Comparison of Techniques for Isolating Equatorial Rossby Waves in Synoptic Studies

Carl J. Schreck, III Department of Earth and Atmospheric

SciencesUniversity at Albany, SUNY

Introduction

● This study investigates Equatorial Rossby (ER) wave activity in the Central Pacific during the September-November 1997, the peak of a very strong El Nino.

● These waves might be associated with three events of twin tropical cyclogenesis.

● Waves are examined using three methods:– Unfiltered data– 15-40 day band-pass time filter– Space-time filter

Space-time Filters● Wheeler and Kiladis (1999; hereafter WK99)

– n = 1 ER wave filter that is bounded by the dispersion curves for U = 0 and 8 m < h < 90 m

– Separates data into symmetric and antisymmetric components about the equator

● Roundy and Frank (2004) – ER wave filter expands beyond dispersion curves– Enforces no symmetry constraints

● Yang et al. (2003)– Data are broadly filtered into eastward and westward

propagating components– Each component is then projected onto the meridional

structure from shallow water modes– Allows for U ≠ 0

Unfiltered Hovmöller

•850 hPa u contoured every 3 m s-1, with westerlies in red and easterlies in blue, averaged 2.5°S-2.5°N

•Tb below 290 K shaded in 25 K intervals, averaged 10°S-10°N

•Locations of tropical cyclogenesis within 10.5° of equator, with open symbols for SH

15-40 Day Hovmöller

•850 hPa u contoured every 1 m s-1, with westerlies in red and easterlies in blue, averaged 2.5°S-2.5°N

•Tb shaded every 5 K, with negative anomalies in warm colors and positive anomalies in cool colors, averaged 10°S-10°N

•Locations of tropical cyclogenesis within 10.5° of equator, with open symbols for SH

WK99 ER band filtered Hovmöller•850 hPa u contoured every 0.5 m s-1, with westerlies in red and easterlies in blue, averaged 2.5°S-2.5°N

•Tb shaded every 2.5 K, with negative anomalies in warm colors and positive anomalies in cool colors, averaged 10°S-10°N

•Locations of tropical cyclogenesis within 10.5° of equator, with open symbols for SH

15-40 Day Hovmöller

•850 hPa u contoured every 1 m s-1, with easterly anomalies dashed

•Black line indicates objective wave tracks

15-40 Day Hovmöller

•850 hPa u contoured every 1 m s-1, with easterly anomalies dashed

•Black line indicates objective wave tracks

•Magenta line indicates linear best-fit of wave tracks

•Green numbers show best-fit phase speed (m s-1)

•Blue numbers show r2 values

-3.480.71

-4.820.89

-4.670.92

WK99 ER band filtered Hovmöller

•850 hPa u contoured every 1 m s-1, with easterly anomalies dashed

•Black line indicates objective wave tracks

•Magenta line indicates linear best-fit of wave tracks

•Green numbers show best-fit phase speed (m s-1)

•Blue numbers show r2 values

-5.400.99

-4.731.00

-4.531.00

Interaction with Background Convergence

Time-filtered variance

Background zonal wind

Background zonal divergence

Space-timefiltered variance

Unfiltered Map•850 hPa wind vectors

•850 hPa heights contoured every 10 m

•Tb below 290 K shaded in 25 K intervals

Time-Filtered Map•850 hPa wind anomaly vectors

•850 hPa height anomalies contoured every 3 m, with positive anomalies in red and negative anomalies in blue

•Tb shaded every 5 K, with negative anomalies in warm colors and positive anomalies in cool colors

Asymmetric Space-Time Filtered Map•850 hPa wind anomaly vectors

•850 hPa height anomalies contoured every 2 m, with positive anomalies in red and negative anomalies in blue

•Tb shaded every 2.5 K, with negative anomalies in warm colors and positive anomalies in cool colors

Symmetric Space-Time Filtered Map•850 hPa wind anomaly vectors

•850 hPa height anomalies contoured every 2 m, with positive anomalies in red and negative anomalies in blue

•Tb shaded every 2.5 K, with negative anomalies in warm colors and positive anomalies in cool colors

Summary

● Unfiltered Data– Strengths

● No concerns about artificial signals● Easy in real-time

– Weaknesses● Waves obscured by other features

Summary

● 15-40 Day Band-pass Filtered Data– Strengths

● Wave packet shows clearly in Hovmöller● Variance increases in a region of background zonal

convergence● Wind-height relationship resembles idealized mode

– Weaknesses● Difficult to determine phase speed

Summary

● WK99 Space-time Filter– Strengths

● Clearly separates each wave type● Smooth westward propagation● Waves can be clearly identified in maps, particularly

with symmetry constraints– Weaknesses

● Low amplitude● More difficult to examine the relationship between

filtered variance and zonal divergence