Modelling the influence of weed competition on growth of juvenile Pinus radiata at a dryland site
Degree GrantorUniversity of Canterbury
Degree NameDoctor of Philosophy
The reduction in growth of commercially grown conifers by weeds over the juvenile phase is an important management issue. This thesis investigates the mechanisms of interaction between juvenile Pinus radiata and the woody weed broom (Cytisus scoparius L.) at a dryland site, and uses this information to develop processbased models of tree growth, sensitive to competition from weeds. All the results presented in this thesis were taken from an experiment established on a drought prone site, located on the Canterbury Plains, New Zealand. Within the experiment a comprehensive and detailed set of measurements were taken from trees growing weedfree and with broom competition, from the beginning of the second year to the end of the third year after planting. The presence of broom severely retarded the growth of trees over the two year measurement period. Treatment differences in basal area, which became significant one month after measurements started, increased from six-fold at the end of the first year to 12-fold after two years. At the end of the second year, trees growing without broom were two-fold taller, and had a crown diameter four-fold greater than trees in plots with broom. Above-ground biomass was the most sensitive indicator of competition. Compared to trees in weed-free plots, above-ground biomass of trees in plots with broom was reduced by 25-fold after two years. During the first year, the strong link between predawn needle water potential, ѱe, and fractional biomass growth for trees in plots with broom (r²=0.91) suggests growth losses in this treatment were almost entirely attributable to seasonal water deficit. Although ѱe was also strongly related to fractional biomass growth, 𝑓b, over the second year (r²= 0.71), values of 𝑓b across a comparable range of ѱe, were on average 15% lower than during the first year. Inclusion of a simple shading index explained this difference, suggesting broom restricted tree growth during the second year through competition for both light and water. Nitrogen fixation by broom, and the extent of fixed nitrogen transfer to Pinus radiata was assessed from measurements of the natural abundance of the ¹⁵N isotope, 𝛿¹⁵N. Values of 𝛿¹⁵N were highest in Pinus radiata growing without broom, intermediate in Pinus radiata growing with broom, and lowest in field grown broom. Broom was an effective fixer, deriving 81 % of nitrogen in above-ground tissues from the atmosphere, which was equivalent to 111 kg N ha⁻¹ yr⁻¹. The intermediate values of 𝛿¹⁵N found for trees growing with broom suggest that there was some transfer of fixed nitrogen from the broom to the Pinus radiata. However as soil uptake by the broom (29 kg N ha⁻¹ yr⁻¹) considerably exceeded estimated rates of nitrogen transfer (2 kg N ha⁻¹ yr⁻¹) growth of trees was not enhanced by this recycling of fixed nitrogen. Compared with Pinus radiata in weed-free plots, trees growing with broom allocated dry matter preferentially to stems (+8%) at the expense of foliage (-4%), roots (-2%) and branches (-2%). These findings on allocation confirm derivations based on pipe model theory and indicate that the root to shoot ratio in juvenile Pinus radiata remains relatively stable even under severe water deficit. A water balance model was developed to partition water loss and explain treatment differences in tree basal area growth, over the first year. In both treatments, modelled root-zone water storage (W) corresponded closely to measured values over the course of the year. Total annual evaporation in the treatment with broom was 25% higher than that for the treatment without broom. This higher rate was primarily due to wet canopy evaporation from the broom and, to a lesser extent, from increased rates of total transpiration within this treatment. The integral of root-zone water deficit was strongly related to fractional basal area growth of trees in plots with broom. These findings highlight the importance of root-zone water storage in regulating productivity on dryland sites and indicate that water balance modelling provides a useful approach for predicting the influence of weed competition on tree growth. A generally applicable model of juvenile tree growth was developed and tested at this dryland site. This model predicts tree growth by reducing growth from an empirically determined optimum rate (weed-free) using a seasonally estimated competition modifier, which accounts for the degree of weed competition for both water and light. Test results showed good correspondence between measured and predicted tree diameter growth in treatments with and without broom, over the two year period. As this model was developed using process-based modifiers for water and light it is applicable to sites covering a wide range of climatic and edaphic conditions.