Sustainable forest management has been and still is a central objective and a recurrent slogan in forestry. Since the birth of forestry as a science, the ideal of ‘sustained yield’, i.e. maintaining a continuous flow of goods and services from the forest, has occupied a central place in forestry thinking.
This way of thinking anticipated the recent view of sustainability, which assumes that there are desirable states for ecosystems, e.g. managed forests, that humans can maintain indefinitely.
The raising awareness that ‘we must face the impossibility of even defining—let alone pursuing—agoal of ‘sustainability’ in a world characterised by extreme complexity, radical uncertainty, and unprecedented change” is pushing towards a shift in the discussion on forestry not only from a scientific and technical point of view, but also from a policy perspective.
Forestry is now facing a paradox which consists in aiming at sustainability in a changing environment and in a shifting perception of the relationship between ecological and social systems.
Natural systems and social systems are complex systems in themselves. The term social-ecological system (SES) was used to emphasise the integrated concept of humans-in-nature and to stress that the delineation between social and ecological systems is artificial and arbitrary.
SESs are composed of (1) biotic agents ranging from microbes to plants to humans, each with a different degree of information-processing capacity; (2) a set of allowable actions related to their physical or behavioural characteristics; and (3) a physical substrate that includes chemicals, light, and water.
The interactions among these agents and their interactions with the substrate generate dynamic social-ecological systems.
In a complex SES, subsystems such as resources, users and governance systems are relatively separable but interact to produce outcomes at the SES level, which in turn feedback to affect these subsystems and their components, as well other larger or smaller SESs.
Complex adaptive systems theory has been suggested as a means for better understanding forest ecosystem functioning and shaping more effective management approaches, forest ecosystems exhibit all the characteristics of complex adaptive systems because they are heterogeneous, highly dynamic and contain many biotic and abiotic elements which interact across different levels of organisations with various feedback loops.
Forests are non-linear systems, highly sensitive to initial conditions, which makes precise predictions about their future behaviour very difficult. They also show a hierarchical organization: elements at different levels interact to form an architecture that characterises the system.
Forest ecosystems are open to the outside world exchanging energy, materials and/or information. Forest ecosystems’ components and processes interact with each other and with the external environment in many different ways and over multiple spatial and temporal scales.
The need for taking into account the social and economic issues in forest management is not a new concept, but in practice this has usually been translated into adjusting management so as to shape forests to respond to these issues. Therefore, the traditional forestry vision still interprets social and ecological components of forest dynamics as inhabiting fundamentally separate domains.
Instead, interactions with the social and economic systems, such as industries, governments, local communities, and other users of forest products and services together with their cultural backgrounds form an integral part of forests.
This is a particularly relevant issue for managed forests which have been, are and will continue to be profoundly impacted by changes in the social and economic systems, while at the same being themselves drivers of change in these systems, e.g. forest degradation as a cause of poverty, or, at the other extreme, mature, diverse and socially attractive forests in parks and nature reserves as factors of economic development.
Similar to the environmental components, the social, economic and policy elements of the social-ecological system formed by forestry can be assessed at various levels, so that the broader ‘landscape’, defined on a social and/or environmental basis, where forests and forestry are just a part, and whose boundaries and size may change depending upon the issue, constitutes the right level to which forest management, policy and governance issues must be addressed.
All these considerations point out the need for integrating the social and ecological dimensions of forestry into a single framework.
In the conventional forest management approach, multi-functionality, i.e. the provision of multiple goods and services to society, has been based on the ‘wake theory’ which states that if forests are efficiently managed for wood production, then all the other forest utilities will follow.
Dynamics and reactions from other interacting systems have been ignored and the consequences have often been and still are harsh conflicts (e.g. between wood production, landscape and nature conservation, recreation and related stakeholders).
In addition, recent examples show that societal preferences and values can change drastically in a relatively short time radically altering the social environment for forest management.
When considering forests as adaptive systems interacting with the economic and social systems, the concept of multi-functionality changes from a sum of different outputs of forest management activities to a set of complex interactions between various sub-systems.
Forests As Social-Ecological Systems
Resilience of social-ecological systems has been defined as the capacity of the system to absorb disturbance and reorganize while undergoing change so as to still retain essentially the same function, structure, identity, and feedbacks. However, during adaptation a system can undergo major reorganisations: it can still sustain major functions but it may lose its identity.
Adaptability is the capacity of actors in a system to influence resilience, to avoid crossing into an undesirable system regime, or to succeed in crossing into a desirable one. In a social ecological system, this amounts to the capacity of humans to manage resilience.
This leads to consider forest adaptation to disturbances as a dynamic process which involves the system’s resilience and flexibility, not only from the ecological point of view but also from the social-economic one, i.e. concerning both the provision of forest goods and functions and the relation with society’s value system.
The key to resilience in any complex adaptive system is in the maintenance of heterogeneity, the essential variation that enables adaptation.
Conventional forest management, aimed at maximising wood production and based on a command and control approach, has simplified the structure and composition of forest. This simplification, reducing response diversity, makes these systems fragile, more vulnerable to stress, such as parasites, climate change, etc. and therefore, more prone to collapse because unable to respond in an adaptive way.
Forestry Decisions As A Product Of Interactions
In forestry, an increasing part of decisions emanates from a series of interactions between actors that engage with/and react to/one another.
Any decision-making procedure has to cope with these interactions, addressing the following questions: (i) who/what interacts?; (ii) what drives or shapes the interactions (global context, actors’ interests, values, perceptions, knowledge and resources)?; (iii) what are the mechanisms/pathways of interaction (networks, communities, communication patterns)?; (iv) what is the character of the interactions (competition, coordination, co-optation)?; (v) what are the effects of interactions (stringency or homogenization)?; and (vi) how do interactions change over time?
Those social/institutional interactions permanently interfere with the ecological changes, giving the whole system a dynamic that is constitutive of what is called forestry, and resulting in an instability which the actors tend to reduce or master. Although whilst doing so, they introduce new perturbations leading to new feedbacks defining new demands for change.
Changing The Forest Management Paradigm
Forest management has long been dominated by the reductionist and mechanist paradigm founded on two basic principles: (i) perpetuity of the forest based on an equilibrium between standing volume, standing volume increment and allowable cut; (ii) constrained optimisation of productions (marketable or not).
This forest management paradigm considers population and ecosystem dynamics as if they were acting in an invariable environment and according to predictable trajectories.
In such an approach, silviculture aims at controlling natural processes, and cultivation methods try to obtain forest regeneration according to a predefined stand structure model: even-aged or uneven-aged. Forest management tends towards a regulated distribution of age or diameter classes.
This means that silviculture acts as if forest ecosystems could be governed by controlling a few key variables, while other aspects are practically ignored and classified as casual effects. In this approach forest ecosystems are supposed to be totally understood in their functioning and thus shaped so that future results meet management aims.
Yield tables for even-aged stands, or norms for uneven-aged forests, are the main expression of the classical idea that, in principle, by managing forests precisely following such ‘optimal’ schemes, forest growth will probably match managers’ expectation.
This paradigm inherently assumes that: (i) forest ecosystems react to management in a predictable manner; (ii) it is then expedient to anticipate predicted consequences of decisions (i.e. anticipatory management: once all necessary information is gathered to make a scientific forecast, the ‘right’ decision can be made).
Undoubtedly, in former times this approach has contributed to regulate forest exploitation and slow down forest destruction. But classical silviculture and management, with the aim of predicting regeneration rate and producing a constant yield of merchantable wood, have in practice transformed complex ecosystems into simplified systems.
Examples of forest stand simplification in the past centuries are the conversion from mixed forest types with prevailing hardwoods to prevailing conifer forests to foster commercial timber production in Central Europe, or the conversion of mixed forest types with prevailing conifers into hardwood coppices to foster fuelwood production around villages in the Italian Alps and Pre-Alps.
Successfully managing a forest to maximize production of a service (or set of services) may lead to a less resilient and more vulnerable system, not only from the ecological but also from the institutional perspectives.
Actually, a vast bulk of evidence from operational forest stand management shows that predicted outcomes are rarely achieved, at least for naturally originated forests.
Already in 1993, there has been assertion that classical silviculture is based on a short-term perspective and on a paradigm which considers the forest as being in a constant state and with a constant production.
With a full appreciation of the natural complexity of forested landscapes, these assumptions become untenable because processes appear linear and states appear constant only over a limited spatial and temporal field: therefore, foresters must shift their emphasis from maintaining the forest in a given state to maintaining particular processes, and change from concentrating on trees to concentrating on the ecosystem.
Issue Of Low Predictability
When dealing with complex adaptive systems only hypotheses can be drawn about the effects of management practices. Forest functioning and structure, specifically forest reactions to management, are neither completely predictable nor completely random: like many complex systems, forests are characterised by multiple feedback links and close dependency on initial conditions, so that prediction has only a weak power and there is always a high degree of indetermination.
Prediction reliability is conventionally considered an essential feature when applying models as quantitative tools in forest management. Under this view, forest ecosystem processes (e.g. growth, regeneration, succession) are supposed to be fully predictable and, thus, can be manipulated so that forest responses to silvicultural treatments meet management expectations.
As far as natural and semi-natural forests are concerned, even for the simpler volume or increment forecasting, the predictive power of models is so dramatically limited by structurally unpredictable events to be practically constrained to theoretical productions.
Wind breaks, ice storms, diseases, drought, forest fires, pollution, dieback, etc. are examples of phenomena that cannot be considered as having marginal effects on stand development in natural and semi-natural forests. A vast bulk of evidence from literature and practical experience highlight that moderate-to-severe disturbances, both natural and human-caused or human-induced, recur relatively frequently in such forests, falling within the spectrum of chronic and acute effects that stochastically drive ecosystem patterns and processes.
Since the 1970s, there has been a growing effort in developing the modelling approach to growth-and-yield studies. Conventional yield tables have been criticized and there have been great advancements, e.g. in the development of process based models.
The current portfolio of advanced modelling techniques is wide and also focused on complex systems, so that hypothetical indications on future states of forest stands and landscapes can be actually provided over a wide range of conditions, and robust statistical tools, like e.g. Bayesian model averaging, have been developed to better assess the uncertainty of model predictions.
But what is attainable even by these advances is just hypothetical accuracy and uncertainty assessment, since modelling basically assumes that no other factor has an influence on the modelled phenomenon at hand beside those considered by the model itself.
When the variables of a model (or of a set of models) are identified in relation to the observed properties of the phenomenon, many aspects of the phenomenon itself are reduced to only one or few variables or not even represented in the model at all. How then can it be said that “the model is a model of that phenomenon”?
This is constrained by a list of assumptions which specify the model and allow for the logical connection between the model and reality. But no list of assumptions can ever be complete: there is no way of excluding the possibility that some disturbing factor, other than the ones explicitly considered, might have an influence.
And this is even more evident under the acknowledgement that no stationary state can be claimed for environmental conditions. Only if this list is complete we can be sure that the model is a model of the phenomenon at hand. Therefore, a blanket assumption must be implicitly incorporated in the construction of a model: ‘no other disturbing factor is operative’.
Of course, this is a problematic assumption since there is no way of guaranteeing its truth!
This is the reason of the practical poor usefulness of the anticipatory approach based on model forecasting for the management of natural and semi-natural forests where the effects of disturbing factors is never negligible.
While ‘models may be deficient instruments for the reduction of uncertainty as to future system behaviour’, they may serve heuristic and theoretical functions and may outline the space of possible behaviour.
Distinctively, they may contribute to operational guidelines by supporting field management with the objective of performing a few, crucial experiments whose results can eliminate a large number of alternative assumptions, according to an adaptive framework.
In the light of this, we acknowledge that a combination of hypothetical-deductive modelling with guided experimentation may help in understanding certain aspects of how forest ecosystems function and provide context-sensitive knowledge, and that the hypothetical ability of models to asses uncertainty under certain conditions has potential uses in management, but this perspective is very different, both conceptually and operationally, from founding the management of natural and semi-natural forests on the alleged full predictability of their development as embedded in the anticipatory management approach.
Adaptive Forest Management
Adopting an adaptive management approach explicitly considers the system’s low predictability as a value, as its capacity to react to impacts, and requires learning from system reactions to support its resilience.
Silvicultural interventions, artificial by definition, impacting on the structure of forest stands, provoke a certain level of stress in the system: artificial impact must be constrained within the limits of the forest ecosystem’s resilience.
Understanding that natural systems are able to preserve their internal organisation, withstanding even major structural modifications, helps finding key elements for management. Shifting methodological focus from a priori determination to a posteriori assessment implies a heuristic approach or a system theory of trial and error.
In this, we agree that “incremental and adaptive management strategies directed towards feedback mechanisms and reflexive learning processes seem the proper way to cope with an undetermined future and the problems of risk, uncertainty, ignorance and indeterminacy”.
Therefore, successive forest transformations resulting from human interventions, whether of structural or marginal nature, must be observed and interpreted considering the complex interactive relations linking the management subjects (forest and humankind).
Under such a framework, management is urged to move from approaches based on forecasting (i.e. the root of the anticipatory management idea) to approaches based on monitoring: focus is not on the prediction of the effect of each intervention, but rather on the reaction to it as tracked by relevant indicators.
This means moving from a strictly ruled forest planning to adaptive management where, generally, indicators (e.g. regenerative success, net annual increment, or the proportion of healthy individuals) are not intended as reference thresholds but instead as parameters to measure changes over time.
Using sets of criteria and indicators has become a common way to evaluate aspects of sustainable forest management, and several approaches used for certification issues, assessment of forest conditions or adaptability to climate change are described in the scientific literature.
However, forest management is not just an ecological, silvicultural or harvesting issue, because there is a network of ecological, technical and socio-economic aspects which increases problem complexity.
Within a system analysis approach, indicators should allow referring to the context of forest management, measure the quantity and quality of the actions taken and the feedbacks of the forest system to these actions.
The consequence of the change in paradigm has been the proposal of a systemic approach in silviculture and management, with the definition of ‘systemic silviculture’.
With systemic silviculture, management strategies are based on an adaptive approach and continuous monitoring of the reactions of the forest to silvicultural interventions.
Management proceeds along a co-evolutionary continuum between human intervention and reactions of the system, which de facto excludes the typical finalism of linear processes which leads to the normalisation of the forest.
Compared to other recent alternative forest management approaches which have been proposed during the past years in various parts of the world, such as close-to-nature forestry, variable retention forestry or ecosystem management, systemic silviculture takes into account most of the characteristics of forests as complex adaptive systems.
Viewing forests as complex adaptive systems shifts management from the stand level to the landscape level. But because systems have a different behaviour at the different scales which interact in complex ways, the stand level is also relevant from the silvicultural point of view.
When dealing with management, the ‘regulated forest’ is the reference model of the conventional normalizing approach, i.e. a sum of forest compartments where the silvicultural intervention in each compartment is defined by the general model, in a top-down approach; the aim is to ‘homogenise’ composition and structure in each compartment, thereby conceptualising stands of trees as uniform management units and stands are actually ‘the building blocks of sustainable, regulated forests’.
On the contrary, with the systemic approach each stand is treated according to its own specific characters. In this sense, we agree that there is an ecological basis for stands in that disturbances, and we add, natural self-organising processes in a forest, may form discrete stand structures or uniform groups of contiguous trees.
With the systemic approach, silviculture does not aim to homogenise these differences within a compartment, but tends to follow and adapt interventions to the response of each stand.
In other words, there is no set rule or recipe which can be applied uniformly to the different parts of a forest, only a detailed analysis of the characters of each stand can show how to manage the whole forest. In this approach, management must proceed cautiously so that the reactions of the system can be analysed and management adapted so as to take these reactions into account. The landscape level becomes a useful scale from which to analyse the reactions of the forest as a whole.
This relation between the different scales of forest management (from the stand to the landscape) is also determining in defining the adaptive approach of systemic silviculture.
Adaptive management of natural resources is not a novel concept. The adaptive approach applied to forest management has been discussed by several authors, although with different meaning.
Finally, systemic silviculture considers the forest an entity with intrinsic value, shifting from the anthropocentric-ecocentric dichotomy to an integrated and respectful perception of the forest-humankind relationship.
This means taking into an account the ethical dimension, which has recently been pointed out as being critically underdeveloped in the discussion of new approaches in ‘ecological forestry’.