In this paper, productivity will usually mean the cubic meters of harvestable wood that can be grown per year on a forested site. But it is more complicated than that. Productivity not only covers harvestable wood, but the quality of that harvested wood for various purposes. Even more broadly, it includes the productivity of other goods and services an indigenous or plantation forest can, or could, provide. In some cases, all of these elements of productivity are improved or harmed by a particular practice; in other cases, tradeoffs among them must be considered.
Many of the factors in this section have been studied in various settings, and the results are often situation-dependent. For example, fertilizing on an already fertile site often has little effect, while the same amounts of fertilizer on a nutrient-deficient site can have dramatic positive effects, while that level of fertilization on yet a third site may trigger an epidemic of some previously-mild endemic pest attracted to the better nutrition then provided in the trees foliage. Most of these many factors interact with each other. Many have been studied separately, or in groups of combinations (for example, Bower 1999), but rarely are all interacting factors accounted for. Getting it wrong in one factor can negate the good that excellent inputs in several other factors might have been expected to achieve (nicely illustrated in Davidson 1996; see also Section 2.12 below).
In the topics that follow in Section 2, each factor will be briefly discussed, but precise quantification of changes in productivity with particular inputs will not be presented. A comprehensive review is beyond the scope of this paper; furthermore, many of the available quantifications are based on young trees, and they would probably be misleading. Some remarkable changes in wood and non-wood productivity will be presented in the case examples in Section 3. It is important to acknowledge here that establishing and managing a productive forest plantation requires high levels of professional knowledge, experience and judgement.
Indigenous forests are often remote or even inaccessible. Since people need to get to a forest plantation site in order to plant it, access to most forest plantations is usually available for their later management and harvest. The relative availability of access does not directly affect productivity. It can indirectly affect it by its influence on the ability of management to exert leverage through such things as post-planting release, thinning and pruning, and other management activities. Furthermore, accessibility affects the energy it requires to deliver forest products to their users, and thus the economic feasibility of using them.
As agriculture increasingly mechanizes, it has been noted that forest-management activities are becoming relatively more labor-intensive compared to agriculture. Thus, it makes increasing sense to have forests near villages and agricultural operations more distant from them. This change carries an added benefit for villagers who use wood as fuel for cooking and/or heating their dwellings, particularly because food can usually be more easily transported long distances than can the wood to cook it (Smith 1981).
A common spatial relationship has been one of villages surrounded by forest clearings growing crops, with forests at ever-greater distances from the villages as the villages increase in population and more land is cleared. As this relationship inverts, forests will probably be re-established near the villages as forest plantations, although natural regeneration could be used to maintain them once established.
Of all the factors in this section, it appears that site quality has the greatest and often longest-lasting effect on differences in the productivity of various indigenous and plantation forests.
The important components of site quality are: soil depth and drainage; soil physical and chemical composition (including pH); amount and pattern of yearly soil moisture availability; frequency and nature of common and occasional winds, storms and fires; and the general climate of the area. The presence and importance of such things as competing vegetation, and of populations of animals, insects and microorganisms that are either damaging or beneficial to trees, can and should also influence selection and later management of forest plantation sites. The size and health of trees already present on the sites are good but imperfect indicators of site quality. Different species of trees on candidate sites will index them differently, and genetic differences within the same species can also result in substantial differences in derived site-indexes between similar sites, or vice versa.
In most forested regions of Earth, people have recently and historically cleared forests for pasture and crops. In retrospect, many such cleared sites have been too erosive for continued pasturage or cropping, or have proved to be (or been degraded to become) marginal or sub-marginal for their intended agricultural uses. The good news is that such erosive or marginal and sub-marginal agricultural sites often prove to be good or even excellent sites for forest plantations. Furthermore, some non-forested sites very unpromising for agriculture, such as the coastal dunes of southwestern France, northwestern New Zealand, eastern China and western Senegal, have been planted with forest trees and are now productive forest plantations. Finally, many forest plantations are established for a variety of important primary purposes other than supplying wood (Palmberg 1989; Laarman and Sedjo 1992), for example, to rehabilitate sites degraded by uses such as strip-mining. Where such sites are not evaluated and chosen for their potential wood-productivity, any wood eventually produced by them may be viewed as a side-benefit.
The first decision to be made about each forest plantation site is whether it will be a mono-species plantation, or whether it will be planted to two or more species in mixture. European tradition has favored mono-species forest plantations, largely based on European experience in trying to domesticate their indigenous species. However, if the forest plantation is to provide aesthetic, wildlife-habitat, biological diversity, and other services in addition to wood-production, multi-species forest plantations will often be favored. If the forest plantation species of choice are (is) indigenous to the region, then assigning a species or mixture of species to available forest plantation sites can usually be done with reasonable accuracy.
In many regions, introduced species outperform the local species in wood production by multiples of 2, 5, 10 or even more. Extensive species-introduction trials should precede widespread commitment to even the most promising introduced species. But even with such species-introduction trials indicating outstanding performance, small-plot performance in such trials is not always repeated in widespread forest plantations. This may happen, for example, when some pests or pathogens miss the early small trials but later find the large areas of host trees. This can be particularly damaging if the pests or pathogens are also exotic, and arrive without their natural biological controls.
During the past several millennia, a remarkably short list of plants and animals has been successfully domesticated from a remarkably long list of plants and animals that were hunted and gathered for food or befriended as pets and houseplants. The same may or may not be true for forest trees (Hansen and Kjaer 1999).
If indeed many species of trees are to be domesticated, rate of progress on many of them is likely to be even slower than occasioned by the constraints of their large size and long generation times. A list of the species that currently seem likely to be domesticated would include many members of the Pinaceae, surely several pines (Pinus radiata, P. taeda, P. pinaster, P. caribaea, P. patula and others), perhaps a few spruces (Picea abies, P. sitchensis and others), maybe Douglas-fir (Pseudotsuga menziesii), likely several members of the Cupressaceae (Cryptomeria japonica, Cunninghamia lanceolata, Sequoia sempervirens and one or more Cupressus) and perhaps others (one or more Taxodium, the hybrid Cupressus macrocarpa by Chamaecyparis nootkatensis, and Sequoiadendron giganteum). Few or no Southern Hemisphere conifers are likely to make that list. Among temperate and boreal angiosperms, several poplars (Populus) and the many promising hybrids among them, several willows (Salix) and their hybrids, walnut (Juglans nigra and perhaps others), and maybe some cultivars from genera such as Quercus, Betula and Acer seem possible or likely domesticates. Among subtropical and tropical species, Eucalyptus grandis and several other members of the promising Eucalyptus genus, several Acacia species, Gmelina arborea, Tectona grandis, Swietenia macrophylla, Paulownia tomentosa, one or more members of Leucaena, Prosopis, Casuarina, Grevillea and Dalbergia, and a few others are likely candidates for that short list. Some of these will no doubt fall out of favor as attacks by pests and pathogens result in unacceptable levels of loss, or as difficulties such as recalcitrant reproductive biology or weedy invasiveness, become apparent.
Domestication implies genetic changes in populations to suit human needs or purposes. But it is more than that. Domesticated plants and animals rarely reach anywhere near their full potential productivity or performance without husbandry that is adapted to the domesticates. Forest trees will probably be no different.
To some people, natural selection as a force for adaptation and change in nature is good, or at least acceptable, while human-directed selection is suspect. But (at least until modern biotechnology came along) the rules were fundamentally the same. Humans, as a component of the environments of the plants and animals being domesticated, changed how some traits influenced reproductive success. But most of the rules of nature were and are still enforced, and that will be particularly true with respect to domesticated forest trees. Since most forest plantations are left pretty much on their own most of the time, plantation trees with human-directed trait-changes must still function well in what is mostly still the domain of nature.
For any breeding program using parent trees not indigenous to the local region, a first step is to learn or determine the species pattern of provenance variation. Efforts are then concentrated on selecting parent trees from populations that are already well-adapted to the region. Detailed examples for several tropical species are given in Vichnevetskaia (1997). In the first generation of most tree-domestication (sometimes referred to as tree improvement) programs, much attention is given to finding and breeding trees that grow much faster and are individually much larger than average. In the next few generations, the focus usually shifts to finding and deploying families or clones that produce more wood per hectare.
Although some people believe that most organisms are perfectly adapted to their indigenous environments, that in fact is rarely true. Most are imperfectly adapted to the recent environments in which their ancestors survived and reproduced. In order for tree-breeders to increase the arboreal biomass produced per hectare per year in forest plantations, they must select trees that: (a) better and more-completely use those forest plantation environments; and/or (b) that function well with reduced rates of catabolic respiration; and/or (c) that lose less of their biomass to things that damage them or eat parts of them. In short, they breed them for better and/or more efficient adaptation to the forest plantation environments. Many tree-improvement programs have had substantial success in increasing cubic meters of wood produced per hectare per year, with per-generation increases in the neighborhood of 10% or more commonly reported in the early generations. It seems reasonable to expect diminishing returns in later generations. However, environments can be expected to continue changing. The pursuit of adaptations to these changing environments will continue, and humans can assist in it. Thus, continued attention will be given to adaptive traits, but the greatest gains in adaptation will probably be recorded in the first few generations of tree-breeding.
A second strategy, widely used in agriculture, is to improve the so-called harvest-index. In this strategy, biomass is reallocated within crop plants such that the percentage and/or value of the harvested portion of the plant is increased within a fairly constant total biomass produced per hectare per year. The families and clones deployed to forest plantations in later generations will be selected from highly-adapted breeding lines, and they will have combinations of additional selected traits that confer much higher harvest-indexes than are typical of the breeding line as a whole.
The combination of adaptation and allocation strategies has allowed impressive gains over many generations in many domesticated animals and agronomic crops. We may expect similar substantial and sustained success when domesticating trees for forest plantations (Libby 1987).
For a while, it was popular to expect that many of the potentially most-productive families or clones of forest trees would be interactive prima donnas, capable of outstanding performance under conditions favorable for them but likely to perform poorly where conditions were sub-optimal for them. Deploying these to the right sites seemed likely to be the best strategy with respect to harvest productivity. However, this was easier said than done. One major problem was foresters inability to adequately characterize sites in advance of planting, compounded by the rather fine-grained heterogeneity of many forest plantation sites. We also began to wonder if meaningful climate changes might occur at a time-scale shorter than the period from planting to harvest. These factors have combined to shift the preference from potentially highly-productive interactive families and clones to a more-conservative strategy of selecting for broad adaptation.
One sometimes hears the argument that clones are more interactive than seedlings. Although it is hard to prove, this seems unlikely. Genotype-by-environment interaction can be clearly demonstrated for a clone because it can be grown in more than one environment, while a seedlings interaction can be investigated only indirectly by the performance of its relatives on contrasting sites. Zobel (1993, and follow up personal communication) has noted that about 50% of the productivity increase in the second round of the eucalypt program of the Aracruz company in Brazil was probably achieved by deploying proven broadly-adapted clones instead of selected families of seedlings, many of which were probably interactive and by chance planted in the downside environments for their genotypes.
Naturally-regenerated forests usually have family structure, with sibs, cousins and other relatives growing near each other. When these related trees mate with each other, their offspring are to some degree inbred. In the great majority of forest-tree species, inbred trees are less healthy and grow more slowly than do their outbred relatives. This is generally not a problem in naturally regenerated forests, given that the numbers of seedlings established substantially exceed the numbers that will survive the early competition among them. Natural selection weeds out the less-adapted inbreds, and most of the ultimate over-story trees will be the healthier faster-growing outbreds.
Such inbreeding can be a problem if seeds are collected from naturally-regenerated forests, grown in a well-run nursery, and planted into a site-prepared forest plantation with good post-planting care. In this case, many of the inbreds will escape the early harsh natural selection that typifies abundant natural regeneration. However, they will not grow as well and they will be less healthy than outbred trees. If the forest plantation is thinned, most of the inbreds that have survived to that point will probably be removed in the thinning. But this either requires a higher planting density in anticipation of the general poor performance of the inbreds, or it loses some options for the forester during the thinning selections. In some cases, forest plantation practices will be good enough and initial spacing wide enough, or the frequency of inbreds in the planting stock high enough, so that some or even many of the planted inbreds will still be present among the crop trees. If so, that will be to the substantial detriment of forest plantation productivity.
Single-tree-selection harvesting is often aesthetically preferred to clearcutting. However, the trees that reach harvest size sooner, and/or (under market or other production pressures) the better trees, are often harvested while the slower-growing trees and those of less value are left to subsequently reproduce. In such cases, the systems of logging employing partial cutting and natural regeneration are to some degree highgrading the forests, to the long-term genetic detriment of the quality and wood-productivity of the naturally-regenerated forest.
Under domestication, the selection and breeding of trees for forest plantations reverses past and current dysgenic trends associated with highgrading followed by natural regeneration. By assuring that no related trees are parents in seed-orchards, or by deploying only outbred control-pollinated pedigreed families and clones to forest plantations, inbreds will not be planted in those plantations. The joint effect of eliminating inbreds and reversing highgrading is responsible for a large proportion of the first-generation gains being reported by tree-improvement programs, particularly when they contrast the performances of their first-generation trees to those of trees from seeds collected in naturally-regenerated forests.
Appropriate nursery practices include the shipped nursery stock having: reasonably uniform and appropriate propagule size; good root configuration and proper balance between roots and shoots; root systems physiologically ready to commence active growth once planted; the propagules free of harmful insects and diseases when they leave the nursery, but sometimes well-infected with beneficial mycorrhizae; and accurate records as to provenance, family or clone, so that the propagules are correctly deployed to their planned sites. For many species, appropriate nursery practices have been the difference between good survival and early commencement of rapid growth of most planted trees, or slow and erratic commencement of growth or even forest plantation failure.
If the planted trees uniformly commence rapid growth soon after planting, this has several beneficial effects on productivity and efficiency. One effect is to shorten the time until harvest, thus increasing productivity per unit time. It increases efficiency by reducing the need for post-planting care or even eliminating it, with the secondary benefit of thus requiring less herbicide and pesticide during the post-planting period. Uniform establishment will reduce differences in the size of the trees, both at harvest and at the time of thinning. If the trees reach thinning age free of serious lingering nursery effects on size, more attention can be paid to differences in tree form and to genetically-controlled differences in growth rate when selecting the crop trees and removing the poorer trees. If a higher percentage of crop trees has reached the desired size at the time of harvest, that should increase harvest-index. These latter two effects should increase both absolute productivity and product quality.
Adequate site-preparation generally involves reducing or removing competing vegetation. It may also involve bedding or improving drainage on wet sites, breaking up a water-repellent surface layer, or shattering a hardpan layer. Reducing logging slash on the ground during site-preparation removes cover for animals that injure newly-planted trees and improves access and visibility for the planters, thereby increasing their efficiency and perhaps their effectiveness. Good site-preparation confers many of the same benefits to productivity and efficiency as does good nursery stock, and for the same reasons.
Spacing in forest plantations is usually uniform, with the number of trees planted per hectare being only a fraction of the numbers that typically establish with abundant natural regeneration.
While trees are sometimes planted into the understory of established forests (with mixed success), they are typically planted into open or clearcut sites. If a mixed-species forest plantation is being established, sometimes propagules of one or more species are planted in advance of the other(s), or older and larger planting stock is used for some than for others, both strategies generally aiming for similar sizes at about the time of crown-closure. One effect of these latter two practices is the appearance of an even-aged stand even though there may be two or more slightly different age cohorts in the forest plantation. Conversely, poor nursery stock, site-preparation, planting practices and/or aftercare may result in the appearance and structure of a somewhat uneven-aged stand, even though all planted trees in the forest plantation are exactly the same age.
As compared to uneven-aged silviculture or abundant even-aged natural regeneration, a larger fraction of the typical site is unoccupied by trees during the first years after site-preparation and planting. This results in a small-to-modest decrease in the biomass productivity of the tree component of the forest plantation ecosystem. However, planting relatively fewer evenly-spaced propagules usually results in more-uniform tree size and trees more similar in form at the time of harvest, with harvest-index likely increased as a result (for example, Bower 1999). With moderately-spaced trees of similar size, there are some efficiencies available to management with respect to thinning and pruning operations, to other treatments that may be applied, and to harvesting.
The need for rigidly-uniform spacing is probably overrated, and may not only be unnecessary but counterproductive. For example, at a public forum in New Zealand in 1998, it was found that most of those present approved in principle of having forest plantations, but they objected to the straight-row cornfield appearance of the young forests. Research with some species indicates that, if planted trees are each provided an adequate area in which to grow, say 10 square meters, it makes little difference to their growth and form whether each tree is in the exact center of its area, or whether some are closer together and others farther apart, as long as the crowns and roots of each have access to about 10 square meters of forest plantation space. More on this under planting, below.
While there are management advantages to even-aged stands, there are some possible downsides as well. If the species is susceptible to damage from different pests and diseases at different stages of its maturation, then whole stands will be susceptible to the same damaging agents at the same time, possibly leading to buildup or even epidemics of damaging insects and diseases. This of course may also allow management to focus on problem areas with countermeasures if such buildups occur.
Some forest plantation managers give in to the temptation to use untrained and inexpensive planting crews. These short-term cost-savings may be false economies. For example, a recent study of 24 Pinus taeda plantations in Georgia USA found that incorrect planting of both seedlings and stecklings led to an average 9% loss of height growth, an average 7% loss of diameter growth, and a greatly increased incidence of stem crookedness during the first 10 years on site (Harrington and Gatch 1999). Surprisingly, the incidence of fusiform rust, tip-moth injury and bark-beetle injury were all greatly (60% to 400%) increased among the incorrectly-planted trees.
The Swedish University for Agricultural Sciences has studied the effects of carefully training planting crews in Swedish forest plantations. These crews are not only trained to plant the trees correctly, they are trained to find a favorable microsite within the area allocated to a tree, and to plant it there. This practice breaks up the rigid uniformity of rows and inter-row spacing, resulting in a more natural- appearing young forest. It costs more, but it also results in better survival and earlier, more-uniform commencement of rapid growth.
In short, good planting has similar positive effects on forest plantation productivity, harvest index, and management efficiency as do good site-preparation and good nursery stock.
The need for post-planting care ranges from unnecessary to crucial, depending on the site and local biota of pests and competing vegetation. Release from competing vegetation serves in a similar way the same productivity-increasing purposes as do good nursery stock, good site-preparation and good planting, discussed above. Release from competing vegetation also decreases the time and lowers the probabilities that the newly-planted trees may be killed or damaged by such fauna as browsing animals, tree-girdling rodents, seedling-attacking weevils, etc. Control of competing vegetation by physical or chemical tools is probably the most common post-planting care, but direct protection from animal damage may be provided by screening individual trees or fencing the forest plantation, and baiting or insecticides may be used to reduce damage from rodents and insects.
Soil moisture and fertility are two important parameters of site quality, and they may be effectively modified on less-than-ideal sites. For example, fertilizer may be applied to augment nutrients found to be in short supply. (It is more difficult to offset overabundant elements, such as magnesium in ultramafic soils.) Drainage may be provided for waterlogged sites, and irrigation may be done when moisture is in short supply. Irrigation and fertilization, both expensive, are more feasible if wastewater from dwellings or industrial operations can be used. This latter is another advantage of having forest plantations or forested areas located near towns and villages. Short of irrigation, reduction or elimination of competing herbs, brush and unwanted volunteer trees is the most effective way to increase and conserve soil moisture for use by the planted trees (for examples, Powers and Reynolds 1999)
Both of thinning and pruning, which are common silvicultural operations in some countries and in some species, are likely to reduce total wood productivity. But, by reallocating growth onto and within thereby-larger trees, they usually improve average crop-tree form, and improve value and harvest index (e.g. Bower 1999), thus resulting in improving and even increasing net wood productivity compared to unthinned and unpruned stands.
Both thinning and pruning improve management access and reduce the risk of stand-destroying crown fires. The general health of the thinned stand is often improved by having removed diseased and/or insect-infested trees, and by the better vigor that is maintained or develops among the released leave trees.
If all the above tree improvement features are brought together, a rising trend in productivity can be expected in a forest plantation development program. However, as mentioned in Section 2.0 above, if any one is neglected, it is likely that the whole will suffer disproportionately. For example, operations should not exclusively minimize harvesting costs, but those of harvesting, re-establishment and initial weeding as a holistic activity and without impairing vigour. Evidence of a rising trend reflecting the interplay of these gains is seen below as experienced and predicted for Usutu Forest in Swaziland, 1992-2028, based on research results and field trials for each component (Evans, 1999, based on the work of Morris, SAPPI).
Figure 1. Pinus patula, Usutu Forest, Swaziland, estimated yield gains, 1992-2028 (Current base is taken as second-rotation crop yields)
Source: Evans, 1999.
Holistic management also embraces active monitoring of pest and disease levels, and researching pest and disease biology and impacts will aid appropriate responses such as altering practices, e.g. delayed replanting to allow weevil numbers to fall. Careful re-use of extraction routes to minimize soil compaction and erosion is a further example.
Forest plantation managers are increasingly aware of public and regulatory concerns about biological diversity, and are responding to them. Forest plantations usually have less total biological diversity than do indigenous forests, and their associated biota are also different in composition from those of indigenous forests in the same area. However, most forest plantations host much greater biological diversity than do most agronomic crops. If indigenous forests are common in the region, as they should be, then the different suites of biota sustained in forest plantations add to regional biological diversity. This is a clear positive benefit of forest plantations.
Dense natural reproduction, or Central European-style forest plantations with 2000 or more trees planted per hectare, have been referred to by some advocacy groups as, biological deserts. The dense stands of young trees may exclude almost all other vegetation, and thus the many other organisms that would live in or on the associated flora are also absent. Thinning these dense stands and forest plantations will normally allow these associated understory plants to resume their presence. A major advantage of deploying reliable families or clones to forest plantations is that planting densities can be lower, often 1000 or fewer trees planted per hectare, to still achieve similar or even better final-harvest stands. Lower planting density not only reduces planting and later thinning costs. It also leaves much more open niche among the planted trees, where a rich associated biota develops soon after planting and persists through much or all of the rotation.
If indeed carbon credits become available, they will probably be largely provided for two major forest-management options: (1) increasing rotation length in established forest plantations and indigenous forests; and (2) establishing new forest plantations on land that is not now forested. The first will serve to subsidize a likely increase in productivity as the forest plantations fully occupy the sites longer before harvest. It will also increase the value of the wood when harvested from either plantation or indigenous forests. The second, serving essentially to provide cash-flow for new plantations, allows management to better afford some early and mid-rotation activities in the factors above that will increase productivity (for more information on the issue of forest planatations and carbon credits, see Forest Plantations Thematic Papers FP/12, FAO 2001).
 Vegetatively produced
 Swedish University for Agricultural Sciences, Box 7054, S-750 07 Uppsala, Sweden.