NASA Technical Reports Server (NTRS) 19960016108: Topographic forcing of the atmosphere and a rapid change in the length of day

4 Pages·0.35 MB·English

by  NASA Technical Reports Server (NTRS)

)
NASA-CR-20OIO1
!
Reprint Series SCIENCE
15 April 1994, Volume 264, pp. 407-409
)
/, i , -- v
(-!"k , : ,, ) , .
_'+': ,. c__
] ///
/
Topographic Forcing of the Atmosphere and
a Rapid Change in the Length of Day
David A. Salstein and Richard D. Rosen
Copyright © 1994 by the American Association for the Advancement of Science
Topographic Forcing of the Atmosphere and ists now in the measurement of coherent
fluctuations between AAM and l.o.d., the
a Rapid Change in the Length of Day
changes captured in Mw + 0.7 Mp (6)
during the intensive period explain most
David A. Salstein and Richard D. Rosen (64%) of the variance in simultaneous mea-
surements of l.o.d. (7) (Fig. 1A). Most of
During June to September 1992, a special campaign was held to measure rapid changes this agreement is associated with changes in
in Earth's rotation rate and to relate these measurements to variations in the atmosphere's Mw; during this period, fluctuations in Mp
angular momentum, due principally to changes inzonal winds. A strong rise in both length contribute little to the variance in l.o.d.
of day and atmospheric momentum during a particular 6-day subperiod isdocumented, and Therefore, study of the NMC daily wind-
this example of a short-period perturbation is identified with a specific regional coupling based momentum data further to determine
mechanism. Mountain torques within the southern tropics appear to account for most of the source and character of momentum
the rapid momentum transfer between the solid Earth and atmosphere, with those across variability during the intensive period suf-
South America especially important. fices for a broad understanding of the
Earth's rotational behavior during that time
as well.
The relative angular momentum of the
Recent advances in the accuracy of space planet's mean rotation, Mp,were computed atmosphere within 46 equal-area latitude
geodetic techniques used to measure the every 6 hours according to belts (8) was calculated from NMC zonal
Earth's rotation, as well as the availability of wind analyses for the intensive period.
improved calculations of global atmospheric Upon subtraction of the means for that
MW = a3g- lf _ _u cos2 ,:bdk dC dp (1)
angular momentum (AAM), have aided the period from all corresponding daily belt
understanding of dynamic interactions values, we derived a time-latitude diagram
among the planet's solid portion, atmo- (Fig. 1B) of the temporal variability of belt
MP = a4 _ g- lf f PsCOS3qbdk dqb (2)
sphere, and ocean. On time scales be- momentum anomalies. The relative contri-
tween roughly a fortnight and several years, where a is Earth's mean radius; g is the bution of each belt to the behavior of Mwin
changes in the angular momentum of the acceleration due to gravity; u is zonal wind; Fig. 1A is shown in Fig. 1C in terms of the
solid Earth, manifested as variations in the a, _b, and p are longitude, latitude, and fractional covariance between the time se-
length of day (1.o.d.), are almost entirely pressure (Ps is pressure at Earth's surface), ries of the individual belt and global Mw
accounted for by changes in AAM (1, 2). respectively; and _ is Earth's mean angular anomalies. Although a strong signal exists
Discrepancies in this balance at high fre- velocity. Integrals over the global atmo- near 45°S, the broad expanse of positive
quencies have been noted, and details of the sphere were calculated from operational covariance values in low-latitude Southern
momentum exchange mechanisms have not wind and surface-pressure analyses pro- Hemisphere belts, peaking near 10°S, indi-
been fully understood. To address these is- duced by several of the world's weather cates that variations there are most respon-
sues, a campaign was planned to determine centers. These values were collected by the sible for the Mwfluctuations: The temporal
Earth orientation and AAM parameters Sub-bureau for Atmospheric Angular Mo- oscillation between negative and positive
with the most accurate systems available. mentum of the International Earth Rota- anomalies at these latitudes mimics the
This experiment, named SEARCH'92 (3), tion Service for analysis and distribution global behavior.
was conducted from June to September (4). During the intensive period of the Although Fig. 1 isolates the important
1992, with a special period of intensive campaign, AAM appears to undergo anear- centers of momentum variability during our
measurements held from 25 July to 8 August. ly 10-day oscillation superimposed on a study period, it does not reveal the mecha-
During the intensive portion of the cam- (seasonal) upward trend, as shown (Fig. nisms responsible for these variations. The
paign, values of the relative component of 1A) by estimates from the U.S. National development of a new data set for the two
AAM about the polar axis due to winds, Meteorological Center (NMC) using anal- torques that link the atmosphere and Earth
M and of the component due to the yses on a 2.5 ° latitude-longitude grid and 12 (9) allows us an unprecedented opportuni-
TM,
vertical pressure levels up to 50 mbar (5). ty, however, to identify the manner by
Atmospheric and Environmental Research,Inc., 840 Although the oscillatory component is near which the momentum exchange occurs on
MemorialDrive,Cambridge, MA02139,USA. the high-frequency limit that generally ex- the time scales dealt with here. One torque
SCIENCE ° VOL. 264 ° 15APRIL 1994 407
results from differences in normal pressure To study the character of the torques in The torques over South America in
forces across mountain barriers, effecting a greater detail, we focus now on the 6-day particular appear to play a special role in
transfer of momentum from one component period from 31 July through 5 August, the explanation of much of the behavior of
to the other. The second torque arises from when the total torque was at its peak, the globally integrated mountain torque
tangential frictional stresses at the atmo- corresponding in Fig. 1A to the time of the during the 31 July to 5 August study period.
sphere's lower boundary, which transfer most rapid rise in AAM. Of the two This result is evident from Fig. 3B, in
momentum with the land or ocean below. torques, only the mountain one is positive
Ifapplied over the ocean, this stress can be during this 6-day period, and it determines
transmitted within a day or so to the solid the sign of the total torque. Hence, the A..... i_ i
_,_6
Earth through the development of sea-level acceleration experienced by the global at-
_3
differences at the continental margins on mosphere during this period must be due
opposite sides of a basin, as the ocean to interactions with Earth's topography.
= 0
adjusts barotropically to the wind-forcing Therefore, we examine the mountain torque o
field (10). more closely by decomposing its global _-3-
Global values of these two torques have mean value for 31 July to 5 August into 8,
c
been computed four times per day for the contributions made at individual latitudes ,3B >
intensive period of SEARCH'92 as part of (Fig. 3A). During this period, the main
an ongoing effort at NMC (11). Mountain contribution to the global mountain
torque terms are calculated from the surface torque appears to come from the region
E=o0 ",
pressure and topography fields, whereas fric- between the equator and 30°S (14).
I ""\ ,/ I I
tional stress torques are calculated from The torque data set now available per- "
la-3-1 \ .// I I
the physical parameterizations of the fore- mits us to isolate, in turn, which portions of
I
cast model (12). Global values averaged the 0° to 30°S region are most responsible _-61
25 2I6 27I 2I8_ 29 3I0 3I1 1.... 2 3 4 5] 6I 7I
daily of these two torques are displayed in for the maximum there, and we do so by
July 1992 August
Fig. 2. Mountain torques are positive dur- dividing the region into three longitudinal
ing most of the period and dominate the sectors on the basis of the distribution of Fig. 2. (A) Daily means of mountain (solid trace)
and friction (dashed trace) torques from the
frictional stress term, which is negative continents. The net mountain torque
NMC analysis-forecast system during the inten-
and relatively steady throughout (13). across a landmass is generally the small
sive period of SEARCH'92. (B) Time derivative
The sum of the mountain and stress difference between large values of opposite
of the relative plus planetary AAM (solid trace)
torques should, in principle, be identical sign; its evaluation, therefore, involves a compared with the sum of mountain and friction
to the time derivative of global momentum. sensitive calculation (15, 16). For the 6-day torques (dashed trace).
The agreement in Fig. 2B, though very period of rapid atmospheric acceleration
good, is not perfect because momentum is studied here, southern tropical Africa and
computed from the wind and pressure anal- South America dominate (17) the transfer
yses of the NMC system, whereas the torque of momentum from the solid Earth to the
calculations involve a parameterization of atmosphere both within the Southern
friction and other approximations, such as Hemisphere tropics and, therefore from Fig.
the use of discrete temporal representations. 3A, for the globe as a whole.
Fig. 1. (A) Comparison of relative lo A
plus planetary AAM (thick solid E"
trace) and of relative momentum E_o_ 5 j--- /'x N" . 0.1
alone (thin solid trace) with the /",, m ,,...... i " 30S 0 30N 60N 90N
length of day during the intensive ====o Latitude (degrees)
period of SEARCH'92. The atmo- -5
o,,,; II //" II oo.ol _s="g
spheric series are computed from
once daily (00 UTC) NMC analyses x_'_10,
',.. /" , I ,
ofsurface pressure and winds to the I I I I--;I I I I I -0.2
50-mbar level and are reported here
90NB ....... _ C I
in momentum units along the leftas
60__EEi i!i_,,48r_:::_g:::X -- _
well as equivalent I.o.d. units along I
F__,_::: :::::_x_..... _-...... :S:: )
the right, with the planetary compo- _:!:_,i'e- _ ,'8::-"_ ---_u_'¢_¢> ..... _:--_ 5r_::::3::::Q_
nent of momentum weighted by a k "5 _-'g>_ <_g_#ii_i_#'_<;_t::#ii_iY i::!ili_
i
0.7 factor relating to the Earth's re- 1 2 3 4 5
sponse. The I.o.d. values (dotted August
_:::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
trace) are based on Very Long ::::::::::::::::::::::::::::::::::::::::::::::::::::::____ ._> • Fig. 3. (A) Profile in latitude of the z()nal mean
Baseline Interferometry measure- __ 3_........ ._.........:_........................_.__--- mountain torque averaged during 31 July to 5
ments. Means have been removed August 1992. The integral under the curve
from allseries. The straight dashed equals the mean global mountain torque over
lines inthis and the next figure de- 25 26 27 28 29 30 31 1 2 3 4 5 6 7 8-0,2 0,0 0.2 the period, that is, the mean for the mountain
note the period during which the July 1992 August torque curve in Fig. 2 over these 6 days. (B)
atmosphere underwent a rapid acceleration and that we focus on to determine the mechanism Four times daily series of mountain torque for
causing the acceleration. (B) Time-latitude diagram of the difference between the relative angular the globe (solid trace) and forthree continental
momentum within 46 equal-area latitude belts during the intensive period of SEARCH'92 and the regions inthe 0°to 30°S band during 31 July to
mean for each belt during the period. Values, given _n1024kg-m2s-1, are based on once daily (00 5 August 1992: South America (thick dashed
UTC) NMC winds to 50 mbar. Easterly (negative) anomalies are shaded. (C) Fractional covariance trace), Africa (thin dotted trace), and East In-
of anomaly belt momentum values with Mw,the global sumof the values during the intensive period. dies-Australia (mixed dotted and dashed
The sum of the 46 numbers comprising the curve isunity. trace).
408 SCIENCE • VOL. 264 • 15APRIL 1994
,!
which we have plotted time series of the simulations of the atmosphere's general REFERENCES AND NOTES
global mountain torque and the torque circulation (21) support this result. The
1. R. D. Rosen, D. A. Salstein, T. M. Wood, J.
across continental portions of the 0° to 30°S present case study, of course, illustrates an
Geophys. Res.95, 265 (1990); R.Hide and J. O.
belt for this period. Although the torque instance when the mountain torque across Dickey, Science 253, 629 (1991).
across the southern tropical portion of Af- a single continent is highlighted. 2. Contributions ofthe oceans and core are small at
these time scales; those due toexternal tides can
rica is on average larger than that across Synoptic maps of mean sea-level pres-
be significant, but because they occur atspecific,
tropical South America, the latter is much sure based on NMC analyses reveal that a well-known frequencies, they are easily isolated
better correlated in time with the global strong high-pressure center moved to the and so are removed here.
torque, explaining 69% of the variance in east of the Andes near the beginning of 3. J. O. Dickey, Adv. Space Res. 13, 185(1993).
the global time series during these 6 days. August 1992, a shift that appears to ac- 4. D. A. Salstein, D. M. Kann, A. J. Miller, R. D.
Rosen, Bull. Am. Meteorol. Soc. 74, 67 (1993).
Indeed, when this comparison is extended count for the acceleration of the atmo-
5. C!ose agreement among AAM values from allthe
to all of South America, this proportion sphere then. A strong zonal pressure gradi- meteorological centers helps tojustify theuse ofa
rises to 88%. Fluctuations in torque with ent associated with this progression became single AAM series here. Although higher frequen-
cydata are available, theonce dailyvalues shown
time determine the shape of temporal vari- centered over the spine of the Andes on 1
in thefigure are appropriate for representation of
ations in angular momentum, so from this August (Fig. 4), corresponding to the in- the oscillation ofinterest.
perspective conditions around South Amer- crease in the acceleration of the atmosphere 6. The 0..7factor applied to the pressure term in-
ica during 31 July to 5 August 1992 were that began on that day (Fig. 1). Investiga- volves a Love number correction toaccount for
the response ofthe nonrigid, solid Earth toatmo-
especially important in determining the tion of the role of other synoptic events in
spheric loading.
character of AAM then. the global momentum budget will be part of 7. The measurement technique forthedetermination
The relative importance of mountain a more general study of "1- to 2-week" of Earth's rotation here is Very Long Baseline
versus friction torques in the planetary waves in angular momentum evident Interferometry [D. S. Robertson, W. E.Carter, J.
Campbell, H Schuh, Nature 316, 424 (1985)].
momentum cycle appears to be a function throughout much of our record. In these
8. R.D.Rosen and D.A. Salstein, J. Geophys. Res.
of time scale (18), and the dominance of studies, the manner in which various time 88, 5451 (1983).
the former in our case study is consistent scales of behavior are separated can be 9. G. H White, in "Proceedings of the American
with other studies of short-period changes important in the interpretation of results. Geophysical Union Chapman Conference on
Geodetic VLBI: Monitoring Global Change," Nat.
in AAM. Thus, the mountain torque was For example, when trends are also removed
Oceanic Atmos. Adm. Tech. Rep. NOS 137NGS
more responsible for the variations ob- from the belt momentum series in Fig. 1B, 49 (1991), p.262.
served in AAM within the 2-month spe- so as to isolate a 10-day fluctuation, varia- 10. R.M. Ponte, J. Geophys. Res. 95, 11369 (1990).
cial observing periods of the 1979 Global tions near 40°S become much more crucial • 11. G. H.White, in"Research Activities inAtmospher-
icand Oceanic Modeling," CAS/JSC Work. Group
Weather Experiment (16) and explains for global Mw variations than those near
Numer. Exp. Rep. no. 18 (World Meteorological
much of the momentum anomaly observed 10°S (as in Fig. 1C). Similarly, when mean Organization, Geneva, 1993), 2.3-2.4.
at the height of the E1 Nifio-Southern torques for the period are removed (equiv- 12. Inthe NMC model, surface momentum exchange
Oscillation event in January 1983 (19). alent to the detrending of AAM), frictional iscalculated with abulk aerodynamic formula in
which theexchange isproportional tothevertical
Torques over land at submonthly periods torques over 40°S must be invoked to help
momentum gradient, with the proportionality de-
are more important than those over the account for the rising portion of the AAM pendent on wind speed and surface layer static
ocean, according to a recent study (20) "wave," and contributions from mountain stability [M. Kanamitsu, Wea. Forecasting 4,335
that included satellite-based measure- torques over several regions besides those (1989)].
13. Friction torques over both ocean and land are
ments of ocean stresses, and multiyear highlighted here become more important. each negative, with meanvalues of-0.53 x 1019
and -1.46 x 1019kg.m2s-2, respectively, during
the intensive period.
Fig. 4. Maps of sea-level pressure, 10N 14. Substantial frictional torques also transfer mo-
in millibars, based on NMC analy- A_ J <3 alauly1992 mentum into the Southern Hemisphere tropical
atmosphere, but these are more than compensat-
ses in an area near South America 0 ed by negative frictional torques in the southern
for 00 UTC on (A) 31 July and (B) 1 •:fill:: o_/ extratropics.
August 1992, just before and dur- 15. J. M.Wahr and A. H.Oort, J.Atmos. Sci. 41, 190
ing the peak acceleration of the 10S (1984).
atmosphere. A high-pressure sys- 16. R. Swinbank, Q. J. R. Meteorol. Soc. 111, 977
tem moves just to the east of the 20S (1985).
Andes, resulting in a strong zonal 17. Mountain torques computed for southern tropical
pressure gradient across the moun- portions of Africa, South America, and East In-
dies-Australia are 1.94 x 1019,1.54 x 1019,and
tainous topography. _- 50S -0.52 x 1019kg.m2s 2,respectively, based on
thedivision ofthe 0°to30°Sregion into sectors at
g'
4os 0°,90°E, and 120°W longitudes.
¢ t0N 18. R.D.Rosen, Surv. Geophys. 14, 1(1993).
19. W.L Wolfand R.B.Smith, J.Atmos. Sci.44,3656
_' o (1987).
20. R.M. Ponte and R.D. Rosen, J. Geophys. Res.
98, 7317 (1993).
lOS 21. G.J. Boer, ibid. 95, 5511 (1990).
22. Wethank G.White ofthe NMC and D.Robertson
of the National Ocean Service for supplying
2os torque and I.o.d. data for this study, U. Kann for
valuable help in preparing the AAM data forthis
period, and K. Cady-Pereira and P. Nelson of
50s Atmospheric and Environmental Research, Inc.,
for providing excellent programming support.
Supported by NASA under both its Geophysics
4os Program (contract NASW-4751) and Earth Ob-
100w 8ow 6ow 40w 20w serving System Program (contract NAGW-2615).
Longitude (degrees) 22 November 1993; accepted 24 February 1994
SCIENCE • VOL. 264 • 15 APRIL 1994 409