estimation of divergent wind from olr data for use in objective...

8
, \la"5a"" (1993)• .J4, I. 551. 50 I . R: 551. 521. 3 : 551. 509 . 313 Estimation of divergent wind from OLR data for use in objec tive analysis over the Indian region P. L. KULKARN I. D. R. TALWALKAR. S. NAIR " ",I S. RAJA:\lA NI Indian Institute 0/ Tropical ,Heleo/' olog l', PUfI( ' (R eceived JJ November 199 1. Modifi('d 15 J UII (' 1992) 'fI'1.: - 3i'f?I:TtP. Wflf7: 1JT 'jfl,: 'fi"Tf"ilf ff et:r q 'lVf ::j; '4f lf n jfr. -mT-it i !;JlfJ1T t'! >f.T f'Pfr rnfl I Pi -q m it i! fiR T-r 3iA:t:R ir (", In ;ito- ':R if llR 'PIT q;fr;r,T'll'J IfTT fiR" t rr.rr 1J" ':rf I 00 Q. tr. J.i ,.. "fl , .f q s 1Fflf 9"f, tr Tif 'f." T;:H', jf 'llfl:l1J' , f :nv.r 3fT: 'T"{;Jf i, f'Pir I J:;T.n ir 1"" Hri f«t , <Mi jf rf:: =i: iJ ("f iT trt ;{)7 if fJPl Tm ,h,f :'fI. :t r.r RA 9" i:f,"T";{ til :rr"J1m l in: I "?i .r.r if :f; f; "'T ir h:m l'f.'lf I qA i 6fTfRif J:.f:f ;f.' 'F.Tf if X5t1 0:", qt, 'T7 : 'J:l" 'R'f J:;f.:t 'f"I f-:r,;:r T l'J1f 1 W .,\ -q S lNfrt 1 :::r<r..,?r 'f." An: fn9-'J'lT t 'lf7'l'Jffil ij f'f; f:r,ifT'f' WfIfM;t A 'i: ti rnf t JrPir fl ' iWI t t AHSTRAC r . A ...chcrm i... formulated for the use of Ol R data in the es timat ion or ver tical velocity. tli- vergence and then the divergent pan of the wind ove r Indian region. In thi... scheme, ascending mot ion over cloudy region i" estimated from an empirical relation between the cl oud top te mpe r ature and descending 11 10- rion over cloud -free region is estimated f rom the t hermody namic ene rgy equation and both arc blended. Fr \1I11 this blended vertical velocit y field, uivcrgencc. velocity po tent ial and divergent wind" at all standard levcl-, f rum .t to ,.. 1979 at OJ UTe arc com puted. Th ese fields a rc co mpa red with sat ellite cloud pict u res , ra infa ll CIC and they arc found to be realistic in depicting the synoptic co nditions. Total wind is comput ed a'i sum of the evtirnated divergent compone nt and rotational co mponen; comp uted from obse rved wind field. For asses- SOlen! of the scheme. thi ... tota l wind field at 850 hPa is used as initial field in univariate op timu m inter - polation sc he me and a nalyses were m3(] C for the period 4 to 8 July 19 79 , Results show that scheme i" a ble (0 produce reali stic analyses whic h include d dive rgent part of the wind. 2. 1. D('linealion of cloud)' alld clolld-frc(, regiol/ 2. :\Iethodolog)- K"sahar a ('I 01. (1988) h"d sugges ted a scheme in which the ascending motion was mor e dir ectl y c omput ed f rom OLR data and the dzscending mo tion f rom th e thermodynamic en-:rgy equation. Following th is scheme, in the present study we have c omput l.:d th ': vert i- c,,1 moti on "nd subsequently estimaled the divergent wind over the Indi"n region. Thi s diverg ,n t wind has been included in the "nalysis thu s improving the initial specification of wind field for NWP. In this scheme, the vertic"lIy int egmt ed divergence is m"d e zero with lhe constmint that vert ical velocity is zero "t 1000 " nd 100 hPa. Mord.. - Divergent winds, OL R, Objective analysis, NWP. Equivalent blad bod y le mp:rat ur( , W" S sep"ratcly comput ed the requirement thaI the vertically integra ted divergence is zero. was not ensured in this study. 1. Inlrodaction Initi al specification of divergence is import ant for reali stic forecast of precipitati on r"te par ticularly in the tropics. Due to non-divergent in iti,,1 conditi ons precipitation mt es "re und erf orecast in the first few hours of forecast bec"use model requires several h our s to develop a consistent vertical velocity field . Thi s 'spin up' pr ob lem in f orecast models m"v be due to ina ccurate specificaion of divergence, and thermal fields elc ( K"sa ha m ('I 01. 1988). Many stud ies ha ve been m"d e 10 incorporat e the diverge nt wind in "naly sed wind s so th at specific" tion of divergence in the initial condition is impro ved in order to reduce the spin up pr obl em . Sumi el al. (1979), Sumi (i9 81), Tarbe ll ('I 01. (1981), Juli "n (1984), Krishnamurti el 01. (1984), Kri shn"murt i & Low-N"m (1986) and S"lmon & Thomas (1986) h"ve "tlempted to specify the divergent part by differ ent meth ods. Most of these studies have used OLR d"t a to estimate divergent p" rt of th e wind. In our earlier study (Kulk"rni ('I 01. 1992) also using regression rel"t ion between OLR d ata and the divergence, lhe J"tler w"s estimated. Then the divergent part of wind The cloudy region is " regi on of ascending moti on was obtain' d "nd included in the initial guess field of the and the rest of the cloud- free region is likely to be of analysis scheme. Since the divergence at every level descending mot ion . In this scheme the descending (77)

Upload: others

Post on 26-Feb-2021

0 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Estimation of divergent wind from OLR data for use in objective …metnet.imd.gov.in/mausamdocs/544113.pdf · 2019. 2. 21. · the spin up problem. Sumi el al. (1979), Sumi (i981),

,\la"5a"" (1993)• .J4, I. 77 -~.t

551. 50 I . R: 551. 521. 3 : 551. 509 . 313

Estimation of divergen t wind from OLRdat a for use in objective analysisover the Indian region

P. L. KULKARN I. D. R. TALWALKAR. S. NAIR " ",I S. RAJA:\lA NI

Indian Institute 0/ Tropical ,Heleo/'olog l', PUfI('

(R eceived JJ November 1991. Modifi('d 15 J UII (' 1992)

'fI'1.: - 3i'f?I:TtP. ~Jj , Wflf7:1JT 'jfl,: 'fi"Tf"ilfff '~Jf et:r q 'lVf ::j; ~-.fJfR:1 '4f lf n ~ r >:f,<:R ~ <i;~ , jf r. ~~. ~rT. -mT-i t i!;JlfJ1T t'!~~~" >f.T~ f'Pfr rnfl I Pi !{~A -q m it ~ i! :)frf~ fiR T-r 3iA:t:R ~~t1f~ ~ ir (",In ~pn ;ito­zl~ ':R if ;{<Ti~ llR ~r ~I~;;r.t ~'l'lf.fl, 'PIT q;fr;r,T'll'J IfTT fiR" trr.rr ,, ~n ~'Rl:it f~r f~r 1J"':rf I fllh~",

~~P"~1'f J:;f.:t ~ 00 Q. tr . J.i,.. "fl , .f q s 1Ffl f 1 ~' i ~ 9"f, ~ tr Tif'f." T;:H', jf 'llfl:l1J' , ~'f] f:nv.r 3fT: ~'111 rP7f 'T"{;Jf i,~r>:f,r.'Ti'f f'Pir tt;~ r I~ J:;T.n ir "'.~r 1""HriJj~ f«t, <Mi jf rf:: =i: iJ ("f iT trt ~ ;{)7 if fJPlTm ,h,f:'fI.:t r.r RA 9"i:f,"T";{ til :rr"J1m'H~ l in: I "?i~ .r.r ~f~ ~~ if ~'f:ffi f~ ~~rfT"f ~. "f~frll q~ :f;i~ f; "'T ir ~ i'f: f'"f:f

h:m l'f.'lf I ~ qA i ZJ~T-" ir f~l'!, ~ !~"A';n'1~~mQf;'fl~: ~fT~~ 6fTfRif J:.f:f ;f.' 'F.Tf if X5t1 0:", qt, f~ .

'T7 : 'J:l" ~ 'R'f J:;f.:t 'f"I~lT f-:r,;:rT l'J1f1 W .,\ -q S lNfrt 1~79 :::r<r..,?r 3l'~iil 'f." An: fn9-'J'lT~ lT~ t 'lf7'l'Jffi l ij ~R

~ jfr ~ f'f; ~'l~A ~p",f~ f:r,ifT'f ' ~WfIfM;t A 'i:tirnf ~ ~lfq'~ t JrPirfl 'iWI lff.~ t t

A HST RAC r . A ...chcr m i... fo rmu lated for the use of Ol R data in the estimat ion or ver tical velocity. tli­vergence and then the divergent pa n of the wind ove r India n region. In thi... schem e, ascendi ng mot ion o verclo udy region i" est ima ted fro m an emp irica l r ela tion bet ween the cloud top te mpe rature and descending 1110­

rion over cloud -fre e region is estimat ed from the t hermody namic ene rgy equ a tion an d both arc blended . Fr \1I11thi s blended ver tica l velocit y field, ui vcrgencc. velocity po tent ial an d d ivergent wind" a t a ll standa rd levcl-, frum.t to ,.. J u ! ~ 1979 a t OJ U Te a rc com put ed . Th ese fields a rc co mpa red with sat ellite cloud pict ures , ra infa ll CICand they arc fo und to be rea list ic in de picti ng the syno ptic cond itions . To tal wind is computed a'i l h~ sum ofthe evtirnated divergent compone nt and rotational co mponen; computed from obse rved wind field. Fo r asses­SOlen! o f the scheme. thi ... total wind field at 850 hPa is used as ini tial ~ UNS field in univaria te optimu m inter ­polation scheme and a nalyses were m3(] C fo r th e period 4 to 8 July 1979, R esults sho w that scheme i" a ble (0produce reali stic analyses whic h included dive rgent pa rt of the wind .

2. 1. D('linealion of cloud)' alld clolld-frc(, regiol/

2. :\ Iethodolog)-

K"sahara (' I 01. (1988) h" d sugges ted a scheme inwhich the ascending motion was mor e dir ectly computedfrom OLR dat a and the dzscend ing motion fromth e thermodynamic en-:rgy equation. Following th issc heme, in the pre sent study we have computl.:d th '': ve rt i­c,,1 moti on "nd subseq uently estimaled t he divergentwind over the Ind i" n region. Thi s diverg,nt wind hasbeen included in the "nalysis thus improving the initialspecificat ion of wind field for NWP. In this scheme, thevertic" lIy integmted divergence is m"de zero with lheconstmint that vert ical velocity is zero "t 1000 " nd100 hPa .

Mord.. - Dive rgent winds, OLR, Objective analysis, NWP. Equivalent blad bod y lemp:ratur( ,

W" S sep" ratcly computed the requirement tha I thevertically integrated divergence is zero. was not ensuredin this study .

1. In lrodaction

Initi al specification of divergence is important fo rreali stic forecast of precipitati on r"te particularly inthe tr opics. Due to non-divergent in iti,,1 conditionsprecipitation mt es "re und erforecast in the first fewhours of forecas t bec"use mod el requires several hoursto develop a consistent vertical velocity field . Thi s'spin up' pr ob lem in forecast mod els m"v be due toinaccurate specificaion of divergence, mo~isture andth erm al fields elc (K"saha m ('I 01. 1988). Many studiesha ve been m"de 10 incorporate the diverge nt wind in"naly sed wind s so th at specific" tion of divergence inthe initial co ndition is impro ved in order to reducethe spin up pr obl em. Sumi el al. (1979), Sumi (i981),Tarbell ('I 01. (1981), Juli "n (1984), Krishnamurti el 01.(1984), Kri shn"murt i & Low-N" m (1986) and S" lmon &Thomas (1986) h" ve " tle mpted to specify the divergentpart by different methods. Most of these studies haveused OLR d"ta to es timate d ivergent p" rt of the wind.In our earlier study (Kulk" rn i ('I 01. 1992) also usingregression rel"t ion between OLR data and the divergence,lhe J"tler w"s estimated. Then the divergent pa rt of wind The cloudy region is " region of ascend ing motionwas obtain'd "nd included in the initial guess field of the and the rest of the cloud- free region is likely to be ofanalysis sc heme. Since the divergence at every level descending mot ion . In this scheme the descending

(77)

Page 2: Estimation of divergent wind from OLR data for use in objective …metnet.imd.gov.in/mausamdocs/544113.pdf · 2019. 2. 21. · the spin up problem. Sumi el al. (1979), Sumi (i981),
Page 3: Estimation of divergent wind from OLR data for use in objective …metnet.imd.gov.in/mausamdocs/544113.pdf · 2019. 2. 21. · the spin up problem. Sumi el al. (1979), Sumi (i981),
Page 4: Estimation of divergent wind from OLR data for use in objective …metnet.imd.gov.in/mausamdocs/544113.pdf · 2019. 2. 21. · the spin up problem. Sumi el al. (1979), Sumi (i981),
Page 5: Estimation of divergent wind from OLR data for use in objective …metnet.imd.gov.in/mausamdocs/544113.pdf · 2019. 2. 21. · the spin up problem. Sumi el al. (1979), Sumi (i981),
Page 6: Estimation of divergent wind from OLR data for use in objective …metnet.imd.gov.in/mausamdocs/544113.pdf · 2019. 2. 21. · the spin up problem. Sumi el al. (1979), Sumi (i981),
Page 7: Estimation of divergent wind from OLR data for use in objective …metnet.imd.gov.in/mausamdocs/544113.pdf · 2019. 2. 21. · the spin up problem. Sumi el al. (1979), Sumi (i981),
Page 8: Estimation of divergent wind from OLR data for use in objective …metnet.imd.gov.in/mausamdocs/544113.pdf · 2019. 2. 21. · the spin up problem. Sumi el al. (1979), Sumi (i981),