The paper which this Addendum refers to is available here: Western gall rust - A threat to Pinus radiata in New Zealand

The criteria for accepting or rejecting a technology extend beyond the intrinsic properties of the finished product such as physical performance. There are also extrinsic factors such as the history of the product and trust in the manufacturers and suppliers, as well as the perceptions and risk management strategies of various stakeholder groups. A methodology was trialled to take into account the extended supply chain of the product, while simultaneously engaging stakeholders to determine and to understand their perceptual frameworks. Three pine decking products manufactured using different amounts and types of chemical modification were compared using life cycle assessment and the comments of 114 respondents from six stakeholder groups in New Zealand. The results of the perceptual research include a quadrant diagram which allows a visual comparison of the responses of different stakeholders to actual or hypothetical products, aiding the identification of when and why certain technologies may be disqualified from acceptability or become the topic of public debate.

The possible impact of Nectria fuckeliana Booth on the forests and forest industries of New Zealand, a significant exporter of industrial roundwood, was estimated for different scenarios of the spread of the fungal pest and trade measure responses in export markets. An economic model was used to assess the direct effect of the pest and the potential impact of trade bans and phytosanitary regulations to prevent pest arrival in New Zealand's major export markets - China, Japan, and South Korea. Depending on the assumed area affected by N. fuckeliana, the net present value of New Zealand's forest sector gross revenue was US$34 million to US$612 million lower, due to reduced harvest and log exports, even without foreign trade measures. A possible measure, requiring the debarking of New Zealand log exports, would reduce the present value of New Zealand growers' revenues by US$ 1,200 million, even if N. fuckeliana were confined to the already affected area. If China, Japan, and South Korea banned imports of New Zealand logs altogether, and the pest continued to spread at historical rates, the present value of New Zealand growers' revenues would decrease by US$8,200 million. Estimated losses to growers could be, to varying extents, offset by increased domestic production of processed wood products, under both trade measures. The debarking and import ban policies would increase gross revenues for producers in China and South Korea, but also increase the cost to consumers of wood products.

Adaptive control, in conjunction with dynamic programming, has been shown in earlier research to provide superior results from stem and stand log bucking when the stand is subject to order book constraints. Adaptive control can be achieved by adjusting relative prices and small-end diameter specifications as the harvesting operation moves through the stand. We examined the effects on market fulfilment of varying the frequency with which adjustments are made. Apportionment degree was used as the metric for market fulfilment. We found that there was a significant positive relationship between adjustment frequency and apportionment degree; apportionment degree increased as the size of the harvest area between adjustments decreased from 1.2 to 0.2 ha.

Volume data from young trees were combined with predictions from existing individual-tree functions in older stands and used to develop new individual-tree and stand-level volume functions for Pinus radiata D. Don in New Zealand. As part of this process, several new individual-tree volume datasets were compiled to examine variation in stem form factor after accounting for tree height. Analysis of data from a national trial series which included nitrogen fertiliser treatments and controls without fertiliser found no treatment-related differences in form factor at stand age 5 years, except at a nitrogen-deficient sand site located in Woodhill Forest. Analysis of a seedlot comparison experiment, which included unimproved through to highly improved genetic stock, found no differences in form factor at stand age 3 years. Analysis of a stocking rates experiment found no significant differences in form factor when age was taken into account. However, across the combined datasets, form factor decreased strongly with increasing tree height. This relationship between form factor and height was incorporated into new volume functions enabling them to perform well for trees ranging from establishment to clearfelling age, across a range of stocking rates and site types, excluding coastal sand sites.

Sustaining site productivity over multiple timber harvesting rotations requires an understanding of changes in the soil nutrient supply to the forest. To maintain soil fertility the amount of available limiting nutrients removed in each cropping cycle should be matched by inputs to the site. Weathering is one key source of nutrient inputs that has not been well quantified. Mineral weathering at six sites in New Zealand was examined where multiple-rotation forests were being studied to relate nutrient supply to nutrient removal with harvesting. Three sites were located in the North Island of New Zealand at Woodhill, Tarawera, and Kinleith Forests, and three were in the South Island at Golden Downs, Burnham, and Berwick Forests. Growth of Pinus radiata D. Don was used to estimate annual nutrient uptake rates. Weathering rates were determined using two methods: the PROFILE model with appropriate site data for all six sites, and leaching of soil columns from two sites. Annual weathering rates for calcium, magnesium, potassium, and phosphorus were 3-24, 3-10, 3-31, and 0.2-2.5 kg/ha respectively. Column-leaching estimates of weathering resulted in comparable weathering rates to that of PROFILE when used with a soil that had a consistent texture throughout the soil profile, as occurred at the Kinleith site. The Burnham soil had a fine texture over coarse and data did not compare well with the PROFILE estimates, probably due to irregular water flow in the soil column. PROFILE weathering rates appeared reasonable for all sites, indicating its usefulness for predicting nutrient supply by weathering over multiple rotations if sufficient soil and site data are available. Weathering rates calculated using PROFILE and leaching of nutrients from the O horizon suggest that both weathering and O horizon mineralisation are critical to sustainable production over multiple rotations. Without addition of fertiliser some nutrients may be depleted at current rotation lengths.

The coelomycetous fungi described in this paper have been recorded from New Zealand but not fully described. The fungi are:

Corticolous coelomy cetes: Coryneum umbonatum Nees on Castanea sativa Miller and Quercus robur Linnaeus; Stegonsporium pyriforme (Hoffmann) Corda on Acer pseudoplatanus Linnaeus.

Foliicolous coelomycetes: Phyllosticta abietis Bissett & M.E. Palm on Cedrus atlantica (Endlicher) Manetti ex Carriere; Phyllosticta sp. on Pittosporum ralphii Kirk.

Streamflow data from a catchment study located in the central North Island of New Zealand were analysed to determine the effect of land-use and forest management on water yield. Land-uses compared included pasture, pasture converted to Pinus radiata D. Don, and evergreen indigenous forest. The catchments ranged in area from 6 to 37 ha. Rainfall and streamflow were monitored from 1969 to 2000. Leaf area index of three pine stands with different silvicultural management regimes was measured from 1975 to 1985.

Annual flows from pine catchmen ts were lower than from pasture, with an average maximum difference of approximately 400 mm after canopy closure. Thinning of the pine stands reduced the difference and harvesting reversed this trend, with the flow from pine catchments exceeding flows from pasture for the 3 years after harvesting of the pines. Annual flows varied over time, depending on tree age, the silvicultural regime, and variations in annual rainfall (approximately 1200-2100 mm/year).

Evapotranspiration and interception losses increased linearly with the increase in canopy leaf area index of the pine stands, but transpiration was not significantly related to leaf area index or rainfall. Annual variation in rainfall accounted for approximately 60% of the variation in flow across all land-uses. A simple model incorporating pine leaf area index in addition to rainfall accounted for 91% of the variation in streamflow in pasture and pasture/pine catchments.

Water yield averaged approximately 160-260 mm/year, less from pine than pasture, depending on the silvicultural regime. Water yield from the pine forest averaged around 100 mm/year less than from indigenous forest.

Mid infrared (IR), near infrared (NIR), and ¹³C nuclear magnetic resonance (NMR) spectra were recorded on solid soil samples from three neighbouring sites with similar history but different current land use - namely, native podocarp forest, pasture land cleared from native forest, and exotic softwood plantation converted from improved pasture. Principal component analysis and projection to latent structures regression of the spectra showed changes in carbohydrate and phenolic fractions reflected the current land use, with the introduction of pasture and plantation forest decreasing the aliphatic and increasing the aromatic fractions of the soils. Fractional areas for NMR spectra ranged between 0.16 and 0.34 for aliphatic, 0.33 and 0.50 for carbohydrate, 0.12 and 0.40 for aromatic, 0.04 and 0.11 for phenolic, and 0.01 and 0.17 for carbonyl carbon. High carbon concentrations in native forest micro-sites exhibiting low pH occurred predominantly as aliphatic and carbohydrate fractions, even at depths down to 0.5 m. Soil carbon protection following land use diversification was evident for at least 80 years.

The biomass of Pinus radiata (D. Don) root and shoot systems was measured in a number of studies to determine the effect of stand age and environmental factors on the root/shoot ratio. These and other published studies were carefully assessed to determine firstly those which provide unbiased data for estimating below-ground biomass from shoot data and, secondly, the precision of the root/shoot ratios. Because previous studies were designed for a number of different purposes they were assessed in relation to the methodological requirements that we considered had to be met in order to provide accurate root/shoot biomass information.

New data from an effluent irrigation trial at Whakarewarew a Forest near Rotorua, designed to assess the impact of water and nutrient additions on tree growth at stand age 5 years, showed that both the above- and below-ground biomass approximately doubled relative to untreated controls. Consequently, the root/shoot ratio was not significantly affected by effluent type or irrigation rate, and averaged 0.17 across all treatments. At Rangiora near Christchurch, similar results were obtained at an agroforestry trial to assess the impact of moisture and nitrogen supply on tree growth at stand age 5 years, where the mean root/ shoot ratio averaged 0.21. These values were similar to the mean ratio (0.20) in a 42-year-old P. radiata stand at Woodhill Forest. Studies of nursery seedlings ready for planting gave lower root/shoot ratios of approximately 0.15, though the root systems had been trimmed in the nursery.

Analysis of published data suggested a wide range in root/shoot ratios in P. radiata, ranging between 0.12 and 0.58 across all studies. However, we observed a number of limitations in some studies which we believe precluded their use for root biomass estimation and prediction purposes. The root/shoot ratio in studies that met the accuracy requirements necessary for our purpose ranged between 0.13 and 0.24, and in typical P. radiata stands the root/shoot ratio averaged 0.19 (95% CI = 0.023) over all ages and sites. The majority of the root system of trees, other than seedlings, was composed of coarse roots and the root stock, and these were largely extracted in studies we considered as suitable. Root biomass procedures generally resulted in the loss of fine roots in most studies, and a slightly increased root/shoot ratio of 0.2 is therefore recommended for carbon estimation in P. radiata stands across all stand ages and sites. This value is consistent with findings reported for various other coniferous species.

The mean density of stem wood growth sheaths laid down annually in Pinus radiata D. Don stands was modelled as a function of site mean annual temperature, soil nitrogen fertility, ring age, and stocking. The model was based on measurements of mean outerwood basic density at breast height of 30 trees per stand from a single seedlot established at 17 sites located throughout New Zealand. Soil and climate data were obtained from each site, and stem wood disks were sampled at 5-m intervals along the stem from 10 trees per stand from 15 of these sites. Previously reported breast-height basic density data from a comprehensive national survey were used to examine ring age trends from pith to bark, supplemented with data from four trials to determine the influence of tree stocking on outerwood density. The model was tested using data from an independent study of mean outerwood basic density at breast height (30-120 trees per stand) undertaken at 21 stands selected to cover a wide range in site fertility, temperature, tree stocking, and stand age.

Site mean breast-height outerwood basic density ranged betwe en 356 and 494 kg/m³ in the model development dataset, and 316 and 482 kg/m³ in the validation dataset, and increased significantly with site mean annual temperature (T), mineral soil adjusted carbon /nitrogen (C/N) ratio, and stocking. Breast-height density of annual growth rings from pith to bark increased with ring age, and this pattern was consistent at all heights within the stem. The ratio of sheath density to breast-height ring density varied with ring age and increased with increasing nitrogen fertility. To predict the density of annual stem wood growth sheaths, the model firstly estimates the effects of site mean annual temperature, soil nitrogen fertility, and stocking on mean outerwood density at breast height. Secondly, the effect of ring age on annual ring density at breast height from pith to bark is taken into account. Finally, the ratio of sheath density to breast-height ring density for each ring is estimated as a function of stand age and outerwood density.

The national wood density model explained 93% of the variation in outerwood density at breast height for the model development dataset, with model predictions within 0.2% of the measured values. The model explained 86% of the variation in breast-height ring density in the model validation dataset, with predictions from the model averaging 3.2% higher than the measured values. Seedlot differences in breast-height outerwood density contributed in part to the greater variability evident in the validation dataset. The modelled ratios used to predict the density of annual growth increments were not directly tested. However, analogous ratios of whole stem wood density/breast-height outerwood density were derived by site density class for stands across a range of age classes using an independent national dataset, and these were consistent with those predicted using the model. Areas for further model testing and development were identified.

The model can be applied to predict stem biomass and carbon sequestration in P. radiata stands from the increment in stem wood volume. The wood density model has been incorporated in a carbon modelling system (C_Change) to facilitate the prediction of carbon stocks and changes in New Zealand's exotic plantation forest estate.

Understanding decay rates of above- and below-ground coarse woody debris pools is necessary for quantifying forest ecosystem carbon storage and cycling processes. Decay rates were determined from both time series and chronosequence studies. Time series studies, using measurements of wood samples of known initial mass and volume over time, provide more reliable data than chronosequence studies. However, the latter allow more rapid determination of decay rates which is an important factor since the decay of coarse woody debris is a slow process. Most studies indicate that, for both above- and below-ground material, between 30 and 200 years are required to achieve 95% decay. Roots generally take a longer time to decay than above-ground coarse woody debris of similar dimension. Factors controlling decay rate include temperature, precipitation, species, substrate quality and composition, moisture content and dimension of the material, whether the material is suspended or in contact with the soil, and characteristics of the decomposer community. Studies of above-ground coarse woody debris decay rates for Pinus radiata D. Don, New Zealand's dominant plantation species, have been limited to relatively low rainfall locations in New Zealand and Australia. No data are available on this species for below-ground coarse woody root decay rates.

To help avoid conflict and even market rejection, the design and introduction of new technologies can be aided by consideration of the potential concerns of different stakeholder groups. An assessment of their desired goals for technological change may be useful for pre-empting changes in the parameters of acceptability for technologies. As part of a research programme evaluating the parameters of acceptability for bio-based technologies using life cycle assessment of products and an analysis of the perceptual frameworks of stakeholders, respondents from four stakeholder groups in New Zealand were interviewed about the desirable and undesirable trajectories for chemical modification technologies. Three examples of pine decking products derived using different amounts and types of chemical modification served to help explore the contemporary criteria and rationale for acceptability. The responses of the 70 respondents in those four groups indicated that new technologies need to be able to prove their worth with regard to both tangible and intangible qualities if they are to be accepted into the market in the place of either the existing product or an inorganic competitor.

Wood stiffness, or modulus of elasticity, can be measured or predicted using a variety of methods including standard bending tests, machine stress-grading, stress wave methods, or analysis of X-ray diffraction data from SilviScan. Each of these methods was applied to the same wood samples and the inter-relationships were determined.

Dynamic modulus of elasticity (MOE) values from stress wave and SilviScan methods produced very similar results but both these methods produced higher values than the static modulus of elasticity from the traditional bending test. Results from all methods were highly correlated and simple regression equations were developed for converting results between methods. Machine stress-grade was more strongly related to SilviScan modulus of elasticity than to either density or microfibril angle.

New tools have been developed to map the three-dimensional variation of wood properties within tree stems in order to support research into modelling tree growth and stem wood properties. Digital imaging, image processing, and data visualisation techniques have been integrated into software tools (G2Ring and G2View) that allow cost-effective three-dimensional stem analysis.

Specialised image-processing algorithms have been developed to accurately delineate annual ring boundaries and sectors of compression wood on images of cross-cut discs. Discs are cut by chainsaw, wetted to improve contrast, and imaged in the field. The image-processing algorithms handle the significant image "noise" resulting from the surface roughness of the discs. Annual ring detection is sufficiently robust that localised surface markings or defects (such as knots and cracks) do not substantially affect accuracy and the algorithm for detection of compression wood addresses the problem of false classification of latewood. The software, developed on images of discs from Pinus radiata D. Don, uses an efficient combination of automatic detection with operator guidance and override to maximise accuracy and efficiency.

The software incorporates a facility to reconstruct three-dimensional stem models by assembling data from multiple spatially registered discs. Optional information on three-dimensional stem shape, obtained from the existing image-based PhotoMARVL system, allows inclusion of stem sinuosity in the three-dimensional stem model. Custom visualisation software (G2View) has been developed to examine the three-dimensional distribution of measured data, such as ring width and compression wood, within the stem. Visualisation of the stem analysis data is a useful analytical precursor to modelling stem structure.

PhotoMARVL, G2Ring, a nd G2View thus provide a set of tools, able to be used individually and in combination, to collect data useful for analysis and modelling of stem shape and internal structure.

A new photogrammetric image-based dendrometry system called TreeD is particularly suited for measuring stem features on individual standing trees. There are two phases in the application of the system. In phase one, stereo digital images are taken of the sample trees and seven field parameters associated with that particular image/object environment are measured. In phase two, the parameters are used to register the images in a computer where they can be viewed stereoscopically. The dimensions of any feature visible on the tree stem can then be measured, and the position of the feature located in three-dimensional space.

Validation of the system was carried out using images of a radio mast of known dimensions; results showed that for heights less than 20 m, characteristics such as stem diameters, branch diameters, internodal distances, and (implicitly) stem sweep can be measured to an accuracy of ±1 cm. Absolute height measurements were measured to an accuracy of better than 10 cm.

The system can provide essential information for ste m and log characterisation in pre-harvest assessment. It also provides a useful image archive of all sample trees and data capture points and is routinely being used in research trials to improve understanding of stem quality.

Western gall rust (Peridermium harknessii J. P. Moore (syn. Endocronartium harknessii (J. P. Moore) Y. Hiratsuka) is potentially a serious threat to exotic Pinus radiata D. Don plantations of New Zealand although the pathogen has not been recorded here. Mechanisms that may have prevented invasion of the pathogen include geographic isolation, biological characteristics of the fungus, stand management, and regulatory mechanisms affecting transport and establishment. Major factors may include a low probability of importation of infected seedlings, unlikely spore transport in the atmosphere across the tropics, and asynchrony of rust sporulation and pine susceptibility in North America and New Zealand. The outbreak or "wave year" phenomenon in the native range of western gall rust demonstrates that both biological and microclimatic conditions must be suitable for establishment to occur. We conclude that the probability of invasion of New Zealand by western gall rust is very low

The rate of production of oil and turpentine in the effluent from the plug screw feeder at the exit from the presteamer of a fibreboard pilot plant was determined after a steady state of operation had been reached. From these data, the recovery of turpentine from the plug screw feeder effluent, as a proportion of that available in the wood chips, was shown to be 27%. Turpentine levels in wood of Pinus radiata D. Don were determined and compared with previous data, along with causes of variations in these data. The presteamer plug screw feeder effluent from a medium-density fibreboard production line appears to be a useful source of turpentine.

Growth of Pinus radiata D. Don was examined after a harvesting and tillage trial at Taringatura in the South Island of New Zealand; in that experiment forest soils had been subjected to various intensities of traffic by a range of harvesting machinery, followed by ripping of the soil. Harvesting traffic disturbed but did not compact the soil. Soil penetration resistance was markedly decreased by ripping but not significantly affected by traffic intensity. Ripping increased and traffic reduced the stem volume of P. radiata.

Most notably, there was an interaction between ripping and traffic: stem volume increased with the number of passes of harvesting machines on ripped soils but decreased with the number of passes on non-ripped soils. This was explained by traffic reducing the competitive weeds in the ripped treatments.

Acoustic velocity, measured using time of flight tools, provides a non-destructive measure of wood stiffness in standing trees. In order to assess how best to assess a stand of trees to a given level of precision, acoustic velocity was measured across three sites in New South Wales, Australia. All sites had similar climatic conditions and had had the same silviculture but different establishment dates. Analysis of variance and regression analysis were used to determine the components of variation and relationships between acoustic velocity, diameter at breast height, and age.

The variation within a stand was greatest between trees, followed by that between sides within trees, between plots, and within each side of a tree. There was a significant positive relationship between acoustic velocity squared and age, but little relationship between acoustic velocity squared and diameter at breast height. An optimal sampling strategy was developed that involved sampling four plots per stand, each plot containing 12 trees, and acoustic velocity measurements being taken four times on each of the two sides of the trees being sampled. With this strategy the mean acoustic velocity squared of a stand can be estimated to within at least +10% of the mean.

Patterns of change from pith to bark and the distribution of variability in density across forest areas, sites, and trees were determined using data from breast-height cores taken from harvest-age Pinus radiata D. Don (ages 28 to 37 years) as part of a resource survey of four forest areas in the south-west slopes area of New South Wales. Seventeen of the sample sites had been thinned, one site was a thinning trial, and the other two sites were unthinned but matched to two of the thinned sites to determine the effects of thinning. Pith to bark cores were cut into five-ring segments and extracted basic density was determined for each segment.

Little difference in average basic density was found between the forest areas, particularly for the first 10 rings adjacent to the pith. The major source of variability was between trees within each site - this accounted for 77-92% of the total variation at each age. Thinning had no discernible significant effect on density at any age. Density in the outermost five-ring segment (corresponding to ages 23-28 to 32-37 years) was poorly predicted by juvenile core density (inner 10 rings), indicating little opportunity for forward or backward prediction of density.

Forest health monitoring is essential to sustainable management of Pinus radiata D. Don plantations. Conventional survey techniques such as aerial sketch mapping are qualitative and subjective, their effectiveness depending on the skill of the surveyor. In contrast, digital remote sensing has the potential to provide quantitative and objective data on the location, extent, and severity of crown damage at a range of spatial scales. Decision tree analysis can incorporate both categorical and continuous data and is inherently non-parametric. Decision trees were used to model the crown condition of P. radiata plantations in southern New South Wales in three situations involving discoloured leaves, stunted crowns, and transparent crowns associated respectively with the Diplodia pinea (Desm.) Kickx fungus, nitrogen deficiency, and the pine aphid Essigella californica Essig. Spectral indices and fraction images derived from linear spectral mixture analyses of remote sensing scenes were used to classify crowns into either two or three condition classes. The best performing model was obtained for D. pinea with a two-class classification of crown discoloration (overall accuracy [OA] 92%

Preliminary stand growth and yield models were constructed for even-aged Acacia melanoxylon R. Br. plantations in New Zealand. Models that predict mean top height and basal area growth, initial basal area, post-thinning basal area, mortality, and total standing volume per hectare were fitted to permanent sample plot data biased towards younger ages. The data consisted of 1722 plot measurements from 229 sample plots. Site index estimates predicted using a polymorphic Chapman-Richards difference equation fitted as a nonlinear mixed model ranged from 11.2 m to 35.5 in mean top height at age 30. The average site index was 24 m. All suitable data were used to fit models, preventing separation of independent validation data. The models were tested as a system of equations by comparing total standing volume predictions with data used to fit the models. Volume predictions were imprecise but unbiased on average. Predictions of standing volume and tree size development demonstrated the influence of site quality and stocking on A. melanoxylon growth and yield. Stands located on average sites, thinned to 200 stems/ha at age 10, were predicted to have 290 m³/ha total standing volume and 49 cm average diameter at breast heighnt at age 35 years, giving a mean annual volume increment of 8.3 m³/ha.

A new open-pollinated breeding strategy for Eucalyptus nitens (Deane & Maiden) Maiden in New Zealand was explored using microsatellite markers to reveal the parental identity of forwards selections. Microsatellites are the preferred markers to reveal genetic relationships between individuals, largely owing to their co-dominant inheritance.

Forwards selection of individuals for the breeding population and future deployment was simulated using 10 open-pollinated seedling offspring from each of 10 clones in a clonal seed orchard. A set of 15 microsatellite markers was chosen from the 41 initially tested. Ninety of the 100 progeny sampled matched consistently to a single mother and father and 13 of these were evidently selfs. Eight had a maternal match only; this would indicate that either there was contamination by pollen from outside the orchard or there was an occasional failure by the marker set to identify the orchard pollen. One seedling had no maternal match and it was not possible to discriminate between two fathers in another.

There was a broad level of outcrossing at the individual and provenance levels, and there appears to be little indication that individual clones favour specific pollens. Estimates of the coefficient of inbreeding and coefficient of co-ancestry were derived for the seed orchard and forward selections.

Maintaining the stability and productivity of residual forest after harvesting is a crucial requirement of sustainable forest management. We assessed the long-term (43-year) impact of group selective harvesting using ground extraction of one-third of merchantable volume on 30 ha of mixed conifer/hardwood forest at Pureora, North Island, New Zealand. Selective harvesting did not substantially alter canopy composition or the population structure of conifers in the two 15-ha harvested blocks. It did not adversely affect the stability of the residual forest, with similar mortality rates in merchantable conifer trees in the 15-ha control (0.7%/year) and residual conifers in the two harvested blocks (0.6%/year, 0.8%/year). Tree mortality in harvested blocks was not related to harvesting disturbance. Productivity was somewhat reduced by selective harvesting (0.3,0.2 m³/ha/year in harvested blocks; 0.4 m³/ha/year in the control) because of reduced tree densities. Net merchantable volume increment was negative in conifer species in all blocks, ranging from -0.4 to -0.6 m³/ha/year. Although regeneration of major canopy species had largely recovered within a decade of harvesting much more time will be needed for the forest as a whole to recover fully.