The response of an open stratified bay to wind forcing

Abstract

Current and temperature measurements collected during the summers of 1974, 1975 and 1979 are used to investigate the wind‐induced response of St Georges Bay, Nova Scotia. A multivariate‐frequency response analysis shows that temperature and the along‐bay component of current in the lower layer are coherent with the local wind stress at periods of 2–6 d, with the wind stress accounting for 35–65% of the observed variances. Winds are also coherent with the surface currents but account for only 20–25% of the variance. Two dynamically different regions are identified. Near the entrance to the bay, cross‐bay wind stresses (τ_x) produce Ekman drift with compensating flow in the lower layer. In the interior of the bay, the near‐bottom currents are coherent with along‐bay (τ_y) wind stresses and are directed upwind. These currents are topographically steered and a response to surface pressure gradients set up by the wind. The surface gradients are believed to be part of the wind‐induced set‐up within the southern Gulf of St Lawrence.     

Subtidal response of the Scotian shelf bottom pressure field to meteorological forcing

Abstract

Data collected during the Canadian Atlantic Storms Program (CASP) show subtidal variations in subsurface pressure (SSP) to be highly coherent throughout the Scotian Shelf region, and well correlated to fluctuations in the alongshelf component of wind stress (τ^y). Analysis using a frequency‐dependent multiple regression model verified that τ^y is the primary source of local forcing to the SSP field, although non‐locally generated variations in SSP are also important. The two components of local wind stress and a non‐local SSP term combine to explain over 90% of SSP variance on the Scotian Shelf.

Statistical results describing the response to τ^y change dramatically depending upon the inclusion of non‐local forcing. In a model including both types of forcing, the SSP response to local forcing behaves like the solution to a dynamical model forced by time‐dependent wind stress with sea‐level prescribed to zero at the eastern cross‐shelf boundary. Local τ^y forcing becomes more effective to the west and onshore, whereas the phase suggests propagation to the west. The importance of τ^y is reduced at higher frequencies. Describing SSP with a statistical model containing local forcing alone may lead to an incorrect interpretation of SSP dynamics, particularly in the synoptic band where the wind variance is greatest.

Energy originating from a non‐local source is most obvious at ω > 0.5 cpd and at locations on the eastern half of the shelf, but plays an important role at all sites and at all frequencies. These variations propagate to the west at speeds of 6.5 (ω < 0.2 cpd), 25–33 (0.2 cpd < ω < 0.5 cpd), and 12–17m s^?1 (ω > 0.5 cpd). The exponential decay scales at all frequencies are ～900 km in the direction of phase propagation. The non‐local response is consistent with theoretical estimates of first‐ and second‐mode shelf waves for this region and represents the most direct evidence of shelf wave activity on the Scotian Shelf to date.     

Inertial oscillations near the coast of Nova Scotia during CASP

Abstract

Analysis of current, temperature and salinity records in the nearshore region of the Scotian Shelf during the Canadian Atlantic Storms Program (CASP), reveals that the inertial wave field is highly intermittent, with comparable amplitudes in the surface and deep layers. Clockwise current energy in the surface layer is concentrated at a frequency slightly below inertial, consistent with Doppler shifting by the strong mean current and/or straining by the mean flow shear, whereas the spectral peak in deep water is at the local inertial frequency. Clockwise coherence is high (γ² ≥ 0.8) horizontally over the scale of the array (60 km × 120 km) and in the vertical, with upward phase propagation rates of 0.15–0.50 × 10^?12 ms^?1, inversely proportional to the local value of the Brunt Väisälä frequency. Clockwise current energy decreases in the onshore direction and appears to be completely inhibited on the 60‐m isobath.

A case study of the response to the CASP IOP 14 storm indicates that the inertial waves may be generated by a strong wind shift propagating onshore at a speed of 10 ms^?1. On the eastern side of the array (Liscomb line), clockwise current oscillations propagate onshore in the surface layer at a rate (8.1 ± 0.9 m s^?1) comparable with the speed of the atmospheric front, while waves in the pycnocline move offshore at a lower (internal wave) speed (1.8 m s^?1). Furthermore the temperature and salinity fluctuations are in (out) of phase with longshore current in the deep (surface) layer. However, on the western side of the array (Halifax line), the inertial waves are more complex. A sharp steepening of phase lines at the coast indicates that the phase speed of clockwise current oscillations is considerably reduced and the evidence for offshore propagation of internal waves is less clear. The discrepancies between observations on the two lines suggest that the internal wave field is three‐dimensional.

Results of simple mixed‐layer models indicate that the inertial response near the surface is sensitive to the accurate definition of the local wind field, but not to certain model physics, such as the form of the decay term. The observations also show some qualitative similarities with models for two‐dimensional response to a moving front (e.g. Kundu, 1986), but the actual forcing terms are more complicated, based on IOP 14 wind measurements.     

风应力对热带斜压海洋的强迫

利用一个线性的具有不同密度、温度的热带海洋两层模式,分析了热带西太平洋对纬圈风应力的响应.解析地求得热带西太平洋温跃层厚度、洋流及海温分布.结果表明次表层温度变化明显要比表层海温变化大,同时在大洋西部次表层发展起来的扰动向东传播能引起海温分布形态的异常.理论结果支持观测己表明的热带西太平洋物理量的变异在ElNino/La Nina事件中起着重要作用的事实.     

Propagation of coastal‐trapped waves under an ice cover in Hudson Bay*

Abstract

Three arrays of current‐meter moorings were deployed under landfast sea ice in southeast Hudson Bay for eight weeks in spring 1986. Spectral analysis shows low‐frequency signals with periods of 3 to 11 days. These signals are interpreted as being due to coastal‐trapped waves propagating cyclonically in Hudson Bay; their theoretical dispersion relations and corresponding modal structures are presented for winter stratification and are compared with observations. At a period of 3 days both the modified external Kelvin wave and higher mode continental shelf waves may be important in describing the observed low‐frequency variability, whereas at a period of 10 days the Kelvin wave appears to be the dominant mode. The generation mechanisms for these coastal trapped waves are also investigated. Two sources have been studied: the longshore atmospheric pressure gradient and the average atmospheric pressure over the ice cover in Hudson Bay. Coherence and phase analyses performed with time series of longshore current and atmospheric forcing data reveal that both the average atmospheric pressure and the longshore atmospheric pressure gradient are important in explaining the observed low‐frequency variability, without indicating which one is the most important.     

Laterally‐propagating Rossby waves in the easterly acceleration phase of the quasi‐biennial oscillation

Abstract

It is shown that oscillating mean flow solutions exist in the one‐dimensional Holton‐Lindzen (1972) model in the presence of a single Kelvin wave, mean flow diffusion, and an easterly zonal force per unit mass that is constant in height and time except at those points in the time‐height cross‐section where the latitudinally‐integrated mean flow is less than some prescribed easterly value. The latter forcing is intended to crudely represent the absorption of quasi‐stationary planetary Rossby waves at the tropical zero‐wind line. Our results suggest an alternative, and somewhat simpler, possible interpretation of the quasi‐biennial mean zonal wind oscillation in the equatorial lower stratosphere.     

Sea‐level response at Nain,Labrador, to atmospheric pressure and wind

Abstract

Two years of subtidal sea‐level data from Nain, Labrador, are analysed in terms of local atmospheric pressure and the two components of geostrophic wind or stress. Frequency‐dependent response coefficients are determined by multiple regression analysis involving inversion of the cross‐spectral matrix of the inputs. At very low frequencies the response to pressure is isostatic and the wind stress coefficients are consistent with those determined by Thompson et al. (1985) from analysis of a longer series of monthly means. There is very little change in the response between icy and ice‐free seasons. The wind, or stress, coefficients correspond to geostrophic set‐up by a narrow longshore current but do not show as much of an increase of phase lag with increasing frequency as expected. The pressure response is less than isostatic and lags as the frequency increases from zero to about 0.02 cph. Possible reasons for this are discussed. Removal of wind as well as pressure effects ffom the sea‐level data makes only minor changes to the monthly mean residual sea‐level.     

Five years’ central pacific sea level from in situ array,satellite altimeter and numerical model: Research note

Abstract

Temporal and spatial features of central equatorial Pacific Ocean sea‐level variation appear similar, in measurements from two very different systems (one in the ocean and one carried on a satellite), and in results from a numerical model of the region. In particular, there is an interannual cycle: during El Nino, Kelvin waves appear at the equator, and the sea‐surface ridge associated with the equatorial current system shifts southward; in non‐El Nino years, instability waves appear at 6°N (strongest around the end of each calendar year), and the ridge shifts to the north. This three‐way comparison gives support to both measurement systems and to the numerical model.     

Mechanism of Kelvin and Rossby waves during ENSO events

Summary ?The fields of sea-level height anomaly (SLHA) and surface zonal wind anomaly (SZWA) have been analyzed to investigate the typical evolution of spatial patterns during El Ni?o-Southern Oscillation (ENSO) events. Sea surface temperature (SST) changes during ENSO events are represented as an irregular interplay of two dominant modes, low-frequency mode and biennial mode. Cyclostationary principal component (PC) time series of the former variables are regressed onto the PC time series of the two dominant SSTA modes to find the spatial patterns of SLHA and SZWA consistent with the two SSTA modes. The two regressed patterns of SLHA explain a large portion of SLHA total variability. The reconstruction of SLHA using only the two components reasonably depicts major ENSO events. Although the low-frequency component of SST variability is much larger than the biennial component, the former does not induce strong Kelvin and Rossby waves. The biennial mode induces much stronger dynamical ocean response than the low-frequency mode. Further decomposition of the SLHA modes into Kelvin and Rossby components shows how these two types of equatorial waves evolve during typical ENSO events. The propagation and reflection of these waves are clearly portrayed in the regressed patterns leading to a better understanding of the wave mechanism in the tropical Pacific associated with ENSO. A close examination suggests that the delayed action oscillator hypothesis is generally consistent with the analysis results reported here. Rossby wave development in the central Pacific in the initiation stage of ENSO and the subsequent reflection of Kelvin waves at the western boundary seems to be an important mechanism for further development of ENSO. The development of Kelvin waves forced by the surface wind in the far-western Pacific cannot be ruled out as a possible mechanism for the growth of ENSO. While Kelvin waves in the far-western Pacific serve as an intiation mechanism of ENSO, they also cause the termination of existing ENSO condition in the central and eastern Pacific, thereby leading to a biennial oscillation over the tropical Pacific. The Kelvin waves from the western Pacific erode the thermocline structure in the central Pacific preventing further devlopment of ENSO and ultimately terminating it. It should be emphasized that this wave mechanism is clear and active only in the biennial mode. Received August 15, 2001; revised March 6, 2002     

Some effects of truncation on topographic instability

Abstract

The topographic stability of forced planetary waves in α β‐channel is investigated using a barotropic model. The equilibrium forced waves are the result of the interaction of a constant mean zonal wind over finite‐amplitude surface orography. Small‐amplitude perturbations of the equilibrium flows are considered that have a wavy part with the same zonal wavenumber as the forcing but an arbitrary meridional structure. The mean zonal part of the perturbations is also taken to be arbitrary. This configuration allows us to (1) isolate those instabilities that depend crucially on topography through form drag and (2) investigate non‐topographic effects on topographic instability that arise from the convergence of Reynolds stresses. A numerical stability analysis is then performed wherein the effects of truncation are emphasized.

This numerical approach casts doubts about the results obtained from some earlier studies involving various ad hoc assumptions. We find, in particular, that unstable long waves (i.e. waves with the zonal wavelength greater than the meridional wavelength) exist under superresonant conditions. This contradicts some previous results that suggest long waves are unstable only when the flow is subresonant. Further, our model reveals the existence of some interesting travelling instabilities. The latter are shown to depend on both form drag and Reynolds stresses in that these two mechanisms alternate in time in supplying the perturbation with the required energy to maintain the exponential growth.     

Some studies of tropical/equatorial waves over Indian Tropical Middle Atmosphere: Results of tropical wave campaign

Summary During the last phase of the Indian Middle Atmosphere Programme everyday launchings of high altitude balloons were carried out at three locations i.e. Trivandrum (8.5°N, 77.5°E), Hyderabad (17.2°N, 78.3°E) and Bhubaneshwar (21.3°N, 85.5°E) for measuring winds and temperature between 1 and 30 km altitude in a campaign mode from 23 January 1989 to 31 March 1989. The data thus obtained have been examined to determine the characteristics of tropical/equatorial waves. Spectral analysis of the time series (68 points) of both zonal and meridional wind components using Maximum Entropy Method (MEM) reveal the presence of waves with periods between 4–30 days.Strong oscillations centered around 5 days and 18 days seem to dominate in the upper troposphere and lower stratosphere at all the three stations. While 5 day wave has an amplitude of about 2 m/s, the 18 day wave has an amplitude between 8–10 m/s in the zonal and 5–6 m/s in meridional component around tropopause. Its amplitude is maximum over Hyderabad and decreases somewhat on either side i.e. over Trivandrum and Bhubaneshwar. Weekly rocket wind data from Balasore near Bhubaneshwar show that 18–20 day wave continues to propagate vertically in the altitude range of 30–60 km. Temperature data also exhibits similar oscillations with amplitude of about 1 K for 4–5 day wave and 2–3 K for 18 day wave maximising just above tropopause ( 18 km).It is found that some of the observed wave modes, particularly the 18 day wave have characteristics matching those of forced Rossby wave rather than Kelvin wave while the 5 day and 9 day waves have characteristics matching those of mixed Rossby-gravity waves. The latter may be generated due to convective forcing in the troposphere while the former may be penetrating from the midlatitudes.With 15 Figures     

Spatio‐temporal variability in a mid‐latitude ocean basin subject to periodic wind forcing

Abstract

The mid‐latitude ocean's response to time‐dependent zonal wind‐stress forcing is studied using a reduced‐gravity, 1.5‐layer, shallow‐water model in two rectangular ocean basins of different sizes. The small basin is 1000 km × 2000 km and the larger one is 3000 km × 2010 km; the aspect ratio of the larger basin is quite similar to that of the North Atlantic between 20°N and 60°N. The parameter dependence of the model solutions and their spatio‐temporal variability subject to time‐independent wind stress forcing serve as the reference against which the results for time‐dependent forcing are compared.

For the time‐dependent forcing case, three zonal‐wind profiles that mimic the seasonal cycle are considered in this study: (1) a fixed‐profile wind‐stress forcing with periodically varying intensity; (2) a wind‐stress profile with fixed intensity, but north–south migration of the mid‐latitude westerly wind maximum; and (3) a north–south migrating profile with periodically varying intensity. Results of the small‐basin simulations show the intrinsic variability found for time‐independent forcing to persist when the intensity of the wind forcing varies periodically. It thus appears that the physics behind the upper ocean's variability is mainly controlled by internal dynamics, although the solutions’ spatial patterns are now more complex, due to the interaction between the external and internal modes of variability. The north–south migration of wind forcing, however, does inhibit the inertial recirculation; its suppression increases with the amplitude of north–south migration in the wind‐stress forcing.

Model solutions in the larger rectangular basin and at smaller viscosity exhibit more realistic recirculation gyres, with a small meridional‐to‐zonal aspect ratio, and an elongated eastward jet; the low‐frequency variability of these solutions is dominated by periodicities of 14 and 6–7 years. Simulations performed in this setting with a wind‐stress profile that involves seasonal variations of realistic amplitude in both the intensity and the position of the atmospheric jet show the seven‐year periodicity in the oceanic circulation to be robust. The intrinsic variability is reinforced by the periodic variations in the jet's intensity and weakened by periodic variations in the meridional position; the two effects cancel, roughly speaking, thus preserving the overall characteristics of the seven‐year mode.     

Annual ENSO simulated in a coupled ocean–atmosphere model

Using an output from 200-year integration of the Scale Interaction Experiment of EU project-F1 model (SINTEX-F1), the annual ENSO reproduced in the coupled general circulation model is investigated, suggesting the importance of reproducing an annual cycle in realistically simulating ENSO events. Although many features of the annual ENSO are reproduced, the northward expansion of sea surface temperature anomaly (SSTA) in the eastern tropical Pacific stays south of the equator. It is suggested that this model bias is due to the excitation of the too strong Rossby waves in the southeastern tropical Pacific, which reflect at the western boundary and intrude into the eastern equatorial Pacific. The zonal wind stress anomaly along the equator also plays an important role in generating the equatorial Kelvin waves. The amplitude of SSTA for the annual ENSO mode is reproduced, but its variance is only 20% of the observation; this is again due to the lack of northward migration of seasonal SSTA in the equatorial region and weaker coastal Kelvin waves along South America. Remedies for the model bias are discussed.     

Frequency and directional evaluation of the ocean data gathering program (ODGP) wave spectra at Hibernia

Abstract

The frequency and directional wave‐modelling capability of the Ocean Data Gathering Program (ODGP) deep water spectral wave model is assessed through comparison with WAVEC data gathered at Hibernia. Both qualitative and quantitative analyses indicate better agreement with observations during storms and with the wind‐driven component of the wave spectra. There is statistically poor modelling of the swell. A coherence analysis on derived wave vectors indicates that the ODGP model does not simulate geophysical variability with time‐scales less than about 30 h for overall spectral energy and less than 24 h for wave energy of frequency greater than 0.6 rad s^?1 (0.095 Hz). The signals associated with swell waves are incoherent at nearly all time‐scales.     

Decadal and long-term sea level variability in the tropical Indo-Pacific Ocean

In this study, we analysed decadal and long-term steric sea level variations over 1966–2007 period in the Indo-Pacific sector, using an ocean general circulation model forced by reanalysis winds. The simulated steric sea level compares favourably with sea level from satellite altimetry and tide gauges at interannual and decadal timescales. The amplitude of decadal sea level variability (up to ~5 cm standard deviation) is typically nearly half of the interannual variations (up to ~10 cm) and two to three times larger than long-term sea level variations (up to 2 cm). Zonal wind stress varies at decadal timescales in the western Pacific and in the southern Indian Ocean, with coherent signals in ERA-40 (from which the model forcing is derived), NCEP, twentieth century and WASWind products. Contrary to the variability at interannual timescale, for which there is a tendency of El Niño and Indian Ocean Dipole events to co-occur, decadal wind stress variations are relatively independent in the two basins. In the Pacific, those wind stress variations drive Ekman pumping on either side of the equator, and induce low frequency sea level variations in the western Pacific through planetary wave propagation. The equatorial signal from the western Pacific travels southward to the west Australian coast through equatorial and coastal wave guides. In the Indian Ocean, decadal zonal wind stress variations induce sea level fluctuations in the eastern equatorial Indian Ocean and the Bay of Bengal, through equatorial and coastal wave-guides. Wind stress curl in the southern Indian Ocean drives decadal variability in the south-western Indian Ocean through planetary waves. Decadal sea level variations in the south–western Indian Ocean, in the eastern equatorial Indian Ocean and in the Bay of Bengal are weakly correlated to variability in the Pacific Ocean. Even though the wind variability is coherent among various wind products at decadal timescales, they show a large contrast in long-term wind stress changes, suggesting that long-term sea level changes from forced ocean models need to be interpreted with caution.     

Some aspects of internal gravity waves in the multicell-type convective system

Summary Some aspects of internal gravity waves in the multicell-type convective system are examined using a linear theory and a nonlinear numerical model. The basic-state wind is assumed to increase linearly with height and then remain constant.In the theoretical part, the two-dimensional, linear, steady-state response of a stably stratified atmosphere to specified cooling representing the evaporative cooling of falling precipitation in the subcloud layer is analytically considered. It is shown that there exist an updraft on the upstream side of the cooling and a downdraft on the downstream side. As the wind shear increases enough, the magnitude of the updraft decreases. This is because a large portion of the specified cooling is used to compensate for the positive vorticity associated with the positive wind shear and accordingly the effective cooling necessary to produce perturbations is reduced.In the numerical part, a two-dimensional version of the ARPS (Advanced Regional Prediction System) that is a nonhydrostatic, compressible model with detailed physical processes is employed. Results from the dry simulation, in which the steady cooling is specified in the model, show that the simulated quasi-steady field resembles the linear, steady-state solution field because the nonlinearity factor of thermally-induced waves in this case is small. For the moist simulation, the quasi-steady perturbations obtained from the dry simulation are used as initial conditions. It is shown that gravity waces can effectively initiate convection even with small amplitude and that updraft at the head of the density current somewhat resembles the linear, steady-state response of a stably straified flow to the specified cooling. The updraft, that is, forced internal gravity waves, at the head of the density current is responsible for the initiation of consecutive convective cells that move downstream and develop as a main convective cell. This study suggests that internal gravity waves play a major role in the initiation of consecutive convective cells in the multicell-type convective system and hence in its maintenance.     

El Nio事件发生和消亡中热带太平洋纬向风应力的动力作用 Ⅱ.模式结果分析

为了分析 ＥＩ Ｎｉｏ事件发生和消亡中热带太平洋纬向风应力的动力作用，建立一个类似于Ｚｅｂｉａｋ的简单热带海洋数值模式，在观测到的风应力异常的强迫下，模拟赤道太平洋地区 １９７１年１月至 １９９８年８月海表温度异常的变化。结果表明，模式对观测的Ｎｉｏ３区海表温度异常（ＳＳＴＡ）有很好的模拟能力。模拟和观测Ｎｉ区ＳＳＴＡ之间的相关系数可达 ０．９０。模式对 Ｅｌ Ｎｉｏ事件期间赤道太平洋海表温度异常随时间变化也有较好的模拟能力。为了分析Ｅｌ Ｎｉｏ期间ＳＳＴＡ的空间分布及其随时间变化的动力学机制，还对１９８６～１９８９年 ＥＮＳＯ循环期间赤道太平洋地区观测的 ＳＳＴＡ的传播特征及其形成机制进行了分析。模式较好地模拟出了观测到的赤道太平洋地区ＳＳＴＡ的传播特征，即从１９８６年底至１９８７年 ４月， ＳＳＴＡ具有向东传播的特征，从 １９８７年 ６月至 １９８８年 ２月具有向西传播的特征。动力学分析的结果表明，赤道中西太平洋地区的缔向风应力异常对 Ｅｌ Ｎｉｏ事件的发生和消亡具有重要作用。赤道中西太平洋地区的西风异常可强迫出东传的Ｋｅｌｖｉｎ波，这个东传的 Ｋｅｌｖｉｎ波对正 ＳＳＴＡ的东传起主要作用，当这个东传的 Ｋｅｌｖ     

Vorticity waves in a shallow basin

Observed oscillatory current patterns in the southern basin of Lake Michigan, with a distinctive peak in the energy spectrum at a period of about 90 h, are simulated using a linear potential vorticity conservation model. Solutions of the forced vorticity equation in a paraboloidal basin show rotational, oscillatory motions tuned to the low-frequency topographic modes that are very similar to the observed flow patterns. Topography-controlled vorticity waves are excited most effectively by wind episodes with frequency nearly in resonance with the topographic modes. Bottom resistance has no significant effect on the frequency equation; it simply decays the waves slowly in the open lake and more quickly near the coast. Flow patterns of both the gravest free vorticity wave and the corresponding forced wave consist of two opposite circulation cells separated by a null streamline through the center of the basin and rotating cyclonically near the free wave and atmospheric forcing frequencies, respectively. Interactions between the forced and free waves result in an apparent rotational pattern with a frequency the median of the two. A combination of elliptic—paraboloidal basin and shorter period forced modes can approximate the observed Lake Michigan response. Doppler shift, due to the persistence of cyclonic vorticity in the flow field, is also determined to be a factor in shifting the elliptical basin mode to a higher frequency.