Analyses and Comparisons of Two Reanalyses for Tropical Monthly v-component wind field

Qin Yujing ,Wang Panxing

(Nanjing University of Information Science and Technology, Nanjing 210044, China)

Wind field is the most key element in tropic weather and climate

analysis. Monthly mean v-component wind field is of importance for

it represents the climatological and anomalous features of tropic meridional

cell. Hadley Cell (hereafter H.C.) is the most intensive

and extensive  meridional circulation whose common

upwelling branch's location and intensity are synchronously changed with those

of the intertropical convergence zone (hereafter ITCZ). In this paper, we use

monthly mean OLR minimum band representing ITCZ

in order to compare H.C. upwelling branch determined

respectively by v component Wind field of ECMWF and NCEP/NCAR

reanalysis (E data and N data ). Three datasets are chosen in tropic area (30ºS°ª30ºN)

from 1979 to 2001:1) NOAA Interpolated

Outgoing Longwave Radiation (OLR) (2 spatial

dimensional)series; 2) ERA-40 monthly mean v component wind field (3 spatial

dimensional) series; 3)NECP1 monthly mean v component wind field (3 spatial

dimensional) series.

      On

the basis of Lorenz decomposition principle,

OLR of the given  month

and 500hPa vertical speed , (computed according to E data and N data separately)

determined by mass stream function  are decomposed into 4 independent part (

as shown in table 1); among which the Climatological

zonal unsymmetrical component()is called stationary wave and

the anomalous zonal unsymmetrical component () is called stationary wave anomaly.

Table

1 components of OLR, ,

 in  derived from Lorenz

decomposition

classes
Climatological zonal symmetrical
Climatological zonal unsymmetrical
Anomalous zonal symmetrical
Anomalous zonal unsymmetrical

Analyses are done from two perspectives

(climatological and anomalous) and we put emphasis on the relationship between ITCZ given by OLR and 500hPa H.C. upwelling branch given by ,(locations and intensities). Main

results are as followings:

1)The

status of peak location of H.C. upwelling branch given by ,can be divided into two types: single-peak

pattern and double-peak pattern as well as that of ITCZ given by (fig.1). Also, the months

when they appear respectively are totally consistent with those of ITCZ.

Statistical analyses show that the average latitude departure that E data maximum point departs from minimum point is 1.25º, less

than the value 1.875º computed from N data  and  . Thus, the climatological

analysis portion indicates that  component wind field are intimately

related to ITCZ, implying that the tropical v wind field of E data is subtlely

superior to that of N data.

Fig.1 The relationship between induced by (thick solid line) and the

ascending branch of H.C..

a,  and ; b,  and .

2)Using

singular vector decomposition(SVD) method, the correlation between  and

() in January and July are investigated. In the first spatial modes of and , the latitude location of valley (peak) corresponding

to the maximum (minimum) value of , demonstrating consistence anomalous location

between ITCZ and ascending branch of H.C.. Meanwhile the time coefficients series

of the first spatial modes vary synchronously, with peaks (valleys) according

to simultaneous El Nino (La Nina) events, explaining that tropical SSTA is a

key factor for anomalous ITCZ and H.C., and the correlationship in January is

better than that in July.  

3)Using

SVD the relationships between and °¢over

the tropic Indian Ocean, western and middle Pacific Ocean(30ºS°ª30ºN,60ºE°ª120ºW)in January and July are analyzed. The first SVD

spatial pattern indicate that Negative s (dark shaded) basically

overlap with positive s,

which mainly assemble in mid-Pacific in January and mitigate to tropical

western Pacific, ocean to the east of the Philippines, coastal regions in

south-east China and the Bengal Bay. The peaks (valleys) of the first SVD modes

time coefficient series are also consistent with

El Nino (La Nina) events (except for La Nina in January 1985). The first SVD

spatial pattern and time series of and are similar with those of and . These imply that the relationships between

OLR and 500hPa from

two analyses which caused by regional difference of also have obvious physical background.