|
<br />~~.q- --- :~-
<br />.,~
<br />,~<_:-., ';~ ~c
<br />t':,:.. -;
<br />
<br />
<br />,"
<br />
<br />
<br />'~h,.".."""'~....a.-::. " ""="<li~."'",",,""', '.'''~'~'W"=,.,.,~<'::;'';~'''''"'''''a'.~,;I'''.12Z,,,,,. ~_"&'V;Ji:.....,0.<~., ~1
<br />
<br />
<br />
<br />t
<br />,
<br />t
<br />I
<br />t
<br />,
<br />I
<br />
<br />I
<br />f
<br />+
<br />
<br />.....~~.,~...-.-.l-...~~...."'-"...~ f-.;!Q;:"~~.3'1"'.7~.=
<br />
<br />1178
<br />
<br />JOURNAL OF APPLIED METEOROLOGY
<br />
<br />VOLUME 19
<br />
<br />
<br />,
<br />
<br />Many CC's were verified as such by the review of
<br />logs, aircraft and ground-based photography, and
<br />satellite imagery. In fact. it was this review, cor-
<br />related with the PPI displays, that suggested the 30
<br />dBz and 9 km MSL thresholds used in the definition.
<br />Many CC's were excluded because they did not
<br />develop entirely within the 150 km radius area of
<br />radar coverage, moved within 25 km of the radar
<br />where ground clutter is a problem so no data are
<br />recorded, developed after dark, or merged with
<br />another CC in their first 40 min of existence with
<br />radar reflectivity ~20 dBz. A total of 103 CC's
<br />formed the final sample.
<br />
<br />b. Rainswath samples
<br />
<br />. Rainfall accumulation patterns for each of the
<br />103 CC's was estimated from digitized radar data
<br />measured at the lowest (l0) tilt. The Z-R relationship
<br />used was developed near Miles City in western
<br />North Oakota:3
<br />
<br />Z = 155Rl.88.
<br />
<br />c. Raingage network simulations
<br />A computer image of any simulated uniform rain-
<br />gage network can be constructed by specifying three
<br />parameters:
<br />
<br />" Smith, P. L.. Jr.. D. E. Cain and A. S. Dennis, 1975: Deriva-
<br />tion of an R-Z relationship by computer optimization and its use
<br />in measured daily rainfall. Preprints 16th Radar Meteorology
<br />Con/.. Houston, Amer. Meteor. Soc., 461-466.
<br />
<br />. SIMULATED
<br />~IN GAGE
<br />
<br />~x = north-south gage spacing I
<br />~y = west-east gage ,spacing
<br />az '= eastward shift of successive rows
<br />
<br />(2)
<br />
<br />Fig. I illustrates the use of these parameters. For a
<br />, hexagonal array
<br />
<br />~y = 2M ]
<br />~x = .....r3~y/2. "
<br />
<br />(3)
<br />
<br />(1)
<br />
<br />The area represented by one gage is ~x~y which is
<br />the same as would be found by applying the Thies-
<br />soen (1911) polygon.
<br />Rainswaths were superimposed on simulated net-
<br />works. Each gage's accumulation was estimated
<br />soley from the radar range bin (0.5 km by 1.00)
<br />located directly above it. Each simulated network
<br />was placed under each CC once; i.e., there was no
<br />random shifting about to deriv~~ multiple samples
<br />because this would not simulat~~ an actual experi-
<br />ment. Ten hexagonal raingage configurations were
<br />llsed. Except for the 41.6 km2 per gage, the area
<br />represented by each gage was double the area of that
<br />for the next smaller simulated network. The simu-
<br />lated networks had a sufficient areal extent to in-
<br />dude all CC rainswaths. "Ground-truth" rainfall
<br />accumulations in this simulation were assumed tobe
<br />the contributions summed from all range bins.
<br />It is recognized that the use of radar-estimated
<br />rainfall amounts as "true" valu~:s at each raingage
<br />
<br /> Gf .
<br /> AZ
<br /> ax M-~ .
<br />. .i !
<br /> !
<br /> ;
<br />
<br />.
<br />
<br />.
<br />
<br />G
<br />
<br />\..
<br />
<br />/
<br />
<br />FIG. 1. Configuration of convective complex s.arnpling.
<br />
<br />"
<br />
<br />i.
<br />
<br />,.....,...,
<br />
<br />-:<~'.~":f?'-i'.,,,!'i
<br />
<br />
<br />'-";";<"~''''''':;'~:~'~';'. .'Y"'..,
<br />
<br />.','" ;,~"'~'::-"~~iP~~/?~~~r~,,-~.<!,"~,~!"~--t.~-:'~~~~f~J~"P~~~';r'-'~.~':i'?1~~~~!:~~'\..
<br />
|