Search

Tropical cyclones are some of the most devastating natural hazards and the “three beasts”—Harvey, Irma, and Maria—during the Atlantic hurricane season 2017 are recent examples. The European Centre for Medium-Range Weather Forecasts (ECMWF) is working on fulfilling its 2016–25 strategy in which early warnings for extreme events will be made possible by a high-resolution Earth system ensemble forecasting system. Several verification reports acknowledge deterministic and probabilistic tropical cyclone tracks from ECMWF as world leading. However, producing reliable intensity forecasts is still a difficult task for the ECMWF global forecasting model, especially regarding maximum wind speed. This article will put the ECMWF strategy into a tropical cyclone perspective and highlight some key research activities, using Harvey, Irma, and Maria as examples. We describe the observation usage around tropical cyclones in data assimilation and give examples of their impact. From a model perspective, we show the impact of running at 5-km resolution and also the impact of applying ocean coupling. Finally, we discuss the future challenges to tackle the errors in intensity forecasts for tropical cyclones.

Minimum miscibility pressure (MMP) is an important parameter when designing a miscible gas flood. Traditionally, the MMP is evaluated using slimtube tests, which are time-consuming and expensive. Since it is typically not feasible to test more than a few injection gases on a subset of relevant reservoir fluids, reliable methods are required to estimate the MMP. The most popular correlations are often based on data from low-temperature reservoirs and are not always very reliable.In this work, we make use of the new, comprehensive ADNOC PVT database, which contains more than 100 slimtube MMP measurements for a variety of injection gases, including sour gases. We then complement this data source by performing equation of state (EOS) based miscibility calculations covering a large temperature range with appropriately tuned EOS models. The combination of measured and simulated data is then used as input for development of a general correlation.The new correlation is a modification of the Eakin-Mitch formulation. It has been tuned to a large variety of injection gases and reservoir fluid compositions, and covers a wider temperature range. The average deviation is 5% and the maximum error is less than 20%. Results show that the MMP exhibits a maximum versus temperature, a feature also noticed previously during development of CO2 miscibility (Yuan et al., 2005; Alshuaibi et al., 2019).One of the novelties of this work is the identification of a more complicated temperature-dependency of the MMP. Furthermore, the new correlation considers not just lean gases but also rich gases and sour gases, which develop miscibility based on the combined condensing-vaporizing mechanism. We believe that this model is more robust because it covers a much larger parameter range compared to existing correlations.

Long horizontal or highly deviated wells require lower completion liners to improve well and reservoir management of the carbonate reservoirs in ADNOC's portfolio. Optimum well productivity depends on effective stimulation post drilling. Traditionally, a complicating factor in a matrix-acid stimulation treatment has been poor control of the acid placement along the reservoir section to total depth (TD) without stage equipment or chemical diverters. The limited-entry liner (LEL) lower completion addresses the shortcomings of other stimulation techniques. It consists of a number of unevenly spaced holes with the purpose to distribute acid evenly along the reservoir section to be stimulated. ADNOC is planning to execute a series of pilots to test and possibly expand the operational envelope of the LEL technique.Design of a the hole-size distribution requires a mathematical model, which accounts for the various pressure drops encountered in the system as well as the impact of the chemical reaction between acid and rock. In this work, we present a comprehensive algorithm for designing a hole-size distribution which yields a user-specified acid coverage and also provides the best possible wormhole penetration, subject to a number of constraints related to surface pressure, pump rate, and acid volume available. We make use of findings published over the past 15 years in the SPE literature to arrive at an enhanced design philosophy, which is appropriate for the carbonate reservoirs in ADNOC's portfolio.The design model takes into account a number of important parameters, including fluid and reservoir properties, as well as completion information such as tubing ID, liner ID, and the LEL hole sizes. A check is made using nodal analysis software to ensure that the LEL completion does not act as a flow restriction during back-production.

Miscible gas injection is a proven enhanced oil recovery method for medium-light oils. Miscibility is typically assessed with the slimtube experiment, which aims at identifying the minimum pressure (MMP) above which the displacement process is multi-contact miscible and leads to low microscopic residual oil saturation. During implementation of a miscible gas injection scheme, there will often be areas close to producing wells where the reservoir pressure is below the MMP and the question then arises whether the process can still be considered miscible. Another complicating factor occurs in reservoirs with significant lateral fluid property variation, where it is not intuitively clear how the MMP varies aerially.We employ compositional reservoir simulation to investigate the impact of flowing bottom-hole pressures below MMP on the development and propagation of a miscible front. Starting with 1D simulations, we track a number of important properties such as the effluent gas composition over time, the component K-values at distinct positions along the displacement front, as well as the pressure at the front over time.Due to the high gas mobility, the pressure drop per unit distance is greatly reduced in the areas flooded by gas. This means that the reservoir pressure at the gas-oil front is higher than the pressure at the same position before arrival of the front. Therefore, a miscible displacement can be maintained even when not only the flowing bottom-hole pressure but also the average reservoir pressure is below MMP. A key requirement is, of course, that the gas is injected at a pressure above MMP. We evaluate various parameters in terms of their ability to distinguish between miscible and immiscible displacement and find that the phase density difference versus pressure provides a unique advantage for estimating the MMP and we believe it can be used not only in 1D slimtube simulations but also in other thermodynamics-based algorithms. We also find that the vapor mole fraction, inferred from GOR measurements, is a better metric than GOR itself and that molar ratios applied to the effluent gas are very useful for tracking miscibility.

ADNOC has embarked on the second phase of its ambitious integrated capacity model (ICM) project with the overall aim to optimise its fluid production portfolio from the well level to the processing facilities. A key feature of the new software tool is the ability to track and predict fluid properties over time across the entire production network, comprising thousands of wells and a myriad of pipelines.The reservoir fluid composition is assigned at well level for each producing reservoir. The compositional tracking over time is straightforward for many wells, but complicating factors do arise, such asLateral compositional variation related to complex reservoir charging history Vertical compositional gradients, especially for near-critical fluids The presence of initial and secondary gas caps, resulting in gas coning Injection of miscible gas for enhanced oil recoveryThe fluid systems range from medium-API oil to gas condensates and the key chemical components vary as follows: C1 [5-80%], CO2 [0.5-8%], and H2S [0-35%].Mixing of pressurized fluids with different compositions at various junctions in the network requires a robust thermodynamics model to capture the associated variation in fluid properties, particularly density and viscosity as a function of pressure and temperature. We demonstrate that it is possible to constrain one unified equation of state applicable to all fluids, as long as the fluid systems used for the tuning span the entire range of compositions observed. Mixing of fluid streams is computationally much simpler if each stream is made up of the same components (although in different amounts) with the same component properties. On average, the predicted fluid density is within 1% of the measured value from a multi-stage separator test.

High GOR is a constraint for many production wells. Foam has been proved able to reduce GOR in several field production well treatments. The primary goal of this study is to qualify foams for use at reservoir conditions (280 bars, 103 °C). Foam will block gas and thereby reduce Gas-Oil Ratio (GOR) and improve field oil production. Bulk foam experiments and dynamic foam core flooding experiments have been performed to qualify foam for gas shut off at reservoir condition. Foam has been studied in static foam tests using different surfactants, and in limestone core flood experiments at reduced conditions 30 bars and 50 °C. Further foam generation and gas blocking test at reservoir conditions have been conducted both in outcrop and in reservoir material. Dynamic core flooding experiments at 30 barg and 50 °C have shown generation of moderate to strong foams in carbonate limestone core material both with Nitrogen and Methane as the gas phase. Strong foam with mobility reduction factors around 65-70 have been generated in reservoir core material at reservoir conditions. The presence of oil in the carbonate material is a challenge for in-situ foam generation. One solution to the problem has been identified by use of a pre-flush with FS-500 (fluorinated surfactant), prior to alpha olefin sulfonate (AOS). CO2-foamers such as Duomeen has been tested, but gave no foam with Nitrogen as gas phase. Higher concentration of AOS surfactant was able to generate in-situ foam without assistance of fluorinated surfactant. The generated foam is stable also with lower surfactant concentration after first generation. The mobility reduction factor is constant for a range of surfactant concentrations. Strong foam has been generated in carbonate reservoir core material at reservoir conditions. The presence of oil in carbonate material prevents in-situ foam generation at low surfactant concentration, but gives strong foam at higher surfactant concentration.

This paper describes preparations and planning for a campaign of foam gas shut-off pilot operations in a large carbonate reservoir located offshore Abu Dhabi containing an oil column in equilibrium with a large gas cap. Throughout the field history and due to the heterogeneity (permeability ranges from 5 mD to 1 D), the major challenge to produce the oil rim independently from the gas cap was how to control premature gas breakthrough in the oil producers. Mechanical interventions in high gas-oil ratio wells are particularly complicated due to the risk of losing oil potential and are generally unsuccessful.Injection of foam for gas shut-off (FGSO) is a near-wellbore treatment, which has been trialed elsewhere in the industry with some success. Foam can act as an auto-selective agent to shut-off confined gas inflow through a gravity-controlled source like coning or cusping, while oil breaks the foam, resulting in preferential oil flow and reduction in gas-oil ratio. In addition, this type of operation has been identified as an EOR enabler, because it can help prepare for the technical and logistical challenges of using EOR chemicals in the field, generate data useful for the modeling of surfactant and polymer under reservoir conditions, and mitigate early gas breakthrough in the case of gas-based EOR developments.For the reservoir in question, a key complicating factor was to identify a surfactant, which could generate strong foam in-situ (mobility reduction factor of 50) at harsh reservoir conditions (temperature of 220-230 °F and water salinity above 200,000 ppm, including 20,000 ppm divalents), with an acceptable level of adsorption. The candidate selection process took into consideration overall behavior of the reservoir as well as performance of the individual high-GOR wells. Target well selection criteria included homogeneity of permeability, an understanding of gas sources and their movement, and observation of a rate- or draw-down-dependent GOR.The experimental lab program involved testing several surfactant formulations in bulk as well as in corefloods with and without the presence of reservoir oil to evaluate foaming ability and level of gas flow reduction. One formulation showed the right level of in-situ mobility reduction, in addition to stability and moderate adsorption at the prevailing reservoir conditions, and was therefore selected for a pilot test involving four wells.

Minimum miscibility pressure (MMP) is an important parameter in the design of a successful miscible gas injection project. It is common industry-practice to infer MMP from a series of carefully designed slimtube tests, but there are cases where a quick and rough estimate of MMP is required to screen for potential application of miscible gas injection.Mathematical correlations are attractive because they require few input parameters, little information about the fluids, and are quick and easy to use. Whereas MMP using hydrocarbon gas or nitrogen can be predicted with relative confidence using standard correlations, the same level of accuracy is not found for correlations predicting CO2 MMP, in part because of a more complicated temperature-dependent mass transfer.This study is part of an ongoing effort to evaluate miscibility conditions using a variety of injection gases across the ADNOC reservoir fluid portfolio. We have tested the most common existing CO2 MMP correlations against measured MMP data from slimtube tests for a large number of Abu Dhabi reservoir fluids. The results show that none of the existing correlations can accurately predict MMP for these crude oils and we believe that this is attributed to the fact that most available correlations were established based on reservoir temperatures much lower than the typical temperature conditions of 250 °F found in many Abu Dhabi reservoirs.We therefore propose a new correlation for estimating CO2 MMP with particular emphasis on high-temperature fluid systems. The new formulation has been developed based on multiple-parameter regression of measured data as a function of temperature, saturation pressure and reservoir fluids composition. In addition to the existing data sets available in the literature, we performed 17 additional slimtube experiments on different Abu Dhabi reservoir fluids to enlarge the parameter envelope and increase robustness of the model. The proposed new correlation predicts CO2 MMP within 6.4%. To constrain the correlation, we performed a series of EOS-based MMP calculations for several fluids in the temperature range of 100 to 350 °F. These calculations confirm that MMP attains a maximum value around 320 °F.

Many of the oil reservoirs in ADNOC’s portfolio have been producing for decades and continue to deliver their target rates, thanks to development schemes centred on pressure support from peripheral water injection and/or crestal gas injection, where applicable, or from line-drive patterns.As these reservoirs mature, a number of enhanced oil recovery (EOR) techniques are being evaluated to increase the ultimate recovery factor. Choosing one or several appropriate EOR methods starts with a robust screening methodology, which in this case has to apply across an entire reservoir portfolio, rather than just to a single asset.The key objective of the screening efforts presented in this paper is to estimate the EOR potential from each reservoir in a systematic manner to allocate the right resources, to the right fields, at the right time, with the right technology. Therefore, the screening methodology must take into consideration the following aspects: The focus of the screening is on identifying opportunities, which are or will become technically feasible by the time deployment is requiredThe screening procedure must remain sufficiently high-level to be able to deliver an outcome within a short time-frameA single EOR method is not necessarily applicable across an entire reservoir; hence several EOR methods can potentially be implemented in the same reservoir but in different areasThe thought process (workflow) has to be properly documented so that the screening exercise can be updated whenever new technologies or new relevant field data become availableThe new EOR screening tool considers more EOR methods compared to previous screening efforts: It takes a much broader view on potential source gas options for miscible gas injection, namely enrichment of sweet hydrocarbon gas, sour gas streams, CO2 with or without impurities, as well as nitrogen for some high-temperature reservoirs containing volatile oil. This step involved estimating miscibility conditions for hundreds of combinations of injection gas compositions and reservoir fluids using equation-of-state based MMP calculations in addition to common MMP correlations.The reservoirs are all carbonate formations and are characterized by high temperature and high salinity. These conditions have traditionally been a challenge for chemical EOR methods involving polymers and surfactants. However, recent R&D progress has opened up for opportunities not previously considered. Both polymer and foam agents are currently being piloted, which have the potential to change the EOR landscape in Abu Dhabi and perhaps elsewhere as well.As reservoirs mature and breakthrough of injection fluids begin to occur, the need for improved reservoir conformance becomes evident. The success of any EOR technique relies on the right well placement with the right monitoring in place and the right level of injection profile control.

Gas injection is a proven EOR method in the oil industry with many well-documented successful field applications spanning a period of more than five decades. The injected gas composition varies between projects, but is typically hydrocarbon gas, sometimes enriched with intermediate components to ensure miscibility, or carbon dioxide in regions such as the Permian Basin, where supply is available at an attractive price.Miscible nitrogen injection into oil reservoirs, on the other hand, is a relatively uncommon EOR technique because nitrogen often requires a prohibitively high pressure to reach miscibility. Unlike other injection gases, the minimum miscibility pressure for nitrogen decreases with increasing temperature. In fact, in deep, hot reservoirs containing volatile oil, nitrogen may develop miscibility at a pressure similar to the MMP for hydrocarbon gas or carbon dioxide. The phase behavior is more complicated than what can be captured by correlations and hence requires equation-of-state calculations.Results from a recent EOR screening study in ADNOC indicate that a couple of high-temperature oil reservoirs in Abu Dhabi may be potential targets for miscible nitrogen injection. This paper discusses key aspects of the EOS modeling. Advanced gas injection PVT data are available to enable a fair comparison between nitrogen, carbon dioxide and lean hydrocarbon gas. In this work, we have modelled and analyzed the phase behavior of two volatile oil systems with respect to nitrogen, hydrocarbon gas, and carbon dioxide injection, as part of a reservoir simulation study, which will be covered in a subsequent publication; see Mogensen and Xu (2019). Nitrogen behaves differently from hydrogen carbon gas, despite the fact that the two gases lead to similar minimum miscibility pressures. At the prevailing reservoir pressure, the swelling factor with hydrocarbon gas is four times higher than for nitrogen. Furthermore, the reservoir fluid density increases during swelling with nitrogen, whereas it decreases as a result of hydrocarbon gas swelling. The same trend is observed for viscosity. Injection gas blends with various proportions of nitrogen and carbon injection shows that the MMP is constant when more than 35-40% nitrogen is present in the blend.

A large portion of the remaining oil world-wide is located in tight formations, often of carbonate origin, with substantial heterogeneity. In this review article, we aim to study EOR opportunities from a reservoir characterisation point of view. The main categories chosen include heavily-fractured reservoirs, communicating reservoir layers with large permeability contrast, low-permeability reservoirs and finally reservoirs with large oil-water transition zones. The literature sources, which are critically reviewed, have highlighted suitable methods in each of the four categories. In many cases, field implementation suggests that EOR is indeed economically feasible and a number of key learnings can be extracted from these successful projects. In general, the earlier EOR can be implemented, the better. This reduces the incremental cost required for EOR and provides an earlier pay-back in terms of higher production. The increment due to EOR may become more difficult to quantify since the baseline case would have to be entirely based on numerical simulation. A low baseline recovery without EOR shifts the economics in favour of EOR. Use of the existing well stock is not always sufficient or suitable; in most cases EOR must be complemented by drilling of new wells with appropriate well type, well placement, sufficient monitoring and surveillance. Large fields offer more EOR opportunities because injectant breakthrough can be better managed by redirecting injectant to poorly swept areas. EOR can also be taken offshore; industry experience shows that if the size of the prize is big enough, EOR projects are being sanctioned. Gas injection has the broadest range of applicability for a number of reasons. Occasional lack of market to export produced gas from remotely-located fields provides a strong incentive for gas re-injection. Furthermore, gas offers better injectivity compared to water or chemical EOR methods. Gas injection does not require development of miscibility to be successful, although miscibility helps. Swelling of highly under-saturated fluids leads to oil viscosity reduction and oil saturation increase and delays premature gas breakthrough. However, gas based EOR may suffer from poor sweep efficiency due to geological heterogeneity, gravity override and viscous fingering leading to unfavourable mobility ratio and premature gas breakthrough. Hence, gas-based EOR will require mobility control to improve sweep efficiency. Appetite for risk taking varies among operators. Unique problems require unique solutions. Operators who demand firm proof before field testing may have to adjust their mind-set and pilot a new concept in a statistically significant number of wells to fully evaluate the potential.

Gas injection is a proven EOR method in the oil industry with many well-documented successful field applications spanning a period of more than five decades. Miscible nitrogen injection into oil reservoirs is a relatively uncommon EOR technique because nitrogen often requires a prohibitively high pressure to reach miscibility. In this paper, we investigate the potential for applying miscible nitrogen injection for a high-temperature, low-permeability carbonate reservoir, which contains a volatile oil with some H2S and CO2. The field is still in its early development phase but suffers from relatively low throughput rate because of low permeability. At the prevailing reservoir pressures, extensive phase behavior studies confirm that the reservoir fluids develop miscibility with nitrogen, carbon dioxide, and hydrocarbon gas. Nitrogen behaves differently from hydrocarbon gas, despite the fact that the two gases lead to similar minimum miscibility pressures. At the prevailing reservoir pressure, the swelling factor with hydrocarbon gas is four times higher than for nitrogen. Furthermore, the reservoir fluid density increases during swelling with nitrogen, whereas it decreases as a result of hydrocarbon gas swelling. The same trend is observed for viscosity. Injection gas blends with various proportions of nitrogen and carbon dioxide shows that the MMP is constant when more than 35–40% nitrogen is present in the blend. The simulation studies involve a number of sensitivity runs performed on sector models, which are sufficiently fine-gridded to capture the compositional transition zone propagating between injector and producer pairs. Miscible nitrogen injection comes out as a viable option with the potential to increase recovery by 10–20% above the current water flood development scheme. The significantly improved sweep and displacement efficiency are due to N2 miscibility with reservoir oil under reservoir conditions, possibility of increased PV injected in a N2 WAG scheme, and the ability to maintain the initial reservoir pressure.

The ECMWF OCEAN5 system is a global ocean and sea-ice ensemble of reanalysis and real-time analysis. This paper gives a full description of the OCEAN5 system, with the focus on upgrades of system components with respect to its predecessors, ORAS4 and ORAP5. An important novelty in OCEAN5 is the ensemble generation strategy that includes perturbation of initial conditions and a generic perturbation scheme for observations and forcing fields. Other upgrades include revisions to the a priori bias correction scheme, observation quality control and assimilation method for sea-level anomalies. The OCEAN5 historical reconstruction of the ocean and sea-ice state is the ORAS5 reanalysis, which includes five ensemble members and covers the period from 1979 onwards. Updated versions of observation data sets are used in ORAS5 production, with special attention devoted to the consistency of sea surface temperature (SST) and sea-ice observations. Assessment of ORAS5 through sensitivity experiments suggests that all system components contribute to an improved fit to observation in reanalyses, with the most prominent contribution from direct assimilation of ocean in situ observations. Results of observing system experiments further suggest that the Argo float is the most influential observation type in our data assimilation system. Assessment of ORAS5 has also been carried out for several key ocean state variables and verified against reference climate data sets from the ESA CCI (European Space Agency Climate Change Initiative) project. With respect to ORAS4, ORAS5 has improved ocean climate state and variability in terms of SST and sea level, mostly due to increased model resolution and updates in assimilated observation data sets. In spite of the improvements, ORAS5 still underestimates the temporal variance of sea level and continues exhibiting large SST biases in the Gulf Stream and its extension regions which are possibly associated with misrepresentation of front positions. Overall, the SST and sea-ice uncertainties estimated using five ORAS5 ensemble members have spatial patterns consistent with those of analysis error. The ensemble spread of sea ice is commensurable with the sea-ice analysis error. On the contrary, the ensemble spread is under-dispersive for SST.

This paper investigates the Atlantic Ocean influence on equatorial Pacific decadal variability. Using an ensemble of simulations, where the ICTPAGCM (“SPEEDY”) is coupled to the NEMO/OPA ocean model in the Indo-Pacific region and forced by observed sea surface temperatures in the Atlantic region, it is shown that the Atlantic Multidecadal Oscillation (AMO) has had a substantial influence on the equatorial Pacific decadal variability. According to AMO phases we have identified three periods with strong Atlantic forcing of equatorial Pacific changes, namely (1) 1931–1950 minus 1910–1929, (2) 1970–1989 minus 1931–1950 and (3) 1994–2013 minus 1970–1989. Both observations and the model show easterly surface wind anomalies in the central Pacific, cooling in the central-eastern Pacific and warming in the western Pacific/Indian Ocean region in events (1) and (3) and the opposite signals in event (2). The physical mechanism for these responses is related to a modification of the Walker circulation because a positive (negative) AMO leads to an overall warmer (cooler) tropical Atlantic. The warmer (cooler) tropical Atlantic modifies the Walker circulation, leading to rising (sinking) and upper-level divergence (convergence) motion in the Atlantic region and sinking (rising) motion and upper-level convergence (divergence) in the central Pacific region.

We discuss the state of Arctic sea ice in the global eddy-permitting ocean reanalysis Ocean ReAnalysis Pilot 5 (ORAP5). Among other innovations, ORAP5 now assimilates observations of sea ice concentration using a univariate 3DVar-FGAT scheme. We focus on the period 1993–2012 and emphasize the evaluation of model performance with respect to recent observations of sea ice thickness. We find that sea ice concentration in ORAP5 is close to assimilated observations, with root mean square analysis residuals of less than 5 % in most regions. However, larger discrepancies exist for the Labrador Sea and east of Greenland during winter owing to biases in the free-running model. Sea ice thickness is evaluated against three different observational data sets that have sufficient spatial and temporal coverage: ICESat, IceBridge and SMOSIce. Large-scale features like the gradient between the thickest ice in the Canadian Arctic and thinner ice in the Siberian Arctic are simulated well by ORAP5. However, some biases remain. Of special note is the model’s tendency to accumulate too thick ice in the Beaufort Gyre. The root mean square error of ORAP5 sea ice thickness with respect to ICESat observations is 1.0 m, which is on par with the well-established PIOMAS model sea ice reconstruction. Interannual variability and trend of sea ice volume in ORAP5 also compare well with PIOMAS and ICESat estimates. We conclude that, notwithstanding a relatively simple sea ice data assimilation scheme, the overall state of Arctic sea ice in ORAP5 is in good agreement with observations and will provide useful initial conditions for predictions.

A new eddy-permitting ocean reanalysis has been recently completed at ECMWF. It is called Ocean ReAnalysis Pilot 5 (ORAP5), and it spans the period 1979–2012. This work describes the new system, evaluates its performance, and investigates how the estimation of climate indices are affected by the assimilation system settings. ORAP5 introduces several upgrades with respect to its predecessor ORAS4, including increased horizontal and vertical resolution, an prognostic sea-ice component, new versions of the ocean and data assimilation system, revised surface fluxes, new version and treatment of satellite sea surface height data, and assimilation of sea-ice concentration, among others. ORAP5 shows similar performance to ORAS4, with improvements in the northern extratropics (especially in salinity), and slight degradation in the Southern Ocean, probably because the observations are insufficient to constrain the increased level of variability in ORAP5. The sensitivity experiments show that superobbing of altimeter data and correlation length-scales of the background errors have a visible impact on the time evolution of global steric height and its partition into thermo/halo-steric contributions. The sensitivities are especially large in the pre-Argo period, when there is the risk of producing unrealistic steric height variations by overfitting the altimeter data. Compared with a control run without data assimilation, all the assimilation experiments also show stronger variability in the halosteric component in the pre-Argo period. The results highlight the importance of sub-surface observations to assist the assimilation of altimeter data, and the need of using a variety of metrics for evaluating ocean reanalysis systems.

Quantifying the effect of the seawater density changes on sea level variability is of crucial importance for climate change studies, as the sea level cumulative rise can be regarded as both an important climate change indicator and a possible danger for human activities in coastal areas. In this work, as part of the Ocean Reanalysis Intercomparison Project, the global and regional steric sea level changes are estimated and compared from an ensemble of 16 ocean reanalyses and 4 objective analyses. These estimates are initially compared with a satellite-derived (altimetry minus gravimetry) dataset for a short period (2003-2010). The ensemble mean exhibits a significant high correlation at both global and regional scale, and the ensemble of ocean reanalyses outperforms that of objective analyses, in particular in the Southern Ocean. The reanalysis ensemble mean thus represents a valuable tool for further analyses, although large uncertainties remain for the inter-annual trends. Within the extended intercomparison period that spans the altimetry era (1993-2010), we find that the ensemble of reanalyses and objective analyses are in good agreement, and both detect a trend of the global steric sea level of 1.0 and 1.1 +/- 0.05 mm/year, respectively. However, the spread among the products of the halosteric component trend exceeds the mean trend itself, questioning the reliability of its estimate. This is related to the scarcity of salinity observations before the Argo era. Furthermore, the impact of deep ocean layers is non-negligible on the steric sea level variability (22 and 12 % for the layers below 700 and 1500 m of depth, respectively), although the small deep ocean trends are not significant with respect to the products spread.

A new operational ocean reanalysis system (ORAS4) has been implemented at ECMWF. It spans the period 1958 to the present. This article describes its main components and evaluates its quality. The adequacy of ORAS4 for the initialization of seasonal forecasts is discussed, along with the robustness of some prominent climate signals. ORAS4 has been evaluated using different metrics, including comparison with observed ocean currents, RAPID-derived transports, sea-level gauges, and GRACE-derived bottom pressure. Compared to a control ocean model simulation, ORAS4 improves the fit to observations, the interannual variability, and seasonal forecast skill. Some problems have been identified, such as the underestimation of meridional overturning at 26°N, the magnitude of which is shown to be sensitive to the treatment of the coastal observations. ORAS4 shows a clear and robust shallowing trend of the Pacific Equatorial thermocline. It also shows a clear and robust nonlinear trend in the 0–700 m ocean heat content, consistent with other observational estimates. Some aspects of these climate signals are sensitive to the choice of sea-surface temperature product and the specification of the observation-error variances. The global sea-level trend is consistent with the altimeter estimate, but the partition into volume and mass variations is more debatable, as inferred by discrepancies in the trend between ORAS4- and GRACE-derived bottom pressure.