My WebLink
|
Help
|
About
|
Sign Out
Home
Browse
Search
WMOD00531
CWCB
>
Weather Modification
>
DayForward
>
WMOD00531
Metadata
Thumbnails
Annotations
Entry Properties
Last modified
7/28/2009 2:40:36 PM
Creation date
4/24/2008 2:53:23 PM
Metadata
Fields
Template:
Weather Modification
Title
Raingage Network Requirements from a Simulated Convective Complex Weather Modification Experiment
Date
7/14/1980
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.
/
8
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 />,J~ <br /> <br />~ ." <br />,:.t;J .. <br />'~i <br />~::.~ <br /> <br />'. <br /> <br /> <br />1180 <br /> <br />,..~:.....,...l;o;l:...... . <br /> <br />JOURNAL OF APPLIED METEOROLOGY <br /> <br />VOLUME 19 <br /> <br />y <br />I <br />t <br /> <br />. <br /> <br /> e <br />2.0 2.0 <br /> La~ 5 I <br /> Lare. half <br /> 1.5 <br /> All 103 CC'll ~. <br /> 7 <br /> 1.0 <br /> An 103 CC'. <br /> I <br />.a 1.& <br /> SMail half Sill. II half ~ <br /> <br />o ,', <br />o 10 100 1000 <br /> <br />AREA PER GAGE (km9.) <br /> <br />FIG. 3. Means j, standard deviations s( 10" m3). and coefficients of variation S"ly2. for rain swath <br />accumulation versus lraingage spacing, <br /> <br />which represent among-storm.s plus sampling vari- <br />ances, also decreased for the same reason. The de- <br />crease of y at large gage spacings is due to increas- <br />ingly more CC's being missed entirely. From about <br />100-800 km2 per gage, y increases for the larger <br />CC's (also reflected in combined 103 samples) be- <br />cause the rainfall per gage changes little while the <br />area per gage increases. The variances generally <br />show an increase for wide gage spacings. The coef- <br />ficient of variation, S2/y2, which is an index of re- <br />quired sample size, shows a sharp increase for gage <br />spacings greater than about 300 km2. For the smaller <br />half of the CC' s this occurs before the 100 km2 point. <br /> <br />4. CC's required to detect a treatment effect <br /> <br />The Wilcoxon (Mann-Whitney) rank-sum test <br />(Noether, 1967) was applied to test for significance <br />of treatment effect. Sample CC's were generated by <br />randomly choosing a storm and then randomly seed- <br />ing or not seeding it (1: I ratio). Each CC was then <br />"returned" to the data base so it could be chosen, <br />. at random, again. The rank-sum test was first applied <br />after 15 samples were generated in this fashion, and <br />then in increments of five samples until the required <br />one-tailed a-level was reached. If more than 350 <br />samples were needed, the process was halted to con- <br />I serve computer time. <br />The convergence of four simulated experiments <br />. to an a-level of 0.05 (= 1.645 standard-normal <br />deviates of Wilcoxon rank-sum test statistic) is <br /> <br />; <br /> <br />. <br /> <br />2.5 <br /> <br />2.5 <br /> <br /> <br />o <br />0/0 <br /> <br />100 <br /> <br />1000 <br /> <br />" <br /> <br />shown in Fig. 4. Rainfall accumulations summed <br />from all range bins were used in this figure. An as- <br />sumed treatment effect of i5 = 25% was added to the <br />total rainfall volume of each randomly "seeded" <br />CC. This figure exemplifies the hazards of detecting <br />a seeding treatment with a large noise level. In one <br />case only 25 storms were required, whereas one <br />reGJ..1ired over 350. If two of these experiments were <br />terminated early, the conclusion would be a nega- <br />tive treatment effect! <br />This process of achieving a specified a-level was <br />repeated' 100 times for each assigned value of 0 to <br />give a frequency distribution of the number of CC' s <br />required. This was used to estima.te {3( = 1 - power). <br />The number of storms required to reach a {3-level <br />was the value separating I OO( 1- {3)% of the totals <br />from the remainder of the 100 n~plications. <br />Figs. 5 and 6 iIIustrate the number of CC's re- <br />quired to satisfy various a, {3 and I) constraints. In <br />several examples, more than the maximum allow- <br />able 350 storms were needed. Fig. 5 presents exam- <br />ples for all 103 cases for a = 0.05 and 0 = 100%. <br />It can be seen that increasing th(~ area per gage also <br />!increases the number of CC's required over most <br />of the range <:onsidered. However, the increase in <br />required CC::'s is rather gradual until about 300 km2 <br />per gage is exceeded. <br />For Fig. 6 the 103 CC's were, stratified into two <br />halves according to total rainfall volume. As might <br />iQe anticipated, gage spacing is more important with <br />lthe smaller storms, becoming critical for values <br /> <br />q'r <br /> <br />7 <br /> <br /> <br /> <br />0.5 ,', <br />010 <br /> <br />, <br />tOO <br /> <br />.,..-- <br />11000 <br /> <br />"'. '''- _ .".'1 ~. <br /> <br /> <br />r-~"~~''':;;,~ ,,:~?:~~~~-~'~~-~.~~~~~~~'~-'" -~~J~::~0~'~.'..J~:~?;.~"?:::'~"~'~ ! <br />
The URL can be used to link to this page
Your browser does not support the video tag.