Timo Pukkala

A heuristic optimization method for forest planning and decision making

The presented planning method searches for an optimal combination of treatment schedules for forest compartments by maximizing, in an iterative manner, an additive utility function. The variables o...

Multi-objective forest planning

Introduction to multi-objective forest planning T. Pukkala. Measurement of preferences in multiple criteria evaluation J.M. Altho, P. Korhonen, P. Leskinen. Multiple criteria decision support methods in forest management. An overview and comparative analyses J. Kangas, A Kangas. Multidimensional measurements and approaches to forest sustainability assessments G.A. Mendoza, R. Prabhu. Using GIS to support multi-objective decision making in forest management. An experience from Ghana, West Africa P.A. Kwaku Kyem. Heuristics in multi-objective forest management J.G. Borges, H.M. Hoganson, A.O. Falcao. Spatial optimisation in forest planning. A review of recent Swedish research K. Ohman. Measuring non-wood forest outputs in numerical forest planning. A review of Finnish research T. Pukkala.

Eucalyptus globulus in Ethiopian forestry

Abstract Eucalyptus globulus Labill. has been planted as the main tree species in the central highlands of Ethiopia since 1895. The success of the 100 000 ha planted so far is based on the adaptability of E. globulus to the highland climate and soil conditions, its vigorous coppicing and non-palatability for livestock, and its suitability for fuel and small construction poles. The mean annual growth of E. globulus in fuelwood plantations varies between 10 and 30 m3 ha−1 year−1. Other equally productive fuelwood species have not been identified for the Ethiopian highlands. Despite alleged ecological demerits, like high consumption of water, additional planting of E. globulus is justified until the current fuelwood shortage is eliminated. In the long run, parts of Ethiopian Eucalyptus plantations may naturally alter in composition into stands of indigenous highland species, e.g. Juniperus procera, Podocarpus gracilior and Olea africana.

Prediction models of landscape preferences at the forest stand level

Abstract This study estimated quantitative models for Finns’ forest stand level landscape preferences, using a new modelling technique. The highly automated forest management planning systems used in commercial forestry could benefit from the quantification of scenic beauty. A total of 137 judges with different backgrounds made paired comparisons of photographs of 100 stands with known growing stock characteristics. The dependence of a stand’s priority with respect to scenic beauty on the growing stock characteristics was modelled using a regression formulation of the analytic hierarchy process. The results indicated that the priority of a stand increases with mean tree height, skewness of the height distribution, and volume of large pines and birches. The priority decreases with an increasing number of trees per hectare. Men and women, as well as forest owners and non-owners, have slightly different preferences. Therefore, a landscape preference model was estimated that accounts for the effect of both the stand characteristics and judge’s background on the priority of stand with respect to scenic beauty. The model explained 24% of the variation in the priority ratio of two stands as evaluated by an individual judge. The same stand characteristics that were used in the model were able to explain 83% of the variation in the mean priority of a stand, among the population of judges.

A method for stochastic multiobjective optimization of stand management

Abstract The paper describes a way of simultaneously optimizing several stand management objectives differing in measurement units and character. The simulation-optimization system accommodates, besides multiple objectives, the time and risk preferences of the decision maker. Multiple objectives are integrated in optimization through an additive utility function. Time preferences are dealt with by discounting, and the discounting rate is specified separately for each objective. Objectives related to products are described by the sum of the discounted values while objectives related to the status of the forest are described by the integral of the discounted value of a state variable. Different units are made commensurable via subutility functions that scale all objective variables between zero and one. Risks associated with the future growth level and timber prices, and with the preferences of the decision maker, are taken into account by stochastic optimization based on the scenario technique. With this technique, each management alternative produces a distribution of utility indices. This distribution is used to compute a utility index corrected by risk preferences; different parts of the distribution are weighted in a manner that reflects the decision makers attitude toward risk. A case simulation-optimization system is described, based on the above techniques, to deal with multiple objectives, time preferences, risks and risk preferences. Calculations using this system indicated that security and recreational amenities lengthen the optimal rotation of mixtures of Scots pine and Norway spruce, compared to a situation in which the maximal soil expectation value is the only objective. In a multiobjective case, stochastic growth, timber price and preferences increased the rotation length by 15 years (17%) from the deterministic optimum. Attitude toward risk also affected the optimal stand management.

Effect of crown shape and tree distribution on the spatial distribution of shade

Abstract A method for calculating the light extinction probability caused by a forest canopy is presented. With the calculation procedure, it is possible to examine the effect of crown shape, stand density and spatial distribution of trees on the spatial distribution of light extinction probability or on the total shaded area caused by the canopy. At low sun elevations, the momentary projection area of a single crown is greater the more vertically extended the crown is, if the crown volume is held constant. When a longer time period is concerned, the area where the average extinction probability exceeds some arbitrary value is greatest for umbrella-like, horizontally extended crowns. The same is true for a single tree and for a forest stand. When the stand density is low or the tree crown is narrow, the spatial distribution of trees has only a small effect on the amount of shaping; the total amount of shading in random distribution is almost the same as in systematic distribution. In a stand where the total horizontal projection area of crown cones is high, the spatial distribution of trees should be systematic for effective light interception.

Prediction of the amenity of a tree stand

Slides and computer drawings of 100 managed stands were evaluated for scenic beauty and recreational value by 121 people. The size and age of trees increased the beauty and recreation scores as well as a big share of pines and birches. A great number of stems per hectare indicated low amenity. Characteristics describing the tree population explained 70 % of the variation in slide ratings. Ratings of computer illustrations of stands accounted for about the same proportion of the variation in slide ratings indicating that they can be used as successfully as the regression models for estimating the amenity of a tree stand. The ratings of forestry students did not differ from those of biology students nor the ratings of foresters from those of urban citizens.

Relationship between radiation interception and photosynthesis in forest canopies: effect of stand structure and latitude

Abstract Interception of radiation and the consequent potential photosynthesis was studied, by sing a simulation model, in structurally different forest stands at latitu udes 40° and 60°. The studied stands were of two different types with respect to the leaf-area distribution: horizontally homogenous canopies and canopies with an aggregation of leaves into individual crowns. The effect of canopy structure on interception of radiation and photosynthesis was studied by varying leaf area index, stand density, and crown size and shape. In none of the studied cases was the relationship between accumulated radiation interception and photosynthesis strictly linear, but on a longer time-scale (one growing season) this non-linearity was not very pronounced. Neither canopy structure nor latitude substantially affected the slope of the relationship. In conclusion, while properties of canopy structure and incoming radiation determine the actual amount of radiation intercepted by the canopy, the conversion efficiency between intercepted radiation and photosynthesis appeared to be rather insensitive to differences in canopy structure and in properties of incoming radiation.

Integrating scenic and recreational amenities into numerical forest planning.

This study proposes an operational way of computing the amenity value of a forest area that consists of several stands or compartments. Such a method produces an amenity index suitable for numerical optimization. With this method of computation, amenity is divided into scenic and recreation values. Scenic value of the forest depends on the visibility, area, and scenic beauty score of different compartments. Recreation value of the forest is a weighted mean of variety and the area-weighted mean recreation value of compartments. Variety is described by the total length of boundaries between distinctly different compartments. Recreation value of a compartment is the product of accessibility and recreation score of the compartment. Recreation and scenic beauty scores are computed from stand characteristics using empirical models. The study describes a complete planning system that uses the developed amenity index as a decision criterion, and relates a case study to illustrate the effect of amenity objective on choice of treatments for compartments.

Recommendations for Standardized Documentation and Further Development of Forest Growth Simulators

Summary:The present paper documents the practical applicability of growth simulators, gives an overview of current model approaches, defines standards for the description and evaluation of growth models and growth simulators and indicates research needs. The recommendations aim to give users of growth simulators confidence in the transition to modern prediction systems and to increase the level of acceptance of new information technologies. The recommendations provide developers with guidelines for model description, model evaluation and software development and thus contribute towards efficient co-operation.Zusammenfassung:Der Beitrag dokumentiert die praktische Relevanz von Wachstumssimulatoren, gibt einen Überblick über aktuelle Modellansätze, definiert Standards zur Beschreibung und Evaluierung von Wachstumsmodellen und -simulatoren und weist auf Forschungsbedarf in diesem Bereich hin. Die Empfehlungen sollen Anwendern von Wachstumsmodellen Vertrauen beim Übergang zu zeitgemäßen Prognosesystemen vermitteln und die Akzeptanz neuer Informationstechnologien erhöhen. Indem dieses Papier Modellentwicklern Leitlinien für die Modellbeschreibung, Modellevaluierung und Softwareentwicklung gibt, kann es zu einer effizienten Kooperation aller Beteiligten beitragen.