Combining Argo and remote-sensing data to estimate the ocean three-dimensional temperature fields—a first approach based on simulated observations
Introduction
To achieve a better understanding of the ocean circulation and its role in climate, we need to observe and model it at high spatial and temporal resolution (e.g. Wunsch, 2001). Such a high-resolution description is also crucial for most of the operational oceanography applications that are, for example, envisioned by the Global Ocean Data Assimilation Experiment Global Ocean Data Assimilation Experiment, 2001, Le Traon et al., 2001. Smith and Koblinsky (2001) provide a comprehensive description of the likely adequate global ocean observing system. It relies both on in situ data (Argo, XBT, TAO/TRITON/PIRATA, etc.) and remote-sensing data (satellite altimetry, sea surface temperature, etc.). A central assumption is that combining the high-resolution remote-sensing data with the sparse in situ data should allow us to describe the ocean variability at a broad range of space and time scales. The Argo global array of temperature and salinity profiling floats (Argo Science Team, 2001) has been designed with this objective in mind. However, the best use of Argo data will be when they are combined with other data sets, in particular, satellite altimetry (Jason-1), and models through effective data assimilation techniques (GODAE Strategic Plan, 2001). Given the complexity of the data assimilation problem (in particular, at eddy-resolving scales), it is useful and necessary to analyze first the contribution of the data themselves using simple techniques such as optimal interpolation (e.g. Festa and Molinari, 1992, Smith and Meyers, 1996). This is the approach adopted in this study. Our general framework is the evaluation of the present global ocean observing system (in situ and remote sensing) in representing the ocean 3D thermohaline variability.
In a recent study, Guinehut et al. (2002) have analyzed the contribution of Argo data alone from simulated measurements as provided from a North Atlantic numerical model. Their study suggested that a 3° array of profiling floats cycling every 10 days (Argo “nominal” resolution) can retrieve most of the variance of the large-scale and low-frequency temperature and salinity signals as observed by a 1/3° primitive equation model. In this paper, the Guinehut et al. (2002) study is developed further to analyze the contribution of the combination of sparse in situ temperature profile data as given by Argo with high-resolution sea level and sea surface temperature satellite data. The study is focused on the reconstruction of the large-scale, monthly mean, 200-m depth temperature fields but could be applied to any depth and also to the salinity field. We will use outputs and profiling float simulations from the 1/6°-resolution CLIPPER primitive equation model of the North Atlantic Ocean. From an “Argo configuration” (i.e., a 3° array of profiling floats cycling every 10 days) of simulated temperature profiles, the large-scale and low-frequency variability of the temperature field at 200-m depth will be first reconstructed using an optimal interpolation method as in Guinehut et al. (2002). The contribution of the merging of simulated sea level, sea surface temperature and temperature profiles data to the reconstruction of instantaneous but also large-scale and low-frequency temperature fields at 200-m will be then quantified.
The paper is organized as follows. Data are presented in Section 2. Statistical results for the “Argo” profiling float array are presented in Section 3. Section 4 shows the contribution of the merging to the reconstruction of instantaneous fields and large-scale and low-frequency fields. Main conclusions and perspectives are given in Section 5.
Section snippets
Data
We use model outputs and profiling float simulations from the French CLIPPER project (Treguier et al., 1999). The model is a primitive equation model of the North Atlantic Ocean, based on the OPA 8.1 code (Madec et al., 1998), with a 1/6° horizontal resolution and 43 vertical levels. It has a realistic mesoscale variability energy distribution (Treguier et al., 2003). The model is initialized using the Atlantic Ocean climatology especially developed by Reynaud et al. (1998) for the CLIPPER
Contribution of the Argo profiling float array
Our first objective is to reconstruct the large-scale and low-frequency variability of the temperature (T) fields at 200-m depth from simulated T snapshots on a 3° array. This is similar to the Guinehut et al. (2002) study, but here we used a higher-resolution model (1/6° versus 1/3°) and a longer duration (4 years versus 1). The main issue is to analyze how an a priori defined large-scale and low-frequency signal can be mapped from sparse measurements with a low signal-to-noise ratio (mainly
Combining altimeter, sea surface temperature and profiling float data
We now analyze the contribution of high-resolution satellite data [sea level and sea surface temperature (SST)] to the reconstruction of the large-scale, monthly mean, 200-m depth temperature fields when combined to the sparse in situ temperature data as given by Argo. The main issue is to reconstruct 200-m T snapshots at high temporal and spatial resolution and thus improve the representation of the large-scale and low-frequency T fields at the given depth.
Conclusions and perspectives
Our study suggests first that a 3° array of profiling floats cycling every 10 days can retrieve a large proportion of the large-scale and low-frequency variability of the 200-m T fields. No significant degradation of results over time is noted, which is very encouraging for the Argo project. This confirms the preliminary results presented in Guinehut et al. (2002).
The main contribution of this study is, however, to show that the accurate but sparse Argo in situ data can be effectively merged
Acknowledgements
We would like to thank the CLIPPER team, particularly Jean-Marc Molines and Bernard Barnier, for providing us with the model and float simulations and an anonymous reviewer for thoughtful comments on the first version of the manuscript. The study was partly funded by CNES (contract CNES/DSO310/02/0655/00) and the European Commission as part of the Gyroscope project (EVK2_CT_2000_00087).
References (21)
- et al.
A technique for objective analysis and design of oceanographic experiments applied to MODE-73
Deep-Sea Res.
(1976) - et al.
Design of an array of profiling floats in the North Atlantic from model simulations
J. Mar. Syst.
(2002) - et al.
Agulhas eddy fluxes in a 1/6° Atlantic model
Deep-Sea Res.
(2003) - et al.
Argo: the global array of profiling floats
- et al.
Dynamics of eddy motions in the eastern North Atlantic
J. Phys. Oceanogr.
(1985) - et al.
Reducing orbit error with an inverse method to estimate the oceanic variability from satellite altimetry
J. Atmos. Ocean. Technol.
(1995) - et al.
Kinematics of the pacific equatorial undercurrent: an Eulerian and Lagrangian approach from GCM results
J. Phys. Oceanogr.
(1997) - et al.
An evaluation of the WOCE volunteer observing ship-XBT network in the Atlantic
J. Atmos. Ocean. Technol.
(1992) - Global Ocean Data Assimilation Experiment, 2001: Strategic Plan. GODAE Report No 6., Published by the GODAE...
- et al.
The midlatitude resolution capability of sea level fields constructed from single and multiple satellite altimeter datasets
J. Atmos. Ocean. Technol.
(1997)
Cited by (109)
Feature-oriented reconstruction of vertical temperature profile: A feasibility study in the Northwest Pacific Ocean
2024, Deep-Sea Research Part I: Oceanographic Research PapersImpact of ocean fronts on the reconstruction of vertical temperature profiles from sea surface measurements
2022, Deep-Sea Research Part I: Oceanographic Research PapersMechanisms of interannual variability of deep convection in the Greenland sea
2021, Deep-Sea Research Part I: Oceanographic Research PapersCitation Excerpt :A relatively robust estimate of oceanic advective fluxes is possible since 1993, as well. In ARMOR3D, “synthetic” temperature and salinity at different water levels are first derived from altimeter and sea surface temperature (SST) anomalies through a multiple regression (Guinehut et al., 2004). The regression coefficients are regionally obtained by combining the satellite and all available sub-satellite historical data.
An analytical model of open-ocean deep convection with multiple steady states
2020, Ocean ModellingMicronekton distribution in the southwest Pacific (New Caledonia) inferred from shipboard-ADCP backscatter data
2020, Deep-Sea Research Part I: Oceanographic Research Papers