidate and deepen as it moves into southeastern Colorado over the next 24 hours.

18. MesoEta surface MSLP, precipitation, temperature and winds from 12:00 UTC 17 March 2003 going out 33 hours. Sfc low in southeast Colorado is quasi-stationary and remains deep. The deepening low in this region extends through nearly the entire depth of the troposphere. Northeast Colorado is forecast to experience very deep easterly (upslope) flow. The output shows precipitation throughout north-central CO for the entire period. The forecast temperatures suggest that the freezing line remains over the mountains until the next morning while the temperatures are below freezing throughout most of southeast Wyoming. Thickness analysis (not shown) suggests rain/snow line approximately 6000 feet throughout the period. If snow does occur it will be heavier than the climatological 15:1 ratio. Heavier precipitation is forecast to develop over OK/KS by 00:00 UTC and expand with time through 12:00 UTC.

19. GOES water vapor loop 04:30 UTC 17 March 2003 – 00:00 UTC 18 March 2003 – A lot of the value of satellite imagery is in supplying context on the larger scales to help the forecaster understand the NWP output. In this loop, notice the developing warm conveyor belt (WCB) that stretches from Oklahoma into Kansas. It is marked by deep convection associated with a jet max over KS moving into NE. This precipitation seems to be developing in the same location as that forecast by the Eta.

Satellite imagery should not be restricted to supplying synoptic overviews, but should also be used to identify sub-synoptic components embedded in the larger scale. In addition to the jet max in Kansas, there is a small shear zone rotating up from north-central New Mexico into central Colorado. Finally, a jet max becomes apparent as it rotates around the base of the trough from southern New Mexico / northern Mexico into west Texas at the very end of the loop.

20. MesoEta 310K potential temperature surface with pressure (mb), winds and relative humidity (shaded) from 12:00 UTC 17 March, 2003. The Eta shows weak isentropic lift at 00:00 UTC on the 18th. By 06:00 UTC the isentropic lift intensifies over north central CO but by 12:00 UTC it weakens again.

(Remember that isentropic analysis is used to represent the thermal advection contribution to QG omega. Thus, even though we’re utilizing isentropic analyis in this case, both thermal and vorticity advection can contribute to vertical motion. Also, remember that just because we’re using isentropic coordinates instead pressure coordinates, QG assumptions still apply. Finally, remember that isentropic surfaces are not static in time and space, and should not be thought of as ramps that are frozen in time and space along which parcels ascend and descend).

21. Eta (yellow) and NGM (cyan) forecast MSLP 12 hour forecast valid 00:00 UTC 18 March, 2003. The Eta forecast the low to be in southeast Colorado while the NGM. The NGM is chosen here just to make the point of comparing diverging models with satellite imagery. The NGM has the low further north. In order to assess which model is closer let’s go to the visible imagery in the next slide.

22. GOES-10 visible imagery shows the low in extreme southeast Colorado meaning that the Eta was more accurate in its MSLP forecast.

23. Colorado topography. Air moving in from the east will undergo forced lift of

roughly 3,500 ft as it moves from the Colorado border into the Front Range Range. Also, as the low-level air encounters the mountains it will also be partially blocked. This effect causes the boundary layer right along the foothills to deepen and the flow to accelerate (towards the south). This is the phenomenon referred to as the barrier jet earlier.

Remember that the flow through the whole mid- to lower-troposphere was easterly as well, implying that the upslope flow was deeper than the mountain barriers. In a normal upslope situation, the upslope flow is contained by the mountain barriers and snowfall totals west of the Front Range are light. The deep upslope in this case improves chances for significant snowfall in these normally shadowed areas.

24. Eta MSLP and Precipitation forecast from 12:00 UTC 17 March, 2003 out to 18:00 UTC 18 March. Notice the precipitation band stretching from Kansas across southwest Nebraska and wrapping back around into north-central Colorado. Note also the forecast position of the surface low. We can compare the position of the low with surface observations and the precipitation forecast with IR satellite data to get an idea of how well the model is doing.

25. Surface observations from 00:00 UTC 18 March, 2003 – 05:00 UTC 18 March, 2003 - The surface analysis for late afternoon shows pretty good agreement with the MesoEta winds but the temperature forecast was a little high over north-central / northeast CO as well as southeast WY. Note the strong northerly winds in north central CO. This is a result of a barrier jet developing along the Front Range of the Rocky Mountains. The barrier jet causes the boundary layer to deepen and the winds to accelerate as they are deflected southward.

Throughout the evening temperatures remain in the low 40s and high 30s along the northern Colorado Front Range, while southeast Wyoming observations showed low 30s and snow most of the afternoon. Note the intensifying northerly winds and changeover to snow at both Laramie and Cheyenne, WY by 01:00 UTC. Notice that the low pressure and circulation center indicate that the low is in reasonable agreement with the Eta. The changeover from rain to snow took place in Front Range cities that were 5000 feet or greater elevation around 05:00 UTC. Note this is the time of the arrival of the shortwave seen on IR imagery.

26. GOES-12 10.7 um IR loop 20:15 UTC 17 March 2003 – 06:00 UTC 18 March

2003. Early in the loop we can see north-south bands of colder cloud tops in north-central CO. These tops were associated with north-south oriented squall lines (radar loop soon). From 02:30 – 05:30 UTC the short wave moving into Colorado from the southeast causes the areal coverage of the colder cloud tops to expand due to upward motion. Notice that the cold cloud tops associated with the WCB now extend from southwest Nebraska through northeast CO. During this time, the precip type along the Front Range was rain or rain mixed with snow. Recall the Eta had the warm conveyor belt (WCB) further north and east, but satellite shows the WCB is a little further south than the forecast. As the shortwave moves into the Front Range the precip area broadens and precip rates increase. The short wave observed in west Texas continues to expand in areal coverage as it moves north.

27. Platteville, CO profiler data for 19:00 UTC 17 March – 07:00 UTC 18 March 2003. A reflection of the changing dynamics can be seen in the vertical wind structure as the shortwave moves into the area. This corresponds with the WCB enhancement noted on the IR. The winds at all levels increase substantially between 02:00 and 04:00 UTC. The areal coverage of the precipitation increases substantially during this time, as well. Between 05:00 – 05:30 UTC observers from Fort Collins through the Denver area report a changeover from rain/snow to snow (coincident with the arrival of this disturbance). The changeover probably occurred as a result of the continuous and deeper precipitating layer. With precip falling from higher levels, there would be greater cooling of the lower levels.

28. KFTG 0.5 degree tilt reflectivity for the period 00:47-10:27 UTC 18 March 2003. During the period 00:04 through about 02:30 UTC the precipitation takes the form of discrete, north-south lines of intense convection during which severe weather was observed on the Plains. The convective lines were associated with periods of relatively heavy rain along the Front Range. Beginning shortly after 02:30 UTC precipitation becomes widespread (associated with the arrival of the shortwave and enhanced WCB) and transitions to snow between 05:00 – 05:30 UTC along the Front Range. Notice the small circulation moving in from the southeast. This was the small wave that moved up from New Mexico (referred to in the water vapor imagery discussion earlier).

29. GOES water vapor loop 00:30 UTC 18 March 2003 – 12:00 UTC 18 March 2003. Between 11:00-12:00 UTC we notice the thick clouds associated with the first period of heavy snow move off to the northwest and weaken. Fort Collins had on average 8” of snow reported by 14:00 UTC. However, note the intense shortwave/jet max in southwest TX that moves northeastward into western Kansas by 12:00 UTC. This feature was not evident on the model output from the morning of the 17th. New convection can be seen forming as the shortwave reaches the WCB. What is the forecast implication of this? (It is key to holding the low in place, allowing heavy snow to begin agan, and continue for nearly 24 more hours).

30. GOES IR loop 10:15 UTC 18 March – 01:15 19 March 2003. The new convection seen late in the previous loop is indeed associated with a second shortwave, and an invigorated WCB. The intensified WCB moves into northern Colorado during this period, leading to a long duration of heavy snowfall starting around 19:30 UTC (18 March) and continuing through the morning of the 19th.

31. GOES-10 IR loop for the entire event. This loop reviews the entire event from 00:00 UTC 18 March to 23:45 UTC 19 March, 2003. Remember that the imagery tracks the development and movement of WCB’s and provides nowcasting guidance for the beginning or ending of the various phases of this event. The Eta represents the first and second shortwave/WCB’s discussed, however in both instances the model has the max precip amount associated with the features too far north. The role of satellite data here is critical as you can see how the Eta forecast precipitation fields are evolving and adjust your forecast positioning and consequent QPF amounts utilizing both in tandem.