■ ■- - .... ■—^n—»^-—^*

A0-A015 375

SYNOPTIC PERFORMANCE CHARACTERISTICS OF THE TWO-LEVEL ATMOSPHERIC MODEL

C. Schutz

RAND Corporation

Prepared for:

Defense Advanced Research Projects Agency Office of Naval Research

August 1975

DISTRIBUTED BY:

urn National Tethnical Information Service U. S. DEPARTMENT OF COMMERCE

MM«M—i 1 ■- m^mm

T "^^ ■ Ulflil II .Wll« I.IHUIHB M I I I I II IWIU^fF^W""—- -■■' ■PP" «I II

•'.:.■ :

I .

y ■ ■■■ »,;„■•■ n «i -r

n /.■ I ...'

^ I

The lesea,, h described in this Repoil was sponsored bv Ibe Defense Advanced Research Proiecis Agency under conlracl No. DAHC15-73-C-0181 Reports o(

Jonsors^Zr^h"0' neCe!SanlV 'e"eCI *» ■""— » P-SS 5*1

!>

■- - J

T I ■ II ■ IH ■ I I I ■•• mwm, m . ., — ^Pi«mv«iiucmpmnipnrT"«««Hi ■ n an.

I UNCLASSIFIED ICCURITV CLASSIFICATION Or THIS PAOCrfhr- "«r« Krf.r.rf)

This analysis of the two-level atmoipheric model

compares daily changes during January from three

experiments and five years of observed data for

surface air temperatures at Columbia, Mo., and

five basic characteristics in the vicinity of

the Icelandic and Aleutian Lows. Strong diurnal

changes were found in all simulated temperatures,

but low and high diurnal ranges that would accom-

pany specific synoptic events in nature are not

well portrayed. The positions of the lows tested

based on the lowest daily pressures were clumped,

and in the North Atlantic were located southwest

of the climatological center. Mean January pres-

sures from these centers showed that the simulated

IcelanHic Low is ^.3 mb less intense, while the

Aleutian Low is 1.9 mb more intense than observed.

When all analyzed low pressure center, were com-

bined, simulated speeds were slower and durations

longer than observed. Finally, the simulated

tracks were shown to be farther north. Refs.

(Author)

(O UNCLASSIFIED SECURITY CLASSIFICATION OF THIS P*Ot'Wt,*n Dmtm Enffd)

■ - - _ MMMM

rr 4 -^pmmvmmit] n PII P*W«—■ .^ —■*** «'•»■

UNCLASSIFIED SECURITV CLASSIFICATION OF THIS PAGE fWh»n O.i« 1 ntmrmd)

j REPORT DOCUMENTATION PAGE READ INSTRUCTIONS BEFORE COMPLETING FORM

1. REPORT NUMBER

R-1689-ARPA 2. GOVT ACC« SSION NO. 3. ScECIPlENT'S CATALOG NUMBER

4. TITLE fand Sublll/*;

Synoptic Performance Characteristics of the Two-Level Atmospheric Model

5. TY'E OF REPOHT ft PERIOD COVERED

Interim • ■ PERFORMING ORC. REPORT NUMBER

7. AuTHORfaJ

C. Schutz

ft. CONTRACT OH GRANT HUMBERfJj

DAHC15-73-C-0181

S PERFORMING ORGANIZATION NAME AND ADDRESS

The Rand Corporation 1700 Main Street Santa Monica, Ca. 90406

iO. PROGRAM ELEMENT. UROJECT. TASK AREA ft WORK UNIT NUMBERS

11. CONTROLLING OFFICE NAME ANO ADDRESS

Defense Advanced Research Projects Agency Department, of Defense Arlington, Va. 22209

«2. REPORT DATE

August 1975 13. NUMBER OF "AGES

U. MONITORING AGENCY NAME A ADDSlESV" dllUrmtt Irom Conlroltlnt Olltc) 15 SECURITY CLASS, (cl Ihlm raport)

UNCLASSIFIED

15«. DECLASSIFICATION. DOWNGRADING SCHEDULE

1«. DISTRIBUTION STATEMENT (ol (hi, Rtpori)

Approved for Public Release; Distribution Unlimited

17. DISTRIBUTION STATEMENT (of ihm mbilrael mni..*d In Bloc* 30, tl dlllmfnl Irom Roport)

No restrictions

It. SUPPLEMENTARY NOTES

19. KEY WORDS (Conllnum on rovorf »'dm it nmcmmtmry and Identity by block numbmr)

Atmospheric Modelr. Climatology Atmospheric Circulation

20. ABSTRACT (Conllnur on rmvmrmm mUm II nmcmmmvy and Idmnuty bf block numbat)

see reverse side

W) I JAM 73 W3 EOfTICM OF I NOV 6» li, OBSOLETE I L'NCLASSIFILD SCCURITY CLASSlFlCATrON OF THIS PACE (Whti Dmtm Entmrmd)

. . - _., .____ ■ - - - -

■" ■ ■ ■! ii im«

jj- ARPA ()R1)[R NO : -;H9-1

6P10 Intormdtion Processing Tcchr^iues Ottice

R-1689-ARPA August 1975

Synoptic Performance Characteristics of the Two-Level Atmospheric Model

C. Schulz

A Report prepared for

DEFENSE ADVANCED RESEARCH PROJECTS AGENCY

fefcand SANT\ MONICA, CA. 9040<.

APPROVED FOR rUBLIC RELEASE; DISTR'ÖUTION UNLIMITED

- J

-— ' ■' ' »

-111-

l'REFACE

This report is concerned wxth the daily synoptic performance dis-

played by the Mlntz-Arakawa two-level atmospheric general circulation

model (CCM). Such presentations, which afford a comprehensive and

nearly instantaneous picture of the state of the atmosphere at a given

time each day, can help to identify pecularities of the model, and may

serve as a guide for future adjustments. Others working in the fifld

may also find these results useful for comparison with their results.

This report is part of the work of the Rand/ARPA Climate Program,

one of whose aims is the systematic comparison of model simulations

with obse»-v.3d climate. Other Rand publications related b'<' sulject to

the present report are R-1005-ARPA, which presents a systematic com-

parison of the observed climate with the January control integration

jsed here, and a report on the performance of a revised version of the

model for both January and July, which is in preparation.

These publications and the project of which they are a part are

sponsored by the Defense Advanced Research Projects Agency.

•^—-^____ ,..__^______

UPPMPimVpnW-pnm^Mm.ivpi i i ■l.m.WII | *.«■•■•—• ■" » i ■• wuiiiuji i IVMIIWIIII I I I mm, ««T . i w «•..«*■ i >m^

-v-

SUMMAkY

Tht present analysis of the Mintz-Arakawa model comppres the

synoptic characteristics of the simulated temperature and pressure in

three January experiments with five years of observed January data.

Tests are made of the average surface air temperature at the model grid

point near Columbia, Missouri (CBI), and five basic characteristics of

the circulation across the North Atlantic and North Pacific in tht

vicinity of the Icelandic and Aleutian Lows. Since these ocean regions

are prime weather producers, the degree of agreement between their

modeled and obseived characteristics has global significance for t^e

general circulation i.xide 1 (GCM).

While the model correctly simulates many observed features of the

monthly average Icwge scale surface air temperature, present analysis

shows that the local mean January surface air temperatures at CBI are

also in agreement with observation. However, further analysis of the

daily and hourly simulated temperatures reveals some disagreement in

the simulated diurnal pattern of warming and in the locally-simulated

standard deviation of the temperature.

The characteristics ov the North Atlantic (Icelandic) and North

Pacific (Aleutian) Low;, in January were judged from a comparison of

Synoptic Weath-r Maps at 1230Z each day during 1949-1953, and those

simulated daily by the numerical experiments at 0000Z. These maps

served to identify the positions of all closed lows each day, the depth

or intensity of the low at those positions, and the related speed, day?,

of duration, and track of the lows based on changes in the. daily posi-

tions. In the Atlantic, positions of the mean lowest pressi.rer. for

January were southwest of the cilmatological center documented in

Schutz and Gates (1971). The daily frequency distribution of the lowest

pressures gives a mean depth 4.3 mb higher than that observed in the

Atlantic, and 1.9 mb lower than that observed in the Pacific. This

individual or synoptic characteristic of the low pressure centers is

not evident when only the monthly average simulated pressure is examined.

The analysis of cyclone speed, duration, and track was based on the

daily positions of all closed low pressure centerr across the North

Preceding page blank

T v^mm*^mmr-m^

-vi-

Atlantlc and North Parifii-. It is shown that on the average the model

has fewer centers and moves these cyclones at a slower speed than is

observed. More cyclone speeds above 45 knots are observed than simu-

lated, which may be t^e result of a more irregular movement in nature

during cyclogenesis and cyclolysis. The analysis of cyclone duration

shows that the cyclones were simelar.ed to have slightly longer dura-

tion than is observed. The tracks of all cyclones used for the speed

and duration calculations show thit in both the North Atlantic and

North Pacific, simulated tracks are farther north than those observed.

-vii-

ACKNOWLEOCMENTS

Sincere appreciation is extended to the Environmental Technical

Applications Center (ETAC) of the U.S. Air Force for iti cooperation

in making available the latest magnetic tapes of hourly synoptic-

weather reports at Columbia, Missouri. Thanks are a'so extended to

Rand colleagues for their assistance: to W. L. Ga'.es, who originally

suggested the investigation, and who, along with R. R. Rapp and N. A.

Hanunian, made a careful review of ehe manuscript and extended mary

valuable suggestions; to E. Rodriguez, who developed the programs

necessary to interrogate the relevant weather data bank; and to C. R.

Huber, D. S. Pass, R. L. Mobley, A. B. Nelson, and J. J. Simac, who

produced the various maps and grid point data.

■ —

imT^mewmrvm^

ix-

CONTENTS

PREFACE iii

SUMMARY

ACKNOWLEDC.MENTS

FIGURES

TABLES

Preceding page blank

V

vii

xi

xi ii

Sect ion I. INTRODUCT ION !

[ 1. SURFACE AIR TEMPERATURE 2

III. SEA-LEVEL PRESSURE: ICELANDIC AND ALEUTIAN LOW CHARACTERISTICS 5

Posi t ions 5 Intensity IQ

Speed 14 Duration ^5 Tracks ^7

IV. CONCLUSIONS 21

APPENDIX DATA TABULATIONS 23

REFERENCES 29

T ' ■ "■ i..... . . —— m ■

-xi-

F1GURES

i. Average Dally and Hourly Variations of January Surface Air Temperature, 1950, vs. Experiment 1 3

2. Frequency Distribution of Daily Surface Air Temperature, January 4

3. Mean January Sea-Level Pressure 6

4. Mean and Daily Positions of Lowest Pressure, Iceland:! c Low 8

5. Mean Positions for Lowest Pressure Mean, January 9

6. Mean and Daily Positions of Lowest Pressure, Aleut ian Low 11

7. Daily Variations of Lowest Pressure Observed vs. Experiment 1, January 12

8. Frequency Distribution OL Combined Intensities Based on Daily Lowest Surface Pressures, January 13

9. Frequency Distrioution of Combined Speed Based on Daily Positions of All Low Pressure Centeis, January 16

10. Frequency Distribution of Combined Duration Based on Daily Position of All Low Pressure Centers, January 18

11. Combined Low Tracks Based on All Closed Low Centers (not on lowest pressure centers alone) 19

Preceding page blank

- -

"^ »■■ I ■ '■" i ■ minim wfrnrnt^ammi \ m im n v,**

-xiii-

TABLES

1. Temperature (F0), Columbia, Mo., January 2

2. Pressure (mb) and Mean Position, January 2

3. Speed (knots), January 15

4. Duration (days), January 17

5. Low Tracks, January 20

A-l. Average Daily Surface Air Temperature (0F), January .. 24

A-2. Observed Hourly Surface Air Temperatures (0F), Columbia, Mo. (3858N-9222W) 25

A-3. Simulated Hourly Surface Air Temperature (0F), Columbia, Mo. vgrid point) 26

A-4. Position and Intensirv of Lowest Pressure Center Daily across North Atlantic, January 27

A-5. Position and Intensity of Lowest Pressure Center Daily Across North Pacific, January . 28

Preceding page blank

mmm

-l-

INTRÜDUCTION

For our present purposes, only the temperature simulated by the

model every six hours at the grid point near Columbia, Missouri (map

designation CB1), and the daily changes in closed surface lov pres-

sures across the North Atlantic and North Pacific were used to examine

the model's perTormance from a synoptic viewpoint during January. The

model's performance over three separate January experlnents has been

selected for comparison. The January control experiment itself has

been analyzed by Gates (1975).

Simulated data are presented in tablt^s, graphs, and histograms

for comparison with similar daca observed in January of the years

1949-1953. Hourly observations at CBI each day during the five years

were obtained on tape from the Environmental Technical Applications

Center (ETAC). The daily serits of Synoptic Weather M£ps for 1949-

'953 published by the U.S. Weather Bureau were used to determine daily

changes in the five variables related to closed surface lows across

the North Atlantic and North Pacific.

>• •■"■ I II I 111

-2-

H^. SURFACE AIR TEMPERATURK

The average daily surface air temperatures used for this analysis

were extracted every six hours-0000-0600-1200-1800 local standard

time (LST)-from the 31-day January simulations of three separate runs

of the same control experiment; the runs differed only to the extent

of small differences in their initial conditions. The air temperature

computed by the atmospheric general circulation model (GCM) is described

In Gates et al. (1971), and the model's climatological performance in

tta control integration has been summarised elsewhere (Gates, 1975).

The average daily temperatures observed at Columbia, Missouri, were

taken from the ETAC observations of hourly weather by the method de-

scribed in Rodriguez and Huschke (1974).

From these .emperature data (tabulated in the Appendix), the

monthly mean maxi-rom and minimum temperature, and the mean, range, and

standard deviation of the daily averaged temperature were determined

as shown in Table I. The average daily and six-hourly variations of

January lf50 surface air temperature were calculated as shown in Fig. 1.

The observed mean and standard deviation from the combined 1949-1953

Table 1

TEMPERATURE (F0), COLUMBIA, MO., JANUARY

(Computed daily from 0000-0600 - 1200-1800 GST data)

Observed Experiments

1949 1950a 1951 1952 1953 J_ 2 3

Maximum 58.0 58.3 53.0 58.3 54.0 43.5 47.6 44 9 Minimum 2.0 6.3 1.5 12.8 15.0 19.5 23.4 21.0 Range 56.0 52.0 51.5 45.5 39.0 24.0 24 2 23 9 Mean ^.9 31.9 3^.2 33.7 33.6 32.8 34.6 35.2 Standard

Deviation 12.5 13.4 13.3 12.5 8.1 5.2 5.9 5.5

Synoptically representative year.

wmimmmm^wm mmm imimim***mn' ' mun -■— - —

-3-

ObjerveJ 1950

L J Computed Exp. I

Columbic, Mo.

(38058N-92022W)

"ig. 1 —

10 20 January

Average daily and hourly variations of January surface air temperature,

1950, vs. Experiment 1

- - ■

B^PW.«-.^.*.-*"«.».«■•*... ^p ^n ^m mm^ww'Wy •'•" -■■^■^'-■■-

-A-

data indicate that 1950 Is m<re synoptically representative in the

sense that It most closely aporoximates mean conditions. In these

1950 data, marked excursions of daily mean temperature occur througii-

out the month, which are also reflected in the daily six-hour obser-

vations (see Fig. 1), In the model, while tnere is a strong diurnal

change of all temperatures, the occurrence of the particularly low

and high diurnal ranges that wouj-i accompany specific synoptic events

is not well portrayed. Nevertheless, the simulated average surface

air temperature is within 10F of the mean for 1950 and equally close

to that from approximately 30 years of recorded data (Schutz and Gates,

1971).

The histograms n Fig. 2 for all five years of observed data and

for the three experiments show other differences. Ir general, the

experiments' minimum temperatures are higher and their maximum temper-

atures are lower than those observed, while the experiments' standard

deviation of 5.70F is significantly less than the 12.30F which occurred

in nature for the years 1949-1953. Thus, although the model may ade-

quately simulate the monthly average of such features as the surface

air temperature, the local daily and hourly variations are such that

there is a systematic underestimate of the observed variability.

Columbia, Mo. (38° 58N-920 22W)

Experiments Observed (1949- 1953)

Mean - 31 .50F S.D. - 12.30F

715 r Mean = 34.20F

rreq

uen

cy

(Av

g.

rep

ort

s/m

: ,Jr L-1

S. D. = 5.70F

■9

) 20 i

41 3 -1 1

60

Tempeature (0F)

Fig. 2—Frequency distribution of daily surface air temperature (0F), January

wmfM .......,.,,. .v..a«

-5-

III. SEA-LEVEL WUKSUjLSj ICELANDIC AND

ALEUTIAN LOW CHArACTERISTICS

The mean climatolosical position and intensity of the Icelandic

and Aleutian Lows are shown in Fig. 3, taken from Schutz and Gates

(1971), and are based on approximately 30 years of January data. These

centers are compared with the daily observed and simulated January

values of the lowest pressures and their means. From synoptic data it

is difficult to determine the position and intensity of tbe lowest

daily pressure centers observed or simulated across the North Atlantic

and North Pacific, due botli to the inevitable errors of synoptic analy-

sis and to rhe model's limited grid point resolution. The lowest daily

simulated pressure at a grid point may not be the lowest implied pres-

sure that could occur within the grid of 4° latitude and 5° longitude.

Also, the "observed" lowest pressure depends on the analyst's inter-

pretation of the available data; isobars may be added simply because

of wind speed and direction at a single station some distance from the

actual low center drawn. Therefore, no effort was made to adjust either

the simulated grid point values or to consider interpolating observed

values at the gr\d points used by the model.

POSITIONS

For the North Atlantic the mean January positions of the Icelandic

Low based on lowed pressures for each test year and each experiment

were obtained from the daily data in the Appendix, and are summarized

in Table 2. The average pressure, standard deviation, and range for

the years 19A9-1953 again make 1950 appear synoptically representative,

as was the case for surface air temperature at Columbia. Before aver-

aging, the daily low positions from 1950 formed a distinct envelope of

points around the mean climatological center from Schutz and Gates (1971)

at 62oN-30oW (Fig. 4). Experiment 1, on the other hand, had a well-

defined envelope of points southwest of the observed center. The mean

January positions of the lowest Icelandic low pressure developed from

each year's daily data and from the three January experiments fall into

similar distinct groups (Fig. 5).

"^

-6-

i = -e R-

■o

v a

■

-Q —

J» §

If? y. - ■' _, Ü ^ I ,. fl) "t

"O c

tt) c o

^°? . «) _3

C *- CO

0) 0 § i- i- D w- U. a 0

I o ^ o _ -o

IE a> o

o 0) I

o o \l o

11 ro

en

rp»

-7-

^4 X 3 00 z u CN r-- m r^ -C o> o m in oo cN m <r <r

y^s • • • • • • • ^-\ • • • • ■ ■ •

M m 00 r^ o m m ro in N n r^ rH m r-t N Li -t

|

j> vO «N r^ m >0 3 Q Q O

oo in c^ r^» i/1 r>. rH

>-' r~. rH \o m

■ 3 *—• O rH CT\ CN

H Z U CN O ■ • • • • • • ■ , • ■

4-J IM o fM ao a> m n oo 4-1 CN -» -t Ov -* f^ m vO

1 ON r^ ,_4 r^ in m a ON m co r^ ON ON ON

m r- H

-1 vC 2 3 m a Z H 1 c4 <t CO r-i —i vO 00 V oo in co ON •» CO st a. > • • • • • • a. • ■ • • ■ a

X 1—t vo -* iH m m CN O ■< r-H so -3- PNI m r^ to CO u oo vO (N r^ m <r UJ ON m -T r^

ON ON ON rH >- JJ B 3 Z 3 p -n O ^H o m t-i r^. v£> l*\ O O O vO r^ rH VT i^ ^ m • ■ ■ • • ■ in . • • ^ ■ cr in ^ o r^» O <-t X) c^ in o in oo 00 t^- CN •-j 0/ H ON -3- N* r~. •H vO >l iH oo in ro so

ON ON ON rH •• 3 i5 H 8 N 3 3 Z 3 >^ tl r^j o o o <r o-- ro in CM o o o m 00 CN C H I- m • • • • • • • m i—i a a^ m o m o m ro <r cr- in o in m CO co r-. u ^ ^-i (7N >* m r-» H vO CN P^ ON m ^ r>. m NO c 4-1 c >^v &* & o^ 3 —s ON ON ON rH pu ■ J N C 1

(N B u O Z 3

N3 <fl Z 3

"J 4C 0 •H CN .—i O O O tN r-- ON vT a . >J H O O O 00 4 O 00 •-I "-U —1 -3 i—i in • • • • • • • 3 ^H m ■

J3 ?: C ■*^ o> o o o ^ JN 00 -* •H s-^ o> o m m rn o rH rH M >. tt ^H (^ in ^3- r^ m n 41 H Ov -3 ~T r^ rH m i^- H 1 H •H

i •T3

> <y* cy^ o>

Z 3 < i >

ON ON ON rH

Z M -^-v 1 ■ O O O m 9 O r^ <u O O O CN r^ o o -5 c ■ o • • • • • « o . • • e 0 Xi in O in in f*^ N d <N JO m m o m cN CO vo m . ^-^

•a c

i—t o ^ in r^ rH vO m o

rH O m in oo O ON ON

rH -a- NO u UJ u ^ IH 01 S n >, s B c/; DO Z 3 Z M ■ 'X ■ O O O ^H a- CTN r^ CO O C-" ON r^- ^- CO > 3 c^ • • « • • • • o> • • • • . •H

<« o o o o fi n ^ <r 00 O CO n rH m m 4-1

H

Maximum 10

0 Minimum 95

Range 5

Mean 97

C o •H

-a 4-i u a

C "J n) a w M

Mean

( Latitude 6

Position

\ Longituda 1 9«

Maximum 99

Minimum 95

Range 4

Mean 97

Standard

Deviation 1

Mean

( Latitude 5

Position

\ Longitude 17

CO XJ

c o to (U

a c

<-H rH « U

■H 4_)

U. O c

V)

™ ' ' I ' ^-_-,— ,, , |, panaEaarMvan ^*^^mm^^*^ww^^m^^^^mm*^^^~^^^^

90 30oW

62° N

Jonuory 1950 12^02

30oW

O M«a»i of daily positions

• Climatologicol mean (Fig. 3)

—620N

January - Experiment 1 0000Z

•30 -10 10 30

F-g. 4 — Mean and daily positions of lowest pressure, Icelandi c .ow

wn^m^w*^ ^^m^im^m^^iH^^^^m* n i • HMM uiuiwpw-n ^i« ■ mm\ mumi ii-.u-.--"

CT' "" 1 L i 1 1 r 3 4 = ^

o 00

s o

o c o

O —1

c O o CM 0)

E O ■ ■«t D

*o 6 s Q.

1

g 6 o

8 k.

o i iC

R c o 4-

o

I c D e 5

IT)

aa^^^aK—a—a^a—al^ ■ - - - -■ ■

n<<"i ' ii ii im i ii • i ii i anOTmai

-10-

Although the experiments indicate the Icelandic Low to be south-

west of the moan climatological position, a chec . cf Icelandic Low

positions for Janjaries from 1899-1960 reveals that only about 10 per-

cent were observed in that locale (Bjerknes, unpublished). These

anomalous observed positions of the Icelandic Low in January are gen-

erally near the positions simulated by the experiments.

The mean January positions of the Aleutian Lov for each test year

and each cxpe-iment were similarly determined from the daily data (see

Appendix) an 1 are also summarized in Table 2, The average pressure,

standard deviation, and the rang'' for the years 1949-195j nuke 1951

the most synoptically representative of the test years. Before aver-

aging, the daily positions from 1951 and from Kxperiment 1 each formed

envelopes of points which covered the mean climatologica] center from

Schutt and Gates (1971) at 50oN-170oE (Fig. 6). When daily positions

of the lowest pressure from all the experiments and test years are

averaged and plotted, the resulting centers show good agreement (Fig.

5), even though the centers fi the experiments are more tightly

grouped.

INTENSITY

From the d^a in the Appendix, the highest and lowest of the lowest

pressures, the range, mean, standard deviation (Table 2), and the daily

variations of the lowest pressure were determined. Over the North

Atlantic in the vicinity of the Icelandic Low the observed mean and

standard deviation from the years 1949-1953 indicate that 1950 is

synoptically representative of the intensity. The intensity range of

the daily lowest pressures in 1950 was 55 mb, while in Experimen 1 it

was 21.8 mb. The excursions of pressures within these ranges through-

out January are shown in Fig. 7. In general, the simulated daily varia-

tions of pressure fall vithin the pattern established during January

1950. The histograir.^ of the frequency distribution shown in Fig. 8 of

Intensity data show rhat within the range of lowest pressure the experi-

ments give low pressures that are higher and high pressures that are

lower than those observed. Druyan (1974) found that thr» GISS model al o

shoved central predsures that were systematically too high in January.

Th«. fr-equency distribution of these daily January data for the North

— ■ — — - —

■^■■■^■f HIII, \\mrwimrmmm*'r*mrm*m^*m*w*^*m**^*m*mw mmmwinm .MPJiimnii-x ■> » «■■"■ ■—-•- ■ «■•■ i • ■> -

■11-

90 • • 170oE

I

I

170oW t • • » • • ■

* » - -

• « •

January 1951 1230Z

— 50oN

—46° N

90

« «

• # ff **

* m » m » 0

■' m * *

* • •

* • • *

• * •

Q Mran of daily positions

% Climalological mean (Fig. 3)

« « r «

t • • • • • *

70 January - Experiment 1

0000Z

Fig. 6 — Mean and daily mitions of lowest pressure, Aleutian Low

MAMMM^UA-'. .. ■ "—~ -■

«II ■" ' 1^^^^^——■•> •*> ■" wfm^mm^^-w.^^^^^^^^^Fmmmmmmr^^'^^^m

-12-

Observed (U L f] Computed Exp. 1

Icelandic LOA ( 1950) Aleutian Low ( 1951 )

January 10 20

January 30

Fig. 7—Daily variations of lowest pressure observed vs. Experim..nt 1, Januar)

Atlantic indicates that the simulated pressures are also much less vari-

able than those observed.

Over the North Pacific in the vicinity of the Aleutian Low, the

observed mean and standard deviation from 19A9-195i indicate that 1951

is synoptically representative of the intensity. In 1951 tb ■ daily

lowest pressure had a range of 4S mb, while in Experiment 1 the range

was 42.3 mb. This similarity can be seen in the curve representing

the plot of these daily data in Fig. 7. However, the histograms in

Fig. 8, based on the frequency distribu.ion of intensity data, show

that within the range of lowest pressures, the experiments give low

pressures ^hat are higher and high pressures that are lower than

those observed. As with the position data previously analyzed, there

is better agreement between the experimental and observed intensity in

the North Pacific than in the North Atlantic.

The difference in mean pressure shows that the simulated Icelandic

Low is 4.3 mb less intense than is observed, while the simulated Aleutian

—m mmmmm

wmm:^mmimmnmtmmfm ■ ' ■ ■' ' ' N IMII "•^■■■(»•WW^P—»«—«•

-13-

F.xperiments

Mean ■ 976.8 mb S.D. = 5.8 mb

Observed

972.5 mb 13.0 mb

10r

-i—r—T—i—i—r Pressure (mb)

Icelandic Lows

Mean = 973.1 mb 975.0 mb S.D. ■ 7.5 mb 11.7 mb

I0r

o

—I 1 1 1 1 1 1 1 1 1 1 1 r^—i 950 970 990 1010 950 970 990 1010

Pressure (mb)

Aleutian Lows

Fig. 8—Frequency distribution of comMned intensities

based on daily lowest surface pressures vrrb), January

■

■14-

Low is 1.9 mb more intense than is observed on the average. When

monthly means from the model are compared with long-term jlinatology,

such as that froiu Schutz and Gater- (1971), the model-sin..ilated lows

in Experiment 1 are IZ-IS mb oceper than normal. This disagreement

is due to the ract that the mean monthly p .jss'-re results from aver-

aging L^e pressure for the monLo at each grid point without regard to

the positions of individual lows. For example, if an exemplar grid

point is chosen, then the average pressure there will takj account of

the occasional readings of relatively high pressure depending on the

mobility of .he low center. If the center is quite m bile, as was

shown ';y ^he changing observed pc-.-'tions of the Icelandic Low in 1950

(Fig, 4), the sele-ted grid point will experience many days of rela-

tively high pressures, which implies a fairly high average pressure

for the ir.jiith; but if the low center hovers over and around the

sele_ted grid point, as in the case of Experiment 1 (Fig. 4), each

day's reading will be Irw, from which it follows that t h^ month's

average value will be low. It car be recognized, .herefore, that such

an aggregate monthly statistic might mike it appear that modei results

do not match their natural counterparts, v'.,.'(e daily comparisons at

the geographic centers of the individual lows could reveal a different

result.

^PEED

Across the oceans, calculations for the mean and standard devia-

tion of low center speeds were based on t-he daily positions of all

"closed" low pressure centers, and not exclusively on the movements of

the louest pressure centers just discussed (Table 3). Also, only ocean

points were used so that surface effectt would be standard throughout

the calculations For both the North Atlantic and North Pacific, the

calculation began when a low was at a land or ocean point nearest the

western shore o: the ocean. They were then continued across the ocear.

until cyclolysis occurred, or until an ocean or land point nearest the

eastern shore of the ocean was reached. Although the calculation did

not always begin at the time of cyclogenesis , there wer2 instances w'.ere

cyclogenesis and/or cyclolysis occurred over the ocean while the low

was being tracked.

T »■»^^^f-ww^w-^ "■H '"" """

I

-15-

Table 3

SPEED (KNOTS), JA^mARY

(Based on daily positions of a7! pressure centers)

Icelandic Low

Observed (1230Z)

19A9 1950 1951 1952 1953

Experiments (0000Z)

M^an 23.7 22.1 18.0 20.6 19.3 16.8 14.4 17.2 Standard Deviation 11.1 11.0 9.3 11.7 11.5 10.0 9.6 9.3

Aleutian Low

Obc^rved (1230Z)

1949 1950 1951 1952 1953

Experiments ^0000Z)

Mear' 22.4 20.2 18.7 23.; 22.6 21.5 19.5 19.7 Stanf4 rd Deviati-n 10.1 11.3 11.1 12.2 12.9 8.6 9.3 7.7

The frequency distribution of speed (knots^ shown on the histo-

grams (Fig. 9) is based en the distance a closed low moves each 24

hours. The distance between analyzed low centers from one map to the

next is called a "period." Figure 9 indicates (1) that more periods

were observed than were simulated and (2) that on the average the two-

level model moves the lows slower than is observed. This same feature

of model simulations has also been noted by rirown and Fawcett (1972)

and Druyan (1974).

DURATION

Duration usually indicates the lifetime of a closed low center

from cyclogenesis to cyclolysis. However, since these calculations

were based on the sam.i daily data used for the evaluation of speed,

only data over the oceans were used and these did not always include

a cyclone's complete cycle. In general, the experiments across both

I I ^|HV«V«^«P^

-16-

Experiments

Mean = 16,7 knots

S.D. - 9.4 knots

Observed

8 20

■- 15 -o '-'

"8 lo

CL.

Mean = 20.4 knots S.D. ■ 8.3 knots

'ill''1—M ' ' ] 20 40 60

20.5 kr.ots 11.1 knots

1 ! ' i ' ' n • i i i i i i i i i ( i Speed (knots)

Icelandic Lows

21.3 knots 11.8 knots

Speed (knots)

Aleutian Lows

■ i i i i i i i i i n 20 40 60

F'9' 9—Frequency distribution of combined speed (knots)

based on doily positions of j3M_ low pressure centers, January

■■" ' '*F9Kmmmn*m

-17-

uceans showed that the lows continued slightly longer than those ob-

served; longer duration being, of course, consistent with their lower

speed.

Over the North Atlantic and North Pacific, the mean and standard

deviation of duration shown in Table 4 indicate good agreement during

the individual years observed and the experiments. The frequency dis-

tribition of the combined data given in Fig, 10 shows that the observed

means are lower; however, the standard deviations are higher in the

Pacific and lower in the Atlantic than those simulated.

Table 4

DURATION (HAYS), JANUARY

(Based on daily positions of all pressure centers)

Icelandic Low

Observed (1230Z) Experiments (0000Z)

1949 1950 1951 1952 1953 1 2 3

Mean Standard

3.5 3.2 3.8 3.2 4.9 4.9 3.9 4.2

Deviation 1.0 1.4 1.8 1.8 3.2 4.0 2.9 2.9

Aleutian Low

Observed (1230Z) Experiments (0000Z)

1949 1950 1951 1952 1953 1 2 3

Mean Standard

3.8 4.3 3.3 4.5 4.0 5.5 4.4 3.9

Deviation 3.2 2.2 2.4 2.6 3.0 1.6 2.5 2.3

TRACKS

The racan tracks or paths of all closed low centers shown in Fig.

11 were a by-product of the speed and duration calculations. They are

presented to provide further insight into the differences between simu-

lated and observed conditions. In both the North Atlantic and North

Pacific, the simulated tracks of low-pressure centers are generally

~9^^^~ II I I * WV W^M-f-V*« --•■ ■—«■^■■(■«■IH <|W>>>« V^'JI ■ PW^I^OTM i i

-18-

| 10 •A-

c

3 5 i

_o

« 0

51

Experiments

Meon =4.3 day:

S . J. = 3.3 days

Observed

3.7 days 2.3 c'ays

I Days

Icelandic Lows

u

I 10

J J o

Mean = 4.5 c'ays S.D. ■ 2.3 days "L

6 12 Days

n» M

3.9 days 2.7 days

18 I I | I I | I I I

^ 12 18 Days

Aleutian lows

Fig. 10—Frequency distribution of combined duration (days)

based on dnily position of all low pressure centers, January

-- ia^»iM

-19-

1 > i

-Q

o

I o

o T

o o t

o TV

I

S

T

7

3

P

3

o

01 c o

_o

8

I V

■•-

c 5

J X)

31 _o u

il c o v u E

•♦-

o

2 15 E o U

en

8 o o n $ J? o o en

■ ■ IM^MalM-MlaB_d^MM

■■^PiPMNMnPMMiHmnpinMPWPV" i .in. iiii^«^«vn««OTv«wi«ianmH«wmnMWMPP'iii ■■■»■■IIBMBWHI !■ 111. mi ■ VHMP^^in^l^nHliW« IW"! > l ■ i «PMfn^P^'^^

-20-

north ot those observed, as may be seen by referring to the cross-

hatch',., grid square representing Iceland and the Aleutians. This is

true not only for the primary track (solid line), but tut the secondary

track (dashed line) as well. A second important difference between the

simulations and obsetvatioi. is the more random distribution of the in-

dividual tracks in the observed data. This is shown by the plotted

nurbers at each grid point representing the number of times a low

passed near that grid point at map time. The track characteristics

in Pig, 11 were determined by subjectively connecting the locations of

the maximum number of such passages. The basic tracks developed by

this procedure did not always go through either the mean center of

lowest pressure (Fig. 5), or the latitude and longitude of the single

lowest pressure for January, A summary of all low-pressure center

tracks used indicates that more low "positions" are observed than are

simulated, as shown in Table 5.

Table 5

LOW TRACKS, JANUARY

(Based on daily positions of a?7, pressure centers)

North Atlantic

Observed (1230Z) Experiments (OO0OZ)

1949 1950 1951 1952 1953 1 2 3

Tricks 15 25 21 20 20

North Pacific

14 16 17

Observed (123nz) Experiments (0000Z)

1949 1950 1951 1952 1953 1 2 3

Tracks 16 24 29 20 25 M 13 '3

FT ^^

■21-

IV. CONCLUSIONS

It has been BD^WTI that daily variations of surface-air temperature

and five synoptic variables related lo closed low-pressure centers

give additional insight to inoOel behavior when compared with repre-

sentative observed data. The principal findings are:

1. Large diurnal variations of surface air tempeiature are

simulated between 0600LST to 1200 LSI at Coluv.bla, Missouri,

and these are systematically larger than t iiose observed.

2. When daily surface p* ..-ssure is used at fixed gcc^prapbical

points ro determine the mean monthly pressure, t'.e simulated

Icelandic Low is less intense than observed, while the

Aleutian Low is more intense than observed, as compared with

th« lowest reported pressure each day.

i. The simulated speeds of all low-pressure centers, including

those with the lowest pressure, are systematic,i.ly slower

than the observed speeds.

A. Both the simulated daily and mean January positions of the

Icelandic low-pressure center tend to occur southwest of

the observed climatological center. There is better agree-

ment, however, when comparing positions in the vicinity of

the Aleutian Low.

It is hoped that future improvements of the model will reduce these

errors in synoptic performance, even though the model is not intended

to be used in short-range forecasting. In view of the systematic char-

acter of many of these errors, hov ,■ er, there is a possibility that they

may affect the model's simulation of climatic statistics at time ranges

well beyond the model's acknowledged predictability limit.

^nmimmmmfmmmrwmtmmmi^m^-^^^m^^^mi^m^mmmmmmmm^^'^^^ n ■ —^mr^m^mmmmmi 11 m\ i ■■" ■

-23-

APPENDIX

DATA TABULATIONS

Preceding page blank

■

1 I ■ wmmmmm* "■ ,.■.,•».....,--.-——-• ■^m^^m*m^mm

•25-

Table A-2

OBSERVED HOURLY SURFACE AIR TEMPERATURES (0F), COLUMBIA, MO. (3858N-9222W)

January 1950

i)a"vSJ :r 0OOOLST

SO

06001.ST 1200 LSI 1 SOHLST

51

Avi

I 49 55 51.3 ■) 45 49 59 62 53.8

3 S] (7 29 17 33.5

A 11 3 4 7 6. 1

5 6 1 14 16 9.3 6 16 16 24 n i 11.5

7 22 21 38 33 28.5 8 29 U 41 43 36.0 4 39 »7 . ■ 51 44.0

10 58 40 31 28 39 . 3

11 22 24 33 34 28. 3

1J 35 39 48 42.5

1 j 51 51 54 35 48.3

u 25 19 38 42 31.0

1 j 41 52 31 23 36.8

It 12 11 V.) 30 20.8

17 27 27 44 47 36.3

18 23 14 16 It 17. 1

19 lb 13 23 24 19.5 20 1 1 20 37 39 29.5

Jl 31 31 38 42 35.5 TT 46 44 49 49 47.0

2 j 47 46 45 19 44.3 24 40 48 73 72 58.3

?5 40 35 62 27 41.0

26 15 10 14 15 13.5

27 11 Hi )1 31 21.3 28 28 28 5i) 53 39.8

29 44 20 IS 15 24.3

30 1 . 1 i 18 16 15.0

il 18 U 20 21 18.3

31.9 TOTAL

■

mi twmwmmf'im*

-26-

Table A-3

SIMULATED HOURLY SURFACE AIR TEMPERATURE (0F), COLUMBIA, MO,

Experiment /'I

(GRID POINT)

Da^\ Hr 0000LST 0600LST i200LST 1800LST Avg

371 18.81 19.01 35.12 26.66 24.90 372 13.78 10.80 30.61 22.91 19.53 373 13.42 11.88 44.46 31.56 25.33 J74 5.95 2.92 53.61 41.71 26.05 375 29.90 28.00 47.27 35.95 35.28 376 25.41 21.91 43.05 38.43 32.20 377 31.28 29.50 60.48 44.50 41.44 378 32.05 30.79 41.74 32.25 34.21 379 22.67 21.15 50.39 40.22 33.61 380 22.66 20.25 59.73 48.10 37.69 381 34.80 31.98 62.06 45.33 /3.54 382 18.42 17.50 54.28 45.61 33.95 383 35.02 28.84 56.85 41.09 40.45 384 21.16 24.82 54.11 44.05 36.04 385 31.92 26.49 43.65 36.21 34.57 386 18.12 7.39 52.74 J9.95 29.55 387 27.63 29.75 58.70 47.4.' 40.88 388 33.67 29.74 52.85 40.^1 39.12 389 24.03 19.58 55.96 40.52 35.02 390 20.49 22.01 48.83 39.20 32.63 391 32.54 26.55 35.40 29.88 31.09 392 20.40 19.77 48.48 38.37 31.75 393 24.30 20.89 43.80 34.80 30.95 39A 22.25 18.50 41.9° 33.57 29.08 395 21.22 17. 31 51.60 39.26 32.35 396 20.34 16.90 51.77 40.06 32.27 39 7 26.65 25.04 44.56 36.60 33.21 3J8 23.52 17.94 50.13 34.18 31.44 399 11.39 6.11 49.01 35.47 25.50 400 20.38 19.18 46.46 37.47 30.87 401 24.95 22.21 44.49 38.05 32.43

32.8 TOTAL

^«(■^■•ffi^^^f^^Birwwi Tf^^^^^m

-28- Table A-5

PUSITIÜN AND INTENSITY OF LUifkST PRESSURE CENTER DAILY ACROSS NORTH PACIFIC, JANUARY

12J0Z i 00007 YR 1949 1950 1951 1952 195! "XP. 1 rvj'. 2 f'XP. 3

Uay Laf1 Prets

Lon mb Lat

Lon "ress

mb l.at Press

Lon mh l.at Press

'■ on mb Lat Press

Lon mb Lat Press

Lon mb Lat Press Lat i'ress

1 4_3176 995.0 56159 990.0 45 970.0

'8161 985-0 59 979.9 54175 965.6 »,75 9f"6-3 »180 9f,',•4

2 Hui 98ü-0

^177 995.0 44172 980.0 ^175 985■(, MlM 975.0 58170971.1 5817097,.9 58^970.3

3 58153 969.0 49157 990.0 50169 970.0 •2ltJ "5.0 «,60 q»-0 50,.t 977.6 145 50145 979.6 50,,c 975.8

145 U 50156 980.0 ♦♦ws 990.0 50 970.0 4916) 970.0 50158 950.0

»155 "»^ »155 9»-1 »155 951-5

5 49157 975.0 «159

970.0 50177 975.0 59172 995.0 55]75 960.0 »170 959-5 I n ) »165 q»-5

6 «167 970-0 39148

99 5.0 47170 980.0 60,.e 990.0 -145 41,,, 970.0

-147 54,75 974.3 4l7J 976-4 »,80 975-"

7 52. 2 975.0 48159 990.0 53168 979.2 6-8140 985-0 46uo 970.0 50170 97,.4 *in 97"-4 50]7(| 967.«

8 »>17S 970-0 »161 970.0 60158 975.0 56144 970.0 *«„, 9 70.0 50170 976.H »m»»-« »WO972-9

9 62175 984.4 »159 992.5 4518n 989.0

»140 q70-,) •^HQ 980.0 54)50 975.8 4XJO "76-0 58 174.«

10 4417i 995.0 3915ü 960.0 »2i$3 985.0 55n7 980.0 nm""* »ws 9U1-5 »Wj »".i 66]5098^.3

11 49180 985.0 44160 950.0 4i,ÄJ 990.0

* Ini} 64175 990.0 »7lM -85.0 5", 7,, 982.5 wi« "74--' »170 U87-3

12 52172 980.0 48162 060.0 59]42 875.0 59148 98',-n ,',I46 q7'J) »wo Q7H-2

»,80 976-9 »MO 974-4

13 56168 985.0 47163 970.0 *9MJ 988.8 56n9 975.0 4-7|h6 r,7^,) »160 970-7 »,65 qf'8•,, »165 967 " u 56165 995.0 47170 985.0

»160 970-0 4s 980.0 511S8 980.0 »,65 969-8 »165 ',61-S 54170«6,.H

15 6ü165 998.3 4?169 985.0 58174 980.0 ■4^975.0 43175 ?75-,J »180 9^.9 »,75 "h'-] if,., 965.7

16 M158 985-0 4:,164 970.0 53J 980.0 54154 950.0 *Il74 960.0 »§us ^7-4 »8l70»7*'S "two H?'*

17 »*,M 370.0 4-i1A« 97U.Ü 57,/r 980.'/ -'45 "l62 98Ü-0 Hrn •""■° »170 ■'74-) Hm 907-:

18 61175 980.0 "163

-/85.0 »1« ^O.') 45)sq J75.0 Hm**.* »•wo »«•» »,75 n7'-' J5,7s n7l-:

19 62170 985.0 %*a 988.4 HUi ^5.0 »166 960-0 52177 969.9 V1,70 ■,70/' ht,60 Q72•:, 5817c 968.8

.0 »165 960-0 42156 980.0 50uo 965.0

»164 975-0 47162 ••3-0 »ltot».3 »,70 9Q4-0 6216597,.2

21 »165 955-0 47158

980.0 53,3, 9»-0 41H 965.0 52U8 965.0 54150 970.7 54165 987.7 46,70 986-8

22 58170 9 70.0 "162 985.0

»160 945-0 4-3160 974-9 48,.r 975.0 54155 972.3 »,60 977-5 »no 982-2

23 »165 990-0 41152

10C3.0 56,,. 975.0 T 74 46156 970.( 41160 9»-0 58170 970.9 54175 973.0

»,55 q67-1

M 52169 985-0 48158 995.0 5115;) 980.0 •In, ■""•" 49164 ,,»■(, »150 981-9 »,80 q73-" »,60 qh6-0

M 56170 990.0 37173 1005.0 44164 990-0 5115^ 971.9

»164 975-0 l*lw 7U.1 »165 978-2 58150 973.7

26 38155 985.0 39179 1000.0 MU3 'm-0 47153 975.0 49,c. 985.0 15'i »r/o974-5

»160 ,>71-4 54150 977.3

27 50167 975.0 ?8173 1005.0 39. M 975.0

-1 /b 47I68974.S 42,,, 985.0

- In i »165 978-5 »170 q75-q »no977-1

28 54175 950.0 48.48 1000.0 46164 965.0 4_7178 975.0 45175 985.0 »160 98n-1 »180 975-3 »,70 Q70-8

29 :S:68 970.0 »155

990.0 49172 96- 'I 48170 972.0 42167 955.0 »160 980'6 »180 077-9 »,80 9ft8-7

:J "166 '■75-0

/ I -170

995.0 49158 97 .0 56H2 975.0 47160 970.0 ■8160 981-4 lll50 980.8 5817() .6,3.t

31 43163 995.0 45145 980.0 55160 988.3 59,/t 978.0

-146 46,,, 970.0 -1 74 ■4165 98,•3 - -

Range 950.0-998.3 950.0- 1005.0 945.0-990.0 950.0-995.0 950.0-985.0 •■54.5-996.8 954.,-994.0 951.5-987. 3 Mean 976.9 982 .2 973.8 973.5 968.6 973.95 974.2 971.2

Standard Deviation 11.7 i3 .7 10.4 8.8 8.7 7.42 7.31 7.58

(l.at Mean',. JLon 5 3.4N

175.3E 46. ON

165.0E 51.ON

111,aw 13.21

16 7.0W 4 7. IN

172.4W 5. ~";

17 3.4,- 55.2;;

,76.0E 55. 4N

174.4F

Lai a Latitude, l.on » Lottflt«|ri«i (minus « west).

■- -