Browsing by Subject "päätöksenteko"

Reliable forest inventory data creates the foundation for quality forest planning. The quality of forest inventory data is emphasized when the planning is tried to make as optimal as possible compared to the aims. It is important for the choice of measures and timing that the forest variables used in the decision making are as accurate as possible. Unreliable information at the starting point and the wrong conclusions as a result may lead to inoptimality losses. Additional forest surveys also bring unnecessary expenses. Forest inventory is a financial investment for forest owners. One should not be content with information that is too incorrect and inexpensive, for the inoptimality losses can rise up to be higher than the investment expenses. The meaning of quality in forest inventory information was studied in this thesis. The aim was to find out how the various precision levels of markings in forest inventory information have an effect on both the choice of cuttings and timing in forests with various structures. This thesis aimed to find limits of quality requirements for forest inventory information which enable forest planning that is compatible with the aims. The quality aims of the planning were set according to the decision maker and the employer of the research, UPMKymmene plc. The research material consisted of a set of 337 stands provided by UPM Forest. The development of the stands was simulated by SIMO software. The simulation was made by assuming that the forest inventory data was flawless and by adding error into it. The forest variables that had error added into them were the basal area, average diameter, average height and site index. The simulations were made with a single error and combination error. In the case of single error one variable was added error systematically per cent by per cent up to -30–30 %. In the case of combination the error in one variable was added systematically and in the others randomly. The assumption was that the errors in different variables will not correlate. Measures and measure timings planned for each stand with reference data were compared to plans received by inaccurate starting point information. The planning period was ten years and the planned actions were thinning, clearcutting and no action. The error in the initial data clearly lessened the quality of planning. Over or underestimations of over 10% simulated to the basal area or average diameter alone led to the average accuracy of cutting measures going below the target level of 90 %. While the error grew the result of the planning weakened even more. The average height’s relevance of error in the quality of planning was minor but, instead, the relevance of error of the site index was significant. The site index error was clearly more damaging in pine forests than in spruce forests. The reason for this is likely to be that in pine forests there are three various thinning models and renewal limitations (ct, vt, mt), while in spruce forests there are only two (mt, omt). On grounds of the results of the research relatively small errors in basal area, average diameter and site index can cause a several years’ deviation in cutting planning. The relevance of error varied a great deal in forests of different structure. In the case of measure planning the forest inventory data gathered from well managed young forests does not need to be especially accurate, for the next measure is usually further in the future than the next investment. The accuracy demand of information is also not great in overly thick forests or in forests which have clearly surpassed the renewal limit. In these cases there is no obscurity about the next measure. In the accuracy of inventory one should pay the most attention on stands with young and grown forest cover which have had their last forestry measure done at least ten years ago. The research problem was approached from the point of view of measure accuracy. However, in the future it would be useful to research the effect of error in forest inventory information from the point of view of gain. As a result, one could have more factors in the research, such as timber logistic and planning of forests to be cut, in which the more accurate forest inventory data would be useful.

There is need for information about stands and their future development in forest planning decision making. This information is collected by inventories. In general inventory is repeated with some before-hand set intervals, irrespective of the method. Between inventories information is updated with growth models. Both inventory and using of growth models causes errors in forest planning results, for example in management options. Erroneous predictions can lead to wrong conclusions and inoptimal decisions. If the optimal result is known, economical losses caused by wrong conclusions can be described with so called inoptimality losses. The aim of this study was to answer the question how long forest inventory information, updated with growth models, can be used in forest planning purposes. Study approach was economical, so evaluation of information`s usefulness was based on inoptimality losses which arise when development of the stand is predicted incorrectly with growth models. The study material included 99 stands. Their development was simulated with the SIMO software for 60 years from present. In the 60 years period influencies of growth prediction errors were studied with inventory periods which lengths were 5, 10, 15, 20, 30 and 60 years. It was assumed that new error-free forest inventory information was received in the beginning of each of the inventory periods. In order to study effects of different inventory periods, it was assumed that the growth models were able to predict the true development of stands. Erroneous developments were yielded with error model which was developed for this study and added to the growth models. Inoptimality losses were calculated with the information derived from the optimization of stands` true and erroneous developments. Inoptimality losses increased when the inventory period became longer. Absolute inoptimality loss was approximately 230 eur/ha when the inventory period was 5 years and approximately 860 eur/ha when the inventory period was 60 years. Relative inoptimality loss was 3,3 % when the inventory period was 5 years and 11,6 % when the inventory period was 60 years. The average inoptimality losses were different between different development classes, site classes and main tree species. Study results show that the length of the updating period has an effect on the developing economical losses. It seems also that the inventory period should be different for example in different development classes. However, it is difficult to specify the optimal updating period because total losses are a sum of losses of inventory errors, losses of growth prediction errors and losses caused by other uncertainty sources. The effects of both inventory errors and growth prediction errors are different in different kinds of stands. So estimation of total losses and estimation of inoptimality losses caused by different error sources requires more research.