Characteristics of the variation pattern of divergence

Temporal change s in t he amount of convergence/di vergence in the dayti me and di urnal variati ons in the amount of convergence/ di vergence are affect ed by diurnal variations in the prevaili ng wind and l ocal wind system devel opment . A si gni ficant spatial variati on patter n was theref ore extracted from t he dail y convergence/ di vergence fiel ds, and the cau ses for the principal variations in t he amount of convergence/di vergence were examined to

classif y the sunny li ght -wind days from t he perspecti ve of the local wind circulation and synoptic field variations. At the same ti me, it i s thought that the converge nce/ di vergence field change s along wit h the decadal changes in the wind syst em; thus, the decadal changes in the appear ance f requency of classified wi nd systems on sunny li ght -wi nd days can also be discussed. The empiri cal ort hogonal function (EOF) anal ysis , also called pri ncipal component anal ysis , was conducted on the hourl y amount of convergence/ di vergence between 09:00 and 17:00 on sunny li ght -wind days in the Kanto Plai n in order t o exami ne the princi pal variatio n pattern of convergence/ di vergen ce fiel d on sunny li ght -wi nd days. The hourl y convergence/ di vergence amount in grid points at an elevation of 200 m or lower was used. Any gri d poi nts where val ues coul d not be obtained due t o missi ng within 10 days at each grid point were inter polated wit h the aver age values of all t he other sunny li ght -wind days.

Fi gur e 8 shows the di stribution of t he fact or loadi ng of the EOF anal ysis for the first t o thi rd principal components, and Fi g ure 9 shows the ti me seri es of the hourl y mean EOF score of each principal component. EOF scores ar e indicated by t he values nor mali zed by the st andard devi ati on of each principal component . The contributi on rate of each principal component was 13.4% for the first principal component, 11.2% for the second pri ncipal component , and 7.5% for the third pr incipal component . Cumulati ve contribution up t o t he third pri ncipal component was 32.1% and was not ver y large. However, since the diurnal variation in the EOF scores for the f ourth

and latt er principal components is unclea r, t he fir st to thir d principal components wer e used in this st udy. Furt her more, the days when the EOF score for each hour was an average value plus st andard deviati on of 0.5 or more (average val ue minus 0.5 standard deviation or l ess) wer e defined as days with positi ve anomal y (or days with negati ve anomal y) for each hour.

For ever y 2 h fr om 09: 00 to 17:00, the distribution of wind and convergence/ di vergence amount on days wi th positi ve and negati ve anom al y for each princi pal component are shown i n Fi g s. 10 to 15.

The first princi pal compone nt had a spatial patter n with negati ve factor loading fr om Kanagawa Prefecture to the prefect ural border of Tokyo and Chiba, and positi ve f actor loading for east ern Sa it ama Pr efect ur e and Chi ba to Ibaraki Pr efectur e s. Thi s di stributi on corresponded t o t he border of convergence and di vergence areas that t ypi call y appear ed i n the dayti me on sunny li ght -wi nd days . EOF scores showed a diurnal variati on wi th a positi ve peak at 10: 00 and a negati ve peak at 16:00. On days wit h positi ve anomal y at each hour ( Fi g. 10) , compared wit h an aver age of sunny li ght -wi nd days (Fi g.

7), southerl y sea breeze was weak. Yet, on days wit h negati ve anomal y (Fi g.

11), sout herl y sea breeze became pr edomi nant and the wind speed increased.

Penetr ation of t he sea br eeze front ont o t he i nland area was f ast as well . Based on t hese factor s, it is thought that t he first pri ncipal component was related to t he diur nal variati on in sea br eeze development and temporal variations of each case, and indi cate d t he nort h–south variations of the convergence and di vergence areas of t he sea breeze f ront .

The second princi pal component had a spat ial patter n with negat ive f actor loading around Tokyo Bay and the nort hern Kant o regi on , and positi ve fact or loading on the eastern part of Chi ba P refecture. This distributi on corresponded to t he di vergence ar ea i n Tokyo Bay and the northern Kanto region that appear ed f rom morning until around 13:00 , and convergence area around Chiba Prefect ure, respecti vel y. EOF scores showed a diur nal variati on in which t he negati ve peak was at 12:00 and the val ues became positi ve from 15:00. On days with positi ve anomal y (Fi g. 12), sea breeze in t he southern Kanto r egion, di vergence area of Tokyo Bay, and t he valley breeze in t he inland and di vergence area associ ated with the vall ey breeze wer e uncl ear. In contrast , on days wit h negati ve anomal y ( Fi g. 13), these appeared cl earl y.

Based on these factors , it is thought that the second principal component represent ed the di vergence in Tokyo Bay that is t he framewor k of t he sea breeze cir culati on, t he strength and weakness of convergence near Chi ba Prefect ure associ ated with t his , and the di vergence i n the centr al part of the Kanto Plain locat ed at the down -fl ow area of valley breeze circul ation .

The thir d principal component ha d a spati al pattern with negati ve factor loading in wester n Kanagawa Pref ecture, and positi ve factor loadi ng in southern and northwestern Chiba P r efectur e, and at the pr efectural border of Tokyo and Saitama . This distribution cor r espond ed to t he di vergence ar ea around Kanagawa Pr efecture, obser ved in the afternoon accompanyi ng the penetrati on of sout h erl y sea br eeze and convergence ar ea around the prefect ural border of Tokyo and Saitama corresponding to the sea breeze

front. EOF scores showed a diurnal variati on with a negati ve peak at 14: 00.

On days wit h positi ve anomal y ( Fi g. 14), t he sout herl y sea breeze fr om Sagami Bay was weak . In cont rast, on days with negati ve anomal y (Fi g. 15) , the sout herl y sea breeze was str ong. In r esponse to that , t he convergence around the pref ectural border of Tokyo and Sait ama was extremel y strong, and the di vergence around Kanagawa Pref ecture was strong as well. This indicates that the thir d principal component was relat ed t o the south erl y sea breeze and expressed the str engt h of convergence ar ound the pref ectur al border of Tokyo and Saitama and around Tokyo Bay.