ECOLOGICAL FACTORS AFFECTING ABUNDANCE OF ADVANCED GROWTH IN JACK PINE (PINUS-BANKSIANA LAMB) STANDS OF THE BOREAL FOREST OF NORTHWESTERN QUEBEC

Natural regeneration in jack pine (Pinus banksiana Lamb.) stands in Qu ebec is only sufficient 4% of the time and up to only 33% of the time in mixed stands. This study evaluates the usefulness of forest ecologi cal types as defined by the Quebec Ministry of Forests in predicting a bundance of advanced growth in pure and mixed jack pine stands of the Abitibi region, in northwestern Quebec. Trees above 1.3 m in height an d up to 5 cm DBH were tallied in 102 quadrats of 20 x 20 metres to eva luate advanced growth densities. No significant difference in advanced growth densities was observed between the two main types of surficial geological deposit in the region, lacustrine clays and glacial tills. Only shallow till and organic deposits over bedrock and fluvio-glacia l sands showed significantly higher advanced growth densities. Predict ion can be improved by using ecological variables such as soil texture , moisture regime, distance from a seed source and stand composition. Abundance of black spruce (Picea mariana (Mill.) B.S.P.) regeneration is positively associated with the proportion of sand in the soil profi le and with drier sites; advanced growth of balsam fir (Abies balsamea (L.) Mill.) and white spruce (Picea glauca (Moench) Voss.) is weakly associated with the distance to a fire-preserved zone containing seed trees. Abundance of pin cherry (Prunus pensylvanica L. f.), probably a ssociated with canopy openings, is positively linked with abundance of balsam fir and white spruce advanced growth. Jack pine regeneration b y means of advanced growth occurs almost exclusively in jack pine stan ds on sand: higher proportions of hardwoods in the overstory appear to be negatively linked to jack pine advanced growth. Although the site variables studied showed some significant relationships with advanced growth abundance (maximum R2= 0.32) which were slightly improved with stand composition variables (maximum R2 = 0.38), no satisfactory predi ctive model could be implemented.