DIGITALNA ARHIVA ŠUMARSKOG LISTA
prilagođeno pretraživanje po punom tekstu




ŠUMARSKI LIST 11-12/2018 str. 45     <-- 45 -->        PDF

the large-scale weather systems, both of which are influenced by seasonal and inter-annual variations.
The total area of the study was 598.051,31 ha, while the total forest area was 255.748,82 ha. Turkish red pine forests constitute 192.501,79 ha of this area, which corresponds to 75.26% of the forested areas within the studied region. An evaluation of forest fires in Antalya between 1979 and 2012 revealed that although crown fires occurred less often, 89.07% of the burnt areas in Turkish red pine forests were the crown fires (Table 1).
Data Collection and Sampling – Prikupljanje i uzorkovanje podataka
In this study, 7 saplings and 77 trees were sampled destructively. Trees were randomly sampled only in fire prone areas of Antalya province. H, CW, CL, age, RCD and DBH were measured as independent variables. Vertex IV was used for the H and CL measurements. Following these measurements, the trees were cut in order to measure their crown fuel load. After all the trees were cut, the fuel biomass in their crowns were destructively sampled. Different sampling methods were used to sample the fuel load of the crowns of large and young trees. The first method was used for large trees with a DBH greater than 8 cm and a crown length of at least 3 m. In this method, the live crown (CL) of the cut trees were measured and divided equally into 21 sections. All living and dead fuel materials originating in sections 1, 6, 11, 16, and 21 were removed and separated by fuel size classes (foliages and fuel branches), and then weighed. Weight values obtained through this method were then multiplied with a factor of 4,2 to sample the crown fuel load (Küçük et al. 2008; Robichaud and Methven 1992). In the second method, all fuel crown biomass of young trees with a DBH less than 8 cm and a crown length below 3 m were removed, and then separated by fuel size classes. The removed and classified material was then weighed. Dead fuel branches were not included in the analyses.
Branch samples of crown fuel biomass were separated into standard fuel size classes (Scott and Reinhardt, 2002). Crown fuel biomass classes are foliage, <0,3 cm very fine branches, 0,3–0,6 cm fine branches, 0,6–1,0 cm medium branches and 1,0–2,5 cm thick branches. The main reason to use this classification that the finest fuels are the first to be consumed before the crown fuel biomass are entirely consumed in the flaming front of a crown fire. In fact, only the finest fuels burn in the short duration of a crown fire (Mitsopoulos and Dimitrakopoulos, 2007b; Stocks et al., 2004). In this way, a separate class designated as “active fuels” (Küçük et al., 2008) and foliage and branches thinner than 0,6 cm were added to the crown fuel biomass classification, to indicate materials that are more predisposed to be consumed. Total fuel biomass was determined by weighing the foliage and different fuel branches classes of each tree. Foliage and branches by diameter classes were taken as samples and weighed once again. In this study, we have used oven-drying of fuel samples to determine fuel moisture content (Matthews, 2010). These samples were dried in 103 ± 2°C ovens for 24 ±1 hours to obtain oven-dried samples. The oven-dried samples were then weighed. Based on the ratio between the live weight and oven-dried weight of the samples, the oven-dried weight of the foliage and fuel branches on every tree was determined. In the study, oven-dried weights were used as fuel load values in tables and figures.
Statistical Data Analysis – Statistička analiza podataka
Correlation and regression analyses were used to examine the relationships between tree properties and crown fuel biomass. Stepwise function and logarithmic linear regression models were used to analyze the relationships between fuel biomass and the measured tree properties. Logarithmic regression is commonly used to estimate the relationships between crown fuel load and the different properties of trees and crown (Mitsopoulos and Dimitrakopoulos, 2007a; Molina et al., 2014). The equation used was [ln(Y) = a + bln(x) + ε], where, Y is the dependent variable (needle, branch or total biomass), ln is the natural logarithm, x is the independent variables, a is the constant, b is the regression coefficients, and ε is the error term. The residual variance of dependent and independent variables is both used in the analysis of allometric relationships. For this reason, logarithmic transformation is necessary to remove residual heteroscedasticity (Socha and Wezyk, 2007).
All selected equations were significant at least at P=0,05 significance level. Regression and correlation analyses were performed using IBM SPSS 20.0 for Windows. The measured tree properties H, CL, CW, RCD and DBH were used as independent variables, while the foliage and branch biomasses were used as the dependent variables. Before the variables were analyzed, there were tested for normality.