Variable retention is a new approach to harvesting and silvicultural systems that was developed by ecologists in the Pacific Northwest region of North America to address a wide array of forest management goals. Variable retention recognizes that natural disturbances, such as fire, wind or disease, nearly always leave some standing structure from the original forest. This structural complexity plays an important role in forest ecosystem function and biological diversity. A new "retention silvicultural system" was defined that leaves trees distributed throughout harvested areas. This system facilitates retention of structural features of old-growth forests, such as live and dead trees of varying sizes, multiple canopy layers, and coarse woody debris. Weyerhaeuser's British Columbia Coastal Group will use the variable retention approach for all harvesting by 2003. More than 75% of the company's coastal harvesting in British Columbia used variable retention in 2001. Company guidelines describe the amount, type, and spatial distribution of retention for groups and individual trees. An adaptive management program is monitoring the amount and type of structural attributes retained in relation to the original forest.Key words: old-growth forests, variable retention, silvicultural systems, biodiversity, landscape zoning La conservation variable est une nouvelle approche aux systèmes de récolte et de sylviculture que des écologistes de la région du Nord-Ouest du Pacifique, en Amérique du Nord, ont mis au point pour atteindre un vaste ensemble d'objectifs pour l'aménagement forestier. La conservation variable reconnaît que des facteurs naturels de perturbation tels que le feu, le vent ou la maladie laissent presque toujours sur leur passage une structure résiduelle de la foret antérieure. Cette complexité structurelle joue un rôle important dans la fonction d'écosystème de la forêt et dans sa diversité biologique. On a défini un nouveau « système sylvicole de conservation » qui laisse subsister des arbres distribués sur les parterres de coupe. Ce système facilite la conservation des caractéristiques structurales des vieilles forêts telles que des arbres vivants et morts de tailles variables, des étages multiples du couvert et des débris ligneux grossiers. Chez Weyerhaeuser, le groupe « BC Coastal » appliquera cette approche à toute la récolte, dès 2003. Pour plus de 75 % de la récolte pratiquée dans la zone côtière de la Colombie-Britannique, cette société employait la conservation variable en 2001. Les lignes directrices que s'est données la compagnie décrivent, pour des groupes d'arbres et des arbres individuels, la quantité, le type et la répar-tition spatiale de la conservation. Un programme d'aménagement adaptatif permet de contrôler la quantité et le type de caractères structuraux conservés relativement à ceux de la forêt antérieure.

Stand-level retention is an important component of sustainable forest management which aims to balance ecological, social and economic objectives. Long-term retention of mature forest structures at the time of harvesting (variable retention) is intended to produce future forest stands that more closely resemble conditions that develop after natural disturbances, thereby maintaining greater diversity of habitats for a variety of organisms. Structure includes features such as live and dead trees representing multiple canopy layers, undisturbed understory vegetation and coarse woody debris. Over the past two decades, variable retention has become common on forest lands in the temperate rainforests of coastal British Columbia (BC) and has been applied to a lesser extent in inland forest types. Our review of studies in BC and in similar forest types in our region indicates that both aggregated and dispersed retention can contribute to biodiversity conservation by providing short-term 'life-boating' habitat for some species and by enhancing the structural characteristics of future stands. For example, greater abundance of species present in the pre-harvest forest have been documented for vegetation, birds, carabid beetles, gastropods, ectomycorrhizal fungi and soil fauna in retention cutblocks compared to clearcuts. There are, however, some negative consequences for timber production such as wind damage to retained trees and reduced growth rates of tree regeneration compared to clearcuts. The authors suggest an adaptive management approach for balancing competing objectives when faced with uncertainty. This includes monitoring the implementation and effectiveness of various strategies for achieving goals. Over two decades of experience applying variable retention harvesting to industrial-scale management of forest lands in BC suggests that it is possible to balance production of wood with biodiversity conservation.

Genetic diversity and mating system were quantified for shelterwood, patch cut and green tree-retention silvicultural systems, and compared to adjacent old-growth. This is a component of a larger study conducted in montane old-growth forests of coastal British Columbia to evaluate the feasibility and ecological consequences of alternative silvicultural systems. The experiment includes replicated treatments representing a range of overstory removal adjacent to old-growth and clearcut areas. Based on 22 electrophoretically assayed loci, the effects of silvicultural systems on genetic parameters of amabilis fir (Abies amabilis and western hemlock (Tsuga heterophylla were assessed by comparing an average number of alleles per locus, the percent polymorphic loci, and observed and expected heterozygosity between parental populations and naturally regenerated progeny as well as among treatments. Genetic variation in natural regeneration was greater than in parental populations, especially for low-frequency alleles. Silvicultural treatments caused no significant differences in amabilis fir genetic-diversity parameters, while the shelterwood system resulted in lower observed and expected heterozygosity in western hemlock. Nei's genetic distance revealed that all parental populations were extremely similar. The two species had contrasting mating system dynamics with amabilis fir producing higher levels of correlated paternity and inbreeding with wider variation among individual tree outcrossing-rate estimates. Western hemlock had significant levels of correlated paternity only for the green tree and shelterwood treatments demonstrating family structuring inversely related to stand density. Inbreeding in western hemlock was significant but lower than that observed for amabilis fir with a J-shaped distribution for individual tree multilocus outcrossing-rate estimates. The pollination and dispersal mechanisms of the two species represent the most-likely factors causing these differences. Artificial regeneration may be utilized to augment the genetic resources of natural ingress.

As part of the Montane Alternative Silviculture Systems (MASS) project, this study investigates limits on the growth of montane conifers resulting from varying overstory retention under conventional and alternative silvicultural systems. After harvesting treatments were complete in 1993, Abies amabilis (amabilis fir) and Tsuga heterophylla (western hemlock) seedlings were spring planted in replicated blocks of shelterwood (SW), patch cut (PC), green tree retention (GT) and clearcut (CC) systems. In addition, sub-plots were established within each silvicultural system in which fertilization (at planting) and vegetation control post-planting treatments were applied alone and in combination to test the extent to which growth limitations are related to nutrient availability and vegetative competition. The impact of overstory retention was most pronounced in the reduced light environment of the SW where height growth after seven years was 26-30% lower in both species compared to the untreated CC, GT and PC systems. Although the effect on growth of both species in the SW was mitigated somewhat by fertilization and vegetation control treatments, amabilis fir did not attain free-to-grow height (1.3 m) regardless of post-planting treatment. Time to free-to-grow height in the more open silvicultural systems was reduced in both fir and hemlock with fertilization and vegetation control alone and in combination, except in the CC where the initial growth response to fertilization alone was diminished by the end of the sevenyear study. In contrast to fertilization, the effects of vegetation control on height growth were not apparent until three to five years and seven years after planting in the CC and GT, PC, SW, respectively. Combining vegetation control and fertilization had an additive effect on growth in amabilis fir but not in western hemlock. The effect of silvicultural systems and post-planting treatments on the two species illustrate that both above-and below-ground resource availability (light and nutrients) availability was potentially limiting to growth, particularly in the shelterwood treatment.Key words: MASS, silviculture systems, regeneration, Tsuga heterophylla, Abies amabilis, shelterwood, patch cut, green tree retention, clearcut, fertilizer, vegetation control En tant qu'élément du projet sur les systèmes alternatifs de sylviculture dans la zone montagneuse, cette étude porte sur la réduction de la croissance des conifères alpins à la suite d'une rétention variable du couvert forestier selon des systèmes sylvicoles conventionnels et alternatifs. Suite aux travaux de récolte complétés en 1993, des semis d'Abies amabilis (sapin amabilis) et de Tsuga heterophylla (pruche de l'Ouest) ont été plantés au printemps dans des blocs d'une étude des régimes de jardinage, de coupe partielle, de coupe par troué, de coupe avec rétention d'arbres verts et de coupe à blanc. De plus, des sous-parcelles ont été établies dans chaque régime sylvicole pour lesquels des traitements de fertilisation (au moment de la plantation) e...