Publikasjoner

NIBIOs ansatte publiserer flere hundre vitenskapelige artikler og forskningsrapporter hvert år. Her finner du referanser og lenker til publikasjoner og andre forsknings- og formidlingsaktiviteter. Samlingen oppdateres løpende med både nytt og historisk materiale. For mer informasjon om NIBIOs publikasjoner, besøk NIBIOs bibliotek.

1994

Sammendrag

The aim of this work is to compare three models with respect to estimation of mean diameter (Dg), number of trees per ha (N), tariff number (Hkl), gross value (Pb) and logging costs (K) in forest stands. For Model I, both the number of trees per ha and tariff number are assumed to be recorded in the field. For Model II only the number of trees per ha are assumed to be recorded, and for Model III only the tariff number are assumed to be recorded. Basal area and basal area weighted mean height (HL) are assumed to be recorded in the field for all three models.Details with respect to functions and assumptions for estimation of the different stand variables are given in chapter 2 Figs. 1, 2 and 3. The models are tested on data from sample plot inventories in 157 stands. The material is described in Table 1. Both systematic errors, i.e. mean differences between stand variables from different models, and random errors, i.e. the standard deviations for the differences, are compared. The differences are settled as recorded value minus estimated value (see Formulae 1 and 2 in chapter 3.2). Table 2 shows that Model III underestimates the mean diameter (Dg) (significant, 3.7%) in Norway spruce dominated stands and overestimates the mean diameter (significant, 7.0%) in Scots pine dominated stands. The standard deviations for the differences are about 7% in spruce dominated stands and about 9% in pine dominated stands. Table 3 shows that Model III underestimates the number of trees per ha (significant, 10.4%) in pine-dominated stands. For spruce dominated stands there are no significant differences. The standard deviations for the differences are 14.5% in spruce dominated stands and about 16% in pine dominated stands. Table 4 shows that Model II overestimates the tariff number (Hkl) both in spruce dominated (significant, 7.1%) and in pine dominated (significant, 7.2%) stands. The standard deviations for the differences are 5.7% in spruce dominated stands and 4.1 % in pine dominated stands (see also Fig. 4). Gross values (Pb) in all models are compared to recorded gross values, i.e. gross values calculated according to the observed diameter distribution. Table 5 shows that there are no significant differences between the recorded value and the model values when the means for all spruce dominated stands are compared. In pine dominated stands there is a significant underestimation (1.6%) of gross value for Model II. For the other models no significant differences appear. Table 5 also shows that the standard deviations for the differences are about 4% for all models in spruce stands and about 5.5% for all models in pine stands. The costs according to Model I are regarded as recorded when the logging costs (K) are compared. In this model both the number of trees per ha and the tariff number are assumed to be recorded (see Fig. 1). Table 6 shows that Model II underestimates the logging costs (0.9%), while Model III overestimates them (2.4%) in spruce dominated stands. In pine dominated stands the logging costs are underestimated both by Model II (1.0%) and Model III (3.1%). The standard deviations for the differences are about 1-1.5% for Model II and about 5% for Model III. Comparisons for development class III basically give the same results as in development class IV-V. However, both the differences between the models and standard deviations for the differences, are larger (see Table 7). A general discussion, and a comparison of the models with respect to all stand variables, are carried out in chapter 5. Table 8 shows all systematic tendencies derived from the comparisons. Model I provide for the best results. Because of high inventory costs, however, this model will probably be out of question in most practical inventories. Model II and III, however, also seem to provide for satisfactory results with respect to gross values and logging costs. If the mean diameter and the number of trees per ha are going to be used as input in long term yield forecasts, Model II provides for better results than Model III. With this background it is therefore recommended to apply Model II rather than Model III. This conclusions holds true if all field sampled data are correct, and if the costs for collecting the data are the same for the two models. Such issues are discussed by Eid (1994).