The instability of the atmosphere places an upper bound on the predictability of instantaneous weather patterns. The skill with which current operational forecasting procedures are observed to perform determines a lower bound. Estimates of both bounds are obtained by comparing the ECMWF operational forecast for each day of a 100-day sequence at one range with the operational forecast for the same day at another range, and with the analysis for that day. The estimated bounds are reasonably close together.

Predictions at least ten days ahead as skilful as predictions now made seven days ahead appear to be possible. Additional improvements at extended range may be realized if the one-day forecast is capable of being improved significantly.

A seasonal, zonally averaged, climate model is developed using a hypothetical meridional circulation, which reproduces the atmospheric dynamical heating of a two-level general circulation model. In order to simulate the seasonal temperature cycle, prognostic equations for the land and ocean temperatures are included in the model. The horizontal heat transport in the ocean is modelled as ordinary diffusion. Small-scale vertical transport of sensible and latent heat is described by simple linear terms. For the long-wave radiation calculations, a simple but efficient emissivity approximation scheme is developed. The short-wave radiation treatment is based on an early version of the Mintz-Arakawa general circulation model.

The meridional and seasonal variation of the surface temperature as well as the temperature at the two tropospheric levels are simulated quite well by the model. Concerning surface albedo and planetary albedo, the model gives results which are in good agreement with observed values. Radiative fluxes of the model also compare well with observed values.

The sensitivity of the model to changes in incoming solar radiation and carbon dioxide content is in accordance with results obtained from other models. The model, with variation of the Earth's orbital parameters, has also been used for some preliminary experiments.

Northern Hemispheric twice daily data from the First GARP Global Experiment (FGGE) year are used to compare available potential energy estimates obtained using the so-called exact (isentropic) form of the equations with two commonly used approximate (isobaric) forms. The two approximate forms, one using a fixed lapse rate and the other a lapse rate which varies in the vertical, overestimate the total available potential energy except in mid-summer.

The greatest difference between the exact and approximate forms is in the eddy term, which is considerably smaller in the approximate forms, especially in spring and summer. Partitioning the eddy term into its standing and transient components shows that in the exact form the standing component exceeds the transient term. This is in contrast to the approximate form in which the transient term is the larger in every month except July. The latitudinal and monthly variations of the standing and transient terms show different patterns, especially in the standing component. In the approximate form the latitude of maximum standing eddy values is near 60° N in winter but shifts equatorward to about 30° N and decreases in summer. While the exact form shows a weak winter maximum near 60° N, the dominant maximum develops near 20° N in spring and shifts poleward to about 30° N and increases in summer.

The hemispheric distribution of the vertically integrated grid point values from which the eddy available potential energy was calculated and the standard deviation of these twice daily values produce results which show that the exact form values better conform to the pattern of cyclone activity than the approximate form. In general, the results indicate that the exact form of the available potential energy equations should be more extensively used in diagnostic work.

The effects of non-zero mass balance, often encountered in diagnostic studies using observed data, are discussed for atmospheric energy budget equations written in sigma coordinates. An example of these effects is given by comparing the annual sensible heat budget over North America calculated from ECMWF grid point analyses and operational forecasts. The systematic height errors of the model appear as a local mass imbalance.

An analysis of surface meteorological data, recorded by the West-European synoptic network in June and July 1976, reveals the existence of a diurnal oscillation of surface geostrophic wind in this part of Europe. The oscillation is similar to that observed in the Great Plains of the United States by several investigators.

The analysis delineates physical factors involved in this diurnal oscillation of geostrophic wind. As found by others, the oscillation is induced partly by the diurnal temperature cycle over sloping terrain. Our analysis suggests that a diurnal variation of the horizontal gradient of the temperature field in the atmospheric boundary layer is another major factor contributing to the observed oscillation of geostrophic wind.

A mesoscale mixed-layer model of the planetary boundary layer is applied to the flow past a mountain island during a cold air outbreak over the Kuroshio Current. Utilizing data taken during AMTEX '75, the governing equations are integrated in time to simulate development of a Karmán vortex street downstream of the island of Cheju-do. The use of open lateral boundary conditions and an encroachment scheme which allows some mountain grid points to experience occasional encroachment of mixed-layer fluid as the mixed-layer deepens (and decroachment as the mixed layer thins) is also discussed.

Comparing various non-dimensional parameters on vortex-street characteristics, it is shown that the simulated vortex street resembles the observed atmospheric one rather well. Uncertainty in formulating the proper value of the eddy viscosity and the implication this has on Reynolds number comparisons are also discussed.

Pibal soundings taken during the period 1929 to 1936 were analysed to augment our knowledge of boundary layer climatology. Use of the geostrophic departure method enables computation of surface stress and the cross-isobar angle as functions of stability and baroclinicity, the latter being assessed from the upper wind shear. The comparative effects of both terms are discussed and the importance of the often-neglected baroclinicity in explaining the scatter of the computed parameters is stressed. Use of similarity theory enables us to achieve some further order in the data and eventually to estimate the universal functions *A* and *B* freed from baroclinic effects.

A case study of nocturnal drainage flow is presented. The flow descends a gentle slope of a few percent extending about one half kilometer down to Roskilde Fjord in northeast Denmark. Opposing larger scale flow initially delays the advance of the drainage flow. After several hours delay, the drainage flow arrives at the coast as a sudden relatively deep surge of cold air.

This surge of cold air significantly disturbs the immediate overlying flow. Even though the surge of cold air is characterized by high Richardson number and weak turbulent activity, this drainage front induces low Richardson number and significant turbulent activity in the overlying flow resulting in an “upside down” structure to the “boundary layer”. The drainage flow following the initial surge is thinner, weaker and nearly stationary except for occasional thicker pulses of cold air.

A similarity approach is utilized to investigate a simple axisymmetric steady-state model of the convergence region of a laboratory vortex. The resulting simplified set of equations are solved for a range of swirl angles by varying the tangential or radial velocity component at the outer rim. By increasing the swirl angle the flow is found to go from a one cell to a two cell configuration, i.e., the vertical velocity changes from everywhere positive to negative in the vicinity of the axis. Correspondingly the vertical vorticity maximum moves from the axis outward toward the radius of maximum tangential velocity, making the flow barotropically unstable with respect to unsymmetrical perturbations.