Search

We have investigated the causes of the sea surface temperature (SST) bias in the Angola-Benguela Frontal Zone (ABFZ) of the southeastern Atlantic Ocean simulated by the Norwegian Earth System Model (NorESM). Similar to other coupled-models, NorESM has a warm SST bias in the ABFZ of up to 8 °C in the annual mean. Our analysis of NorESM reveals that a cyclonic surface wind bias over the ABFZ drives a locally excessively strong southward (0.05 m/s (relative to observation)) Angola Current displacing the ABFZ southward. A series of uncoupled stand-alone atmosphere and ocean model simulations are performed to investigate the cause of the coupled model bias. The stand-alone atmosphere model driven with observed SST exhibits a similar cyclonic surface circulation bias; while the stand-alone ocean model forced with the reanalysis data produces a warm SST in the ABFZ with a magnitude approximately half of that in the coupled NorESM simulation. An additional uncoupled sensitivity experiment shows that the atmospheric model’s local negative surface wind curl generates anomalously strong Angola Current at the ocean surface. Consequently, this contributes to the warm SST bias in the ABFZ by 2 °C (compared to the reanalysis forced simulation). There is no evidence that local air-sea feedbacks among wind stress curl, SST, and sea level pressure (SLP) affect the ABFZ SST bias. Turbulent surface heat flux differences between coupled and uncoupled experiments explain the remaining 2 °C warm SST bias in NorESM. Ocean circulation, upwelling and turbulent heat flux errors all modulate the intensity and the seasonality of the ABFZ errors. [ABSTRACT FROM AUTHOR]

The development of systematic biases in climate models used in operational seasonal forecasting adversely affects the quality of forecasts they produce. In this study, we examine the initial evolution of systematic biases in the ECMWF System 4 forecast model, and isolate aspects of the model simulations that lead to the development of these biases. We focus on the tendency of the simulated intertropical convergence zone in the western equatorial Pacific to drift northwards by between 0.5° and 3° of latitude depending on season. Comparing observations with both fully coupled atmosphere–ocean hindcasts and atmosphere-only hindcasts (driven by observed sea-surface temperatures), we show that the northward drift is caused by a cooling of the sea-surface temperature on the Equator. The cooling is associated with anomalous easterly wind stress and excessive evaporation during the first twenty days of hindcast, both of which occur whether air-sea interactions are permitted or not. The easterly wind bias develops immediately after initialisation throughout the lower troposphere; a westerly bias develops in the upper troposphere after about 10 days of hindcast. At this point, the baroclinic structure of the wind bias suggests coupling with errors in convective heating, although the initial wind bias is barotropic in structure and appears to have an alternative origin. [ABSTRACT FROM AUTHOR]

SST errors in the tropical Atlantic are large and systematic in current coupled general-circulation models. We analyse the growth of these errors in the region of the south-eastern tropical Atlantic in initialised decadal hindcasts integrations for three of the models participating in the Coupled Model Inter-comparison Project 5. A variety of causes for the initial bias development are identified, but a crucial involvement is found, in all cases considered, of ocean-atmosphere coupling for their maintenance. These involve an oceanic 'bridge' between the Equator and the Benguela-Angola coastal seas which communicates sub-surface ocean anomalies and constitutes a coupling between SSTs in the south-eastern tropical Atlantic and the winds over the Equator. The resulting coupling between SSTs, winds and precipitation represents a positive feedback for warm SST errors in the south-eastern tropical Atlantic. [ABSTRACT FROM AUTHOR]

We examine the influence of the South-American land-mass and its mountains on the significant cyclic diurnal and semidiurnal components of the average circulation in the adjacent area of the southeastern tropical Pacific (SEP). Our approach is based on a number of numerical simulations with the regional atmospheric model weather research and forecasting forced by the National Centers for Environmental Prediction's final analysis operational analysis data. In the control simulation the model domain covers the SEP and a large part of South America. In several sensitivity experiments the domain is reduced to progressively exclude continental areas. We find that the mean diurnal cycle is sensitive to model domain in ways that reveal the existence of different contributions originating from the Chilean and Peruvian land-masses. The experiments suggest that diurnal variations in circulations and thermal structures over the SEP (mainly forced by local insolation) are influenced by convection over the Peruvian sector of the Andes cordillera, while the mostly dry mountain-breeze circulations force an additional component that results in semi-diurnal variations near the coast. A series of numerical tests, however, reveal sensitivity of the simulations to the choice of vertical grid, limiting the possibility of solid quantitative statements on the amplitudes and phases of the diurnal and semidiurnal components across the domain. [ABSTRACT FROM AUTHOR]

We present an analysis of the oceanic heat advection and its variability in the upper 500 m in the southeastern tropical Pacific (100W-75W, 25S-10S) as simulated by the global coupled model HiGEM, which has one of the highest resolutions currently used in long-term integrations. The simulated climatology represents a temperature advection field arising from transient small-scale (<450 km) features, with structures and transport that appear consistent with estimates based on available observational data for the mooring at 20S, 85W. The transient structures are very persistent (>4 months), and in specific locations they generate an important contribution to the local upper-ocean heat budget, characterised by scales of a few hundred kilometres, and periods of over a year. The contribution from such structures to the local, long-term oceanic heat budget however can be of either sign, or vanishing, depending on the location; and, although there appears some organisation in preferential areas of activity, the average over the entire region is small. While several different mechanisms may be responsible for the temperature advection by transients, we find that a significant, and possibly dominant, component is associated with vortices embedded in the large-scale, climatological salinity gradient associated with the fresh intrusion of mid-latitude intermediate water which penetrates north-westward beneath the tropical thermocline. [ABSTRACT FROM AUTHOR]

About a third of the El-Niño/Southern Oscillation (ENSO) variability in the HadCM3 coupled general-circulation model is shown to be associated with variability in the south-east tropical Pacific (SETP) area. Sea-surface temperature (SST) anomalies along the east Pacific tend to precede ENSO anomalies. In HadCM3, SST tendencies in the SETP area are controlled mainly by surface latent heat fluxes and short-wave cloud forcing. Interannual SST anomalies in the SETP tend to propagate meridionally. In the winter season (JJA), this is consistent with a wind-evaporation-SST (WES) mode. Coupling with the strato-cumulus cloud (Sc) cover is critical in reducing the evaporative damping of the WES mode, and external forcing is provided by extratropical circulation anomalies. In spring, SETP variability and ENSO are coupled via the low-level circulation, resulting in a mutual reinforcement. Cloud-cover anomalies are not strongly controlled by local SSTs, and appear mainly dependent on atmospheric meridional advection. The apparent association between cold SSTs and Sc cover does not reflect a positive local feedback. These conclusions are not sensitive to the model’s warm SST bias, associated with reduced stratocumulus clouds and weak southerly wind stress, which depends on erroneous near-field orographic forcing of the coastal circulation. Some of our results are supported by similar evidence from observational datasets and other CMIP3 models. [ABSTRACT FROM AUTHOR]

We analyse the differences in the properties of the El Niño Southern Oscillation (ENSO) in a set of 17 coupled integrations with the flux-adjusted, 19-level HadCM3 model with perturbed atmospheric parameters. Within this ensemble, the standard deviation of the NINO3.4 deseasonalised SSTs ranges from 0.6 to 1.3 K. The systematic changes in the properties of the ENSO with increasing amplitude confirm that ENSO in HadCM3 is prevalently a surface (or SST) mode. The tropical-Pacific SST variability in the ensemble of coupled integrations correlates positively with the SST variability in the corresponding ensemble of atmosphere models coupled with a static mixed-layer ocean (“slab” models) perturbed with the same changes in atmospheric parameters. Comparison with the respective coupled ENSO-neutral climatologies and with the slab-model climatologies indicates low-cloud cover to be an important controlling factor of the strength of the ENSO within the ensemble. Our analysis suggests that, in the HadCM3 model, increased SST variability localised in the south-east tropical Pacific, not originating from ENSO and associated with increased amounts of tropical stratocumulus cloud, causes increased ENSO variability via an atmospheric bridge mechanism. The relationship with cloud cover also results in a negative correlation between the ENSO activity and the model’s climate sensitivity to doubling CO2. [ABSTRACT FROM AUTHOR]

Palaeoclimate simulations provide an opportunity for climate model evaluation as well as having a potential role in assigning relative likelihood to different ensemble members in probabilistic climate change prediction, supplementing constraints provided by the instrumental record. Here we take some initial steps towards such an approach by performing ensemble experiments with the Hadley Centre HadCM3 model under pre-industrial and mid-Holocene (6,000 years before present) forcing conditions. We examine the changes in both mean tropical climate and El Niño-Southern Oscillation (ENSO) variability, as palaeoclimate records suggest that ENSO amplitude was reduced in the mid-Holocene. Experiments are performed with perturbations to physical parameters in the atmosphere–surface component of the model, and with different implementations of heat and freshwater flux adjustments. Heat flux adjustments are required to stabilise model versions in which perturbations cause a net radiative imbalance. While we find broad agreement between different model versions in terms of changes in mean climate in the mid-Holocene, a detailed and quantitative comparison with the geographically-sparse palaeo-record is limited by systematic model biases. In the simulations without seasonally-varying flux adjustments there are modest reductions in ENSO amplitude of the order of 10–15%, lower than the range of reductions inferred from coral proxy records. We examine the mechanisms for these changes, and discuss the implications for the design of future ensemble experiments to formally quantify uncertainty in climate change predictions using palaeoclimate simulations. [ABSTRACT FROM AUTHOR]