Prediction of Site Index and Apple Rootstock Performance from Environmental Variables

Abstract

Previously, we developed stability analysis models for nine rootstocks tested over 19 apple prroducing states and demonstrated significant rootstock-site interactions for cumulative yield (CY) and trunk cross-sectional-area (TCSA). A key input in these models is site index (SI), estimated from the mean over rootstocks within site. Our goal was to extend the usefulness of these models by developing further models to estimate SI for untested sites from climate variables. Prediction of SI from mean daily maximum temperature (T) and total moisture (M) (sum of precipitation and irrigation) was evaluated over five periods based on previous work and on approximate phenological phases of the apple tree over the geographic area included in this study: A = December-January (dormant), B = February-April (prebloom), C = May-June (set), D = July-September (growth to harvest), and E = October-November (postharvest to leaf senescence), resulting in ten climate-time variables for model development. Complete records of T and M were available from 11 and 9 states, respectively. SI _TCSAwas not significantly correlated with any T or M variable and therefore may have been influenced more by soil factors or deviations in orchard management practices than by climate. SI _CYwas well correlated with T _C, T _Dand M _C, but T _Cand M _Cwere strongly collinear. Further model development concentrated on prediction of SI _CYfrom T variables over data from ten states (with California removed as an outlier). All one-, two-, and three-variable multiple regression models were evaluated. We concluded that SI _CYwas best predicted from a linear relationship with T _C.