How neighboring SST anomalies influence a particular atmospheric variable, such as precipitation, over the land regions of the Americas, can be complex. While the last El Niņo of 1997 reduced boreal summer Mexican monsoon rainfall, the El Niņo of 1987 had little effect on summer precipitation over Mexico. A comparison of SST anomaly patterns shows that the SST anomaly west of Mexico was positive in 1997 but negative in 1987. We suspect that this SST difference is a significant cause of the differences in rainfall patterns (Fig. 1).

It has been customary for atmospheric modelers to take SST as a given boundary condition. In reality, however, SST itself is affected not only by the ocean but also by atmospheric heat fluxes. Thus, it can be misleading to use SST as a fixed boundary condition.

Conceptually, we distinguish between "passive SST anomalies," purely caused by random atmospheric flux, and "active SST anomalies" which result mainly from changes in the oceanic circulation. While the former is to a large degree unpredictable, the latter may potentially be predictable for long lead times.

PROJECT GOALS:
We clarify the notions of active and passive SST variations by defining a passive ocean as one where the variability occurs only above a fixed interior depth D of the upper ocean, which we take as the annual maximum of the oceanic mixed-layer depth. Variability which is more than passive is called active. Based on this concept of active and passive SST we will

(*) explore where in space and time oceanic thermal variations and transports influence variability in the rainfall in the tropical/sub-tropical Americas,

(*) discern where the ocean's role is more than passive, and in those cases, analyze the circulation modes through which the ocean acts.

If it turns out from our investigations that most of the relevant SST variability in the eastern Pacific and the Atlantic can be regarded as active, then the use of specified SST as the lower boundary condition for GCM and mesoscale modeling can be used without further worry. If distinguishing between active and passive SST variability is found to be important, however, then this would suggest the imposed lower boundary condition method of atmospheric modeling is problematic.

METHODOLOGY:
To explore the cause of SST anomalies and its influence on tropical rainfall over the Americas we use an atmosphere-land model, a slab ocean mixed layer model, an ocean model with which we perform coupled and uncoupled experiments.

The atmospheric component is modeled using the "Quasi-equilibrium Tropical Circulation Model" (QTCM), which is a model of intermediate complexity between a standard general circulation model (GCM) and simpler models. The QTCM is GCM-like in that it is based on the primitive equations, simulates the full tropical climatology, and includes a relatively complete set of physical parameterizations. By incorporating constraints on the vertical structure of temperature that come from quasi-equilibrium convective parameterizations, the model is made numerically fast and easy to analyze compared to a GCM. QTCM is coupled to a land-surface model "SLand" which includes a biophysically base evapotranspiration parameterization. A simple mixed-layer model with specified oceanic heat transports ("Q-flux") is also part of the standard QTCM distribution. As our focus is on the active heat transport present in the ocean, we cannot discard many of the processes included in an ocean GCM. Our specific interest in tropical variability on seasonal to interannual time scales, however, allows us to restrict our attention to the upper ocean. We use an Upper Ocean Model (UOM) that is based on an ocean GCM (NCOM), but fixes the deep ocean baroclinic variables at their climatological values.

When the UOM is forced with observed wind and heat fluxes, the model is able to reproduce the observed SST pattern (Fig. 2). There is a slight cold bias of the modeled SST in the eastern equatorial Pacific, presumably caused by the model's upwelling velocity, which is stronger than observed. Some tuning of the model's near surface viscosity might be necessary to correct this bias.

RESULTS AND ACCOMPLISHMENTS
1. Atmospheric teleconnections
We used the QTCM to analyze the mechanisms that are involved in the El Niņo induced negative rainfall anomalies adjacent to the warm SST anomalies in the central and eastern Pacific. We forced the QTCM with positive SST anomalies in the Pacific from January to March 1998. An ensemble of model integrations with different initial conditions were used to diagnose the behavior of the QTCM and to analyze the flux anomalies that were responsible for the observed precipitation anomalies.

The El Niņo related warming in the central and eastern Pacific leads to an enhanced convective energy flux into the atmosphere over this region (Fig. 3). The anomalous tropospheric warming is spread out by wave dynamics over a wide area. However, the sinking regions, and thus the regions of anomalous dryness, are more confined. In traditional theory, anomalous radiative cooling associated with warm temperature would initiate relative descent. We find, however, that the radiative anomalies associated with the tropospheric warming are quite small, on the order of a few W m-2, and produce only small descent anomalies. Instead, the anomalous cooling which balances the anomalous adiabatic warming in the sinking regions is caused by different mechanisms in different regions. For the PACS-relevant decent area over Venezuela and eastern Brazil, our analysis shows that the cooling is caused by anomalous advection of cold and dry air. In other regions, such as the South Pacific, the anomalous evaporation is the major contributor to the cooling. Using the moist static energy budget, we derived relatively simple theories to match each of the balances which includes the effects of convective, cloud-radiative and land-surface feedbacks.

Another finding of this study is that the observed cold SST anomalies surrounding the warm El Niņo region cannot be reproduced by anomalous atmospheric heat fluxes only, when the QTCM is coupled with a simple mixed-layer ocean in which ocean transport is represented by climatological surface heat fluxes. This suggests that changes in the oceanic circulation and thermocline depth are important for the cold SST anomalies in the western and subtropical Pacific associated with El Niņo.

2. Coupling of QTCM with UOM
The two models are coupled without flux correction over the Tropics (30S-30N). At higher latitudes the forcing fields are prescribed. The coupled system exhibits a warm SST bias in the Tropics and a Pacific cold tongue that extends too far westward, a common deficiency of many state-of-the-art coupled GCMs. The cause of the cold tongue problem is due mainly to the strength and westward extent of the easterly surface winds along the equator. The atmospheric model's 850 hPa winds, however, compare well with observational estimates from reanalyses, which have a better atmospheric boundary layer formulation. The slight equatorial cold bias of the ocean model reinforces this deficiency in equatorial surface winds.

The analysis of Wallace et al. (1989) has shown that the stabilization of the atmospheric boundary layer due to cold SSTs significantly reduces the momentum flux from the atmosphere into the ocean over the eastern equatorial Pacific. We find that implementing this effect in our coupled simulation has a major impact. Incorporating the stability effect into our atmospheric flux calculations and tuning the models surface wind estimates was crucial to obtaining a reasonable climatology of the coupled system. We are still working on a better atmospheric boundary layer formulation and on analyzing the SST cold bias of the oceanic component. A number of annual mean surface fluxes of the coupled system compare reasonably well with the observational estimates by Large et al. 1997. The Indian summer monsoon winds are underestimated, which leads to warm SST bias along the eastern African coast and to increased precipitation over central Africa. For figures and further discussion see the coupled model link below.

The coupled system is being used to study and analyze the oceanic and atmospheric branches of the observed teleconnections of ocean SST anomalies with the PACS region. For more information and more up-to-date results of these coupled simulations, please visit our WWW page on the coupled system (link below).

3. Idealized monsoon study

The seasonal movement of the land convection zones depend on ocean-land contrast, but it is not yet established which coupling mechanisms are the most important in setting this contrast. To study these mechanisms we have begun using an idealized context. The QTCM coupled to the mixed-layer ocean model is used to investigate the processes involved in an idealized monsoon occurring on a single rectangular continent. Idealized divergence of ocean heat transports are specified as an annual average ``Q-flux". In this simple coupled configuration, the mechanisms that affect land-ocean contrast and, in turn, the seasonal movement of the continental convergence zones, are examined. These include soil moisture feedbacks; cooling of tropical oceans by ocean transport; ventilation (defined as the import of low moist static energy air into continental regions from ocean regions where heat storage opposes summer warming); and the ``interactive Rodwell-Hoskins mechanism", in which Rossby-wave-induced subsidence to the west of monsoon heating interacts with the convection zone.

The fixed ocean transports have a substantial impact on continental convection. If Q-flux is set to zero, subtropical subsidence and ventilation tend to substantially limit the poleward movement of summer monsoon rainfall. When land hydrology feedbacks are active, the drying of subtropical continents disfavors continental convection, even in the tropics. When ocean transports are included, tropical oceans are slightly disfavored as regions for producing convection; instead, continental convection is favored. The monsoon circulation then produces moisture transport from the ocean regions that allows substantial progression of convection into the subtropics over the eastern portion of the continent. The western portion of the continent tends to have a dry region of characteristic shape. This east-west asymmetry is partly due to the interactive Rodwell-Hoskins mechanism. The ventilation is of at least equal importance in producing east-west asymmetry and is the single most important process in limiting the poleward extent of the continental convection zone. Current work aims at examining these mechanisms over the PACS region.

FUTURE WORK
In addition to pursuing the work on teleconnections over land regions, and the mechanisms by which ocean-land contrast affects the monsoon, described above, we are conducting a number of experiments aimed at the question of active versus passive SST variability and the role of the ocean in teleconnections.

To provide some specific examples, consider the time varying heat flux from the deep ocean into the upper ocean layer, which on average tends to balance heat exchange with the atmosphere. For brevity, we refer to this as the Q-flux, since it is closely related to the time average divergence of the ocean heat transport often called a Q-flux in mixed-layer models. A passive ocean by definition does not include any anomalies in the Q-flux from the deeper ocean. We will first estimate the Q-flux from an uncoupled ocean integration with observed fluxes. A time average of this Q-flux will provide an estimate of the climatological Q-flux. We will then run a mixed-layer model coupled to the atmospheric model, forced at its base by the time series of the estimated Q-flux. The internal variability of the mixed-layer atmosphere coupled system will then be estimated from a model integration with estimated climatological Q-flux. Ensemble integration will then be used to determine the active SST variability and the associated atmospheric response.

We also plan to perform various coupled experiments to study climate teleconnections via the ocean and the atmosphere. The coupling between ocean and atmosphere will be restricted to specified spatial windows with climatological fluxes outside the window. One of these experiments will be performed as follows: First, a run with an active coupling area over the eastern Pacific will serve as a control run. In this run, ENSO will be suppressed as it arises from wind stress anomalies in the central and western Pacific, and only weak passive SST variability will result. Next we force the ocean with the observed wind stress anomalies in the central and western Pacific. This run will reproduce most of the El Niņo related SST anomalies in the eastern Pacific, but without the strong atmospheric teleconnections stemming from the large convective changes in the central Pacific. The resulting teleconnections to PACS land regions will then be examined.

PUBLICATIONS
Chou,C., J. D. Neelin, and H. Su, 2000: Ocean-atmosphere-land feedbacks in an idealized monsoon, Quart. J. Royal Met. Soc., (submitted).

Danabasoglu, G. and J. C. McWilliams, 1999: Upper Ocean Model, J. Climate, (submitted).

Large, W. G., G. Danabasoglu, J. C. McWilliams, P. R. Gent and F. O. Bryan, 2000: Equatorial circulation of a global ocean climate model with anisotropic horizontal viscosity, J. Phys. Oceanography, (submitted).

References:

Large, W. G., G. Danabasoglu and S. C. Doney, 1997: Sensitivity to surface forcing and boundary layer mixing in a global ocean model: annual-mean climatology, J. Phys. Oceanography, 28, 2418-2447

Wallace, J. M., T. P. Mitchel and C. Deser, 1989: The influence of sea surface temperature on surface wind in the eastern equatorial Pacific: Seasonal and interannual variability, J. Clim, 2, 1492-1499.

Additional publications available at: http://www.atmos.ucla.edu/~csi

CONTACTS

J. David Neelin (PI)
Email: neelin@atmos.ucla.edu
Phone: (310) 206-3734
 Fax: (310) 206-5219

Ning Zeng (Co-PI)
Email: zeng@atmos.ucla.edu
Phone: (310) 825 3439
Fax: (310) 206-5219

James C. McWilliams (Co-PI)
Email: jcm@atmos.ucla.edu
Phone: (310) 206 2829
 Fax: (310) 206-5219

Matthias Munnich
Email: munnich@atmos.ucla.edu
Phone: (310) 794 5899
 Fax: (310) 206-3051

Institution:
Institute of Geophysics and Planetary Physics
University of California, Los Angeles
3845 Slichter Hall
Los Angeles, CA 90095-1567

LINKS
Atmosphere model
http://www.atmos.ucla.edu/~csi/QTCM/

Land model:
http://www.atmos.ucla.edu/~zeng/land.html

NCOM/UOM:
http://www.cgd.ucar.edu/csm/models/ocn-ncom/

QTCM/UOM coupled system: http://www.atmos.ucla.edu/~munnich/TCM

Return to CLIVAR PACS Home Page