CLIMAT AND OROGRAPHIC IMPACT ON VOLUME INCREMENT OF THE FORESTS OF BEECH, FIR AND SPRUCE ON LIMESTONE AND DOLOMITE IN FEDERATION OF BOSNA AND HERZEGOVINA

Research results of the stand volume increment changes in mixed forests of beech, fir and spruce on limestone and dolomite in Federation of Bosnia and Herzegovina (FB&H), is presented in this work. Impact of different factors to volume increment value was investigated: habitat quality, values indicators of structural characteristics particular forest stands, and orographic characteristics if the terrain. Impact of climate differences was investigated trough position of the forests stands in particular ecological-vegetation regions in B&H. Research was done on example of 799 forest stands, trough complex methods of multiple regression analyses, and variance analyses (ANOVA), The results shows that volume increment of the forest stands, besides taxation values: basal area, relative share species in volume, average diameter of the trees, medium height of the trees 50 cm in diameter, significantly depend of the altitude and belonging to particular ecological-vegetation regions in B&H. The result of the analyses is complex model for forest stand volume increment assessment. It was concluded that volume increment of these forest is higher about 1 m/ha, now than 50 years ago, when very similar investigation was done in the same forests.


INTRODUCTION -Uvod
Mixed forest of beech, fir and spruce, are the most important part of forest resources in Bosnia and Herzegovina. According results of the second state forest inventory in B&H, area of these forests covering about 665 300 ha 2 . About 85 000 ha limestone and dolomite in Federation of Bosna and Herzegovina is temporary inaccessible because of mines danger. These forests covering about 371 000 ha land on limestone and dolomite. The most of it (92%) is in state ownership. It is about 22 % of the total high forest area in the state ownership (1652 400 ha) in B&H.
Typical uneven aged structure of the tree crowns has 70 % forest stand; group -uneven aged structure has 22 %, and two-storey structure about 8 %. Regeneration of these stands is almost exclusively based on natural regeneration, while the artificial introduction of seedlings practiced only sporadically, as a complement to the natural regeneration of stands. "Tree selection system of management was used in these forests in the past, in which the dominant regeneration of stands natural (MATIĆ 1977). PINTARIĆ (1991), defines the tree selection forest as follows: "By this we mean such forest in which in a relatively small area existing trees of all heights and thickness. Trees of different sizes can be mixed individually or in small groups, and can be located next to each other or one below the other. This unevenness is a constant in the selection forest". SCHUTZ (1975SCHUTZ ( , 2001 points out that the individualization of production and process of self-care, is the main feature of the tree selection forests stand 3 .
During the 50s, 60s and 70s of the last century, were carried out extensive research sizes taxation elements of forest stands, as well as their interdependence, of the the economically most important tree species in Bosnia. These studies were first started in the managed forests of beech and fir with spruce (MATIĆ 1959). Results of these studies are presented in tables (MATIĆ ET ALL. 1963, DRINIĆ ET ALL 1980, 1990 and used for the design of the normal composition of the selection stands (MATIĆ 1963). MATIĆ (1959MATIĆ ( , 1971, conducted research sizes taxation elements in managed forests (mixed and pure) of fir, spruce and beech, in Bosnia. On the basis of data collected from 383 temporary sample plots, laid out in these forests, from 1952 to 1958, the made a comprehensive analysis of the natural process of changing the size of the main taxation elements (measurement characteristics) in the selection forest stands, using the method of multiple regression and correlation analysis. Pure beech and mixed forests of beech and fir with spruce on different soil types and rocks were taken as a unique whole (one population). Regression equations for estimation of the volume increment, diameter increment and volume of fir, spruce and beech in mixed uneven aged forest stands, was the results of analyses.Then follow the research of the dynamics of growth and development of beech and fir forest with spruce and fir and spruce, in the established permanent sample plots (PAVLIČ 1966;DRINIĆ 1972DRINIĆ , 1976BOZALO 1980BOZALO , 1985.

Ahmet Lojo
In this regard, the first significant researches, to study the growth and development of selection stands, were investigation by BURGER (1950BURGER ( , 1951BURGER ( , and 1952. He presented the results of the volume increment size dependence of other taxation elements sizes in the selection forest stands of beech, fir and spruce. On the study and the separation of basic and productive forest types in the mixed forests of beech, fir and spruce in Bosnia and Herzegovina, worked ĆIRIĆ, ET ALL (1971). They are, inside these forests in B&H, isolate and describe 37 basic types of forests within the framework of mixed forests of beech and fir with spruce. Of these, 16 basic types of forest is isolated and described on the limestone and dolomites.
STEFANOVIĆ AT ALL (1983a), were singled out and described several basic types of forests in the forest of fir and spruce, spruce and white pine and spruce forests in the climatogenous belt of the beech and fir forest (with spruce). STEFANOVIĆ ET ALL (1983b) conducted mapping and allocation of ecological-vegetation's rayon in B&H. B&H territory is divided into provinces, districts and rayons, based on phytogeographic, climate and vegetation components analysis. The province is represented by the broadest category and rayon closest category in this division. Within this zoning, through the descriptions of the individual regions, the ranges of most regional community forests in B&H are described in details.
HASENAUER (2006), points out that in the last few decades there was a significant change in the conditions of growth of forest tree species in Europe. As a contribution to it, cites researches by other authors (KENKA ET ALL 1991, 1993, SPICKER 1996, KARJALAINEN ET ALL 1999. Notes a positive trend of the increasing the productivity of habitats in the southern parts of Northern Europe, Central Europe and some parts of southern Europe. LOJO (2011), conducted research quality of forests classification to the basic forest management planning units -management classes, exploring, among other things, the process of volume increment changing of mixed stands of beech and fir with spruce on the limestone and dolomites. The results of research, which refers to the size dependence of volume increment of a number of other factors as independent variables, is shown in this paper.
IBRAHIMSPAHIĆ (2013), investigate changes in position of the stand height curves, on the permanent sample plots (plot area 1-2 ha), in the mixed uneven aged forest of fir, beech and spruce on Igman Mountain, based on periodical tree height measurement. She concludes: average tree height, with 50 cm in breast diameter, could give reliable assessment of the habitat quality for fir, beech or spruce. Height curve has the smallest movement (dislocation up or down) in time for tree diameters around 50 cm.
Knowing the laws change increment, among other things, is a necessary condition for a realistic assessment of normal size and structure of the growing stock in a selection stand, or to create realistic cutting plans. Within this, important question is whether the geographic location within the FB&H, has a significant impact to size of the increment of the forest.
Besides the factors, that are measuring characteristics of individual stands, basal area, or the average tree diameter of the stand, the average tree height in the stand as an indicator of habitat quality, the proportion of individual species, the influence of geographical factors: slope, exposure and elevation to the size of volume increment is investigated.
Therefore, the goal of the work is set as follows: Examine sizes and patterns of the volume increment changes in the forests of beech and fir with spruce on limestone and dolomites in the FB&H, according to the structural characteristics of growing stocks, orographic characteristics of terrain and affiliation to different ecological-vegetation's rayon in the FB&H.

THE STUDY SAMPLE -Uzorak
To solve the goal of this study, we used data collected in uneven aged forest stands, within mixed forests of beech and fir and beech and fir with spruce and fir and spruce forests within. Data was obtained from forest inventories in 11 forest management regions (FMR) in the F B&H, which were then selected, analysed and processed at the level of individual stands. Measurements in these areas were carried out during the period 2000 to the 2005 year. From inventory data, the data relating to the above mentioned forests stands, 799 of them, were selected, and compiled into a single database. All selected forest stands are larger than 40 ha, uneven aged structure, and each is uniform per orographic characteristics. That was the criteria for selections The lowest stand is situated at an altitude of 570 m, while the tallest at a height of 1585 m above the sea level. The average altitude of all the stands in the sample is 1094 m. Altitude of the individual stands is determined by "center of gravity" altitude of polygons representing a stand on the forest vector map.
Considering the method of choice stands in the sample, which was independent of the exposure, slope or elevation, we can say that the pattern of these elements represents well whole collective of the mentioned forest 4 . The distribution of stand by terrain slope, expressed in degrees (°) is shown on Chart 3.

Sample structure by rayons in FB&H -Struktura uzorka po rejonima u FBiH
Given that one of the aims of this work is to investigate the possible influence of climatic factors on the size of the volume increment in considered mixed forest stands, it would be defined a regions, more or less uniform in that.  ET ALL 1983). According to the zoning, the largest part of the stands sample, in this work, belongs to the area of internal Dinarides, which stretches from the northeast, from the Plješivica mountain to the Volujak and Maglić mountain, in the southeast part of the B&H. This area is divided into 7 provinces, of which 4 are represented in the sample. Within them, the sample is represented by 5 rayons. One part of the stands sample belongs to Sub-Mediterranean province, within sub-Mediterranean mountainous district (Table 1).
From the description of the climate rayons, we bring the following: Ključ-Petrovac district occupies the northwest part of the internal Dinarides area. Within that area the extreme southwest part (rayon A) is under the influence of the Pannonian climate, warmer during summer, where spruce is absent. Spruce participates in mix with fir and beech in the central and south-eastern part (rayon B) of this district, due to more pronounced continental climate. Influence of the continental climate prevailing in the central part, due to the higher altitude during the winter. Stronger influence of Mediterranean climate is during the summer. The ratio of evapotranspiration and precipitation are favourable.
Glamoč -Kupres rayon (C) is extremely mountainous with altitude above 1000 m, in the southern part of the zone, due to the penetration of the Mediterranean climate, the spruce is also absent from forests of beech and fir. Relationship evapotranspiration and precipitation is favourable.
Sarajevo -Zenica rayon (D) occupies the central part of Bosnia and Herzegovina, which also feel strong incursions of the Mediterranean and Pannonian climate, which is why spruce rarely participates in the forests of fir and beech, except at high altitudes, with a smaller proportion of the stand volume.
Romanija rayon (E) occupies the mountainous regions of eastern Bosnia, and stands from the sample are located in the western part. Character of the climate is the mountainous with the influence of the continental climate in the period November -May. In the forests of beech and fir, spruce is often significantly presented. The ratio of evapotranspiration and precipitation is favourable.
Igman -Zelengora rayon (F), with the average and distribution of rainfall (47% of rainfall in the growing season) also has a pronounced influence of the Mediterranean climate, but during a winter has a strong influence of the continental climate. Relationship of rainfall and evapotranspiration is favourable.
Sub-Mediterranean district (G) is relatively small and is not divided on the individual rayons. It occupies the south-western part of the Dinarides in B&H under the strong influence of the Mediterranean climate. Generally above 800 meters above sea level and terrain is very mountainous. Northern border has mainly coincides with watershed between the Adriatic and Black sea basin. Maritime precipitation regime has expressed and the relationship between evapotranspiration and precipitation is favourable. In the higher mountain regions, where the forests of beech and fir presents, spruce is very rare and only occurs at higher elevations, on the border with of the internal Dinarides area and the smaller microclimatic cooler locations. Structure of the sample, according to habitat quality for certain species of trees (determined bay tree height, expressed by absolute measure -average height of trees 50 cm thick, and by relative quality classes), has presented in Table 2. The 527 stands which consist of all the three main species of trees together, the average quality habitat, expressed by relative classes and by the average height of trees with a breast diameter of 50 cm, are as follows in Table 3. Based on the average height of trees with a diameter of 50 cm, of the particular tree species, average height of trees with a diameter of 50 cm of the all species together (H50-pros), has calculated using the ratio of mixture in stand volume, per tree species as weights.
This measure is used in certain analyses of the structural-production characteristics and interdependencies of the taxation elements of the forest stands, as a very good indicator of habitat quality, since the quality measures values of certain tree species "follow" each other in the same forest stand.  Based on the average height of trees with a diameter of 50 cm, of the particular tree species, average height of trees with a diameter of 50 cm of the all species together (H50-pros), has calculated using the ratio of mixture in stand volume, per tree species as weights.
This measure is used in certain analyses of the structural-production characteristics and interdependencies of the taxation elements of the forest stands, as a very good indicator of habitat quality, since the quality measures values of certain tree species "follow" each other in the same forest stand. MATIĆ (1959) has classified the quality of habitat for the growth of fir, spruce and beech in five site quality classes (relative), based on the possible varying height of the trees in a stand, in relation to the brest diameter of the trees. In order to determine the habitat quality, for practical reasons, for these species of trees, based on the position of the tree height curve , he has formed a graphic scheme of the 5 heigt bands for each tree species.
Thees classes are a relative measures of habitat quality. Particular class is determined by continuous variables (height of trees) and can be set equality: class I = 1.0, II = 2.0, class III = 3.0, etc, but they are still the relative indicators of habitat quality. Variable "H50-pros" is the absolute measure (indicator) of the habitat quality, and is not bound only for B&H and as such suitable for the analysis of productivity of beech, fir and spruce throughout its area distribution range.
In Table 2, we can see that the stands of the average of the first (I) and fifth (V) quality class a bit and put them together in the sample is less than 1% . Even 63% of stands have moderate habitat conditions (class III) for the growth of these species. In general we can say that the variation of habitat quality for fir, spruce and beech on limestone and dolomite is in 3 relative site quality classess (II, III and IV).
The ratio of tree species in stand volumen by group forest association Omjer smjese vrsta drveća po grupama asocijacija šuma By the ratio of tree species in the volume of the stand, the structure of the sample is as follows (

Annual increment of the volume of tree species by types of forests
Godišnji zapreminski prirast vrsta drveća po vrstama šuma Growing stock of the tree species by types of the forests Zaliha (zapremina drveta) vrsta drveća po vrstama šuma  For an explanation of volume increment changes, it is necessary to find a regression model, hich will include all statistically significant factors (variables). In this paper, the goal is not only the best possible estimate of the volume increment in some particular cases, but exploring the impact of this size by certain important factors as independent variables, whether categorical or continuous.

Size of the annual volume increment (all species combined) in depend of the other taxation elements of the stand -Veličina tekućeg zapreminskog prirasta sastojine (svih vrsta drveća zajedno) u zavisnosti od drugih taksacionih elemenata sastojine
To determine the regression model, computer programs "Statistics" 7.0 and "Statgraphic" 5. 0 were used. For better assessment of changes in the size of the annual volume increment (IVHA), selected variables were chosen (Table 7).
Taking into account the criteria of proper procedure of the multiple regressions, some variables used as an independent in the regression equations need to be transformed, by appropriate formula, due to their size structure and type of deviations from normal size distribution. Standardization of all independent variables been imposed as mandatory transformation, in order to avoid problems of variables collinearity (HARTIGAN 1975, PECINA 2006. Variable altitude (VISINA), slope (Nagib), the average height of the tree with beast diameter 50 cm (H50-pros), average diameter of the conifer trees (Dg-CET), the share of fir in stock (OS-Jl), the share of broadleaves in stock (OS-Lis), are standardized, that is, there was a transformation of their original size by the following formula: By this transformation, all variables receive an average size 0 and standard deviation 1 (Table 8).
Soil coverage by tree crowns variable (STZ), which did not have even close to a normal distribution of their data, previously transformed by the formula:   Optimal set of influential variables and their interactions-products is obtained, using GRM (generalized regression models) module in "Statistic 7.0 "Forwardstepwise" selection procedure of the most influential variables, and inclusion of the transformed variables size of individual stands in the sample was used (Table 9, value "p" <0.05). Their influence on the dependent variable IVHA was statistically significant, and the selected variables and their interactions were not collinear with each other (Table 11, inflation factor VIF<5) (PECINA 2006). The significance of the influence of individual variables and their interactions is shown in Table 9 (F ratio of the variances).
The variable slope "Nagib-st" had no statistically significant effect on the size of the volume increment (IVHA), as a single variable, or, in interaction with other variables. In the final step of determining the regression equation, the independent variable -volume increment (IVHA) is also transformed in order to complete homogenization of variance and normalization of residuals.  Table 10 shows the variance explained part of the variation of the dependent variable (Model) and the unexplained part of the variation (Residual) and the ratio of these variances F.
Degree explanations expressed through determination coefficient R 2 = 0.787, R 2 (corrected) = 0.783, while the standard error of the estimate = 1.183 m 3 /ha. Table 11 shows the calculated values of parameters in equation with indicators of their importance. Since the calculated size of the inflation factor VIF <5, for each parameter, and , there is no problem of multi-collinearity between individual variables and their interactions.

Ahmet Lojo
Chart 4 shows the quality of the regression by schedule deviations (residuals) from the estimated actual size of the dependent variable volume increment (IVHA). From this description, and the next, in Chart 5, it can be seen that the schedule of the residual is random throughout the domain of the size IVHA, i.e., there is no systematic deviation of the residuals around estimated value, in any part of the domain changes variable IVHA, and that the variance of the residuals, according to the Chart, 5, is homogeneous as a result of optimization of the regression, using BoxCox transformation of the dependent variable. Charts 6 to 11, shows the individual effects of the independent variables in their transformed size, on the dependent variable -volume increment (IVHA), with average size and the influence of other independent variables in the equation model 1, and the average individual interaction effects of independent variables. From the distribution of residuals cannot be seen that there is a systematic deviation in any part of the domain of independent variables, which indicates the quality of the regression equation and the specified parameters. From charts can be seen the character of the influence of individual variables, i.e., does the size of the dependent variable IVHA increases or decreases with increasing the size of some of the independent variables. To display the real -absolute impact of an individual variables, or explanations of impact character on the dependent variable, it should construct a curve with the original sizes of the dependent and independent variables. It's done for impact of the stand basal area (GHA) variable and shown on Chart 12.

Size of the annual volume increment depending on the location in the FBiH and the exposure of the stand -Veličina godišnjeg zapreminskog prirasta sastojine u zavisnosti o ekološko-vegeticskom rejomu u FBiH i ekspozicije sastojine
When the most influential set of continuous independent variables and their interactions, the size of the volume increment (IVHA) was chosen, trough procedure Forward-Stepwise, categorical variables are then included as independent in the regression model (2) Table 12, 13, and 14. Increasing the number of independent variables (Model 2) slightly increases the degree of explanation IVHA, for about 1%. Degree explanations expressed through determination coefficient R 2 = 0.79, R 2 (corrected) = 0.79, while the standard error of the estimate = 1.16 m 3 /ha. Exposure showed no statistically significant effect on the size of the volume increment (Chart 14), while the influence of stands belonging to different eco-vegetation's rayon of in FB&H, significant (Table 13, the value "F" and "p" for variable EV_R and EKSP). Other independent variables and their interactions still have a statistically significant effect on the dependent variable, except for one interaction (Nd_Vis_st * H50_pr_st * GHA_TS), whose size p slightly exceeds the level of significance of 0.05. The differences in the sizes of the calculated size of volume increment by some of the regions are statistically significant, but such differences do not exist between all regions (Table 13). It is obvious from the Chart 13. In Figure 2, the same color rounded stands within individual regions, among which there is no difference in size of volume increment, (with the same sizes of the other independent variables in the equation model 2).
If we analyse the schedule mutually similar EV_R in Figure 2 (or table 14), it is easy to conclude that grouping those rayons, which are similar in character of the climate. Kljuc -Petrovac (B) and Romanija (E) with a typical continental mountain climate, in which the volume of increment is, in average, higher than the increment of other rayons. Glamoč -Kupres rayon is characterized by its extremely cold and long winters and incursions of the Mediterranean climate in summer. Spruce is on the edge of its own area. The conditions for the growth of stands, according to the results of this analysis, are slightly worse in this rayon comparing to rayons with a continental mountain climate. In other rayons, impact of continental mountain climate is mixing with occasional impacts of the Pannonian -summer drought, or Mediterranean climate. In terms of the conditions for the growth of stands make the transition between the previous two groups of the rayons.
B and E rayons, continental -mountain character of the climate This result, due to the geographical distribution mutually similar regions per volume of the annual increment (IVHA) is a logical. The real and detailed reasons for the observed differences in the size of the increment are not necessary to find, into this work. It is enough to say that there are significant differences in the volume of timber of the same forest stands, in certain rayons of the Federation of Bosnia and Herzegovina.

DISCUSSION AND CONCLUSION -Diskusija i zaključci
The goal of the analysis in this work is focused on exploring the impact of ecological vegetation rayonss (EV_R), exposure of the stand (EXP), and the altitude, on the size of volume increment (IVHA) within mixed stands of beech, fir and spruce. In order to investigate the influence of these factors, previously it was necessary to eliminate the effects of other factors -structural characteristics of the stand, which was done in the first step in forming a regression model 1, which includes all significant continuous measuring variables. In the second step, categorical variables are included (in model 2): exposure of the stands and ecological-vegetation rayons in Bosnia and Herzegovina.
It has been shown that exposure of the stand had no significant effect on the size of the volume increment, while ecological-vegetation rayons had important and statistically significant effect on the size of the volume increment. In Figure 2, homogenous groups of the rayons, (EV_R) which do not differ in the size of the volume increment, are the color-coded, It could be seen that there is a logical arrangement in the space, and the differences in the size of the increment, between individual rayons can be explained by the character of their climate.
Also, effect of altitude is significant. With increasing altitude, size of the annual increment reduces. Increment is reducing about 1 m 3 /ha in the range of height distribution, 600 to 1400 m above the sea level, with the same other sizes of factors affecting the change in increment.
How is the sample for analysis in this paper is representative, showing comparative data from mixed forests of beech, fir and spruce, of this sample and data gained from Second state forest inventory in B&H, conducted in the period 2006-2009  in table 15.  Table 15 shows the average data of the annual volume increment, from the sample of this study (sample), and the average size of the increment of selected plotsdata from Second state forest inventory in Bosnia and Herzegovina (SFI_B&H), for each type of forest. The differences in the estimated increment volume are small and do not exceed the statistical error of the increment volume estimates in SFI_BiH.
MATIĆ (1959) conducted a similar study in the framework of these forests. His sample had included even the stand on silicate soils-rocks in B&H. He also conducted similar researches about character of the volume increment changes of these forests. If the regression models of this study, and regression models obtained by MATIĆ 1959, include in, the average (the same) size of the independent variables (the same habitat quality and share the same species in stock, roughly the same level of land coverage, we get the result -a conclusion that the average increment in the forests of beech and fir with spruce in limestone and dolomite, increasing by almost 1 m 3 /ha per year today, but more than 50 years ago, when the data were collected for Matić's research. We do not know clear reasons for it. MATIĆ (1959) predicted an increase of the increment in the future, because at the time of his research, in the forests, was still significantly presented trees grown in virgin condition, with a small length and volume of the tree crowns.
It is possible, that there had been significant changes in the volumes of the tree crowns, due to intensive forest management after 50 years, but there have not been such investigations of tree crowns in recent years in B&H.