Penobscot Experimental Forest Research Articles

Strategies for enhancing long-term carbon sequestration in mixed-species, naturally regenerated Northern temperate forests

We compared long-term C sequestration in the pools of aboveground portions of live trees, dead wood, and harvested wood products among highly contrasting forest management scenarios on a rotation (30-100 years) and 100-year basis. Average annual net change in C (AAC) and the cumulative sum of net changes in C were calculated using 65 years of data from permanent plots representing contrasting approaches to managing mixed-species stands dominated by shade-tolerant coniferous species on the Penobscot Experimental Forest in Maine, USA. Simulations of tree growth and mortality were used to estimate C pools to 100 years. On a rotation basis and for all pools combined, scenarios with selection cutting had greater AAC than those with shelterwood cutting followed by thinning or with diameter-limit cutting (p < 0.05). For combined pools, the cumulative sum of net changes in C for the unmanaged, selection, and guiding diameter-limit stands was positive for most of the study period. Our results suggest that strategies that maintain overstory stocking levels necessary to regenerate desired species and promote the development of sawlog-sized trees can enhance long-term C sequestration in mixed-species, naturally regenerated northern temperate forests.

Machine Learning Techniques for Tree Species Classification Using Co-Registered LiDAR and Hyperspectral Data

The use of light detection and ranging (LiDAR) techniques for recording and analyzing tree and forest structural variables shows strong promise for improving established hyperspectral-based tree species classifications; however, previous multi-sensoral projects were often limited by error resulting from seasonal or flight path differences. The National Aeronautics and Space Administration (NASA) Goddard’s LiDAR, hyperspectral, and thermal imager (G-LiHT) is now providing co-registered data on experimental forests in the United States, which are associated with established ground truths from existing forest plots. Free, user-friendly machine learning applications like the Orange Data Mining Extension for Python recently simplified the process of combining datasets, handling variable redundancy and noise, and reducing dimensionality in remotely sensed datasets. Neural networks, CN2 rules, and support vector machine methods are used here to achieve a final classification accuracy of 67% for dominant tree species in experimental plots of Howland Experimental Forest, a mixed coniferous–deciduous forest with ten dominant tree species, and 59% for plots in Penobscot Experimental Forest, a mixed coniferous–deciduous forest with 15 dominant tree species. These accuracies are higher than those produced using LiDAR or hyperspectral datasets separately, suggesting that combined spectral and structural data have a greater richness of complementary information than either dataset alone. Using greatly simplified datasets created by our dimensionality reduction methodology, machine learner performance remains comparable or higher to that using the full dataset. Across forests, the identification of shared structural and spectral variables suggests that this methodology can successfully identify parameters with high explanatory power for differentiating among tree species, and opens the possibility of addressing large-scale forestry questions using optimized remote sensing workflows.

Northern mixedwood composition and productivity 50 years after whole-tree and stem-only harvesting with and without post-harvest prescribed burning

Whole-tree harvesting (WTH) is increasingly used to extract forest biomass for energy and commercial wood products. Slash burning, which is used for fuels reduction and site preparation, also reduces aboveground biomass. Yet effects of incremental biomass reduction, from either WTH or slash burning, on long-term forest productivity and composition are poorly understood. This research uses data from a 50-year-old study in northern mixedwood (Picea – Abies – hardwood) stands on the Penobscot Experimental Forest in Maine, U.S.A., to address these concerns. Clearcutting was conducted in 1964–65 with WTH, stem-only harvesting (SOH), and SOH with post-harvest prescribed burning of logging residues (SOHB). Growing stock, composition, and soil properties (O horizon thickness and soil drainage) were measured 50 years after treatment. Hardwood composition (percent of total basal area) increased from pre-treatment levels in all treatments but was greater in SOHB than SOH and WTH. Eastern white pine (Pinus strobus), though a minor species, was significantly more abundant in WTH than SOH or SOHB. Results indicated no other significant differences in species composition, or in stand structure or productivity (total basal area, stem density, dominant height, quadratic mean diameter, and total above-ground carbon stock) among treatments. Independent of treatment, we observed relationships between soil properties and stem density and quadratic mean diameter (qmd), such that lower stem density and greater qmd were observed on sites with greater O horizon thickness. These findings suggest that relative to SOH, WTH and SOHB do not degrade northern mixedwood stand productivity as expressed by stand structure and stocking 50 years after a single treatment, even on a site with low to moderate production potential. Nevertheless, species shifts associated with clearcutting (i.e., shade-tolerant conifer to intolerant hardwood composition) and prescribed burning in this forest type should be considered in light of the potential application of either for intensive silviculture treatments.

Joint hierarchical models for sparsely sampled high-dimensional LiDAR and forest variables

Recent advancements in remote sensing technology, specifically Light Detection and Ranging (LiDAR) sensors, provide the data needed to quantify forest characteristics at a fine spatial resolution over large geographic domains. From an inferential standpoint, there is interest in prediction and interpolation of the often sparsely sampled and spatially misaligned LiDAR signals and forest variables. We propose a fully process-based Bayesian hierarchical model for above ground biomass (AGB) and LiDAR signals. The process-based framework offers richness in inferential capabilities, e.g., inference on the entire underlying processes instead of estimates only at pre-specified points. Key challenges we obviate include misalignment between the AGB observations and LiDAR signals and the high-dimensionality in the model emerging from LiDAR signals in conjunction with the large number of spatial locations. We offer simulation experiments to evaluate our proposed models and also apply them to a challenging dataset comprising LiDAR and spatially coinciding forest inventory variables collected on the Penobscot Experimental Forest (PEF), Maine. Our key substantive contributions include AGB data products with associated measures of uncertainty for the PEF and, more broadly, a methodology that should find use in a variety of current and upcoming forest variable mapping efforts using sparsely sampled remotely sensed high-dimensional data.

Assessing the role of natural disturbance and forest management on dead wood dynamics in mixed-species stands of central Maine, USA

Dead wood pools are strongly influenced by natural disturbance events, stand development processes, and forest management activities. However, the relative importance of these influences can vary over time. In this study, we evaluate the role of these factors on dead wood biomass pools across several forest management alternatives after 60 years of treatment on the Penobscot Experimental Forest in central Maine, USA. After accounting for variation in site quality, we found significant differences in observed downed coarse woody material (CWM; ≥7.6 cm small-end diameter) and standing dead wood biomass among selection, shelterwood, and commercial clear-cut treatments. Overall, total dead wood biomass was positively correlated with live tree biomass and was negatively correlated with the average wood density of nonharvest mortality. We also developed an index of cumulative harvest severity, which can be used to evaluate forest attributes when multiple harvests have occurred within the same stand over time. Findings of this study highlight the dynamic roles of forest management, stand development, and site quality in influencing dead wood biomass pools at the stand level and underscore the potential for various outcomes from the same forest management treatment applied at different times in contrasting stands.

Modeling forest biomass and growth: Coupling long-term inventory and LiDAR data

Abstract Combining spatially-explicit long-term forest inventory and remotely sensed information from Light Detection and Ranging (LiDAR) datasets through statistical models can be a powerful tool for predicting and mapping above-ground biomass (AGB) at a range of geographic scales. We present and examine a novel modeling approach to improve prediction of AGB and estimate AGB growth using LiDAR data. The proposed model accommodates temporal misalignment between field measurements and remotely sensed data—a problem pervasive in such settings—by including multiple time-indexed measurements at plot locations to estimate AGB growth. We pursue a Bayesian modeling framework that allows for appropriately complex parameter associations and uncertainty propagation through to prediction. Specifically, we identify a space-varying coefficients model to predict and map AGB and its associated growth simultaneously. The proposed model is assessed using LiDAR data acquired from NASA Goddard's LiDAR, Hyper-spectral & Thermal imager and field inventory data from the Penobscot Experimental Forest in Bradley, Maine. The proposed model outperformed the time-invariant counterpart models in predictive performance as indicated by a substantial reduction in root mean squared error. The proposed model adequately accounts for temporal misalignment through the estimation of forest AGB growth and accommodates residual spatial dependence. Results from this analysis suggest that future AGB models informed using remotely sensed data, such as LiDAR, may be improved by adapting traditional modeling frameworks to account for temporal misalignment and spatial dependence using random effects.

Evaluation of forest management effects on the mineral soil carbon pool of a lowland, mixed-species forest in Maine, USA

Concerns about climate change have increased interest in ways to maximize carbon (C) storage in forests through the use of alternative forest management strategies. However, the influence of these strategies on soil C pools is unclear. The primary objective of this study was to test for differences in mineral soil C stocks among various silvicultural and harvesting treatments that were initiated in the 1950s and have been maintained since on the Penobscot Experimental Forest in central Maine, USA. Five mineral soil cores below the surface organic horizon to a depth of 1 m were collected from each replicate (n = 2) of selection, shelterwood, and commercial clearcut treatments. For these treatments, the mean mineral soil C stock was 47.7 ± 16.4 Mg ha−1 (mean ± SD). We found no significant differences in average mineral soil C stocks among treatments. However, a post hoc power analysis indicated that the probability of detecting a significant treatment effect was only 6%. We determined that 98 stands per treatment would be required to be 80% certain that the F test would detect a difference in average mineral soil C stocks whenever any pair of treatments had C stocks differing by more than 5 Mg ha−1.

Factors influencing organic-horizon carbon pools in mixed-species stands of central Maine, USA

The overall goal of this study was to evaluate the correlation of multiple abiotic and biotic factors with organic-horizon (O-horizon) carbon (C) content on the Penobscot Experimental Forest in central Maine, USA. O-horizon samples were collected and their associated depths were recorded from stands managed with a range of silvicultural and harvesting treatments (i.e., selection, shelterwood, and commercial clearcut) and an unmanaged control. The overall mean for O-horizon C content from all samples was 25.6±16.1Mgha−1 (mean±SD). The samples were used to develop a pedotransfer function for predicting O-horizon C content from O-horizon depth (R2=0.47, RMSE=1.6Mgha−1) so that an average of O-horizon C content could be calculated for permanent sample plots on which abiotic and biotic factors were quantified. O-horizon depth measurements recorded along transects on permanent sample plots were used to calculate plot average O-horizon C content. There were no significant differences in average predicted O-horizon C content among selection, shelterwood, and commercial clearcut treatments. However, variation in predicted O-horizon C content between stands where the same treatment was applied was statistically significant and was likely due to the timing of harvests and abundance of dead wood buried within O horizons. Depth to redoximorphic features, cartographic depth to water table or saturated zone, drainage class, fine woody debris mass, downed woody debris volume, tree basal area, and the relative basal area of conifer species were not significant predictors of predicted O-horizon C content at the plot level. When the individual predicted values of O-horizon C content were modeled, within-plot variation accounted for 83.8% of the variance. Bryophyte mass, which was predicted from bryophyte cover, only explained 1.2% of the variation in O-horizon C content at the microsite level. These results highlight the sizeable variability in O-horizon C content within and among these mixed-species stands with various forest management and natural disturbance histories.

Dynamic spatial regression models for space‐varying forest stand tables

Many forest management planning decisions are based on information about the number of trees by species and diameter per unit area. This information is commonly summarized in astand table, where a stand is defined as a group of forest trees of sufficiently uniform species composition, age, condition, or productivity to be considered a homogeneous unit for planning purposes. Typically, information used to construct stand tables is gleaned from observed subsets of the forest selected using a probability‐based sampling design. Such sampling campaigns are expensive, and hence, only a small number of sample units are typically observed. This data paucity means that stand tables can only be estimated for relatively large areal units. Contemporary forest management planning and spatially explicit ecosystem models require stand table input at higher spatial resolution than can be affordably provided using traditional approaches. We propose a dynamic multivariate Poisson spatial regression model that accommodates both spatial correlation between observed diameter distributions and also correlation between tree counts across diameter classes within each location. To improve fit and prediction at unobserved locations, diameter specific intensities can be estimated using auxiliary data such as management history or remotely sensed information. The proposed model is used to analyze a diverse forest inventory dataset collected on the United States Forest Service Penobscot Experimental Forest in Bradley, Maine. Results demonstrate that explicitly modeling the residual spatial structure via a multivariate Gaussian process and incorporating information about forest structure from Light Detection and Ranging (LiDAR) covariates improve model fit and can provide high spatial resolution stand table maps with associated estimates of uncertainty. Copyright © 2014 John Wiley &amp; Sons, Ltd.

Silvicultural Rehabilitation of Cutover Mixedwood Stands

We investigated rehabilitation of mixedwood stands degraded by exploitative cutting on the Penobscot Experimental Forest in Maine. Three precommercial rehabilitation treatments were applied: control (no rehabilitation), moderate rehabilitation (crop tree release [CTR]), and intensive rehabilitation (CTR, timber stand improvement [TSI], and red spruce fill planting). Crop trees (primarily red maple, paper birch, spruce, aspen, and eastern hemlock) were selected and released based on their potential for improved growth and value, spacing, and species composition. Rehabilitation reduced sapling basal area, density, and hardwood abundance and increased crop tree diameter increment. Fill planting increased spruce stocking, but many planted seedlings were browsed. Long-term projections suggested that future stand value will repay costs of moderate rehabilitation (CTR); intensive rehabilitation (CTR-TSI-planting) as applied in this study requires greater investment than can be repaid through quality and growth improvements of low-value hardwoods and softwoods. Although simulations suggested no difference in future stand value between treated and untreated stands, improvements in composition, growth, and quality after rehabilitation will facilitate later commercial thinning and shelterwood regeneration in stands which otherwise have few management options.

Assessing the Feasibility of Low-Density LiDAR for Stand Inventory Attribute Predictions in Complex and Managed Forests of Northern Maine, USA

The objective of this study was to evaluate the applicability of using a low-density (1–3 points m−2) discrete-return LiDAR (Light Detection and Ranging) for predicting maximum tree height, stem density, basal area, quadratic mean diameter and total volume. The research was conducted at the Penobscot Experimental Forest in central Maine, where a range of stand structures and species composition is present and generally representative of northern Maine’s forests. Prediction models were developed utilizing the random forest algorithm that was calibrated using reference data collected in fixed radius circular plots. For comparison, the volume model used two sets of reference data, with one being fixed radius circular plots and the other variable radius plots. Prediction biases were evaluated with respect to five silvicultural treatments and softwood species composition based on the coefficient of determination (R2), root mean square error and mean bias, as well as residual scatter plots. Overall, this study found that LiDAR tended to underestimate maximum tree height and volume. The maximum tree height and volume models had R2 values of 86.9% and 72.1%, respectively. The accuracy of volume prediction was also sensitive to the plot type used. While it was difficult to develop models with a high R2, due to the complexities of Maine’s forest structures and species composition, the results suggest that low density LiDAR can be used as a supporting tool in forest management for this region.

Multivariate Spatial Regression Models for Predicting Individual Tree Structure Variables Using LiDAR Data

This study assesses univariate and multivariate spatial regression models for predicting individual tree structure variables using Light Detection And Ranging (LiDAR) covariates. Many studies have used covariates derived from LiDAR to help explain the variability in tree, stand, or forest variables at a fine spatial resolution across a specified domain. Few studies use regression models capable of accommodating residual spatial dependence between field measurements. Failure to acknowledge this spatial dependence can result in biased and perhaps misleading inference about the importance of LiDAR covariates and erroneous prediction. Accommodating residual spatial dependence, via spatial random effects, helps to meet basic model assumptions and, as illustrated in this study, can improve model fit and prediction. When multiple correlated tree structure variables are considered, it is attractive to specify joint models that are able to estimate the within tree covariance structure and use it for subsequent prediction for unmeasured trees. We capture within tree residual covariances by specifying a model with multivariate spatial random effects. The univariate and multivariate spatial random effects models are compared to those without random effects using a data set collected on the U.S. Forest Service Penobscot Experimental Forest, Maine. These data comprise individual tree measurements including geographic position, height, average crown length, average crown radius, and diameter at breast height.

Hierarchical Bayesian spatial models for predicting multiple forest variables using waveform LiDAR, hyperspectral imagery, and large inventory datasets

In this paper we detail a multivariate spatial regression model that couples LiDAR, hyperspectral and forest inventory data to predict forest outcome variables at a high spatial resolution. The proposed model is used to analyze forest inventory data collected on the US Forest Service Penobscot Experimental Forest (PEF), ME, USA. In addition to helping meet the regression model's assumptions, results from the PEF analysis suggest that the addition of multivariate spatial random effects improves model fit and predictive ability, compared with two commonly applied modeling approaches. This improvement results from explicitly modeling the covariation among forest outcome variables and spatial dependence among observations through the random effects. Direct application of such multivariate models to even moderately large datasets is often computationally infeasible because of cubic order matrix algorithms involved in estimation. We apply a spatial dimension reduction technique to help overcome this computational hurdle without sacrificing richness in modeling.

Nonnative invasive plants in the Penobscot Experimental Forest in Maine, USA: Influence of site, silviculture, and land use history1,2

Abstract We investigated the occurrence of nonnative invasive plants on approximately 175 ha comprising a long-term, 60-year-old U.S. Forest Service silvicultural experiment and old-field stands in the Penobscot Experimental Forest (PEF) in central Maine. Stands in the silvicultural experiment were never cleared for agriculture, but have been repeatedly partially cut. Our objectives were to determine the extent of nonnative invasive plant populations in the PEF, and to relate invasive plant abundance and distribution to management history and environmental factors (overstory composition and basal area, canopy openness, and soil characteristics). We found ten invasive plant species in the study area. Very few occurrences of these were in the silvicultural experiment; where present, invasive plants there appear to be associated with proximity to seed source, and a greater degree of recreational or silvicultural disturbance. Ordination showed that the environmental variables which were associated with invasi...

Northern White-Cedar Regeneration Dynamics on the Penobscot Experimental Forest in Maine: 40-Year Results

Abstract The objective of this study was to assess the long-term dynamics of northern white-cedar (Thuja occidentalis L.) seedling and sapling growth and mortality on the Penobscot Experimental Forest in Maine. Data collected between 1965 and 2005 in four twice-replicated partial cutting treatments were analyzed. White-cedar seedlings established in all treatments despite relatively high white tailed-deer (Odocoileus virginianus Zimmerman) population densities. However, although it appears that regeneration cohorts of associated softwoods increased in size over time, the white-cedar cohort did not. Ingrowth of white-cedar from the seedling to sapling stage was lower than the combined rates of sapling mortality and recruitment to the pole stage; sapling density of this species in 2005 was &gt;80% less than it was at the start of the measurement period. Sapling mortality was high, and recruitment to larger size classes was low, although mortality decreased and recruitment increased as sapling size increased. Browsing was prolific; 90% of white-cedar seedlings and small saplings showed signs of browse in 2005. Overall, white-cedar sapling growth was slow, with an estimated 100 years needed to grow from small sapling to merchantable size in the study stands. Efforts to release white-cedar saplings through precommercial treatment and control of browsing pressure are recommended.

Forest management and temporal effects on food abundance for a ground-foraging bird ( Catharus guttatus)

While food abundance is known to limit bird populations, few studies have simultaneously examined both temporal and habitat-related variation in food abundance for ground-foraging birds. We surveyed ground-dwelling arthropods throughout the breeding season and in three forest harvest types (clearcut, selection, and shelterwood stands managed for spruce, fir and hemlock) at the Penobscot Experimental Forest (PEF), Bradley, Maine, USA. Pitfall trapping during 2003 and 2004 was used to measure seasonal and habitat-related variation in abundance of ground-dwelling arthropod food items for hermit thrushes. Arthropods were sorted into relevant taxonomic and size class categories based on the known diet of hermit thrushes. Habitat variables included forest management type, stand-level structure and composition of tree species, and microhabitat measures in a 1 m 2 area around the pitfall traps. We used redundancy analysis to find stand-level and microhabitat measures that best explained variation in arthropod abundance during 2004. The majority of arthropod captures comprised three groups: ants (Formicidae, 44%), ground beetles (Carabidae, 13%), and spiders (Araneae, 11%). There was significant temporal variation in captures for many taxa. Notably, there were as many as six times more Araneae in early compared to late season samples suggesting the importance of Araneae to ground-foraging birds. Carabidae were more abundant in selection versus either shelterwood or clearcut stands on one sample date. Conversely, Formicidae and Curculionidae were more abundant in shelterwood plots on one and three sampling dates, respectively. Overall, our data provided little evidence that forest management type impacts arthropod food abundance, in stands 5–20 years post-harvest. Abundance of Carabidae was positively correlated with the number of softwoods less than 5 cm DBH and the number of hardwoods greater than 10 cm DBH, while abundance of Formicidae and Homopterans was negatively associated with these same measures. Several taxa were negatively associated with percent of the 1 m 2 around the pitfall that was shaded, including Araneae, Carabidae, Formicidae, and Staphylinidae. These results suggest that less-shaded forest microhabitats are important areas of increased food abundance for ground-foraging birds, regardless of forest management type or tree composition. Consequently, forest managers seeking to maximize food abundance for ground-foraging birds in managed spruce-fir forests should consider retaining some large diameter hardwood trees and use management techniques that reduce forest floor shading.

Reference Stands for Silvicultural Research: A Maine Perspective

Abstract Silvicultural experiments should have untreated stand replicates in which development can be tracked over time. Unfortunately, field studies are seldom ideal. This article is one of six in this issue addressing experimental controls. Our focus is the Penobscot Experimental Forest (PEF) in Maine, where a 55-year-old experiment in northern conifer silviculture has an unreplicated, somewhat atypical control. The identification of stands that represent desired endpoints or natural states is another consideration and may be difficult if based on rare conditions. Big Reed Forest Reserve and Maine's Ecological Reserve System are discussed as possible benchmarks for management on the PEF and serve as examples of the opportunities and challenges associated with the identification of benchmark ecosystems.

Comparison of Fixed Diameter-Limit and Selection Cutting in Northern Conifers

Abstract Diameter-limit cutting is a common type of harvest in which all merchantable trees above specific size thresholds are removed. Despite a long history of application, controlled experiments of these harvests are rare and the cumulative effects of repeated diameter-limit cuts are largely unknown. The Penobscot Experimental Forest in Maine is the location of a long-term USDA Forest Service experiment in which both fixed diameter-limit and selection cutting have been applied at 20-year intervals since the early 1950s. After three entries, present value of gross harvest revenue was greater in the fixed diameter-limit than selection treatments. However, sawtimber volume and growth, total and merchantable volume, regeneration stem density, and inventory value all were lower in the fixed diameter-limit than selection stands. Accumulated value (harvest plus residual) and species composition did not differ between treatments. Within-treatment analysis revealed desirable directional changes in cull percentage and species composition in the selection but not diameter-limit cut stands, suggesting trends that may result in greater future treatment disparity. These data confirm the degrading effects of fixed diameter-limit relative to selection cutting, and reveal that greater short-term value removals are offset by lower residual stand volume and value. North. J. Appl. For. 22(2):77–84.

Snag longevity under alternative silvicultural regimes in mixed-species forests of central Maine

Predictions of snag longevity, defined here as the probability of snag survival to a given age, are key to designing silvicultural regimes that ensure their availability for wildlife and form an important component of carbon flow models. Species, diameter at breast height, stand density, management regime, and agent of tree mortality were assessed for their effect on snag longevity in a long-term silvicultural study on the Penobscot Experimental Forest in central Maine. Snag recruitment and fall data from USDA Forest Service inventories between 1981 and 1997 were analyzed using parametric survival analysis. A Weibull model fit the data best, indicating a significant lag time followed by rapid fall rates. Half-times varied among species, with Thuja occidentalis L. having the longest (10 years) and Picea species the shortest (6 years). Snag longevity was significantly greater with increasing diameter and decreased with increasing stand density. Agent of mortality and silvicultural treatment were also significant. Two models were developed for estimating probability of snag survival over time, one that included predictor variables unique to the silvicultural systems study on the Penobscot Experimental Forest and one using predictor variables available in most standard inventories. Snag survival models can easily be incorporated into comprehensive forest dynamics models to facilitate estimates of wildlife habitat structure and carbon flow.

Silviculture alters the genetic structure of an eastern hemlock forest in Maine, USA

We evaluated the influence of long-term silvicultural selection on the genetic structure of an eastern hemlock (Tsuga canadensis (L.) Carr.) forest at the Penobscot Experimental Forest, in Maine, USA. Plots in this forest received one of the following three treatments: (1) selection cuts in which small and poorly formed trees were preferentially removed in 1957 and 1977; (2) diameter-limit cuts in which trees at least 24 cm in diameter were removed in 1952, 1973, and 1994; or (3) no harvesting (an unmanaged control). Because of an association between the occurrence of rare alleles and tree phenotypes, phenotypically based tree removals were associated with a shift in allelic frequency. Where smaller trees with inferior phenotypes were preferentially removed (selection cut), the number of rare alleles and estimates of future genetic potential were lower than in the control group. Because of the theoretical long-term evolutionary benefit of unique gene forms, the loss of rare alleles could diminish the potential of populations to adapt to and survive ongoing environmental change. In contrast, alleles that were rare in the control group existed at a higher frequency in the diameter-limit cut. However, productivity was low in this stand, where the frequency of characteristically rare alleles was artificially amplified.