Importance of seasonal growth patterns in modelling interactions between radiata pine and some common weed species

Abstract

Previously described weed competition models that use competition indices based on weed size relative to tree size (e.g., tree height divided by weed height) require models of weed growth and models of weed-free tree growth. A method is presented to model these using standard sigmoidal growth functions and to incorporate regular seasonal patterns of tree growth using Fourier series approximations. The method is tested against data from a field trial at Rotorua, New Zealand, in which Pinus radiata D. Don was grown both on its own and in competition with several common weed species. Weed and tree height were best modelled by a Weibull function, and tree ground line diameter by a Schumacher function. Seasonal fluctuations in both tree and weed growth were adequately modelled by a single-term Fourier series. All weed species showed very similar, strong seasonal fluctuations in height growth, peaking in February. Seasonal fluctuations in tree height growth, and especially diameter growth, were less marked and peaked earlier, in December. Simulations suggested that it is necessary to account for seasonal effects when modelling competition during the first year of growth, but that seasonal effects have less influence in subsequent years.