EAS 308, Meeting 18
Compare RAMS-predicted with actual convective boundary layer (CBL) evolution
Last meeting, you learned how to use the RAOB program to analyze the soundings for your R4N contest day. As a result of this learning, I assigned you (Assignment 12, due today prior to class) to determine the actual depth of the CBL for your 12, 15, 18 and 21Z soundings. In this meeting, you will learn how to determine the predicted CBL depths.
Background
From the OSTIV paper in our class readings here’s how the depth of the CBL is determined using the RAMS predictions:
Following Olofsson and Olsson7, the predicted surface temperature and dew point values and the predicted environmental sounding were used to determine the height of the cloud-free and cloud-topped CBL as depicted in Figure 3:
The dT term in the figure denotes the buoyancy of the rising air parcel where
dT = 0.4 x (1 + W/200) x 20/ff (B-1)
and W is the net sensible heat flux near ground in W/m2 (if W 200, W = 200) and ff is the wind speed near ground (10 m) in km/h (if ff < 20 (10 kts), ff = 20). It was found by Liechti4that when W was calculated to be greater than 200 W/m2 in reality the value would not exceed 200 as the sensible heat flux is dispersed by thermals. The ff term in Equation (B-1) accounts for the effect of strong winds “shredding” thermals near the ground.
Determining predicted CBL depths
I have analyzed the RAMS weather prediction (using my forecast regions) for all the contest days and sent you your yymmddhh.ram weather file via e-mail prior to this meeting. Now let’s learn how to read the TTC display of your *.ram file. We’ll use my contest day (061014) as anexample.
First, I locate the TTC release of 23 October 2006 (the latest) and launch the application. Remember, my regtherm.cfg file that I have given you (it’s on the web as Assign11_Answer.cfg and must be renamed regtherm.cfg for TTC to recognize the file) must be in the same directory as your TTC. This is because TTC looks for the regtherm.cfg file to display the forecast regions. When you launch TTC and select Mid-Atlantic SA in the upper-right hand window, my forecast regions will be displayed.
Next, display the region numbers by clicking on Specials and on Region numbers. The region surrounding the M-ASA airport (39.7N, 77.3W) is 800. This is the region in which you collected the sounding data from the NOAA-ARL-READY archives. It is the region we will focus on for validating the RAMS CBL-depth predictions.
Continuing, display the 06101400.ram file by clicking on Weather, RAMS and browsing to where you put your *.ram file. Double click on the file icon and the weather should be displayed (but you loose the region numbers, memorize where region 800 is!):
The first features that you see are the colors that correspond to the potential flight distance (PFD). This is the horizontal distance a Standard Class glider can fly if the pilot used the first soarable thermal of the day to the last soarable thermal of the day. You see my contest day is quite good with PFD values of 300 to 500 km.
Now, we turn our attention to the time-height graph to the right. Use the scroll arrows to adjust the height to 5 km and the time to start at 6 and end at 18 (EST). Then put your cursor in the Region 800. You see many things. At this time, for our CBL prediction, we focus on the squiggly grey line going from left to right. This is the top of the CBL in km. To get the depth you must subtract the airfield elevation of 179 m. So, estimating the depths that correspond to the times of the soundings I’ve analyzed here’s what results:
Time (EST) CBL-top (m MSL) CBL-depth (m AGL)
07 est 000 000
10 est 800 621
13 2300 2121
16 2300 2121
I enter these data in my Assignment 13 Excel file (duplicate the Assignment 12 Excel sheet and add in these data):
Time / Time / Est. altitude / Station elevation / actual CBL depth / predicted CBL depth(Z) / (EST) / (m MSL) / (m MSL) / (m AGL) / (m AGL)
12 / 7 / 440 / 179 / 261 / 0
15 / 10 / 1276 / 179 / 1097 / 621
18 / 13 / 2382 / 179 / 2203 / 2121
21 / 16 / 440 / 179 / 261 / 2121
I performed a linear regression analysis because I expect the actual and predicted CBL depth to be positively related. That is, they both should increase in a similar fashion because they are linked through the convection mechanism:
0. Build a Scatter diagram in Excel (see above)
1. Right click on a data point and choose “Add trend line”
2. Choose “linear” type
3. Click on “Options”
4. Choose “Set intercept to 0” and “Show R2 value on chart”
A perfect positive correlation would have a R2 value of 1 and a perfect inverse correlation would have a value of -1. Here we have a value of near zero indicating no correlation due to the poor correspondence between the 16EST actual and predicted CBL depth values. Inspecting the actual 16EST sounding (not shown)revealed an extremely modest surface temperature inversion forming, an inversion not predicted by RAMS. To show the influence of this data point, I removed this 16ESTdatapoint and the correlation increased dramatically:
Thus, the RAMS prediction of the FORMATION of the CBL at the M-ASA airfield on 14 October 2006was quite good. However, the DISSIPATION of the CBL was not predicted well.
What did you find for your contest day? That’s Assignment 13.
18-1