My WebLink
|
Help
|
About
|
Sign Out
Home
Browse
Search
WMOD00562
CWCB
>
Weather Modification
>
DayForward
>
WMOD00562
Metadata
Thumbnails
Annotations
Entry Properties
Last modified
7/28/2009 2:40:53 PM
Creation date
4/24/2008 2:55:57 PM
Metadata
Fields
Template:
Weather Modification
Title
Snow Accumulation Algorithm for the WSR-80D Radar: Second Annual Report
Date
6/1/1997
Weather Modification - Doc Type
Report
There are no annotations on this page.
Document management portal powered by Laserfiche WebLink 9 © 1998-2015
Laserfiche.
All rights reserved.
/
82
PDF
Print
Pages to print
Enter page numbers and/or page ranges separated by commas. For example, 1,3,5-12.
After downloading, print the document using a PDF reader (e.g. Adobe Reader).
Show annotations
View images
View plain text
<br />I-- <br />I <br /> <br />Another approach for a CTF would be to raise the differences to a power like 2.0, which would give even <br />more weight to the heavier hourly accumulations. However, this approach would give undue emphasis <br />to any outliers. The distributions of hourly snowfall accumulations are highly skewed with a least <br />frequency for the heaviest accumulations. When dealing with the heaviest accumulations, it is not <br />obvious what is an outlier and what is a valid measurement. Generally, little difference exists among <br />published Ze-S relations at lighter snowfall rates, say less than 20 or even 25 dBZ, where the large <br />majority of samples exist (see figs. 1 and 2 and table 7 of Super and Holroyd [1996]). But the relatively <br />rare samples at heavy snowfall rates will have considerable influence on calculated Ze -S relations, which <br />can vary substantially beyond about 25 dBZ. Because of the uncertainties with the scarce heavy rates, <br />we have chosen to continue using differences to the power 1.0 for the CTF, giving all samples the same <br />weight. <br /> <br />The optimization scheme used has been found to be reasonably sensitive in selecting the "best" fJ value <br />for most data sets. Figure 1 is a plot of the normalized CTF for fJ values between 1.0 and 3.0 for <br />Cleveland gages 1 and 2 and Denver gages 1,2, and 3. The ranges of these gages from their respective <br />radars are given in table 2. CTF values have been normalized by dividing them by their sample sizes. <br />Figure 1 shows that all gages but Denver's gage 3 have a reasonably well behaved minimum in the <br />normalized CTF near a fJ value of 2. This general pattern for 4 of the 5 gages within 61 km of the radars <br />is encouraging and suggests that the optimization scheme is performing as intended. It is not known why <br />the normalized CTF varies little for Denver's gage 3 for fJ values greater than 2.2. However, the <br />insensitivity of the CTF test for that particular data set suggests its results should be used with caution. <br /> <br />1.5 <br /> <br />I <br />I <br />1.0 l <br /> <br /> <br />1,4 <br /> <br />o <br />o 1.3 <br /> <br />x <br /> <br />Z'1.2 <br />......... <br />..... <br />t- <br />U <br />- '-'" 1. 1 <br /> <br />DEN 1 <br /> <br />0.9 <br />1.0 <br /> <br />1.5 2.0 2.5 <br />Beta exponent <br /> <br />3.0 <br /> <br />Figure 1. - Plots of normalized CTF (criterion function) versus the exponent of equation (1) for data sets from five <br />different gages (see text). <br /> <br />7 <br />
The URL can be used to link to this page
Your browser does not support the video tag.