Link to article
Site productivity, assessed through site index, was modelled using partial least squares regression as a function of soil and climatic variables. Two alternative models were developed: a full model, considering all available explanatory variables, and a reduced model, considering only variables that can be obtained without digging a soil pit. The reduced model was used for mapping the site index distribution in Portugal, on the basis of existing digital cartography available for the whole country. The developed models indicate the importance of water availability and soil water holding capacity for site index value distribution. Site index was related to climate, namely evaporation and frost, and soil characteristics such as lithology, soil texture, soil depth, thickness of the A horizon and soil classification. The variability of the estimated values within the map (9.5–16.8 m with an average value of 13.4 m) reflects the impact of soil characteristics on the site productivity estimation. These variables should be taken into consideration during the establishment of new plantations of cork oak, and management of existing plantations. Results confirm the potential distribution of cork oak in coastal regions. They also suggest the existence of a considerable area, located both North and South of the Tagus river, where site indices values of medium (]13;15]) to high (]15;17]) productivity classes may be expected. The species is then expected to be able to have good productivity along the northern coastal areas of Portugal, where presently it is not a common species but where, according to historical records, it occurred until the middle of the sixteenth century. The present research focused on tree growth. Cork growth and cork quality distribution needs to be further researched through the establishment of long term experimental sites along the distribution area of cork oak, namely in the central and northern coastal areas of the country.