skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Coupled Model Simulations of Boreal Summer Intraseasonal (30-50 day) Variability, Part 1: Systematic Errors and Caution on Use of Metrics

Abstract

Boreal summer intraseasonal (30-50 day) variability (BSISV) over the Asian monsoon region is more complex than its boreal winter counterpart, the Madden-Julian oscillation (MJO), since it also exhibits northward and northwestward propagating convective components near India and over the west Pacific. Here we analyze the BSISV in the CMIP3 and two CMIP2+ coupled ocean-atmosphere models. Though most models exhibit eastward propagation of convective anomalies over the Indian Ocean, difficulty remains in simulating the life cycle of the BSISV, as few represent its eastward extension into the western/central Pacific. As such, few models produce statistically significant anomalies that comprise the northwest to southeast tilted convection which results from the forced Rossby waves that are excited by the near-equatorial convective anomalies. Our results indicate that it is a necessary, but not sufficient condition, that the locations the time-mean monsoon heat sources and the easterly wind shear be simulated correctly in order for the life cycle of the BSISV to be represented realistically. Extreme caution is needed when using metrics, such as the pattern correlation, for assessing the fidelity of model performance, as models with the most physically realistic BSISV do not necessarily exhibit the highest pattern correlations with observations. Furthermore, diagnostic latitude-timemore » plots to evaluate the northward propagation of convection from the equator to India and the Bay of Bengal also need to be used with caution. Here, incorrectly representing extratropical-tropical interactions can give rise to 'apparent' northward propagation when none exists in association with the eastward propagating equatorial convection. It is necessary to use multiple cross-checking diagnostics to demonstrate the fidelity of the simulation of the BSISV.« less

Authors:
;
Publication Date:
Research Org.:
Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States)
Sponsoring Org.:
USDOE
OSTI Identifier:
943813
Report Number(s):
UCRL-JRNL-231507
Journal ID: ISSN 0930-7575; CLDYEM; TRN: US200902%%419
DOE Contract Number:
W-7405-ENG-48
Resource Type:
Journal Article
Resource Relation:
Journal Name: Climate Dynamics, vol. 31, N/A, February 21, 2008, pp. 345-372; Journal Volume: 31
Country of Publication:
United States
Language:
English
Subject:
58 GEOSCIENCES; CONVECTION; EQUATOR; HEAT SOURCES; INDIA; INDIAN OCEAN; LIFE CYCLE; METRICS; MONSOONS; OSCILLATIONS; PERFORMANCE; SHEAR; SIMULATION

Citation Formats

Sperber, K R, and Annamalai, H. Coupled Model Simulations of Boreal Summer Intraseasonal (30-50 day) Variability, Part 1: Systematic Errors and Caution on Use of Metrics. United States: N. p., 2007. Web.
Sperber, K R, & Annamalai, H. Coupled Model Simulations of Boreal Summer Intraseasonal (30-50 day) Variability, Part 1: Systematic Errors and Caution on Use of Metrics. United States.
Sperber, K R, and Annamalai, H. Wed . "Coupled Model Simulations of Boreal Summer Intraseasonal (30-50 day) Variability, Part 1: Systematic Errors and Caution on Use of Metrics". United States. doi:. https://www.osti.gov/servlets/purl/943813.
@article{osti_943813,
title = {Coupled Model Simulations of Boreal Summer Intraseasonal (30-50 day) Variability, Part 1: Systematic Errors and Caution on Use of Metrics},
author = {Sperber, K R and Annamalai, H},
abstractNote = {Boreal summer intraseasonal (30-50 day) variability (BSISV) over the Asian monsoon region is more complex than its boreal winter counterpart, the Madden-Julian oscillation (MJO), since it also exhibits northward and northwestward propagating convective components near India and over the west Pacific. Here we analyze the BSISV in the CMIP3 and two CMIP2+ coupled ocean-atmosphere models. Though most models exhibit eastward propagation of convective anomalies over the Indian Ocean, difficulty remains in simulating the life cycle of the BSISV, as few represent its eastward extension into the western/central Pacific. As such, few models produce statistically significant anomalies that comprise the northwest to southeast tilted convection which results from the forced Rossby waves that are excited by the near-equatorial convective anomalies. Our results indicate that it is a necessary, but not sufficient condition, that the locations the time-mean monsoon heat sources and the easterly wind shear be simulated correctly in order for the life cycle of the BSISV to be represented realistically. Extreme caution is needed when using metrics, such as the pattern correlation, for assessing the fidelity of model performance, as models with the most physically realistic BSISV do not necessarily exhibit the highest pattern correlations with observations. Furthermore, diagnostic latitude-time plots to evaluate the northward propagation of convection from the equator to India and the Bay of Bengal also need to be used with caution. Here, incorrectly representing extratropical-tropical interactions can give rise to 'apparent' northward propagation when none exists in association with the eastward propagating equatorial convection. It is necessary to use multiple cross-checking diagnostics to demonstrate the fidelity of the simulation of the BSISV.},
doi = {},
journal = {Climate Dynamics, vol. 31, N/A, February 21, 2008, pp. 345-372},
number = ,
volume = 31,
place = {United States},
year = {Wed May 30 00:00:00 EDT 2007},
month = {Wed May 30 00:00:00 EDT 2007}
}
  • The boreal summer intraseasonal variability (BSISV) associated with the 30-50 day mode is represented by the co-existence of three components, poleward propagation of convection over the Indian and tropical west Pacific longitudes and eastward propagation along the equator. The hypothesis that the three components influence each other has been investigated using observed OLR, NCEP-NCAR reanalysis, and solutions from an idealized linear model. The null hypothesis is that the three components are mutually independent. Cyclostationary EOF (CsEOF) analysis is applied on filtered OLR to extract the life-cycle of the BSISV. The dominant mode of CsEOF is significantly tied to observed rainfallmore » over the Indian subcontinent. The components of the heating patterns from CsEOF analysis serve as prescribed forcings for the linear model. This allows us to ascertain which heat sources and sinks are instrumental in driving the large-scale monsoon circulation during the BSISV life-cycle. We identify three new findings: (1) the circulation anomalies that develop as a Rossby wave response to suppressed convection over the equatorial Indian Ocean associated with the previous break phase of the BSISV precondition the ocean-atmosphere system in the western Indian Ocean and trigger the next active phase of the BSISV, (2) the development of convection over the tropical west Pacific forces descent anomalies to the west. This, in conjunction with the weakened cross-equatorial flow due to suppressed convective anomalies over the equatorial Indian Ocean reduce the tropospheric moisture over the Arabian Sea, and promote westerly wind anomalies that do not recurve over India. As a result the low-level cyclonic vorticity shifts from India to southeast Asia and break conditions are initiated over India, and (3) the circulation anomalies forced by equatorial Indian Ocean convective anomalies significantly influence the active/break phases over the tropical west Pacific. Our model solutions support the hypothesis that the three components of the BSISV influence each other.« less
  • General circulation model (GCM) simulations of low-frequency variability with time scales of 20 to 70 days are analyzed for the Pacific sector during boreal winter. The GCM's leading mode in the upper-tropospheric zonal wind is associated with fluctuations of the East Asian jet; this mode resembles, in both structure and amplitude, the Pacific/North American (PNA) pattern found in the observation on these time scales. In both the model and observations the PNA anomaly is characterized by (1) a linear balance in the upper-tropospheric vorticity budget with no significant Rossby wave source in the tropics, (2) a barotropic conversion of kineticmore » energy from the time mean Pacific jet, and (3) a north/south displacement of the Pacific storm track. In the GCM, the latter is associated with synoptic eddy heat flux and latent heat anomalies that appear to contribute to a strong lower-tropospheric source of wave activity over the North Pacific. This is in contrast to the observations, which show only a weak source of wave activity in this region. The GCM produces 60% of the total observed Pacific sector low-frequency zonal wind variance. About one-third of the missing variability appears to be due to unrealistic simulations of the Madden-Julian oscillation; the remainder is characterized in the variance spectrum as a deficit in the overall level of [open quotes]background[close quotes] variability. 41 refs., 13 figs.« less
  • No abstract prepared.
  • Infrared radiative cooling rates are calculated over the Asian summer monsoon between 5{degrees}S-2{degrees}N and 40{degrees}-135{degrees}E at a spatial resolution of 5{degrees} x 5{degrees} for the summer seasons of 1984 and 1987. A medium spectral resolution infrared radiative transfer model with specified temperature, moisture, clouds, and trace gas distributions is used to obtain the cooling rate profiles. Cloud distributions for the two summers are obtained from Indian National Satellite measurements. Seasonal mean and intraseasonal variations of clouds and radiative cooling rates over a 21-76-day range of periods are examined. The analysis identifies centers over the central and eastern Indian Ocean, andmore » western Pacific Ocean, along the equator, and along 15{degrees}N, where seasonal mean cloud amounts range from 40% to 80% with cloud tops mostly in the middle and upper troposphere. Intraseasonal variability of clouds is also large over these centers (% variances > 25%). Consistently, seasonal mean cooling rates are at a maximum (3{degrees}-5{degrees} day{sup {minus}1}) in the upper troposphere between 300 and 400 mb, related to cloud-top cooling. The cooling rates below 400 mb are between 1{degrees} and 3{degrees}C day{sup {minus}1}. The cooling rates exhibit intraseasonal amplitudes of 1.0{degrees}-1.5{degrees}C day{sup {minus}1}. It is shown that intraseasonal variability of cooling rates over the Indian Ocean can perturb convective heating by 10%-30% in the upper and lower troposphere. Based on a one-dimensional radiative-convective equilibrium model, it is estimated that the radiative damping timescale over the Indian Ocean region is {approximately}3 days. Based on this damping timescale and in conjunction with a model of equatorial Kelvin waves with first baroclinic mode, it is hypothesized that the variable cloud-radiative cooling rates can alter phase speeds of Kelvin waves by up to 60%. 80 refs., 14 figs., 3 tabs.« less
  • Dust variability in the climate system has been studied for several decades, yet there remains an incomplete understanding of the dynamical mechanisms controlling interannual and decadal variations in dust transport. The sparseness of multi-year observational datasets has limited our understanding of the relationship between climate variations and atmospheric dust. We use available observations and a century-length fully coupled Community Earth System Model (CESM) simulation to show that the El Niño- Southern Oscillation (ENSO) exerts a control on North African dust transport during boreal summer. In CESM, this relationship is stronger over the dusty tropical North Atlantic than near Barbados, onemore » of the few sites having a multi-decadal observed record. During strong La Niña summers in CESM, a statistically significant increase in lower tropospheric easterly wind is associated with an increase in North African dust transport over the Atlantic. Barbados dust and Pacific SST variability are only weakly correlated in both observations and CESM, suggesting that other processes are controlling the crossbasin variability of dust. We also use our CESM simulation to show that the relationship between downstream North African dust transport and ENSO fluctuates on multidecadal timescales and may be modulated by the North Atlantic Oscillation (NAO). Our findings indicate that existing observations of dust over the tropical North Atlantic are not extensive enough to completely describe the variability of dust and dust transport, and demonstrate the importance of global models to supplement and interpret observational records.« less