With the exception of the tropics, nowhere has the relationship between resource management and conservation of biological diversity been more controversial than in the Pacific Northwest region of the United States. Widespread loss and fragmentation of old—growth ecosystems have stimulated critical review and revision of existing forest management policies. However, studies of the consequences of forest management for plant species diversity are sorely lacking. We present data from permanent—plot and chronosequence studies in managed and unmanaged forests of western Oregon and Washington to describe the early responses of understory communities to forest harvest, and to suggest how post—harvest practices that alter natural successional processes may influence long term patterns of diversity and species occurrence. Permanent—plot studies of early succession in old—growth Pseudotsuga forests suggest that changes in understory diversity are fairly short—lived following clear—cut logging and slash burning. Populations of most vascular plant species recover to original levels prior to canopy closure. However, diversity may remain depressed for more than two decades on severely burned sites, and some species may experience local extinction. Evidence of the effects of post—harvest practices on vascular plant diversity is limited by an absence of community—level studies in older, managed forests. Chronosequence studies of natural forest stands indicate that, following canopy closure, vascular plant species diversity tends to increase with time, peaking in old growth. Few understory species are restricted to, or absent from, any stage of stand development (i.e., young, mature, or old growth). However, many species differ significantly in their abundance among stages. A majority of these showed greatest abundance in old growth. Changes in levels of resources (increased shade), changes in the spatial variability of resources and environments (increased horizontal and vertical heterogeneity), and species' sensitivity to fire and slow rates of reestablishment/growth may drive these trends during natural stand development. Silvicultural prescriptions that maintain or foster spatial and temporal diversity of resources and environments will be most effective in maintaining plant species diversity. Practices associated with intensive, short—rotation plantation forestry, that preclude or delay the development of old—growth attributes, may result in long—term loss of diversity. Ultimately, it may be necessary to manage some stands on long rotations (150—300 yr) to maintain understory species that require long periods to recover from disturbance.

Patterns of abundance were examined for vascular plant species during 21 yr of succession in two clear—cur and burned Pseudotsuga forests in the western Cascade Range of Oregon. A majority of forest understory species persisted through disturbance. Most colonizing species established within 2 yr after burning. Individualistic species responses were described by a series of broadly overlapping, unimodal curves of constancy and canopy cover, differing in time of initiation, duration, and magnitude. Thus, early successional change was characterized by gradual shifts in the abundance of generally persistent species. Eleven population patterns (species groups) were identified. Interactions of life history traits and disturbance explain the temporal trends of the most common species. Within the groups of invading species, the timing of initial establishment, as well as the timing and magnitude of peak abundance were related to the origin of propagules, phenological traits, potential for vegetative expansion, and temporal and spatial variation in disturbance. Abundance patterns of invading species were also influenced by stochastic and historical factors. Contrasting responses of species between sites reflected differences in histories of logging and slash burning. Within the groups of residual species, temporal patterns of abundance reflected initial species distributions, resistance to logging and burning disturbance, mode of reproduction, morphological traits, and spatial variation in disturbance intensity. These observations suggest that early secondary succesion in Pseudotsuga forests has a deterministic component, founded in the life history traits of the available species, and a stochastic component reflecting site history and variation in disturbance.

Vegetation changes were studied for 21 yr in two clearcut logged and slash—burned Pseudotsuga forests in the western Cascade Range of Oregon. Detrended correspondence analysis (DCA) was used to examine the successional relationships among six understory communities exposed to a gradient of disturbance intensity. Euclidean disturbances between pre— and postdisturbance samples in ordination space were used to compare community resistance to disturbance and long—term recovery, or resilience. Ordination through time for plant communities revealed a common pattern of rapid floristic change away from predisturbance composition, followed by gradual, unidirectional return. Early, but transient, convergence of successional pathways was common among mesic— and dry—site communities, reflecting the broad distribution of colonizers and the floristic similarity of predisturbance understories. Distinct sequences were observed on moist sites, reflecting more unique residual and colonizing floras. Ordinations also revealed increasing compositional change with disturbance intensity. Successional sequences were dominated by residual species on relatively undisturbed sites and by alternate suites of invading species on moderately disturbed and burned sites. Variation in the response gradient between watersheds reflected the modifying influence of local environment, stand history, and change in succession. Resistance and resilience varied little among plant communities but were generally lowest for the depauperate Coptis community and greatest for the compositionally and structurally diverse Polystichum and Rhododendrom—Gaultheria types. Both measures were strongly influenced by disturbance intensity. The stability of Pseudotsuga understories derives from the moderate tolerance of initial understory dominants to burning and in their ability to subsequently perennate from subterranean structures. Variation in the long—term response of communities reflects complex interactions between species life history, disturbance intensity, and chance, suggesting that both deterministic and stochastic factors must be considered in evaluating community stability and response to disturbance.

Although most temperate forests are actively managed for timber production, few data exist regarding the long‐term effects of forest management on understory plant communities. We investigated the responses of understory communities to a factorial combination of silvicultural‐thinning and nutrient‐addition treatments maintained for 12–16 yr in a set of 21–27 yr old Douglas‐fir (Pseudotsuga menziesii) plantations. The four thinning levels span those used operationally (final stem densities of 494–1680 trees/ha); the two fertilization levels included a control and N addition in the form of urea at ∼60 kg N·ha−1·yr−1, about twice the dosage used operationally. Understory vegetation cover showed significant effects of thinning, with the highest thinning level resulting in the highest observed cover values. However, in some cases low levels of thinning resulted in a reduction in understory cover compared to unthinned controls. Understory vegetation declined dramatically in response to urea fertilization, with up to a 10‐fold drop in herb‐layer cover in unthinned stands. Species richness showed a simpler response to treatments, increasing in response to thinning, but decreasing in response to fertilization. Examination of species–area relationships indicated that effects of thinning and fertilization on species richness were similar across the range of spatial scales examined. Tree canopy cover, assessed by means of hemispherical photograph analysis, increased with fertilization, and estimated understory light levels decreased with fertilization, but neither showed a significant response to thinning at the time of measurement (12–16 yr after tree removal). Thus, treatment effects on understory cover and species richness were not a simple function of canopy cover or estimated light availability. Rather, there was a weak positive relationship between estimated understory light flux and vascular plant cover and diversity in nonfertilized plots, and no such relationship in fertilized plots. The lack of correspondence between treatment effects on canopy cover and understory vegetation may be due to time lags in understory response to changes in canopy cover or to treatment effects not mediated by light availability, such as physical disturbance during thinning operations and toxicity responses following application of urea fertilizer. Species‐specific responses to treatments were in part predictable as a function of plant life‐form and edaphic association: species affinity for high soil moisture was the best predictor of fertilization responses, while life‐form was the best predictor of thinning responses, with ferns and graminoids showing the largest positive responses to thinning. The successional status and stature of understory plant species were not significantly related to treatment responses. In sum, our results indicate that silvicultural thinning and fertilization can have large effects on understory plant diversity and community composition. However, such effects were not a simple function of understory light leve...

Tree mortality is a critical but understudied process in coniferous forest development. Current successional models assume that mortality during early forest development is dominated by density‐dependent processes, but few long‐term studies exist to test this assumption. We examined changes in forest structure and patterns of tree mortality 14–38 years (1979–2001) after clear‐cut logging of two experimental watersheds in the western Cascade Range of Oregon, USA. We sampled 193 permanent plots (250 m2) six times generating 75126 data records and 7146 incidents of mortality. Mean density peaked at >3000 stems/ha (≥1.4 m tall) after 22–25 years; bole biomass increased continuously to >100 Mg/ha. At final sampling, stem density varied by two orders of magnitude and biomass by a factor of 10 among sample plots. Suppression mortality occurred in >80% of plots and was >2.5 times as frequent as mechanical damage (uprooting, stem snap, and crushing). However, biomass lost to mortality via mechanical damage was nearly four times that lost to suppression, a result of episodic storms that created windthrow patches, with some plots losing 30–50% of biomass. Total annual mortality increased from 1.0% to 5.3% of stems over the study period and was highly variable among species. Although mortality rates were highest for sprouting hardwoods (reaching 9.7% in Cornus nuttallii), biomass of most hardwood species increased through canopy closure as dominant stems achieved large sizes. Shade‐tolerant conifers (Tsuga heterophylla and Thuja plicata), typically assumed to be absent or to play a minor role in early forest development, accounted for 26% of stems after 38 years. In regression tree models, environmental attributes of plots had limited ability to predict mortality. Instead, stem density prior to canopy closure was the strongest predictor of cumulative mortality (either suppression or mechanical damage). Our long‐term studies suggest that current models of early forest development are overly simplistic, particularly in their treatment of mortality. Although suppression was the dominant demographic process, mechanical damage yielded greater loss of biomass and greater structural heterogeneity through creation of windthrow gaps. Thus, gap‐forming processes that operate late in succession and contribute to structural complexity in old‐growth forests can also occur early in stand development.