|
<br />".
<br />
<br />co
<br />
<br />I
<br />V
<br />
<br />Before discussing the results of the
<br />model comparison with observations, several points
<br />concerning this phase of the workshop need to be
<br />emphasized. First, it should be recognized that
<br />the data sets were not collected for this purpose
<br />and are, therefore, not as appropriate as they
<br />might need to be. The second point is that model
<br />runs were necessarily limited by the available
<br />input data and program run costs, an expense
<br />which was borne by the sponsors of the partici-
<br />pating modelers. Considering the sensitivity of
<br />model results to input conditions, this is an
<br />important point to remember when examining the
<br />comparative predictions. Lastly, it should be
<br />remembered that a major purpose of this model
<br />comparison was to lay the foundation for designing
<br />a more explicit and meaningful model-observation
<br />experiment that would be conducted in the future.
<br />
<br />Tables 2 and 3 give results of the
<br />various model runs for clouds which fit into
<br />approximately the size of the observed clouds for
<br />that day, except for the slab-symmetric models
<br />which predict their own size clouds. Table 2 is
<br />for August 10 and table 3 is for August 17. In
<br />general, the one-dimensional, steady-state models
<br />or single parcel Lagrangian models, which are
<br />models 1 through 3, predict the taller clouds
<br />observed on the days. Their maximum vrrtical
<br />velocities ranged from 15 to 25 m sec-. Note
<br />the different results for model I (la and lb of
<br />tables 2 and 3). These two sets of results show
<br />the model's sensitivity to different initiating
<br />impulses of 2 and 1 m sec-l, respectively. This
<br />further illustrates the importance of proper
<br />model initialization. In general, all of the
<br />parameters for each of the three models showed the
<br />proper tendency when compared with observations.
<br />That is, the parameter differences tended to be
<br />of the same sign, though not necessarily of the
<br />same magnitude, as the observed differences
<br />be~veen August 10 and 17. The rainfall accumu-
<br />lations from these lD, SS models were not con-
<br />sidered'to be a realistic prediction and were
<br />not taken very seriously by the modelers at the
<br />conference.
<br />
<br />,<::,~' ,O>~.
<br />
<br />,,'
<br />,"
<br />
<br />,,'
<br />,..
<br />
<br />,,'
<br />,+;
<br />
<br />,oA~'
<br />
<br />".
<br />J"
<br />
<br />100.
<br />
<br />..J'..J~"'.J-..,f..J'J'~~~~Y
<br />
<br />
<br />
<br />~. ~~fi~~~~~~~~'::
<br />
<br />
<br />
<br />9Q / ..~t .. / ',' t., ',.", ,', /
<br />
<br />300. <>~:5t~:~.:~;:>~~;:~~:~:.::;:~~~:~;~ "'",
<br />
<br />~oo. .-,- .....--f-.---\~.-r-~ -.--.~__....L..;.o..-~.L--4..~~ ...
<br />~;~. >< ....... ~~..~:... V(.....~<~;.~~~.>;<\........(......~. . ?,()'
<br />500. ~-/-::J-x/ ;r.~<L~~~7' ;:L"-;'-!--'v"c"", './..' ~".
<br />,,'.. ' '.' '. . ". /, ". \ ". '. '
<br />600 %-7':' --':', - ',,-:: -"+7"'-\.."~P-'--.:)...-L'':::' '::-'...
<br />., ~"" ...... .~ ~..<;;,.)~.. ,''-'>}'l.1 ....< \ ....1. ......;1./.... <;/1.<:;)'
<br />700. -,......;- - . .;--,-t;; _ .;k--3....;~~-a.-,~\2'-.:;-'~r ""':.-
<br />
<br />:gg:~7~'" .::~... ':-:Z;;";.~:~.~~~~~~;;.::::.?!:~
<br />
<br />1000. . . . I. .
<br />
<br />STL811
<br />INIT1AL SOUNDING AT 13 HR 0 HIN
<br />Fi(JW'e 2. Skew-T analysis showing terrperatu:r>e and
<br />dewpoint profiles for St. Louis, Mo., on August l7,
<br />Z073. at lJOO L.S.T.
<br />
<br />~~L_~____,-"<-_
<br />
<br />The one-dimensional, time-dependent
<br />(lD, TD) models numbered for the 10th, of. .
<br />which three gave fairly realistic results. For
<br />August 17, there were five runs made with one-
<br />dimensional, time-dependent models. In general,
<br />the one-dimensional, time-dependent models gave
<br />lower cloud tops than the steady-state models.
<br />The bases were predicted by these models and, in
<br />some cases, checked out fairly well, and in all
<br />cases within one grid interval of observed
<br />cloud base. Maximum vertical velocities for these
<br />models on August 10 ranged from 16 to 19.8 m sec~l
<br />and on August 17 from 11.4 to 21.5 m sec-l. The
<br />rain rates on August 10 were up to 54 rom hr-l,
<br />which agreed well with the observations for that
<br />size cloud. Rainfall accumulated on the ground in
<br />these models was, in general, less than observed.
<br />Two of these models did predict hail at the ground,
<br />which was observed, and most of the models
<br />predicted the right tendency, at least partly.
<br />
<br />The two-dimensional, time-depeudent
<br />(2D, TD) slab models predicted a field of clouds
<br />and were successful in predicting the proper
<br />tendency from the 10th to the 17th. In addition,
<br />the cloud tops, vertical velocity maximum" maximum
<br />rain rate, and maximum accumulated rainfall were
<br />within the range of the observations. Neither of
<br />these models predicted significant hail at the
<br />ground, although one of them does include a hail
<br />process.
<br />
<br />t >--
<br />t
<br />
<br />~.r'A
<br />. .
<br />
<br />:/' -"-~,~':
<br />I I I
<br />20 30 40
<br />TIME
<br />AUGUST 17, 1973
<br />
<br />I
<br />'"
<br />
<br />I
<br />"
<br />
<br />I
<br /><0'
<br />
<br />
<br />\
<br />
<br />,1
<br />I
<br />I
<br />I
<br />I
<br />I
<br />/.-f\--r....,
<br />, .
<br />~.
<br />
<br />_......1.--..J
<br />" "
<br />
<br />TIME
<br />AUGUST 10. 1973
<br />OBSERVED RAOAR ECHO TOPS '1$. TIME -NEAR ST. LOUIS, MISSI)URI
<br />
<br />Fi(JW'e 3. Corrparison between observed radar echo-
<br />top height (solid lines) and predicted time-height
<br />cross section (dashed) from three modeZs for
<br />August lO and l7, 1973. Note the similar'
<br />characteristics of models and observations.
<br />
<br />~
<br />
|