Filtering the Merged Files

Description of Scatter Plots FILTERING THE MERGED FILES

One of the goals of the ABFM project is to consider the feasibility of using the radar to imply if the for electric fields in the region are a hazard or not. To do this we have calculated several parameters using the radar data and plotted those parameters on a scatter plot against the magnitude of the electric field as measured on the aircraft. There are some conditions where the radar data are not valid. For example, if the radar has been attenuated due to wet radome or intervening precipitation then the radar during those times should not be used in the analysis. These times should be filtered out (removed) from the data set used to create the scatter plots.

Table 1 is a list of the columns in the merged file that we used for filtering the data. The first column is the column name (in the merged file). The second column discusses the column contents. The third column shows how the filter was applied for the scatter plots.

Table 1: Filter Columns in the Merged File

COLUMN HEADER

HOW TO USE IT

VALUES WE USED

LDARm5 CGm5 LDARpm5 CGpm5

You need to decide how many lightning flashes you will accept. The range is +/- 20 km from the aircraft and -5 minutes.
We have used the m5 (for minus 5, rather than the pm5 for plus and minus 5).

ldarm5 < 2 LDAR (we accept 1 LDAR source)
cgm5 = 0 CGLSS (we accept 0 CGLSS sources)

Cloud_Type

go to Merged File Info button: Cloud Types

0	=	no anvil
20	=	debris
10	=	distinct anvil with base >= 5 km altitude still attached to core
9	=	distinct anvil with base >= 5 km detached from core
8	=	anvil with particles measured but no detectable radar return
6	=	gnd clutter and other low altitude returns present, but still considered an anvil
4	=	Anvil with distinguishable base, but base below 5 km. Cases too far from radar to reliably determine base are excluded./td>

cloud_type &ge 5 (accepting Cloud_type &ge 5)
AND
cloud_type &le 10 (accepting Cloud_type &le 10)

I recommend looking at the 4.5, that is the value I gave the anvil_type if I thought "someone" might consider it an anvil, but I didn't think so.

Pass_num

A sequential number of passes for this flight.
The form is pass#.anvil# e.g.
9.2 = pass #9 in anvil #2
spirals are designated by an S on the end of this parameter,
7.1S = pass # 7 in anvil #1 was a spiral

Did not use Pass_num for filtering

invoid_74C invoid_88D

These are 0/1
Was the aircraft with 10 km of the cone of silence at 10 km altitude? If yes, then invoid = 1

invoid = 0

core_20km

This is 0/1
If the aircraft was within 20 km of a convective core with 35-40 dBZ showing on the 4 km CAPPI or higher then core = 1

core = 0

atten_74C

This is 0/1
If the time coincides with the attenuation times then atten_74C = 1

atten = 0

frac

Using the correct fraction column for the cutoff desired, a filter can be made assuming that a certain fraction of the column contains detectable points.

frac(.05) i.e. at least 5% of the column contains detectable signal : frac &ge 0.05
frac(.10) i.e. at least 10% of the column contains detectable signal : frac &ge 0.10

base

A filter can placed accepting only data with a base at a certain value

Values chosen for our work include:
base &ge 4
base &ge 5

Table 2 shows the filters used for the various scatter plots. Column 1 is the name used in the applet for choosing a particular filter. Column 2 is the identifying name for that filter when looking at a list of files. Column 3 shows what filters were actually used for that scatter plot and the values for those filters. (The invoid filter was used for the appropriate radar. e.g. invoid_74C was used only for the 74C radar). The attenuation filter was only applied to the 74C radar. Column 4 is a clockwise listing of the plots presented.

Table 2: Scatter plot filters

FILTER NAME

FILE NAME

FILTERS USED

PLOTS

Original

fan_clmin

invoid	=	0
pos_alt	&ge	5000
atten_74C	&le	0
LDARm5	<	2
CGm5	&le	0
core_20km	=	0
cloud_type	&ge	5
cloud_type	&le	10
frac11x11	&ge	.05