A method of analysis, initially developed as a method of distinguishing between chaotic dynamics and observational noise in time series, was applied to transects of chlorophyll a from the Southern Ocean. The algorithm works by predicting the behaviour of a data set based on patterns present within the data. This differs from previous analyses in that it permits classification of the dynamics governing the system. In a chaotic system, predictions become exponentially poor as one attempts to predict further ahead, due to the sensitivity of the system to initial conditions. However, in a system governed by stochastic noise, no such deterioration in predictions is observed. This work represents the first application of the algorithm to oceanic transect data, and our results show a less than exponential decline in predictive ability. The behaviour of the prediction curves is closely related to that of the corresponding autocorrelation functions, indicative of a stochastic, but spatially coherent data set with significant positive autocorrelations. Furthermore, successful prediction is correlated with the sum of the chlorophyll power spectrum. We conclude that the predictive ability of the algorithm is greatest when spatial variation in the chlorophyll transects is high.