Neville R. Smith

The Tropical Ocean‐Global Atmosphere observing system: A decade of progress

A major accomplishment of the recently completed Tropical Ocean-Global Atmosphere (TOGA) Program was the development of an ocean observing system to support seasonal-to-interannual climate studies. This paper reviews the scientific motivations for the development of that observing system, the technological advances that made it possible, and the scientific advances that resulted from the availability of a significantly expanded observational database. A primary phenomenological focus of TOGA was interannual variability of the coupled ocean-atmosphere system associated with El Nino and the Southern Oscillation (ENSO).Prior to the start of TOGA, our understanding of the physical processes responsible for the ENSO cycle was limited, our ability to monitor variability in the tropical oceans was primitive, and the capability to predict ENSO was nonexistent. TOGA therefore initiated and/or supported efforts to provide real-time measurements of the following key oceanographic variables: surface winds, sea surface temperature, subsurface temperature, sea level and ocean velocity. Specific in situ observational programs developed to provide these data sets included the Tropical Atmosphere-Ocean (TAO) array of moored buoys in the Pacific, a surface drifting buoy program, an island and coastal tide gauge network, and a volunteer observing ship network of expendable bathythermograph measurements. Complementing these in situ efforts were satellite missions which provided near-global coverage of surface winds, sea surface temperature, and sea level. These new TOGA data sets led to fundamental progress in our understanding of the physical processes responsible for ENSO and to the development of coupled ocean-atmosphere models for ENSO prediction.

A univariate statistical interpolation scheme for subsurface thermal analyses in the tropical oceans

Abstract A scheme for real-time objective analysis and quality control of ocean temperature data using statistical interpolation is presented. The data base comprises BATHY and TESAC message summaries of XBT profiles collected in real-time under the ship-of-opportunity program, as well as directly recorded data which are collected and archived some time after the observation time. The subsurface temperature observations, spread irregularly in time and space and of mixed quality, are first verified against climatology and other observations before interpolation onto a regular grid. The analysis procedure is based on univariate statistical interpolation techniques which are in standard use throughout the meteorological community. Attention is focused on the elimination of bad or inconsistent measurements and unwanted redundancy. A non-trivial observation error autocorrelation function is incorporated to take account of spatial and temporal coherency of geophysical noise along XBT tracks. A novel technique for combining partial analyses from over-lapping sectors is also presented. The efficacy of the scheme is demonstrated through a series of examples based on tropical Pacific Ocean observations for the last several years. The appropriateness of the statistical models and first-guess field is evaluated and the practical implementation of super-observation formation and data checking is illustrated. An assessment of the real-time Pacific Ocean analysis system for two years, through statistics and direct map evaluation, indicates the scheme is performing well. Information is not communicated between analysis periods, a role which we expect statistical and dynamical models to assume in future.

Space and time scales for optimal interpolation of temperature - Tropical Pacific Ocean

Abstract Autocorrelation functions of sea surface temperature (SST) and depth of the 20°C isotherm (D 20 ) are estimated at 95 locations along the tropical Pacific expendable bathythermograph (XBT) ship-of-opportunity tracks, and used to determine statistical parameters required for optimal interpolation. The parameters are variances for signal and noise, and spatial/temporal decorrelation scales. Estimates were made for two periods of time: June 1979 to May 1982 proceding the 1982/83 El Nino, and June 1979 to May 1983 including it. Parameters for the first period indicate smaller scales and weaker signals than the ones for the second. The difference results from El Nino, whose large temperature signal dominates the statistical structure. Understanding the dynamics of the smaller interannual signals and in particular the precursors to El Nino is as important as understanding El Nino. The smaller scales are therefore recommended for optimal interpolation and design of the XBT network. Three summaries of the scales are presented: 1) averages in three degree latitude bands; 2) values in dynamic regions representing the major zonal currents; and 3) median values for 95 estimates in the tropical Pacific 18°N to 18°S. The median values are recommended for design of the tropical XBT network. The median scales for the depth of the 20°C isotherm are: 3° latitude, 15° longitude and 2 months; and the median signal-to-noise (amplitude) ratio is 0.75. Scales for SST are greater than or equal to scales for D 20 , thus subsurface temperature is the limiting factor in designing the network. Analysis of mapping errors suggests that optimal sampling requires two to three samples per decorrelation scale. Two modes of XBT sampling are recommended for TOGA. Broadscale sampling for horizontal mapping of the temperature field is recommended at a density of two samples per scale, which requires an XBT station every 1.5° latitude and 7.5° longitude, monthly. The broadscale mode should be implemented in as large an area as possible with available shipping. Sampling on repeated transequatorial sections for time series studies is recommended at a density of three samples per scale, which requires 18 sections per year, with stations every 1° latitude. The time series mode is useful for more accurate studies of thermal structure and currents, and should be implemented on a few routes in each ocean which transect major thermal features.

Assimilation of subsurface thermal data into a simple ocean model for the initialization of an intermediate tropical coupled ocean-atmosphere forecast model

Abstract An adjoint variational assimilation technique is used to assimilate observations of both the oceanic state and wind stress data into an intermediate coupled ENSO prediction model. This method of initialization is contrasted with the more usual method, which uses only wind stress data to establish the initial state of the ocean. It is shown that ocean temperature data has a positive impact on the prediction skill in such models. On the basis of hindcasts for the period 1982–91, it is shown that NIN03 SST anomaly correlations greater than 0.7 can be obtained for hindcasts of duration up to 13 months and greater than 0.6 up to 16 months. There are also clear indications of skill at two years.

The BMRC Coupled General Circulation Model ENSO Forecast System

An El Nino-Southern Oscillation (ENSO) prediction system with a coupled general circulation model and an ocean data assimilation scheme has been developed at the Australian Bureau of Meteorology Research Centre (BMRC). The coupled model consists of an R21L9 version of the BMRC climate model and a global version of the Geophysical Fluid Dynamics Laboratory modular ocean general circulation model with resolution focused in the tropical region and 25 vertical levels. A univariate statistical interpolation method, with 10-day data ingestion windows, is used to assimilate ocean temperature data and initialize the coupled model. The coupling procedure does not use any flux corrections. Hindcasts have been carried out for the period 1981-95 for each season (60 in all), for up to a lead time of 12 months. This paper will describe these initial experiments and show that the skill of sea surface temperature (SST) hindcasts in the tropical Pacific is comparable to other published coupled models. The skill of the model is strongest in the central Pacific. SST skill tends to be lower during the earlier 1990s than during 1980s in the eastern Pacific but not in the central Pacific. Since the ENSO SST anomaly in the central Pacific is the most important forcing of regional and global climate anomalies, the high SST prediction skill and its insensitivity over the hindcast period in this region in this model give grounds for optimism in the use of coupled general circulation models.

A recent change in the mean state of the Pacific basin climate: Observational evidence and atmospheric and oceanic responses

The first half of the present decade has been characterized by anomalous conditions in a number of Pacific basin atmospheric and oceanic variables. The sea surface temperature (SST), in particular, has been warmer than normal over vast areas of the subtropics and the Gulf of Alaska. The SST anomaly pattern is shown to be part of a long-term trend which began in the early 1970s. This trend is the dominant source of SST variability over significant portions of the basin. Experiments with atmospheric models demonstrate that the atmospheric anomalies of the 1990s are consistent with this change in SST and the mechanisms for this are analyzed. Finally, experiments with an ocean general circulation model are used to investigate which features of the atmospheric anomalies are responsible for the observed SST anomalies. It is determined that wind stress changes and changes in heat flux due to wind speed reductions are the most likely causes. Two potentially important mechanisms for positive ocean-atmosphere feedback leading to the climate anomalies of the 1990s have thus been identified.

An evaluation of expendable bathythermograph and Tropical Atmosphere-Ocean Array data for monitoring tropical ocean variability

Two important sources of information for monitoring and predicting the El Nino-Southern Oscillation are the Tropical Atmosphere Ocean (TAO) buoy array and the volunteer observing ship expendable bathythermograph (VOS XBT) network. The subsurface ocean temperature measurements of these two networks are evaluated in the context of analysis and monitoring of medium- to low-frequency variability in the tropical Pacific Ocean. It is shown that both systems capture the principal features that characterize variability in the equatorial central and eastern Pacific Ocean but only TAO can capture the higher temporal fluctuations in detail. In the western Pacific, the VOS XBT network is able to capture the slowly moving, westward propagating Rossby waves and their reflection and interaction with the western boundary. TAO is limited in this respect because of its equatorial focus. The accuracy of temperature maps is estimated using the rms analysis error variance. The accuracy of the VOS XBT network fluctuates in space and time but is relatively steady averaged over the tropical region, while that of TAO has gradually improved as the basin wide array has been implemented. Case studies using VOS XBT data only, TAO data only, and the combined full data set indicate the networks are complementary and supportive, evaluated basin wide and over the tropical region. The high temporal resolution and regularity of the equatorially focused TAO network are complemented by the broad-scale, irregular tropical VOS XBT sampling in the equatorial region and by VOS coverage outside the domain of TAO. A proxy for net information content is derived from the estimated rms error variance of the analysis. For the equatorial region, TAO now provides in excess of 70% of the information, its dominance beginning in 1992. For the tropical region (20°S–20°N) the net information content of the respective systems is of comparable magnitude, each equivalent to in excess of 100 independent samples per 10-day period in 1994, reinforcing the notion that both systems are valuable sources of subsurface ocean information for monitoring medium- to low-frequency tropical variability.

Stability of North Atlantic Deep Water Formation in a Global Ocean General Circulation Model

Abstract A global ocean general circulation model is forced using mixes boundary conditions (i.e., a restoring condition on the upper-level temperature but using a fixed, specified surface salt flux). Freshwater flux anomalies lasting 5 years are then applied over the western half of the subpolar gyre in the northern North Atlantic. The current climate is found to be stable to anomalies that have salt deficits equivalent to about seven times that estimated for the “great salinity anomaly” of 1968–1982, although this value is a function of the duration over which the anomaly is imposed. Above this level the thermohaline circulation collapses to a state in which the zonally averaged overturning associated with North Atlantic Deep Water formation is only about half its original value, the sea surface temperatures over the North Atlantic are lowered, and both the subpolar and subtropical gyres have weakened horizontal transports. Various atmospheric feedbacks on the momentum and salt flux are then applied und...

Ocean modeling in a global ocean observing system

The oceanographic community is currently contemplating the design of a global ocean climate observing system to help monitor, describe, and understand the seasonal to decadal climate changes of the ocean and to provide the observations needed for climate prediction. This review attempts to define a role for modeling within that system, the central theme being that the observational and modeling elements must be developed in concert, with the presence of one enhancing the value of the other. Three distinct categories of model-to-data interface are identified. In the first class, models and data collection develop separately, being joined only by intermittent validation steps. In the second, and by far most important, class the model and data collection evolve together, either in a time-space data assimilation and prediction system, or through the application of inverse methods. In the final category, model information feeds back to the observing system design, and vice versa, and the model assimilation system provides quality control on the data. The key role of (atmospheric) models in the determination of surface fluxes to drive ocean models is discussed. A nontrivial role is proposed for ocean models whereby they provide additional, and largely independent, constraints on atmospheric forecast system estimates. The role of ocean models in the analysis of surface and upper ocean fields needs to be developed, particularly with respect to salinity and nonphysical fields. The use of models in rationalizing the choice of observation platforms is discussed, together with some of the difficulties in interpreting such studies. The state of tropical ocean prediction is reviewed with particular emphasis on systems that assimilate subsurface temperature data. A range of thermocline models are also reviewed with the emphasis on subduction and the problem of initializing and constraining models that resolve mesoscale eddies. Some of the issues involved in matching the models to particular observational methods, and vice versa, are discussed with respect to tropical ocean and thermocline modeling. The current state of global ocean and coupled climate general circulation models and models for studying tracer circulation and coupled physical-biological systems is evaluated. This prognostic path and the products and knowledge derived from integrations are contrasted with the inverse modeling approach which attempts to infer ocean circulation through a combination of observational and modeling constraints. Again we speculate on the model-data interface and on the different measurement strategies and data requirements. The concept of community modeling and the need for substantial resources and international organization are discussed. A case for global ocean observing system centers with colocated modeling and data collection is made, but with model diversity and individuality emphasized.

An Ocean Observing System for Climate

The conceptual design of an ocean observing system for the routine, long-term gathering and processing of ocean data useful for monitoring, describing and predicting ocean climate and its variability is discussed. The ultimate aim of the system is represented by four application areas; atmospheric prediction; ocean and coupled ocean-atmosphere climate prediction; state-of-the-art ocean climate assessment; and model validation. Models are presented as the unifying glue for the system, providing a means for exploiting observed information in many different ways as well as a means for processing complicated and diverse data sets into a form which has practical applications. Monitoring, description and prediction require different supporting environments in order to exploit this potential. The overall objective of the system is broken down into a series of goals and sub-goals roughly aligned with surface, upper ocean and deep ocean applications and with modelling and information management requirements. A prioritization of these goals is presented and it is shown that ordered implementation of the elements supporting these goals will lead to a sensible, staged implementation of the observing system. The research, development and exploitation of appropriate technology is emphasised. Trade-offs and rationalisation across the elements of the ocean observing system for climate and between other climate components and ocean modules is central to the development and successful implementation. The design is presented as the first stage in a constantly evolving and maturing ocean observing system for climate applications.