The most important hydrogeological finding to come out of the field investigations carried out during the various phases of test pumping of Stage 1 of the Thames Groundwater Scheme was the realisation that the magnitude of transmissivity (T) and storativity (S) decline significantly with depth (Owen et al. 1977). Bearing in mind that the purpose of the scheme is to abstract ground-water at times of drought, the main implication of this finding is that aquifer parameters are likely to be appreciably lower under operational conditions than under normal seasonal conditions. The author employs the Nutbrown inverse method (Nutbrown 1975) to compute the areal distribution of T and S which are then input to the groundwater model. This particular inverse technique is based on a cyclic time integration of the conventional linear groundwater flow equation, consequently, T and S values computed by this method will necessarily correspond to some average seasonal groundwater stage condition. Morel's ‘non-linear’ model is so called because the transmissivity parameter is represented as: T = K a b (1) where K a = K a ( x , y ) ... average hydraulic conductivity, b = b( x , y , h)... saturated aquifer thickness, and h = h( x , y , t) ... groundwater head K a is obtained by dividing the T value computed from the inverse method by 50 m (assumed average aquifer thickness). Storativity is a function of areal position only as computed by the inverse technique. Essentially, the author is modelling the unconfined Chalk aquifer as though it were a vertically homogenous porous medium. Equation (1)