Posteriori adjustments

Another notorious problem of the coupled model is its tendency to drift to the states of atmosphere and ocean that are unlike the present states during a long time integration. In the noncoupled system, the bormdaiy conditions act as an anchor to the long-term trend of the model and the system cannot drift too far away from reality. By contrast, in the coupled system, because of the imperfection of the system, both the atmospheric and ocean models tend to drift toward their equilibrium states, which are sometimes far away fiom reality. This is called climate drift and is a cousin of the spin-up problem. In order to deal with this climate drift problem, techniques have been developed. They are based on the practice of flux correction, which makes a posteriori adjustments of various fluxes at the air-sea interface in order to prevent climate drift. Unfortunately, the correction terms are not necessarily small compared with the fluxes themselves. Therefore, the need of flux correction is indicative of the shortcoming in handling the interface conditions. 

Several procedures have been proposed to avoid, or at least moderate, this effect (Boys Bernardi 1970, Daudey et al. 1974b, Mayer 1983). The problem stimulated a lot of controversy (see, e.g. Gutowski et al. 1986, Collins Gallup 1986)—we shall not jump into this jungle here, since no unique scheme has yet been accepted. Most schemes imply an a posteriori adjustment of the interaction energy. For the present purpose, that is in order to derive an expression for the interaction operator, we shall take advantage of the a priori analysis of Mayer (1983) followed in his Chemical Hamiltonian Approach (CHA). 

A pure phenomenological model of such an intricate process, taking into account all possible reaction steps, is therefore a powerful tool for the scale up and the prediction of performances of trickle-bed reactors. Such a model (20 has proved to be able to correctly reproduce experimental data using only two adjustable parameters. It has been checked in several cases (hydrogenation of alphamethylstyrene (3J, hydrogenation of 2-butanone (, hydrorefining (J5) ), with more or less volatile liquid reactants and it appeared to be also useful to calculate a posteriori the extent of the different types of wetted catalyst area and their different effectiveness factor. 

The next two types of EOS represent a more practical a posteriori approach to C-J calculations. The form of the EOS for an imperfect gas is assumed, and parameters appearing in it are adjusted to reproduce (for a few prototype explosives) experimental D-po curves, and sometimes also C-J pressures. C-J calculations are then performed on related explosives in the hope of achieving good agreement with experimental values. 

---