TECHNICAL REPORTS OF THE METEOROLOGICAL RESEARCH INS士ITUTE NO．20
MEAN STATISTICS OF THE TROPOSPHERIC MRI・GCM−I BASED ON12−YEAR INTEGRATION
BY
FORE（）AST RESEARCH DIVISION，MRI
気象研究所技術報告
第20号
気象研究所対流圏大気六得環モデル（MRI・GCM・1）
による12年間の積分
予報研究部
気象研究所
METEOROLOGICALRESE入RCH
OCTOBER 1986
INSTITUTE
Meteorological Research暫nstltute
Establishe（i in 1946
Director：Dr．Yoshiro Sekiguchi
Forecast Resea．rch Division Head：
Typhoon Research Division Head
Physica1Mete・；・1・gyResearchDivi忌i・n Head：
ApPhedMete・；・1・gyResearc虹Divisi・n Head
Meteorological Satellite Research Dlvision Head：
Seismology and Volcanology Research Division Head：
Oceanographical Research Division Head Upper Atmosphere Physics Research Division Head：
Geochemical Research Division Head：
Dr．Masahiko Aihara Mr．Keizo Masamura Mr．Hachiro Uemura Mr．Tsunehiro Majima Mr．Syoichi Koinuma Dr．Masaji Ichikawa
Dr．Hayato Ii（1εし
Dr．Hisashi Muramatsu Dr．Yutaka Kawarada
1−1Nagamine，Yatabe−Machi，Tsukuba−Gun，『Ibaraki−Ken305，Japan
Technical Reports ofthe Meteorobgical Research口nstitute
E4πoプー吻一〇h招∫ Syoichi Koinuma
E4露073 ：Yasuo Sato Haruo Ohnishi
ア
Tomoaki Yoshikawa Tsutomu Ta、kashima Masahiro Endoh Yukio Makino 版微zg初g E4π073：Keiko Nishida，Yus母i Yuhara
Takayo Matsuo Hiromi Takayama Katsnumi Hirose
Tθ伽卿1鋤碗3φオhθMθずθ070〆09吻撰63躍じh伽」伽6
has been issued at irrgular intervals by the Meteorological Research hlstitute since1978as a medium for the publication of survey articles，technical reports，data reports and review articles on meteorology，oceano言raphy，seismology and related geosciences，contributed．by the members of the MRI．
序
本報告は，気象研究所予報研究部の大気大循環研究にたずさわる諸氏による大気大循環モデル
（MRI．・GCM−1）による12年間に及ぶ長期積分の結果から，各種気象要素の気候値，偏差につい て水平分布，南北断面，時間変動などを図示したものである。
大気の大規模な循環の実態を把握し，その維持や変動の機構を解明することは，気象学の中心課 題の一つである。大循環に関する解析的研究はこれまでも進められてきたが，FGGE以後，特に全 球的資料の蓄積とそれらの解析を通して実態把握は，より定量的かつ詳細になってきた。
一方，大気大循環に関する数値シミュレーションによる研究は，それぞれに特徴を備えたモデル を用いて世界の主要な気象センターにおいて行なわれてきており，その成果はめざましいものがあ る。気象研究所においても筑波移転に伴って電子計算機が設置され，気象研究所で開発された大循 環モデルによる数値実験が本格的に始められた。このモデルの内容の詳細については，既に，気象 研究所技術報告Nα13（1984）に示されている。
現在，世界気候研究計画（WCRP）の実施がWMO，ICSUを中心に企画されており，大気大循 環の季節変動，年々変動の解明は重要な課題の一つとなっている。数値シミュレーションはこのた めの一つの有力な手段であり，現実大気の長期変動の特徴がどの程度再現されるかを調べておくこ とは，特に必要である。
本報告は，この点についてのMRI・GCM−1の特性を示すものであり．，更に今後の改良，発展と 共に，気候研究分野における主要な道具として成果を挙げることが期待される。
昭和61年6月
長彦部
研正 究 報
原予 究所
研相象 気
一i一
Foreword
Thisrep・rtsummarizestheresults・fthepreliminaryan段lyses・fa12−year integration of the atmospheric general circulations （AGC） based upon the
MRI・GCM−L This work has been done by the collaboration of staff members who participated in．the study of the AGC，one of the prihcipal research subjects of the Forecast Research Division．
It is widely recognized that one of the centr＆I problems in meteorology is to understand the causes and physical mechanisms which control the characteristics of the AGC．The amount of knowledge on the beh＆viors of the＆tmospheric circulation
system h＆s been much increased due to the elaborate analyses on the basis of meteorological data obtained from the global6bservation network，notably after the FGGE．On the other hand，＆nother approach bαsed upon numerical experimentation has seen considerable progress at several meteorological research centers ＆nd universities in various countries，and many remarkable results have been published in the journals every year．
A high−speed electronic computer was installed in the Meteorological Research In甜tute（MRI）at the time of its removal from Tokyo to Tsukuba Science City．
Since then，r6search activities on the general circulation＆t th6MRI have been greatly expanded．Details of the numerical model（the MRI・GCM−1）used in this study have already been described in the Technical Report No。13（1984）of this series．
Implementation plans of the World Climat Research Program（WCRP）are now going on under the joint cooperation of the WMO and ICS．U．The purpose oHhis research program is to underst＆nd the climate v＆riability＆nd its causes，so that the
understanding of the long−term fluctuations of the global atmosphere，for ex．ample，
vari＆tions over the time scale from one season to several years，are one of the important subjects．
Numerical experimentation seems to be the most appropriate for achieving this end． However，it is desirable to estimate to what extent the numerical model simu1＆tes the characteristic features of the large−scale circulations in the real
atmosphere．In this report，the performances of the MRI・GCM−I in the evalu＆tions of the climatological averages and fluctuations around them under the amual cycle of
⁝m
climat・1・gicalb・undaryf・rcings且resh・wn。Thr・ughthec・ntinualeff・rtst。impr。ve
thequality・fthepresentMRI・GCM，wearesurethattheMH・GCMwilleventually
make it possible to predict climatic variability in a fairly satisfactory manner．
June 1986
Masahiko Aihara，Head Forecast Research Division
一iv一
Contents
概要（和文）
Page
1
Introduction 3
Chapter l Horizontal distributions of monthly mean quantities（1）
12−year averages Sea level pressure
Geopotential height at300mb Temperature at800mb
Zonal wind at200mb Streamline at200mb Streamline at900mb Velocity potential at200mb
Precipitεltion
Cumulus precipitation 1．10Evaporation
I．11Sensible heat flux
1．12Tota．I diabatic heating rate 1。13Cloudiness
1．14Snow depth 1．15Ground wetness 1．16PBL depth
1．17Water vapor mixing ratio at900mb
11111111111
Chapter2Horizontal distributions of monthly mean quantitles（H）
yeartOyearVariatiOn 2．1 Sea level pressure
2．2 Geopotential height at300mb 2．3 Temperature at800mb 2。4 Velocity potential at200mb
15 15 16 16 17
一V一
2．5 Precipitation
Page
18
Chapter3Latitude−height cross sections of monthly mean quantities Zonal mean temperature
Zonal mean zonal wind Zonal mean meridional wind Meridional stream function
Zonal mean mixing ratio of water vapor
Zonal means of the standard deviations of monthly mean quantities．
Standard deviations of monthly and zonal mean variables Total diabatic heating rate
Chapter4Mean annual variations
1・234 56789
1atitude−time cross sections Solar flux at the surface
Solar heating rate Long−wave・heating rate Total radiational heating rate Cloudiness
Precipitation Evaporation
Sensible heat flux at the surface Total diabatic heating rate 4ユO Surface air temperature 4．1LSea level Pressu「e 4．12Zonal wind 4．13Meridional wind
Chapter5Longitude−time cross sections
5．1Deviation of geopotential height from the zonal mean 52 Anomalies of geopotential height
5。3 Zonal wind at the equator 5．4Meridional wind
99001
り乙り乙QJ3QJ一vi一
5．5 Total d量abatic heating rate
Acknov》ledgemens
References
Page
31
33
35
一v五一
LIST OF171GU1〜ES
Fig．0 Surface boundary condition used in the MRI・GCM−1
Page
39
Globa翌distribution of monthly mean va，lue
﹂2345￡︒7名9JJユ91︒1︒1﹂4＆999＆9＆＆9＆999・9999 Sea level pressure
Geopotential helght at30σmb Temperature at800mb
Wind velocity of the zonal wind at200mb Streamline at200mb
Streamline at900mb
Velocity potential and its diverge興t wind at200mb Precipitation ralte
Preclpitatlon rate by cumulus convection 1．10Evaporation rate
1．11Senslble heat flux from the surface
1．12Vertically integrated net atmospheric heating l．13Cloud三ness
1．14Snow depth 1．15Ground wetness 1．16Depth o｛the PBL
L17Water vapor mixing ratio at900mb
Ensemble average of the monthly mean value，the standard deviation from the ensemble average and deviation of the monthly mean from the ensemble average for each year Fig．
Fig．
Fig．
Fig．
Fig．
Fig．
Fig．
Fig．
Sea level pressure：January Sea level pressure：ApriI Sea level pressure：July Sea level pressure：Oct6ber Geopotential height at300mb Geopotential helght at300mb Geopotential height at300mb Geopotential height at300mb
January・
April July October
1 1 1 1 11 1
一ix一
壌客客︒19●19．19︒19●19−19︐19嬉・19 2．9 Temperature at800mb：January…
2．10Temperature at800mb：、Aかril−
2．11Temperature at800mb二July−
2．12Temperature at800mb l October…
2．13Velocity potential at200mb：January・・
2．14Velocity potential at200mb：April 2，15Velocity potential at200mb：July・
2．16Velocity potential at200mb：October・
2，17Precipitation rate：January 2．18Precipitation rate：April……
2．19Precipitation rate：July 2．20 Precipitation rate ：October・… 一・
P
Latitude−he｝ght cross section Fig．
Fig．
Fig．
Fig．
Fig．
Fig．
Fig．
Fig．
Zonal mean temperature Zonal mean zonal wind Zonal mean meridional wind Meridional stream function
Zona1，mean mixing ratio of water vapor
Zonal mean of sta興dard deviation of monthly mean geopotential height，
meridional wind and temperature
3．7 Standard deviation of the zonal and monthly mean geopotential height，
meridional wind an4temperature 38 Zonal mean diabatic粋ea亘nεrate
3．1
3．2
3．3
3．4
3．5
3．6
184 186 188 190
192 zo、nal wind，
194
zonal wind，
202 210
Time−latitude cmss section of zollal mean value Fig、
Fig．
Fig．
Fig．
Fig．
Fig．
1234醍︾6
444444
Downward solar flux at the surfaceHeating of the total air column by solar radiation Heating of the total air column by long−wave radiation Heating of the total air column by total radiation Total cloudiness
Precipitationr＆te
213 214 215 216 217 218
X『
Fig．
Fig．
Fig．
Fig．
Fig．
Fig．
Fig．
4．7 Evaporation rate
4．8 Upward sensible heat flux at the surface 4。9 Total diabatic heating of the total air column 4．10Surface air temPerature
4．11Sea．level pressure
4．12Zonal wind speed at200mb，500mb and800mb 4．13Meridional wind speed at200mb
Page 220 221 222 223 224 226 232
Hovm611er diagram
Fig．5．1 Deviation of geopotential height from the zonal mean at300mb （at70。N，50。N，30。N，30。S，500S，and70。S）
Fig。52 Geopotential hβight anomaly from the l2−year mean at300mb （at70。N，50。N，30。N，300S，50。S，and70。S）
Fig．5．3 Zonal wind at the equator（at200mb and900mb）
Fig．5．4Meridional wind at（800mb，30。N），（900mb，0。）and（800mb，30。S）
Fig．5．5 Total heating rate of total air column at70。N，50。N，30。N，Oo，300S，
50。S a．nd70。S………・・…
238
262 274 282
294
xi一
概 要＊
気象研究所技術報告第工3号で述べた気象研究所大気大循環モデルー1（MR I・GCM−Dを用 いて12年間の数値積分を行った。モデルは100mbにトップを持ち，上下5層，経度方向に5度，
緯度方向に4度の解像度を持つ。外部データとして必要な海面水温と海氷分布に関しては気候値を 用いた。従ってモデル実験では両者は毎年同じ季節変化を繰り返す。
大気大循環は年々気候値のまわりを変動しているが，モデル大気の長期間積分の平均としで得ら れる気候値とその気候値のまわりの変動がどのように再現されるかを検討することが今回の実験の 主要目的である。
この技術報告では，12年間の積分で得られたモデルの気候値，標準偏差，季節変化等を主要な変 数について図示する。結果についての詳細な解析は現在進行しており，ここでは結果の解釈・議論 はほとんど行なわない。まず結果を図の形で公表するのが本報告の目的である。第1章では，月平 均の気候値の水平分布図を示す。第2章では海面気圧等の5つの量について月平均値の標準偏差の 水平分布図を示す。あわせて各年の月平均値の気候値からの偏差の分布図も示す。第3章は月平均 の気候値の緯度・高度断面図及び月平均値の標準偏差の東西平均図，東西平均量の月平均値の標準 偏差を示す。第4章では，東西平均量の緯度・時間断面図を示す。第5章では，経度・時間断面図 を示す。対応する観測量が得られるものについては，あわせて図示してあるので，観測値のデータ
・ブックとしても利用されたい。
＊時岡達志・山崎孝治・鬼頭昭雄
一1一
Introduction＊
The climate system is composed of theatmosphere，the ocean，the cryosphere，the lithosphere and the biomass，It has various tlme−scales of variation．The shortest one is the upper limit of the predictable period in the sense of short−range weather forecastipg． The largest one is the span of the earth s age itself， Even if we limit ourselves to the shortest time−scale，our present knowledge and understanding of their characteristics are still very fragmentary．As is repeatedly stated in the documents of the World Climate Research Pr・9｝am，m・delresearchis・ne・fthepr・misingapPr。achesf。rtheimpr。vement。f。ur present knowledge and understanding of them．Atmospheric general circulation models have succeeded in reproducing the basic characteristics of the global circulations of the atmosphere．
Of course，the models are still far from complete．Model atmospheres show systematic biases from the real atmosphefe。 For example，see TokiokaθヵαZ．（1985）for the Meteorologlcal
Research Institute model（the MRI・GCM−1）．
Although the model is not identical to the real atmosphere yet，we can still consider the former as a good analogue of th￠latteL Study of long−term variations．of a model atmosphere is therefore useful for the understanding of the variations of the real atmosphere as well as for the understanding of the model climate itself and thus for the improvement of the model．
Manabe and Hahn（1981）first conducted、such a study．They・integrated a spectral general circulation model with rhomboidal truncation at wavenumber15for18model years，and analyzed the15−year data。 Their results show that most variability in the．extratropics is reproduced without year−to−year changes in thq prescribed sea surfすce temperature（SST）．Lau
（1981）、has further surveyed the characteristics of long−term variations of the model atmosphere． Recently Lau （1985）has run two 15−year integrations including ob串erved year−to−year changes of the SST over the tropical Pacific basin．Heミhowed that the temporal variance of the200mb height in the perturbed SST rμn is larger than the corresponding quantity in the climatological SST run by a・factor of2−60ver the tropics；．whereas・the same SST fluctuations are much less effective in enhancing the variability in middle and higher latitudes．At Oregon State University（OSU）a10−year integration has been performed with
‡he OSU two−layer GCM。Kushnir and Esbensen（1985）have studied a subseasonal varia炉ility
＊Pr6pared by T．Tokioka，K．Yamazaki and A．Kitoh．
一3一
for northem hemisphere winter with the use of simulated data。
Recently，a12−year integration has been performed at the Meteorological Research Institute （MRI）with the5−layer MRI・GCM−I without prescribing year−to−year changes in the SST．This technical report is presented for the purpose of showing the mean statistical maps of the MRI・GCM−I based on the12−year run．Therefore detailed descriptions，interpretations and discussions of the results are not included。They will be published in separate papers in near future。
〃 04θZ
The model（the MRI・GCM−1）used for the present study is a tropospherlc five−layer grid model with lts top at100mb．The horizontal resolution is40in latitude and5。in longitude．
Physical processes included are the parameterizations of the penetrative cumulus convection by Arakawa and Schubert（1974）and of the planetary boundary layer based on Randall（1976），
radiation（．Katayama，1972），ground thermodynamics and hydrology（Katayama，1978）． The diumal variation as well as the seasonal variation of solar insolation is included in the model．
There are no interannual varlations in prescribed boundary conditions，乞．召。，the sea surface temperature and the sea ice coverages．The prescribed sea surface temperature and the sea ice distributions together with the model topography used are shown in Fig．0．See Tokiokaθ診
αム（1984）for further details of the mode1．January and July、performances of the model are presented in Tokiokaθ孟αZ．（1986）and Kitoh and Tokioka（1986）．
翫Pθ7伽傭
The run was started at January100Z of year l and was continu6d until March100Z of year13．The initial conditi6n of the run was taken from the atmospheric condition on January l produced by another simulation with a different version of the model． Excluding the first tw・．m・nths・fthesimulati・n・the12−yearsi血ulateddataareanalyzedlnthisrep・rt。Data
were originally sampled for every6hours．In this report，maps are based not on the origlnal but on the10−day mean and the monthly mean data。
Chapter l describes the horizontal distributions of the12−year average of the monthly mean quantities。Chapter2shows the horizontal distributions of the year−to−year variation of the monthly mean quantlties、 Chapter3shows the latitude−helght cross sections of the monthly mean quantities．Chapter4describes the mean・annual variation in the latitude．time
一4一
cross seρtions．Finally，Chapter5shows the longitude−time crgss sections at selected latitudes。
Both the mean annual variation and year−to−year varlations are lncluded．
一5一
L Horizontal distributions of monthly mean quantities（1）
12−yearaverages＊
In this chapter，horizontal distributions of some selected variables are presented．Monthly aver孕ge for12years is taken for each var玉able．Diabatic process related variables are exactly the mean of12years for a corresponding month。On the other hand，adiabatic fields（sea level
pressure，geopotential height，wind field，temperature and water vapor mixing ratio）are stored once every6hours and those、 instantan曾ous data are used to calculate the means．9−point spatial・smoothing is applied to variables plotted in this chapter．
1．l Sea level pressure
Figure1．1shows the mean sea−level pressure for each month．Contour interval is4mb．
Values greater than lo20mb are shadedl those less than Iooo mb dotted．In the Northem Hemisphere，the Siberian high is seen from October through February，although its center is 1・catedfart・thes・uth・fthe・bserved（Fig・2・1・1）・AnAsianm・ns・・nall曾wpressureareais found in July and August．Subtropical highミare Iocated in the east of the Pacific Ocean・and the Atlantic・Ocean during summer．HQwever，the southwestward extension of the ridge of the
high over the Pacific Ocean is not well simulated，resulting in a northward displacement of the ridge near Japan．This is in accord with the excessive rainfall over the Pacific Ocean near 30。．N・180。E（Fl9・2・19・1）・Highs・verthe．eastemPacificOceanand・verN・rthAmericaare
combined to make one anticyclone systenl in the northem winter．Deep lows are situated in the northem parts of the Pacific Ocean and the Atlantic Ocean in winter．There is an
exces＄ive low−pressure zone in high latitudes during summer and autumn．This deficiency of them・delisin『acc・rdwiththeexcessiverainfall（Fig・L8）・theexcessivelywet3r・und
condition（Fig．1．15）in this season and the excessively low temperature over the arctic s6a ice
（Kitoh and Tokioka，1986）．In the S6uthem Hemisphere，a low pressure belt surrounds the Antarctic continent throughout the year．Subtropical highs have their centers over the continent in the southern winter and over the oceans in the southem summer．
＊This，chapter is prepared by A．Kitoh，Forecast Research Division
一7一
Tech．Rep．MeteoroL Res．Inst．No．201986 1．2 Geopotential height at300mb
Figure1．2shows the geopotential height distribution at300mb、Contour interval is80 g．p。m，Additional contours of9720，9780and 98003．p．m．are drawn with dashed lines。Three
centers．of anticyclones are clearly seen over the continents around200S in February．The northem counterpart over the we6tem Pacific Ocean is the most dominant in the northem winter．This center moves its position from『winter to summer between the westem tropical Pacific Ocean and South China．In the Northem Hemisphere，anticyclones over Saudi Arabia
・and Mexico become stronger than the one over South China in summeL The former two anticyclones・appear from July through September．The latter one moves its center・rapidly in May，June and，again，SeptembeL These feat皿es are in good．agreement with，the
observations．
1．3 Temperat皿e at800mb
Figure1．3shows temperature at800mb．Contour interval is5℃．Cold air lower than
−30℃is seen from November to March around eastem Siberia and the adjacent Arctic Ocean with the coldest（一35℃）contour in January．Temperature gradient around Japan is large from October to April．In the northem winter，cold surge around the eastem periphery of the TibetanPlateauis￠vident・butitextendsfarsq翠tht・thelnd・chinapeninsula・The20℃
contour appears in March ih the Sahel region（15。N）and・moves northward with time toFthe Sahara desert（25。N〉until August．High temperatures in the northem summer observed over the Middle East（Fig．2．11．1）are not reproduced in the model probably due．to the wet ground
condition。Temperature in the eastem part of the continents is lower than that in the westem part in the northem winter．This symptom is less clear in the Southem Hemisphere continents，
which is in good correspondence with the observation（Fig．2．9。1）．
1．4 Zonal wind at200mb
Figure1．4shows the wind velocity of the zonal wind at200mb．Contour interval is10m
s『1andvalueslessthanOms−1aredotted．StrongjetsarelocatedoverJapanandeastem
North America with peaks greater than70m s『1and60m s一1，respectively，in・both January and February。Subtropical jets in the Southem Hemisphere are located at30。S and45。S in the southem winter and summer，respectively，The wind velocity of those jets is around40m
s『1an has less variation with season compared to those in the Northem Hemisphere．The maxima of jets tend to lie to．the south or over the Southem Hemispheric contlnents，i．e．，
一8一
Tech．Rep。Meteorol．Res．Inst．No．201986
Africa，Australia and South America．Easterlies prevail in low latitudさs with their maxima in
the・s岬merhemisphericside・alth・ughthereexistwesterliesinthecentralandtheeastem
Pacific Ocean．
L5 Streamline at200mb
Figure1．5shows the streamline at200mb。In the northem winter，there are pairs of anticyclones in low Iatitudes．Anticyclones in the Southem Hemisphere are stronger than those in the Northem Hemisphere over Africa and South America，while、the reverse holds in the
westem Pacific Ocean and Australia。In the northem summer ther6are anticyclones over South Asia（the Tibetan high）and over Mexico with another weak high pressure belt at10。S．
Mid一・ceanictr・ughsarepresent・vertheP註dficandtheAtlanticOceaninb。th、hじmispheres in January，and only in the northem Hemisphere in July．Cross−equatorial northerly flow exists
over the Indian Ocean from May to SeptembeL On the other hand，there exists southeτly flow over the maritime contin6nt，from December to February．
1．6 Streamline at900mb
Figure1・6sh・wsthestrea卑lineat900mb・Overallfeaturesarec・nsistent噂hthesea
level pressure in Fig，1．1。A distinct feature．is the seasonal progression of the monsoonal wind system． During thenorthem winter there is，a strong anticyclonic flow aro丘nd the Tibetan，
Plateau．The resultant easterhes prevall ovgr the Indochina peninsula and the Indian Ocean．
These easterlies weakeロby the end of Apri1，although the子e still exists an皐nticyclonic circulati㎝over the Indochina peninsula。The Mascarene high is seep all year round over the South Indian Ocean around30。S。Easterlies along the northem periphery of the Mascarene high begin to have a southerly component in April at the east coast of Africa。In May the monsoon circulation is clearly seen and in June it is fully established with．a、Strong cross−equatorial flow，the Somali jet and westerlies over the Indian subcontinent。In September dry northeasterlies start to blow over the Indochina peninsula，while there stiH is a cross−equatorial flow from the Southem Hemisphere along the east coast qf Africa。An anticyclonic circulation is established i October with the low一互evel ea．sterlies over the Indi4n subcontinent and the cross−equatorial flow．over the Indian Ocean reverses its direction in November and remains northerly until next March。
The dry northerly flow begins tQ bloW across the、equator in Novemb爾r over the maritim6 continent，It penetrates into Australia in December，accompa耳ying a cyclonic circulation in the
一19一
Tech．Rep．Meteorol．Res．Inst．No．201986
northwest of Australia．This system Iasts for two months and begins to fade out in February．
Strong cross−equatorial flow fromthe Northem Hemisphere to the Southern Hemisphere prevails bver South America from December through March and in the reverse direction至rom May through Auεust．The latter is a part of a large anticyclonic system over the eastem South Pacific Ocean and South America．♪robably due to the insufficient representation of the narrow Andes especially over Pem，easterlies prevail over Brazil and Peru in the southem winter。
L7 Velocity potential at200mb
Figure1．7shows the velocity potential and its di〉ergent wind at200mb．Contour interval
・fthevel・cityp・tentialis106m2sτ ・Thedivergentcenterisl・cated・verthecentraI
equatorial Pacific Ocean near the datehne from October through March and over the maritime continent and the Indochina peninsula in the rest of the year。Distributions of divergent and convergent centers observed in January and July（Figs。2．13。1and2。15。1）are well reproduced in the model1However，those in the transition seasQns in the model are different from the
observed（Figs．2．14．1and2。16．1）．
The rapid transitions of the center occur between March and April・and between September and October in the modeL The former transition is quicker than th Iatter。This transition is also seen from the movement of the heavy cumulus precipitation area in Fig．1．9．
Convergent centers are situated over the Tibetan Pla．teau in t血e northem winter and over the broad area from the eastem Pacific Ocean through the Atlantic Ocean to Africa all the year round，However，the latter center of convergence tends to form a single strong system over the Atlantic Ocean in the horthem summer and tends to split into two centers over the
eastem Pacific Ocean and over Africa in winter．
Strong divergりnt wind is simulated over the westem Pacific Ocean from November through February and over the south Indian Ocean from May through August，corresponding to the A、sian winter and summer monsoons．
1．8Precipitation
Figure L8shows the preciかitation rate．Contours of1，2，5，7．5a！nd10mm day『1are drawn．Values greater than5mm day−1are shaded and those Iess than l mm daジl dotted．
Basic characteristics of observed climatological feat皿es（Figs，2．17．1，2．18．1，2．19．1and2．20．1）
are well simulated l in the modeL However，differences are found regionally，e．g．exaggerated precipitation over the summer continents and precipitation maximum not over Assam but over
一10一
Tech．Rep．MeteoroL Res．Inst．No．201986 thelndianOceaninJuly．
Awell−organizedprecipitationareaisfoundinthetropicsexceptforov弓rtheeastem
Pacific Ocean where sea surface temperature is low．PreCipitation over the Atlantic Ocean is also less compared to that zonally averaged。Heavy precipitation greater than IO mm day−1is simulated over the Indian Ocean，around 10。N in May，June and Julyl over the equatorial Indian Ocean，South Africa and Brazil in the northem winter．In that season are also found large precipitation areas over the northem Pacific Ocean and the northem、Atlantic Ocean，
which correspond to、cyclonically active areas．The precipitation in the baroclinic zone of the Soりthem Hemisphere is uniform in the zonal direction due to the absence of prominent stationary circulations，Excessive precipitation is found over the Eurasian continent and the North America in June and July．Precipitation is also ekcessive around．the dateline between 20。N and30。N in the northem summer，accompanying the northward displacement of the subtropical higlh（Figs．1．1 and 1．6）and its dry zone．
1．9 Cumulus precipitation
Figure1．9shows the precipitation rate by cumulus．convections which have their roots in the planetary boundary layer（pBL）．Contours are the same as in Fig．L8。A co㎎parison of
Figs．，1．8and1．9士eveals that precipitation in low latitudes is mostly・caused by cumulus precipitation．Excessive raipfall over continents．at hi帥1atituqes in中e Nqrt無em Hemispねgre from June through August noted in Chapter L8is due to cumulus precipitation。
1．・10 Evaporation
Figure L10shows the surface evaporation rate with the same contours as in Fig，L8．
Large evaporation is simulated over the central and the westem Pacific Ocean and the Indian Ocean in low・latitudesl Evaporation from the Arabian Sea is、relatively small in・Ahgust，
September and October。Off the east coast of Japan and North．America，there exist heavy evaporation areas due to cold surges in the colder months of the year。The same situation is found with Iess magnitude off the east coast of South Africa and South America．Also noted is a peak、of evaporation to the・northwest of Australia in June and July by the dry southeasterlies from i旦land Australia（see Figs．1．6and1．17）．Evaporation over continents in the summer hemispheres is exaggerated。
一11一
1．11 Sensible heat flux
Figure1．11showsthes6nsibleheatfluxfrom・the−surface。Contourintervalis50Wm一2
and，negative values are dotted．Large positive sensible heat flux is seen off the・east coast of Japan and North America in winter months．Also Iarge are the fluxes over the oceans around the sea ice／ocean boundaries in both hemispheres in the winter畠easons．Over the continents，
large sensible heat flux is simulated in the Sahara Desert throughout the year，the southem part of Africa and South America in winter，and Australia in summer。Negative sensible heat flux is found／over the continents in winter hemispheres and over some parts of the oceans．
1．12・ Total diabatic heating rate
Figure1．12shows the vertically integrated net atmospheric heating rate by『diabatic processes．Contourintervalis50Wm−2．andpositivevalues（atmosphereisheateddiabatically）
are shaded．The condensational heating and the net radiational heating（actually cooling）are main contτibutors to℃he net・atmospheric heating／coohng．As the latter is rather uniform in space，the spatial pattem of the net atmospheric heating r6sembles that of the condensational
he註ting（Fig．1．8）．The sensible heat flux convergence（Fig．1．11）is also substantial regionally，
e。9。thedesertUregi・nsand・fftheeastc・asts一・fthec・ntinentsandthes♀aice／・cean
boundaries in winteL The at卑osphere is diabatica．Ily cooled in polar regions，subtropical high pressure areas，the eastem Pacific Ocean and the Atlantic Ocean in low latitudes and the continents in winter hemispheres．The westem part of the Tibetan Plateau becomes a heat
sou「ceinA叫althoughtheeasteτnpa「tofit「emainsa・heatsinkuntilJunemai軽lyduet・
the late disappearance of sn6w cover（Fig．1．14）．
・1．13 Cloudiness
Figure1．13shows the cloudiness．Two types of clouds are incorporated in the model．
Clouds associated with the grid−scale［supersat皿ation and those with・the sub−grid−scale penetrative cumulus convections。In those cases we assume that the cloud fiUs the grid box entirely and that the cloudiness is LO．Cloudiness of shallow cumuli whose tops are below400 mb is set to O．0．Contour interval is20％．Values greater than80％・are shaded and those less than20％dotted。Generally，cloudiness is lafge in high latitudes and has its mini血a in the
subtropics．The model overestimates the cloudiness in high latitudes due to the excess of shallow clouds by the grid−scale condensation and underestimates it in low latitudes probably due to neglecting the shallow cumulus with its top below400mb．
一12一
Tech．Rep．Meteorol．Res．Inst．No。201986
1．14 Snow depth
Figure1．14shows the snow depth distribution．Contours are drawn for1，5，10and50cm』
ahdvaluesgreaterthanlocmareshaded．N・tethatsn・w−c・veredareaswh。sesn。wdepthis
less than l cm are not shown in this figure．In the Northem、Hemisphere，the snow−covered area is most extensive in March。It begins to retreat in April。It still covers the Tibetan Plat3au，Siberia and Alaska in June。Althoug耳model August is the month when there is no
sn・w−c・veredareawithgreaterthahlcmsn・wdepthintheN。rth益mHemisphereexcept．f。r
Greenland where the permanent ice sheet ground condition is specified，the snow−covered area is the minimum in July，when snow stiU exists ln parts of the Tibetan Plateau，the
northeastem Siberia and Alaska。In model August the snow−covered area begins to spread southward and occupies the major part of the Eurasian continent at the end of October。In the Southem Hemisphere，only the southemmost part of South America ls covered by snow in winter with a depth oHess than l cm．As seen in the maps，snow over Greenland and Antarctica never disapPears even in summer．Actually sn6w depth over those regions are lncreasing during the integration，corresponding to the observation where ice sheets are being formed．
1．15 Ground wetness
Figure1．15shows the ground wetness．Ground wetness is predicted by considering ground hydrology with the uniform maximum water content of15g cm−2in the ground．Contours of 10，30，50，70andl90％are drawn and values less than50％are dotted，Extensive dry areas are found over the continents in the subtr6pics．Over the land which is covered by snow in winter，it generally begins to dry after the complete disappearance of snow cover with the maxi甲um15g cm−2water content．＊It takes time for the ground to be sufficiently dry，even lf the local surfa．ce evaporation surpasses the precipitation．In July the ground in the Northem
Hemisphere is extremely wet compared with the observed，that is，wetter by more than50％。
＊Owing to a coding error，the・ground water content was set to the maximum15g cm−2 whenever a snowfall occurred irrespective of the past history of the ground wetness．This error must have a large influence on the model chmate ov6r regions where snow falls intermittently or snow depth is not enough，through a ground wetness／evaporation／precipita−
tion link．The effect of the error is under investigation．
一13一
Tech．Rep．MeteoroLRes。Inst。No。201986
1．16 PBL depth
Figure1．16shows the depth of the PBL．Contour interval is20mb and values greater than140mb are shaded．Generally，the PBL is deeper over the oceans than over the
continents．The noctumal shallowing of the PBL seems to be responsible for this，because the atmosphere usually becomes stable at night over the−continents．The PBL is deep over areas of large cyclonib activity，i．e。，off the east coast of Japan，North America and South America in their winter months．Deep PBL is located at the Bay of Bengal in winter．It is deep all the
year round over the equator side of the circum−Antarctic low pressure belt。Other maxima are found off Callfornia a．nd off Pem．
1．17Water vapor mixing ratio at900mb．
FigureL17sh・wsthedistributl・ns．・f，thewatervap・rmixingrati・at900mb・C・nt・ur interval is l g kg−1and values greater than ll g kg−I are shaded．In January moisture is
abundant over the Indlan Ocean，the central and southem portions of Africa，the Amazon Basin，the Australian Mediterran♀an Sea（the Timor串ea and the Arafura Sea）and the belt extending from there to the dateline at the equator and east−northeastward，corresponding to the ITCZ．The contour of10g㎏一1moves northward over the westem Pacific Ocean in May。
In June it has a kink to the south of Japan，which resembles the Baiu−front．It reaches Japan in July and August to form the wettest region at this latitude．
一14一
2．Horizontal distributions of monthly mean
year−tO−year VariatiOn＊
quantities（1）
In this chapter the ensemble averages，the standard deviations of the monthly averages and叢he deviations of each year from the ensemble averages are shown for January，April，July andOct・ber，t・getherwiththe・bきervedv41ues，f・rthesむa−levelpressure，thege。P。tential
height at300mb，the temperature at800mb，the velocity potential at200mb and the precipitation．The notation Y13 ，for example，indicates the 13th year o｛the model atmosphere。The maps of ensemble averages are the same as shown in Chapter L
2．l Sea level pressure
The ensemble average and the standard deviation of the monthly mean sea−level pressure
・fthem・delatm・spherさandthedeviati・ns・feachyear fr。mtheensembleaverageare shown in Fig。2。LL−Fig。2．4．4，together with the climate values of the monthly mean sea−level pressure and the standard deviation for January，April，July and October．The climate values are based on the ECMWF analyses，for the period from1980to1984．The contour interval of the lefthand panels in Fig．．2．L1，Fig．2．2．1，Fig。2．3．．1and Fig．2．4．1is4mb．The
area over1020mb is shaded and the area below1000mb is dotted．The contour interval of the righthand panels in the same figure is2mb，and the area over6mb is shaded．The
contour interval iq the rest of the figures is4mb and the negative area is shaded。
The ensemble average for January（the upper left panel of Fig．2．1．1）is slightly different from the map presented in Tokiokaααム（1985）in polar regions，where the standard deviation is largβ（the upper right panel of Fig。2．1．1）。Although distributions of the standard deviation are
different in details between the model and the observed climate，the values are comparable between them except over Antarctica，where the model gives less variation than the observed climate。
In model January（the upper right of Fig．2．L1）the maximum standard deviatiqn in the Northem Hemisphere is found at the Gulf of Alaska．Maxima are also found over the n・rtheastempart・ftheAtlanticOceanandthes・utheastempart？ftheTibetanPlatead・ln
the devitation maps of eac葺year（Fig。2。1．2−Fig．2。L4），north−south seesaw pattems are also
＊This chapter is prepared by T．Tokioka，Forecast Research Division。
一15一
Tech．Rep．MeteoroL Res．Inst．No。201986
evident（θ．g．Y13，YO3，Y10and Y12）．The north−south seesaw pattems are commonly found in Aprll，July and October，although their amplitude are4ess than that in January。
2．2 Geopotential height at300mb
The ensemble average and the standard deviation of the monthly mean geopotential height at300mb o∫the model atmosphere and the deviations of each year from the ensemble average are shown in Fig．2．5．1−Fig．2．8．4，together with the climate values of the monthly mean geopotential height at3 00r血b and the standard deviation for January，April，July ・and October．The climate values are based on the ECMWF analyses．The contour interval of the lefthand panels in Fig．2．5．1，Fig．2．6．1，Fig．2．7．1and Flg．2．8。1is20g。p．m。The contour interval of the righthand panels｝n the same figures is20g．p．m．，and the area over60g．p．m，is
shaded．The contour interval in the rest of the figures is20g，p．m．The negative area ls shaded．
The model has succeeded in reproduclng basic characteristics in both the monthly mean geopotential height and its standard devlation．However，the standard deviation in the model is systematically less than the climate value．The tendency ls clear in relatively high latitudes，
especially4n April and October．This can also be confirmed in the latitude−height cross section
（Fig，3．6。1and Fig．3．62）．In the deviation maps of each year for each month，wavy pattems
are conspicuous at this level compared with those in the sea level pressure。
2．3 Temperature a，t800mb
The ensemble average and the standard deviation of the monthly mean temperature at 800mb of theτmdel atmosphere and the deviations of each year from the ensemble average are shown in Fig．2．9．1−Fig。2。12．4，together with the climate values of the monthly mean temperature at800mb and the standard deviatlon for January，April，July abd October．The climate values are・based on the ECMWF analyses。The contour intervaL of the lefthand panels in Fig．2。9。1，Fig2。10。1，Fig．2。ILI and Fig。2．12。1is5℃．The contour interval of the right hand panels in the same figures is1℃，and the area over3℃is shaded．The contour interval in the rest of the figures is 1℃。The negative area is shaded。
The ensemble average of the modeHs below the climate value in polar regions，especially in July．as pointed out by Kitoh母nd Tok1oka（1986）．The standard deviation of the model is less than the climate value，especlally in high latitudes of the Northem Hemisphere in January．The deviation maps in Fig。2。92−Fig，2．12．4have good correspondence with those of the sea
一16一
Tech。Rep。MeteoroL Res．Inst．No．20・1986
Ieve｝pressure in Fig．2．1．2−Fig．2．4．4．
2．4 Velocity potential at200mb
The ensemble average and the standard deviatlon of the monthly mean velocity potential at200mb of the model孕tmosphere andりhe deviations of each year from』the ensemble average are shown in Fig。2。13。1−Fig．2．16．4，together with the climate values of the monthly mean『velocity botential at200mb and the standard deviation for January，April，July and October、The cl1mate values are・based6n the ECMWF analyses fOr the period from1980 t・1984・Thec・nt・urinterval・ftheleftha血dpanelsinFig．2．13．1，Fig．2．14．1，Fig．2．15．1and
Fig，2．16。1is106m2s臼曽1．一The contour interva1ρf the righthand panels in the same figures is
2・5×105m2s−1，andthearea・ver10×105m2s−1isshad♀d．Thec・nt・urintervalln毛herest・f
the figures ls2．0×105m2s一正．The negative ar6a is shaded．
The model has succeeded in reproduclng basic characteristics in the monthly mean velocity pqtentlal at20σmb（the left panels in Fig．2．13．1，Fig．2．14．1，Fig。2．15．1and Fig．
2・16・1）・H・wever・therearelar＄edifferencesinthestandarddeviati・n（therighthandpanelsin Fig。2。13．1，Fig、2。14．1，Fig．2．15．1and Fig。2．16．1）．The maximum value in the model is about％
to％of the corresponding clima㌻e value。Besides that，the dis廿ibutions of the maxima are very different between the model and the climate．The year−to−year variation in the sea−surface temperatllre is not considered ln the present model．rNeglecting this variation seems to be the most responsible for the difference。The intra−seasonal variations are not simulated well in the present model（Tokioka and Yamazaki，1986）．This may also be responsib．le for the underevaluation of the standard deviation．
The underevaluation of the standard deviation is found．in almost all variables，as mentioned so far．However，the velocity potential field seems to be one of the quantities most sensitive to neglecting the year−to−year variation in the sea sur｛ace temperature．
In Fig．2．15．1 are shown the same quantities as are shown in the lower panels of Fig。
2．15．1but for the data analyzed at NMC for the period from l979to1983．This figure is added to demonstrate large differences in the standard devlation of the velocity potential between the two data sources．The monthly mean field in Fig、2。15．1 is very close to that in Fig．2．15、1．
However，the standard deviation in Fig．2．15．1 is about half of that in Fig。2．15。1，although similarity exists in the location of peaks between them。
一17一
Tech．Rep．MeteoroLRes．Inst．No．201986
2．5Precipitation
The ensemble average and the standard deviation of the monthly mean precipitation of the mod61atmosphere and the deviations of each year from the ensemble average are shown in Fig．2．17．1−Fig．2．20．4，together with the climate values of the monthly mean precipitation for January，Apri1，July and October．The climate values are bas6d on the data compiled by
Shutz and Gates（1971，1972，1973，1974）．Contours of1，2，5，7．5and10mm day−1are shown．
In the lefthand panels of Fig．2．17．1，Fig。2．18．1，Fig．2．19．1and Fig．220．1，the area over5mm daジ1is shaded and the area below l mm dαy−1is dotted，The contour interval of the
、
righthand panel in the same figures is l mm day−1，ahd the area over2mm day−1is shaded．
The・contour interval in the rest of．the figures is lmm day−1．The negative area is shaded，
The model has reproduced well the basic climatological characteristics in the monthly mean precipitation in low latitudes．However，the present model gives excessive precipitation over the continents，especially in the summer hemisphere．Although the climatological standard deviation of the monthly mean precipitation is currently not available，it is highly probable that the standard deviation in the present modehs much less than the actual value，as in the
velocity potential at200mb．However，the maximum standard deviation in Jaunuary is as much as4mm daジ10ver the equatorial Indian Ocean．In the present model，northerlies are too strong near the surfarce along the eastem coast of China，while easteries are also too
strong in the southem periphery of the Tibetan Plateau（Fig．，1．6．1）．This anticyclonic flow is cold and dry，and thus enha．nces evaporation in the Bay of Bengal．and precipitation over the equatorial Indian Ocean（Tokiokaθ診α」．，1985）．The Iarge standard deviation in this area is closely comected with the i血terannual change of the Siberian high，as confirmed by comparing lhe maps of tke sea level pressure d￠viations（Fig．2．12−Fig。2。L4）with tkose of the
precipitation deviations（Fig。2．17．2−Fig．2．17．4）．
一18一
