The Dinaric Mountains versus the Western Carpathians: Is structural heterogeneity similar in close-to-primeval Abies–Picea–Fagus forests?

Abstract

In this study, we analysed patterns of spatial variation in the basal area of live and dead trees and structural complexity in close-to-primeval forests in the Dinaric Mts. The results were compared with an analogous study conducted in the Western Carpathians. The research was carried out in the Janj, Lom and Perucića forest reserves (Bosnia and Herzegovina) in mixed-species stands of silver fir Abies alba Mill., European beech Fagus sylvatica L. and Norway spruce Picea abies (L.) H. Karst. In the core zones of the reserves, concentric sample plots (154 and 708 m2) were set in a regular 20 × 20 m grid covering approximately 10 ha. The analyses revealed varying distribution patterns of live canopy trees, suggesting that these characteristics may fluctuate to some extent at the regional level. At the spatial scale of 708 m2, attractive associations between dead canopy trees were found, but this tendency disappeared with increasing area. Although stands in the Dinaric Mts. are characterized by an almost twofold greater biomass accumulation compared to those from the Western Carpathians, the study revealed analogous bell-shaped distributions of stand basal areas of live trees and a very similar trend of decreasing variation in stand basal area and structural heterogeneity with increasing spatial scale. Nonetheless, the higher growing stocks, lower ratios of dead to live tree basal area and lower proportion of homogeneous structure types found in the Dinaric Mts. may suggest a less severe disturbance history over recent decades in this region compared to the Western Carpathians.

On the Consequences of Using Moving Window Segmentation to Analyze the Structural Stand Heterogeneity and Debatable Patchiness of Old-Growth Temperate Forests

“…strongly depend on the window size and kernel used in the smoothing procedure and that results obtained using different settings of these parameters should not be compared. The results resemble former findings indicating that phase classification outputs depend on the spatial scale being considered [23,68,69] and that rapidly increasing proportions and average patch sizes of the stand fragments with complex dbh structures close to the overall stand-level distribution are a universal feature of multiaged forests with strongly randomized patterns of structural heterogeneity [18,35,43,56,[70][71][72].…”

“…This conclusion well corresponds with research suggesting that structural heterogeneity occurs primarily in small areas and diminishes rapidly with increasing spatial scales [70,71,81]. Although homogeneously structured patches occur, they cover a smaller proportion of the total forest area than those that are heterogeneously structured [34,35,43,[54][55][56][57]86]. Glatthorn et al [87] concluded that forest development takes place on a continuous scale and that discrimination between development stages splits continuous datasets at selected thresholds.…”

“…The Dinaric Mountains harbor one of the most volume-rich mountain primeval forests on the continent, with growing stocks commonly exceeding 1000 m 3 per ha. In contrast, in the Western Carpathians, which are the northernmost region of occurrence of such forests, average stand volumes are only half of those reported in the Balkans, which probably is attributable to a more a severe disturbance regime in this region [43]. Thus, the study sites selected for this research characterize a possibly wide range of climatic conditions and disturbance regimes in which the remnants of the former pristine fir-beech-spruce forests still can be found.…”

“…Insignificant variation of the reference values within the geographic regions (i.e., Dinaric Mts and Western Carpathians) suggests that they provide reliable estimates for hypothetical steady-state conditions. The dbh distributions and selected characteristics of spatial tree distributions in the studied stands (e.g., the randomness and independence of the distribution across tree size categories and the spatial variation of the basal area and dbh differentiation) were published formerly [35,43,44]. Notes: I d -dbh differentiation index calculated for the entire stand (see Methods), BA 75 -75th percentile of the stand basal area (determined on circular plots 154 m 2 ), BAss-threshold maximal basal area for the steady-state phase, dbh BA -median basal area tree diameter.…”

“…On the other hand, areas that are too small involve a risk of overlooking overstory trees growing in the close neighborhood and thus impeding the ingrowth of understory individuals and misinterpreting such fragments as hotspots of generation replacement. From this point of view, a plausible recommendation is to keep the basic window diameter (or plots used for sampling) close to the average area occupied by overstory individuals, assuming their more-or-less regular distribution, as frequently reported by studies of natural forests [31,35,43]. Regarding trees with dbh over the median basal area dbh as a mature generation, the maximal nearest-neighbor distances for their hexagonal distribution in the studied stands were within a relatively narrow range between 14 and 18 m. Under real conditions, such a regular distribution of the canopy trees is unlikely, and the nearest-neighbor distances will certainly be smaller.…”