Enhanced tools for predicting annual stone pine (Pinus pinea L.) cone production at tree and forest scale in Inner Spain

Aim of the study: To present a new spatiotemporal model for Pinus pinea L. annual cone production with validity for Spanish Northen Plateau and Central Range regions. The new model aims to deal with detected deficiencies in previous models: temporal shortage, overestimation of cone production on recent years, incompatibility with data from National Forest Inventory, difficulty for upscaling and ignorance of the inhibitory process due to resource depletion. Area of study: Spanish Northern Plateau and Central Range regions, covering an area where stone pine occupies more than 90,000 ha Material and methods: Fitting data set include 190 plots and more than 1000 trees were cone production has been annually collected from 1996 to 2014. Models were fitted independently for each region, by means of zero-inflated log normal techniques. Validation of the models was carried out over the annual series of cone production at forest scale. Results: The spatial and temporal factors influencing cone production are similar in both regions, thus the main regional differences in cone yield are related with differences in the phenological timing, the intensity of the influent factors and forest intrinsic conditions. A significant inhibition of floral induction by resource depletion was detected and included into the model. Upscaling the model results in accurate prediction at forest scale. Research highlights: [1] The new model for annual cone production surpass the detected deficiencies of previous models, accurately predicting recent decay in cone production; [2] Regional differences in cone production are due to phenological and seasonal climatic differences rather than to between provenances genetic differences Keywords: zero-inflated models; pine nut; conelet losses; Leptoglossus occidentalis; forest upscaling