highly significantP < 0.01
Because of the limited amount of independent data as well as the high number of tests on the same set of plots, the following model restrictions were imposed on models that included extrinsic and intrinsic risk factors: (i) no interaction terms among the main explanatory variables were fitted as adding these to the model would lead to (almost) saturated models and reductions in statistical power for each of the tests; and (ii) only those variables that were highly significant ( P < 0.01) were considered to be the main result in the Discussion section and Abstract.