Atmospheric data assimilation on the equatorial beta plane

Abstract

Kalman filter theory shows great promise when applied to the assimilation of atmospheric observations. Previous work has concentrated on extratropical dynamics, and tropical aspects have not yet been seriously tackled. In this article, a Kalman filter is applied to the linearized shallow water equations on an equatorial beta plane. The system or model error is constructed from the slow eigenmodes of the model and is based on an expansion in parabolic cylinder functions. The resulting second‐moment statistics are discussed in some detail. The Kalman filter is applied to a special observation network that allows the diagonalization of the system. Following Daley and Ménard (1993), it is then possible to obtain the complete space and time solution for the second‐moment forecast and analysis error statistics. The slow (low‐frequency) and fast (high‐frequency) error statistics are examined separately for both the optimal and suboptimal cases.     

Quasi-quadrennial and quasi-biennial variability in the equatorial Pacific

Evaluation of competing El Niño/Southern Oscillation (ENSO) theories requires one to identify separate spectral peaks in equatorial wind and sea-surface temperature (SST) time series. To sharpen this identification, we examine the seasonal-to-interannual variability of these fields by the data-adaptive method of multi-channel singular spectrum analysis (M-SSA). M-SSA is applied to the equatorial band (4°N-4°S), using 1950–1990 data from the Comprehensive Ocean and Atmosphere Data Set. Two major interannual oscillations are found in the equatorial SST and surface zonal wind fields, U. The main peak is centered at about 52-months; we refer to it as the quasi-quadrennial (QQ) mode. Quasi-biennial (QB) variability is split between two modes, with periods near 28 months and 24 months. A faster, 15-month oscillation has smaller amplitude. The QQ mode dominates the variance and has the most distinct spectral peak. In time-longitude reconstructions of this mode, the SST has the form of a standing oscillation in the eastern equatorial Pacific, while the U-field is dominated by a standing oscillation pattern in the western Pacific and exhibits also slight eastward propagation in the central and western Pacific. The locations of maximum anomalies in both QB modes are similar to those of the QQ mode. Slight westward migration in SST, across the eastern and central, and eastward propagation of U, across the western and central Pacific, are found. The significant wind anomaly covers a smaller region than for the QQ. The QQ and QB modes together represent the ENSO variability well and interfere constructively during major events. The sharper definition of the QQ spectral peak and its dominance are consistent with the devil's staircase interaction mechanism between the annual cycle and ENSO.     

Development and propagation of equatorial waves

Development and propagation of equatorial waves are investigated with the model which includes convection -wave convergence feedback and convection-frictional convergence feedback. Two experiments with an initial Kelvin wave (Exp. K) and with an initial Rossby wave (Exp. R) are carried out. The equatorial waves in Exp. R grow much faster than those in Exp. K. The equatorial waves in both experiments follow zonal (eastward / westward) and meridional (poleward) propagation. The equatorial waves can be partitioned into two meridional modes using Parabolic Cylinder Function. An equa?tor mode denotes a wave component with a positive precipitation center at the equator and an off-equator mode rep?resents a wave component with positive precipitation centers off the equator. The equator mode dominates in Exp. K whereeas the off-equator mode dominates in Exp. R. The rapid wave growth in Exp. R is interpreted by analyzing the eddy available potential energy (EAPE) generation. Stronger off-equator mode in Exp. R obtains more EAPE through convection-frictional convergence feedback which results in more rapid wave growth. The relative vorticity tendency is determined by interactions between Earth’s vorticity and lower-troposphere convergence (divergence effect) and between the meridional gradient and lower-troposphere circulation (beta effect). The eastward and poleward propagation of equatorial waves is a result of the divergence effect, and the westward movement is caused by the beta effect.     

Lagrangian sources of frontogenesis in the equatorial Atlantic front

Estimating the processes that control the north equatorial sea surface temperature (SST)-front on the northern edge of the cold tongue in the tropical Atlantic is a key issue for understanding the dynamics of the oceanic equatorial Atlantic and the West African Monsoon. Diagnosis of the frontogenetic forcings on a realistic high-resolution simulation was used to identify the processes involved in the formation and evolution of the equatorial SST-front. The turbulent forcing associated with the mixed-layer turbulent heat flux was found to be systematically frontolytic while the dynamic forcing associated with currents was found to be frontogenetic for the equatorial SST-front. Nevertheless, the low-frequency component of the turbulent forcing was frontogenetic and initiated the SST-front which was then amplified and maintained by the leading dynamic forcing. This forcing was mainly driven by the meridional convergence of the northern South Equatorial Current (nSEC) and the Guinea Current, which points out the essential role played by the circulation in the equatorial SST-front evolution. The quasi-biweekly variability of the equatorial SST-front and its forcings were found to be more strongly coupled to the wind energy flux (WEF) than to the surface wind stress. In fact the WEF controlled the convergence/divergence of the nSEC and Guinea Current and thus the meridional component of the leading dynamic forcing. The WEF explains the equatorial SST-front development better than the wind does because it is a coupled ocean-atmosphere process.     

The observed teleconnection between the equatorial Amazon and the Intra-Americas Seas

Using observations of rainfall and SST analysis it is shown that there is a robust relationship with two-season lag between the austral summer (December–January–February [DJF]) Equatorial Amazon (EA) rainfall and the following boreal summer season (June–July–August [JJA]) Intra-Americas Seas (IAS) Sea Surface Temperature Anomalies (SSTA). It is observed that in wetter than normal austral summer seasons over EA, the SSTA in the IAS are cooler than normal in the following JJA season. This teleconnection also manifests in the ocean heat content of the IAS region. Our analysis indicates that the net surface heat flux into the ocean (particularly the surface longwave and the shortwave radiative fluxes) dictates the strongest influence on the JJA Caribbean SSTA, the core region of the IAS where the observed teleconnection with EA rainfall is strongest. This study also finds that this teleconnection is in fact a manifestation of the remote ENSO forcing on the Caribbean SSTA through its modulation of the EA rainfall anomalies. In a wet DJF year over EA, the Atlantic Inter-Tropical Convergence Zone (ITCZ) moves further southward than climatology. This causes the dry limb of the associated overturning circulation of the Atlantic ITCZ to reside over the Caribbean Sea region in the subsequent March–April–May and JJA seasons. As a result of this large-scale descent in the wet DJF year over EA, there is a net decrease in the heat flux into the ocean from increased emission of surface longwave radiation in the presence of anomalously dry atmosphere. In a dry DJF year over EA the Atlantic ITCZ is nearly co-located in the core region of the IAS, which is northward than the climatological location, resulting in the descending limb of the overturning location to be located further south of the Caribbean Sea leading to warmer SSTA.     

Mean meridional currents in the central and eastern equatorial Atlantic

Ship-based acoustic Doppler current profiler (ADCP) velocity measurements collected by several major field programs in the tropical Atlantic are averaged and combined with estimates of the mean near-surface velocity derived from drifters and Argo float surface drifts (ADCP+D) to describe the mean cross-equatorial and vertical structure of the meridional currents along 23°W and 10°W. Data from moored ADCPs and fixed-depth current meters, a satellite-derived velocity product, and a global ocean reanalysis were additionally used to evaluate the mean ADCP+D meridional velocity. The dominant circulation features in the long-term mean ADCP+D meridional velocity in the upper 100 m are the tropical cells (TCs) located approximately between 5°S and 5°N, with near-surface poleward flow and subsurface equatorward flow that is stronger and shallower in the northern cell compared to the southern cell. The thickness of the surface limb of the TCs decreases and the northern cell is found to shift further south of the equator from the central to eastern tropical Atlantic. Analysis of two-season means estimated from the ship-based ADCP, near-surface drift, and moored velocity data, as well as the simulated fields, indicates that the maximum poleward velocity in the surface limb of the TCs intensifies during December–May along 23°W largely due to seasonal compensation between the geostrophic and ageostrophic (or wind-driven) components of the meridional velocity, whereas the maximum equatorward flow in the subsurface limb of the northern cell intensifies during June–November along both 23°W and 10°W due to the seasonality of the geostrophic meridional velocity.     

Tropical variability and stratospheric equatorial waves in the IPSLCM5 model

The atmospheric variability in the equatorial regions is analysed in the Earth System Model pre-industrial simulation done at IPSL in the framework of CMIP5. We find that the model has an interannual variability of about the right amplitude and temporal scale, when compared to the El-Niño Southern Oscillation (ENSO), but that is too confined to the western Pacific. At the intra-seasonal periods, the model variability lacks of large-scale organisation, and only produces one characteristic Madden-Julian Oscillation every 10 winters typically. At shorter time-scales and in the troposphere, the model has Rossby and Kelvin Convectively Coupled Equatorial Waves (CCEWs), but underestimates the Kelvin CCEWs signal on OLR. In the model stratosphere, a composite analysis shows that the Temperature and velocities fluctuations due to the Kelvin waves are quite realistic. In the model nevertheless, the stratospheric waves are less related to the convection than in the observations, suggesting that their forcing by the midlatitudes plays a larger role. Still in the model, the Kelvin waves are not predominantly occurring during the life cycle of the tropospheric Kelvin CCEWs, a behaviour that we find to be dominant in the observations. The composite analysis is also used to illustrate how the waves modify the zonal mean-flow, and to show that the model Kelvin waves are too weak in this respect. This illustrates how a model can have a reasonable Kelvin waves signal on the velocities and temperature, but can at the same time underestimate their amplitude to modify the mean flow. We also use this very long simulation to establish that in the model, the stratospheric equatorial waves are significantly affected by ENSO, hence supporting the idea that the ENSO can have an influence on the Quasi-Biennial Oscillation.     

Forecasting the equatorial Pacific sea surface temperatures by neural network models

We used neural network models to seasonally forecast the tropical Pacific sea surface temperature anomalies (SSTA) in the Ni?o 3.4 region (6 ^°S–6 ^°N, 120 ^°W–170 ^°W). The inputs to the neural networks (i.e., the predictors) were the first seven wind stress empirical orthogonal function (EOF) modes of the tropical Pacific (20 ^°S–20 ^°N, 120 ^°E–70 ^°W) for four seasons and the Ni?o 3.4 SSTA itself for the final season. The period of 1952–1981 was used for training the neural network models, and the period 1982–1992 for forecast validation. At 6-month lead time, neural networks attained forecast skills comparable to the other El Ni?o-Southern Oscillation (ENSO) models. Our results suggested that neural network models were viable for ENSO forecasting even at longer lead times of 9 to 12 months. We hypothesized that at these longer leads, the underlying relationship between the wind stress and Ni?o 3.4 SSTA became increasingly nonlinear. The neural network results were interpreted in light of current theories, e.g., the role of the “off-equatorial” Rossby waves in triggering the onset of an ENSO event and the delayed-oscillator theory in the development and termination of an ENSO event. Received: 31 October 1995 / Accepted: 25 July 1996     

Heat budget of the south-central equatorial Pacific in CMIP3 models

ABSTRACT Using data from 17 coupled models and nine sets of corresponding Atmospheric Model Intercomparison Project （AMIP） results, we investigated annual and seasonal variation biases in the upper 50 m of the south-central equatorial Pacific, with a focus on the double-ITCZ bias, and examined the causes for the amplitude biases by using heat budget analysis. The results showed that, in the research region, most of the models simulate SSTs that are higher than or similar to observed. The simulated seasonal phase is close to that observed, but the amplitudes of more than half of the model results are larger than or equal to observations. Heat budget analysis demonstrated that strong shortwave radiation in individual atmospheric models is the main factor that leads to high SST values and that weak southward cold advection is an important mechanism for maintaining a high SST. For seasonal circulation, large surface shortwave radiation amplitudes cause large SST amplitudes.     

Cross‐equatorial error propagation: Research note

Abstract

Numerical simulation experiments published in 1974 by Daley have been repeated with a much higher resolution, spectral, shallow water model. With a forecast period extending toll d, it is shown that a global model in which only the largest scales are used at initial time in the Southern Hemisphere yields a more accurate forecast for the Northern Hemisphere than a hemispheric model does. Compared with a uniform high‐resolution, global model, the error in the Northern Hemisphere forecast is high in the ultra‐long waves but decreases rather rapidly while the resolution of the initial Southern Hemispheric data is increased.     

Structure of the atmospheric surface layer over an industrialized equatorial area

This paper is written to report observations of the structure of the atmospheric surface layer over a coastal industrialized equatorial area. The observations were recorded at Prai Industrial Park, Penang (5° 22′ N, 100° 23′ E) a relatively simple terrain area during the south-west monsoon season in the period of three months using slow response systems. The limitations of the instruments used and its effects on the results are discussed. Wind turbulence and temperature were measured on a 10 m tower and analyzed using eddy correlation method and Monin–Obukhov similarity relations to obtain the normalized standard deviation of longitudinal (σ_u/u), lateral (σ_v/u) and vertical wind velocity fluctuations (σ_w/u) with respect to stability parameter z/L. From the results of the analysis, we found that most of turbulence is generated by shear or mechanical force. It was found that the average neutral value of σ_u/u is 2.35, 1.98 for σ_v/u and 1.47 for σ_w/u with a significantly lower than the proportionality to the power of 1/3 during unstable atmospheric conditions, and thus do not obey Monin–Obukhov similarity theory. It was observed that σ_u/u and σ_v/u values increase linearly in the range of 0 < z/L < 2 and fairly well correlated while σ_w/u does not.     

Variability of ozone in the marine boundary layer of the equatorial Pacific Ocean

This study examines the processes controlling the diurnal variability of ozone (O₃) in the marine boundary layer of the Kwajalein Atoll, Republic of the Marshall Islands (latitude 8° 43′ N, longitude 167° 44′ E), during July to September 1999. At the study site, situated in the equatorial Pacific Ocean, O₃ mixing ratios remained low, with an overall average of 9–10 parts per billion on a volume basis (ppbv) and a standard deviation of 2.5 ppbv. In the absence of convective storms, daily O₃ mixing ratios decreased after sunrise and reached minimum during the afternoon in response to photochemical reactions. The peak-to-peak amplitude of O₃ diurnal variation was approximately 1–3 ppbv. During the daytime, O₃ photolysis, hydroperoxyl radicals, hydroxyl radicals, and bromine atoms contributed to the destruction of O₃, which explained the observed minimum O₃ levels observed in the afternoon. The entrainment of O₃-richer air from the free troposphere to the local marine boundary layer provided a recovery mechanism of surface O₃ mixing ratio with a transport rate of 0.04 to 0.2 ppbv per hour during nighttime. In the presence of convection, downward transport of O₃-richer tropospheric air increased surface O₃ mixing ratios by 3–12 ppbv. The magnitude of O₃ increase due to moist convection was lower than that observed over the continent (as high as 20–30 ppbv). Differences were ascribed to the higher O₃ levels in the continental troposphere and weaker convection over the ocean. Present results suggest that moist convection plays a role in surface-level O₃ dynamics in the tropical marine boundary layer.     

On the warming of intermediate and deep waters in the equatorial North Atlantic

In July 2000, a transatlantic hydrographic section was made on board the Russian R/V Akademik Ioffe in the northern equatorial region at ～6.5° N on the WOCE (World Ocean Circulation Experiment) A06 line. A significant warming in the layers of intermediate and deep waters in the interior eastern basin is determined from comparison of the section temperature data and those obtained on the WOCE A06 line in 1993. This result, together with the results of the previous studies, indicates a substantial warming of intermediate and upper deep waters above 2800–3000 m in the eastern equatorial North Atlantic during the second half of the 20th century. In the 1000–2000 m layer, temperature has increased by 0.13–0.14°C since 1957.     

Seasonal variation of the diurnal cycle of rainfall in the eastern equatorial Indian Ocean

Summary The diurnal cycle of rainfall over the eastern equatorial Indian Ocean was studied for the period 23^rd October 2001 to 31^st October 2003 using hourly data from the Triton buoy positioned at 1.5° S and 90° E. An analysis of the active and weak spells of rainfall for different seasons revealed peaks in the late evening hours in Winter, Summer and Fall and in early morning hours (in Spring) in 2002. The active spell of rainfall peaked in the afternoon hours, during Winter, Spring and Summer in 2003, which agrees with the previous results of Janowiak et al. (1994). An analysis of rainfall events showed that Fall 2002 had a maximum number of rainfall events (90) and minimum (60) were observed in Spring 2003. Further it was found that the majority of rain events (>60%) were less than 3 hours in duration throughout the study period. The longer duration rainfall events (i.e. rain events greater than 6 hour duration) contributed significantly to Spring 2002 (20% of the total rainfall) and Winter 2003 (21% of the total rainfall). Harmonic analysis of the hourly rainfall data for different seasons revealed that diurnal harmonic explains more than 80% of the variance for all seasons. Furthermore, the diurnal harmonic has a maximum amplitude for all seasons except summer, where the semidiurnal and six hourly harmonics are significant.     

Role of sea surface temperature on the equatorial waves and intraseasonal oscillations

Theoretical and Applied Climatology - Ocean interactions are known to play a major role in the modulation of intraseasonal variability. The role of sea surface temperature (SST) and major oceanic...     

On the origin of equatorial Atlantic biases in coupled general circulation models

Many coupled ocean–atmosphere general circulation models (GCMs) suffer serious biases in the tropical Atlantic including a southward shift of the intertropical convergence zone (ITCZ) in the annual mean, a westerly bias in equatorial surface winds, and a failure to reproduce the eastern equatorial cold tongue in boreal summer. The present study examines an ensemble of coupled GCMs and their uncoupled atmospheric component to identify common sources of error. It is found that the westerly wind bias also exists in the atmospheric GCMs forced with observed sea surface temperature, but only in boreal spring. During this time sea-level pressure is anomalously high (low) in the western (eastern) equatorial Atlantic, which appears to be related to deficient (excessive) precipitation over tropical South America (Africa). In coupled simulations, this westerly bias leads to a deepening of the thermocline in the east, which prevents the equatorial cold tongue from developing in boreal summer. Thus reducing atmospheric model errors during boreal spring may lead to improved coupled simulations of tropical Atlantic climate.     

The influence of thermocline slope on equatorial thermocline displacement

The influence of a zonal slope in the thermocline on both wave- and wind-induced equatorial upwelling is studied. It is shown that the vertical displacement of the thermocline induced by a long Kelvin or planetary wave varies as the thermocline depth to the power $\text{?3}\text{8}$, as the wave propagates into a region of changed thermocline depth The zonal and meridional velocities induced by a wave vary as the thermocline depth to the power $\text{?7}\text{8}$. The wind-induced upwelling associated with a zonal slope of the thermocline is shown to be strongly confined to the equator. The importance of thermocline slope on the Ekman divergence is found to increase with time from the initial application of the wind stress perturbation. In the Pacific Ocean the meridional velocity may be altered by up to 60% by the thermocline slope before the influence of the boundaries becomes important.