Original scientific paper
Efficiency of Inventory in Uneven-Aged Forests on Sample Plots with Different Radii
Anamarija Jazbec
orcid.org/0000-0003-2888-8518
; Zavod za izmjeru i uređivanje šuma, Šumarski fakultet Sveučilišta u Zagrebu, Svetošimunska 25, 10 000 Zagreb, Hrvatska
Mislav Vedriš
; Zavod za izmjeru i uređivanje šuma, Šumarski fakultet Sveučilišta u Zagrebu, Svetošimunska 25, 10 000 Zagreb, Hrvatska
Mario Božić
; Zavod za izmjeru i uređivanje šuma, Šumarski fakultet Sveučilišta u Zagrebu, Svetošimunska 25, 10 000 Zagreb, Hrvatska
Ernest Goršić
; Zavod za izmjeru i uređivanje šuma, Šumarski fakultet Sveučilišta u Zagrebu, Svetošimunska 25, 10 000 Zagreb, Hrvatska
Abstract
For the purpose of forest management planning in Croatia, forest inventory is performed on sample plots of permanent sizes, generally from 500 to 1,000 m². Such plots have been accepted in practice on the basis of experience. Thus, the plot size has only partially been adjusted to different forest types. Although methods of sample selection are regulated by Forest Management Regulations and are limited to plots of permanent sizes, the Regulations still allow plot sizes to be adjusted to specific stand conditions. This article compares the results of forest measurement on differently sized plots in order to estimate their efficiency in selection fir and beech stands. Research was carried out in Gorski Kotar, where the stand structure was recorded using a sample of experimental plots. Systematic sampling with a random start was used to establish 103 sample plots in the management units of »Delnice« and »RavnaGora«. Breast height diameters over 10 cm and tree heightsweremeasured on circular plots of 20mradius, set up across a 100 x 100 m square grid, and their positions in the space (azimuth and distance from the centre) were recorded. The values of structural elements (number of trees, basal area and volume) were calculated from the obtained data bymeans of a specially constructed computer programme CirCon for the plots and stands. The precision of estimates was expressed as a relative sample error with 95% confidence. Thanks to spatial data for all the trees, a calculation was made of the structure for sample plots in the radius range from 4 to 20 m. The precision of structural elements estimate (sample error) was calculated for each plot size. The average values in differently sized plots were compared with repeated measurement analysis of variance (RM ANOVA) with significance level of 0.05. The results of different plot sizes were subsequently tested by means of the Fisher’s LSD 'post hoc' test. On a sub-sample of 24 standpoints, the time needed for field measurements on differently sized plots was later measured in order to estimate the efficiency of differently sized plots. From these data measurement time (t) for all plot sizes was assessed in dependence on plot area (P) using the equation t=a*PČb. Walking time and distance between the plots were measured, from which the travel speed in the field was calculated. The efficiency of plots of a particular size was evaluated by the precision of estimate and the time needed for measurements. The achieved efficiency index was evaluated as the product of squared sample error and total time expressed in relation to the efficiency of 20mradius plots. The efficiency for differently sized plotswas also assessed by means of the expected accuracy at given sampling intensity of 5% area (5% cruise), as set down in the Forest Management Regulations. The results show the differences in the estimate of structural elements in dependence on plot sizes, especially for tree number. The differences were confirmed as statistically significant (RMANOVA: F = 2.31, df = 102, p < 0.0001). Subsequent testing revealed that on the plots with radii less than 7 m the estimate of tree number was statistically significantly lower than in the larger plots. Asimilar trend did not prove to be statistically significant either for the basal area (ANOVA PM: F = 0.72, df = 102, p =0.8772) or for the volume (ANOVA PM: F = 0.78 df = 102, p = 0.8124). Expectedly a more precise estimate was obtained with an increase in plot size. Considerable improvement in the precision was observed up to plot radii of about 13 m. The trend was slowed down with further increase: consequently, in terms of precision, plots with at least the above radii are recommended. The assessment accuracy of tree numbers in the smallest plots was better by 10% than that of the basal area; however, this difference decreases as the plot size increases, to become equal at plots with radii of about 12 m. The difference in the assessment accuracy of the basal area and volume ranges from about 2%in the smallest plots to less than 1%in the largest plots with 20m radius. In the case of smaller plots, apart from poorer precision of estimates, there were also bigger deviations from the average obtained on larger plots. With regard to the average assessment value of tree number, the basal area and volume, plots with radii greater than 10 m would be acceptable in the study area. The average measurement time increases almost linearly with an increase in plot size. In terms of the achieved efficiency index at the same number of plots, the smallest plots proved to be the least efficient, while the efficiency increases considerably up to 9 m radius, and levels from radii larger than 13 m. Along with the number of plots adjusted to the given sample size intensity of 5%, plots with radii between 8 and 15 m proved to be more precise than smaller and bigger plots. Time expenditure uniformly decreases with the application of a smaller number of bigger plots in relation to a larger number of smaller plots with an equal total area. Hence, the use of bigger plots is more practical since differences in precision are negligible. In terms of the number of measured trees per plot as a practical criterion that affects the quality of the results and of field work, plots with radii between 12mand 15mare recommended, which contain between 20 and 30 trees on average. This research confirmed the assumption that the reliability of estimates depends on plot size. Specific numerical relationships were obtained for the investigated selection stands. Plots of 500 m2, common in forest management practice, proved to be adequate for the investigated stands. Nevertheless, larger plots could significantly contribute to better precision of estimates at identical sample size. In terms of efficiency, these plots are definitely acceptable, taking into account the fact that even larger plots could also be efficient at identical sample size. By adjusting the size and number of plots to the prescribed intensity, standard plot sizes proved to be adequate. Still, plots of larger radii would possibly save the measurement time. Additional research involving a larger sample and broader range of stand conditions would give a clearer insight into the effect of plot size on precision of estimates. It would also enable better adjustment of plot sizes to individual stands for the purpose of forest inventory rationalization. For similar beech-fir selection stands, a comparison of nested plots (concentric circles) that are adjusted to the diameter structure should definitely be made. By doing so, the efficiency of assessment reliability and measurement efficiency could be improved, which is neglected by the present regulations.
Keywords
Forest inventory; sample plots; stand structure; precision; measurement time; efficiency; fir-beech stands
Hrčak ID:
68219
URI
Publication date:
8.4.2011.
Visits: 1.654 *