U.S. Dept. of Commerce / NOAA / OAR / ERL / GFDL *Disclaimer  

 

Table of Contents 

6. OBSERVATIONAL STUDIES

GOALS

6.1 DEVELOPMENT OF OBSERVATIONAL DATASETS AND
ANALYSIS TOOLS

ACTIVITIES FY97

6.1.1 Enhancing the Information Content of Radiosonde Temperature Data

A collaborative effort lead by Dian Gaffen of the Air Resources Laboratory (ARL), and including Ted Habermann of the National Geophysical Data Center (NGDC), is continuing. The purpose of this effort is to utilize both a unique compilation of "meta-data" describing historical changes in global radiosonde instruments (created at ARL) and advanced statistical methods (created at GFDL and NGDC) to improve the quality of radiosonde temperature data as well as the metadata itself. Software has been developed at ARL for the statistical analysis and graphical display of the radiosonde and meta-data. The meta-data developed at ARL has been ported to GFDL and set up for local use.

6.1.2 Acquisition of High-Resolution Satellite Datasets

As part of a systematic effort to assemble satellite datasets for diagnostic studies of various aspects of the atmospheric circulation, the voluminous archives (amounting to more than 1000 gigabytes on 250 exabyte tapes) of the DX dataset from the International Satellite Cloud Climatology Project (ISCCP) have been acquired from NASA/Langley. This dataset covers the period from 1986 through 1992, and includes the data collected by various operational geostationary and polar-orbiting satellites. The archives contain a large variety of cloud variables analyzed at 3hour intervals with a nominal spatial resolution of 30 km. It is anticipated that this data resource will facilitate the analysis of the mesoscale structure of weather systems occurring in various circulation regimes.

6.1.3 Development of Statistical Software

A library of Fortran routines for spectral and cross-spectral analysis of time series has been created. This package is based on the "multitaper method", a recent development in time series analysis which has demonstrated advantages in performance over more traditional approaches. The code is almost entirely portable (machine independent) and is intended for use within GFDL and for outside collaborators.

PLANS FY98

Collaboration will continue with Dian Gaffen of ARL and Ted Habermann of NGDC. This project will utilize "meta-data" and advanced statistical and data analysis techniques in an attempt to improve the quality of upper air temperature measurements and subsequently to estimate temperature trends.

The DX dataset from ISCCP will be examined in conjunction with the concurrent reanalysis products from NCEP for selected synoptic cases. Emphasis will be placed on the mesoscale organization of cloud properties associated with midlatitude cyclones and tropical convective disturbances. Efforts will continue toward expanding the satellite data base at GFDL for use in diagnostic research of the atmospheric circulation.

6.2 CLOUD PATTERNS ASSOCIATED WITH WEATHER SYSTEMS

ACTIVITIES FY97

6.2.1 Mesoscale Organization of Cloud Cover in Synoptic Disturbances

Selected cases of active synoptic development over the continental United States and tropical western Pacific have been identified. The mesoscale distributions (with a typical resolution of 0.5 degrees of longitude and latitude) of various cloud properties from the ISCCP-DX dataset (6.1.2) have been successfully mapped and analyzed for these cases. A classification scheme for assigning cloud types to individual pixels based on cloud top pressure and optical thickness has been designed and implemented. These high-definition cloud products provide detailed information on the fine-structure of the cloud cover near midlatitude frontal bands and convective clusters in tropical maritime areas.

Simultaneous distributions of atmospheric motion, pressure, temperature and water vapor mixing ratio from the NCEP/NCAR Reanalysis Project have been overlaid on the cloud patterns. A notable degree of consistency was found to exist between the atmospheric circulation fields and the cloud patterns. For instance, circulation vortices over the tropical Pacific are collocated with optically thick clouds with high tops, and elongated cloud bands in midlatitudes are aligned with warm and cold fronts (Fig. 6.1). The individual cloud features also exhibit considerable temporal continuity on time scales ranging from several hours to several days.

6.2.2 Comparison of Model-Simulated Cloud Patterns with Observations

The well-organized cloud patterns in synoptic-scale weather systems have recently been documented using satellite and surface observations in a series of studies (1309, dz). A similar methodology has been applied to the cloud data produced by the prediction model of the European Centre for Medium Range Weather Forecasting (ECMWF).

Composite patterns associated with extratropical cyclones in the North Atlantic were constructed using cloud data for one-day forecasts of the ECMWF model. The model result compares well with the observed cloud composite for the same set of reference dates. The model simulates extensive high-top/thick clouds to the east of the low pressure center in accordance with the satellite data. High clouds with less optical depth are seen to occur along the eastern edge of this cloud mass. Both the simulated and observed patterns indicate the presence of low-top clouds in the cold air mass to the west of the low pressure center. When compared with the satellite pattern, the model result shows a smaller longitudinal extent of the high-top/thick cloud mass, a smaller amount of high-top/thin clouds to the east of this cloud mass, and optically thicker low clouds to the west of the low pressure center.

PLANS FY98

The temporal and spatial evolution of the mesoscale cloud properties accompanying active synoptic episodes will be examined in further detail. The relationships between the cloud patterns and the ambient atmospheric structure will be interpreted on the basis of our knowledge of the physical processes operating in such circulation systems. The cloud fields will also be scrutinized in conjunction with any available mesoscale datasets of the atmospheric circulation (with spatial resolution of 50-100 km).

Sensitivity experiments are underway to determine how the cloud simulation at ECMWF depends on model resolution and the extent to which cloud mass (liquid, ice and cloud fraction) is advected by the wind field of the model.

6.3 AIR-SEA INTERACTION

ACTIVITIES FY97

6.3.1 Atmospheric Response to El Niño in a Higher-Resolution GCM

A salient deficiency of low-resolution GCMs (such as the R15L9 model) in simulating the atmospheric response to El Niño episodes in the tropical Pacific is the much lower amplitudes of the anomalies in the model atmosphere compared to observations (1256, 1393, 1437). In order to evaluate the extent to which this shortcoming could be remedied by using GCMs with higher resolution, a series of pilot integrations with prescribed sea surface temperature (SST) forcing corresponding to the 1982/83 El Niño event has been performed using the R30L14 model. Comparison of the extratropical wavetrains generated by the R30L14 and R15L9 models indicates that the amplitude of the ensemble-averaged 500 mb height anomalies is enhanced by a factor of 2-3 in the higher-resolution experiment (Fig. 6.2). In particular, the R30L14 version produces a more realistic simulation of the circulation anomaly over the North Pacific. Parallel experiments for the 1972/73 and 1987/88 El Niño events yield similar conclusions. These findings clearly demonstrate the advantages of studying various aspects of large-scale air-sea interaction using higher-resolution models.

Diagnosis of other model variables reveals that the stronger extratropical response in the R30L14 experiment is linked to larger precipitation anomalies over the tropical Pacific, and to more energetic synoptic-scale disturbances along the midlatitude storm tracks. Both the enhanced condensational forcing in the tropics and transient eddy forcing in the extratropics are conducive to a higher-amplitude seasonally averaged response in the North Pacific/North American sector.

6.3.2 Multi-Decadal R30L14 Experiments with Interannual Variations in SST Forcing

In view of the improvements in the fidelity of the atmospheric response as simulated in the R30L14 model (6.3.1), a new series of long-term integrations subjected to different spatial configurations of observed SST forcing has been launched with this model. The month-to-month changes in the SST conditions at the lower boundary of the model have been prescribed using the SST dataset compiled at NCEP for the 1950-95 period. In order to evaluate the model response to oceanic changes directly related to El Niño, an experiment with interannual variations in SST forcing only in the tropical Pacific has been completed and compared to analogous results with the R15L9 model. To isolate the effects on the simulated variability due to SST changes and those due to internal atmospheric dynamics, a century-long integration with the R30L14 model using climatological SST conditions has also been conducted.

Composite analysis of the 46-year experiment with interannual SST anomalies prescribed only in the tropical Pacific confirms the higher sensitivity of the R30L14 model to El Niño forcing as compared to the R15L9 model. The R30L14 results display discernible differences between the simulated extratropical wavetrains during warm (El Niño) and cold (La Niña) events, in agreement with the observational data. This asymmetry of the model response with respect to the polarity of the tropical forcing is related to differences in the distribution of the precipitation anomalies over the tropical Pacific during opposite phases of El Niño-Southern Oscillation (ENSO) episodes.

6.3.3 Relationships Between SST Variability in Different Tropical Ocean Basins

Additional investigations have been conducted on the temporal relationships between SST changes in the tropical Pacific, and those in other ocean basins using both observational and model data. The temporal lag of SST anomalies in the Indian, South China Sea and Atlantic sectors following ENSO episodes in the Pacific has been linked to the perturbed net heat fluxes in these sectors. Variations in surface heat fluxes are correlated with changes in local cloud cover and the near-surface atmospheric circulation. Some of the latter changes are in turn driven remotely by ENSO processes. A comparative study between these observational findings and the corresponding model results is currently in progress.

6.3.4 Analysis of the Tropical Atmospheric Warming During El Niño

Using a combination of satellite and radiosonde observations, global temperature anomalies have been investigated to understand the lag correlation between tropical atmospheric warming and El Niño. These observations are compared to GCM simulations with an R15L9 model subjected to observed SST forcing for the period 1946-1988. Analysis of both the satellite and radiosonde data indicates that the atmospheric warming does not take place uniformly throughout the tropics, but occurs earliest over the tropical Pacific and later over the tropical Atlantic and Indian Oceans. Comparison of temperature anomalies from a GCM simulation forced with near-global observed SSTs shows that the model is able to accurately reproduce the ENSO-related atmospheric temperature anomalies in all ocean basins. Yet, when the same model is integrated using observed SSTs only in the tropical Pacific and a seasonal climatology everywhere else, the warming of the atmosphere over the tropical Atlantic and Indian Oceans is substantially reduced and the phase lag is shortened. This suggests that the magnitude and timing of the warming of the tropical atmosphere depends critically upon the warming of the remote tropical oceans (such as the Atlantic and Indian), and not solely on the warming of the tropical Pacific. The mechanisms responsible for the warming of the remote tropical oceans (6.3.3) are therefore also important for explaining the changes in tropical atmospheric temperature during ENSO.

PLANS FY98

An extensive suite of integrations will be conducted with the R30L14 model using various combinations of prescribed and predicted SST forcings in selected geographical locations. Multiple runs for the same SST scenario but initiated with different atmospheric conditions will be performed in order to achieve sufficient sampling of the model responses. Particular attention will be devoted to the inter-sample and inter-event variability of the model simulations.

An attempt will be made to diagnose the causes of interannual variability in tropical tropospheric temperatures, and the relationships between such fluctuations and SST variability throughout the tropics. The vertical structure of the tropical temperature variability will be analyzed using both GCM output and radiosonde data at reliable stations.

Further research will be performed to understand the relative importance of various energy fluxes in warming the tropical atmosphere during El Niño. The observations will also be compared with a new set of higher resolution GCM integrations.

6.4 ATMOSPHERIC VARIABILITY

ACTIVITIES FY97

6.4.1 Diagnosis of Low-Frequency Retrograding Waves at High-Latitudes

Westward travelling waves with typical periods of 3-5 weeks have been identified in subpolar regions of both hemispheres. Such features are detectable in observational records for the past several decades, and in a 100,000-day integration of the R15L9 model under perpetual January conditions. This phenomenon is especially pronounced in the sector extending westward from northern Canada to northern Siberia. The structural characteristics and vorticity dynamics of the retrograding waves in this region have been documented in detail by performing composite analyses of various fields and parameters for outstanding episodes. Considerable agreement has been found between the observational and model-simulation results.

The typical life cycle and spatial development of the flow field associated with these westward migrating fluctuations, as inferred from composite analyses, are reminiscent of those accompanying blocking phenomena over the North Pacific and western North America. As its peak, the composite amplitude of the simulated 500 mb height anomaly (as obtained by averaging nearly 100 individual episodes) exceeds 200 m. Such events contribute significantly to the local atmospheric variability on weekly and monthly time scales.

The time-space behavior of various terms in the vorticity budget indicates that relative vorticity advection, which leads to eastward propagation, is uniformly weaker than the sum of the planetary vorticity advection (i.e., beta-effect) and divergence terms, both of which contribute to westward tendencies. The preferred occurrence of the westward travelling waves in the Alaskan/Eastern Siberian sector may be attributed to the climatological minimum in eastward zonal flow intensity in that region, which reduces the effect of relative vorticity advection (and hence eastward tendency) in the vorticity balance.

6.4.2 The "Step-Function Like" Nature of the Extratropical Seasonal Cycle

A project has been initiated to examine the tendency for the evolution of seasonal changes in the atmosphere to occur in a discontinuous fashion. Initial efforts are focused on the evolution of surface temperatures over the conterminous United States. In developing and testing analysis methodologies, a nearby station (New Brunswick, NJ) having a long available record (104 years) has been chosen. It has been found that abrupt changes occur most frequently during the spring and fall transition seasons when the climatological mean is changing most rapidly. Composites of temperature keyed to a common transition date suggest that temperature fluctuates about a nearly constant level for 2-3 weeks prior to and after the transition. On average about three transitions occur in the spring and three during the fall. The frequency of such transitions is much greater than expected from a first order Markov process. These transitions, while ubiquitous throughout the long historical record, display considerable interannual variability in their "sharpness" so that during some years these "steps" "are quite apparent while in other years the seasonal changes are more trend-like.

PLANS FY98

Further diagnosis will be performed on the dynamical processes accompanying the westward travelling phenomena in high latitudes. Particular emphasis will be placed on the kinematics of the flow field and their implications for the divergence term of the vorticity equation. This investigation will be extended to the data products from the NCEP/NCAR Reanalysis Project, and to the output from a 100-year control integration of the R30L14 model subjected to climatological SST forcing.

Analysis of the "step-function like" nature of the extratropical seasonal cycle will be expanded to the domain of the contiguous United States or greater in an effort to identify large-scale patterns of natural discontinuous change. Additional variables in the free atmosphere will be examined in conjunction with the surface temperature data. If results appear promising, similar phenomena will be sought in simulations from GCMs.

6.5 GFDL/UNIVERSITIES COLLABORATIVE PROJECT FOR MODEL DIAGNOSIS

ACTIVITIES FY97

Continuing efforts have been made to support the GFDL/Universities Collaborative Project for diagnosing atmospheric variability using GCM-generated data. Participants in this project include scientists from GFDL and other government and university institutions. The principal goal of this endeavor is to encourage investigators outside of GFDL to pursue various problems of interest by making use of model output produced at GFDL. Project activities at GFDL during FY97 were mostly concerned with the design and implementation of a new suite of higher-resolution experiments designed to improve our understanding of various aspects of air-sea interaction (6.3.1, 6.3.2), and the diagnosis of westward propagating phenomena at high latitudes (6.4.1).

In collaboration with M.A. Alexander at the Climate Diagnostic Center/NOAA, an effort has also been initiated to couple the R30L14 atmospheric model with a more comprehensive mixed layer model of the surface ocean. This mixed layer model has a variable thickness, and is well-suited for simulation of the seasonal and interannual fluctuations of the thermocline depth. Experimentation with this coupled system should provide fresh insight into the nature of midlatitude air-sea interaction, as well as the relationship between tropical and extratropical fluctuations of the climate system on interannual and interdecadal time scales, and teleconnections between SST changes in different ocean basins.

A workshop was organized at the University of Washington, and attended by most of the principal investigators of this project. The activities of the project were reviewed by an external panel during the workshop. A call for new research proposals has been issued through the Office of Global Programs/NOAA, which has provided financial support for this project since its inception.

A review of the interactions between global SST anomalies and the midlatitude atmospheric circulation, with model results drawn mostly from various studies under the auspices of this collaborative project, has been completed (1437).

PLANS FY98

The selection process for a new group of university investigators for this collaborative project will be completed. Research strategies for the coming years will be formulated after consultation with various participants.

Upon completion of the new oceanic mixed layer model, experiments will be performed to study various aspects of midlatitude and tropical ocean-atmosphere coupling with and without ENSO forcings originating from the near-equatorial Pacific Ocean.


 Table of Contents

*Portions of this document contain material that has not yet been formally published and may not be quoted or referenced without explicit permission of the author(s).