Spatial patterns of tree species in Suserup Skov – a semi-natural forest in Denmark
Introduction
Close-to-nature forest management is claimed to protect biodiversity as well as the ecological structures and functions of the forest and thus to provide long term sustainability, while satisfying the economic needs of the forest owner (Hofle, 1995, Larsen, 1997, Nord-Larsen et al., 2003). When developing silvicultural practices that mimic natural forest structures and processes, natural forests have served as basic reference. In natural forests, the spatio-temporal structure of trees results from a large number of feedback loops of processes caused by the underlying disturbance regime in time and space. For example, current structure (i.e. tree size and occurrence of gaps) have direct impact on seeding and rejuvenation processes, which in turn affects future forest structure. Since different processes modify forest structure, assessment of current forest structure may be used for interpretation of the underlying processes which are commonly difficult to measure (Pretzsch, 2010). As such interactions have implications for designing forest management practices a large scientific effort has been devoted to the assessment and analysis of forest structure.
Previously, natural forest structure has been characterized by the mere distribution of tree sizes (Leibundgut, 1993, Korpel, 1995, Emborg et al., 1996, Emborg, 1998) and the degree of size heterogeneity has been characterized by the Gini coefficient (Lexerod and Eid, 2006, Nord-Larsen et al., 2006). However, indices such as the Gini and Shannon diversity index (Shannon, 1948) do not include the spatial structure and hence interactions between individual trees. Spatial functions such as Ripleys’s K(r) function and pair correlation function, g(r), are among several functions that are used for analyzing unmarked point patterns and only describe distribution of trees without consideration of differences in species or size (Ward et al., 1996, Chokkalingam and White, 2001). In contrast, the bivariate pair correlation function and mark variogram, that are used for analyzing marked point patterns, allow simultaneous consideration of positions of trees and their species or size (Stoyan and Penttinen, 2000, Pommerening, 2002, Wiegand and Moloney, 2004, Wälder and Wälder, 2008, Pommerening and Särkkä, 2013). However, all above-mentioned spatial functions require datasets with known tree positions. In contrast, some neighborhood based indices can be used for quantifying forest spatial structure with little effort as by-product during a normal field sampling: the uniform angle index (von Gadow et al., 1998), mingling index (von Gadow, 1993), and dominance index (von Gadow et al., 2012).
In a study of the dynamics in a temperate, semi-natural beech forest in Denmark (Suserup Skov), Emborg et al. (2000), developed a model of the forest cycle that included five sequential phases: innovation, aggradation, early biostatic, late biostatic and degradation. As each phase occurs asynchronously in patches of varying size, the result was characterized as a fine-grained shifting mosaic of successional stages. The characterization of forest structure and thus the underlying processes in Suserup Skov (Emborg et al. op. cit.) and similar examples (Watt, 1925, Watt, 1947, Christensen et al., 2007, Heiri et al., 2009) has served as a reference for the development of close-to-nature forest practices (Angelstam et al., 2004, Larsen et al., 2010, Larsen, 2012). However, the methodological approach in these studies was mainly descriptive and the scientific description and understanding of forest structures was likely subjective. In this study we therefore aimed to quantify the observed spatial patterns using a series of different spatial indices and functions. We believe that knowledge on spatial patterns and the relationship between different species in a natural forest will be useful as a reference to devise strategies for close-to-nature forest management.
The research questions may be formulated as:
- (1)
How is the horizontal distribution of dominant tree species in Suserup,
- (2)
Are there any interactions between different tree species (segregation or aggregation) or are they independently distributed, and
- (3)
Is there any spatial correlation between tree diameters, and if so, at what scale are they spatially correlated?
Finally, we examined the consistency of the results on questions 1–3 obtained by neighborhood based indices and spatial functions at small scales. The simplicity of neighborhood-based indices including uniform angle index, mingling index and differentiation index provide more possibilities for the application in a normal inventory, therefore, they are still being used and the papers based on them are still being published in international peer-reviewed journals (Hui et al., 2011, Li et al., 2012, Li et al., 2014, Zhao et al., 2014, Meng et al., 2016, Li et al., 2017).
It was hypothesized that a deviation from a random distribution towards more aggregated patterns should be found due to gap regeneration processes. Considering the mixed species forest in Suserup, it was expected that the spatial pattern of species and their mingling characteristics would be different, given their differences in terms of requirements and role in the succession. It was also assumed that the results obtained by spatial indices should be consistent with the results from spatial functions at small scales.
Section snippets
Study area
Suserup Skov is a 19.2 ha semi-natural, nemoral beech (Fagus sylvatica L.) dominated forest located on central Zealand (UTM zone 32: E661870, N6139930). The climate is cool-temperate with a mean annual temperature of 8.8 °C and a mean annual precipitation of 674 mm, which is quite evenly distributed across the year. However, the most of precipitation occurs during late summer and autumn (average climatic data 2001–2010 (Wang, 2013)). The soil parent material is a nutrient-rich, calcareous glacial
Non-spatial structural attributes
The species included in the structural analyses made up 90% of the total stem number in the two plots (Table 1). Species and size distributions differed clearly between the two plots. In the less disturbed plot, beech made up 55% of the stems with a mean diameter of 21.8 cm compared to a mean diameter of 42.3 cm and 10% of the stems in the more recently disturbed plot. Oppositely, sycamore maple made up 40% of the stems in the more recently disturbed plot compared to less than 2% in the less
Non-spatial structural attributes
Within plots, trees showed a diameter distribution typical of old-growth stands covering a wide range of diameter classes and showing a typical decreasing trend in the less disturbed plot. In the more recently disturbed plot beech had a bimodal diameter distribution. The increase in the frequency of beech trees with dbh of 45–75 cm in disturbed plot probably reflects a major natural regeneration event in the decades following 1807, when the forest was fenced and cattle browsing stopped (Emborg
Conclusion
Overall analyses of mark variograms in Suserup Skov, indicated that the semi-natural beech forest is mostly composed of fine-grained patches. Based on our analyses in the less disturbed part of the forest, all tree species showed positive spatial correlation of dbh, probably as a result of gap-phase dynamics. The clustering of elm and ash trees occurred at larger scales (∼15 m) than beech and sycamore maple trees (∼10 m), probably as a result of preference and success of elm and ash regeneration
Acknowledgements
This work was supported by the Iranian Ministry of Science, Research and Technology (MSRT) and Department of Geoscience and Natural Resource Management at the University of Copenhagen for a Ph.D. program in forest biometrics.
References (82)
Understorey light conditions and regeneration with respect to the structural dynamics of a near-natural temperate deciduous forest in Denmark
For. Ecol. Manage.
(1998)- et al.
The structural dynamics of Suserup Skov, a near-natural temperate deciduous forest in Denmark
For. Ecol. Manage.
(2000) - et al.
Correct testing of mark independence for marked point patterns
Ecol. Model
(2011) - et al.
Analysing tree species and size diversity patterns in multi-species uneven-aged forests of Northern China
For. Ecol. Manage.
(2014) - et al.
An evaluation of different diameter diversity indices based on criteria related to forest management planning
For. Ecol. Manage.
(2006) - et al.
Spatial structural characteristics of three hardwood species in Korean pine broad-leaved forest-Validating the bivariate distribution of structural parameters from the point of tree population
For. Ecol. Manage.
(2014) - et al.
Spatial associations among tree species in a temperate forest community in North-western Spain
For. Ecol. Manage.
(2010) - et al.
The significance of different indices for stand structure and diversity in forests
For. Ecol. Manage.
(2001) - et al.
Economic analysis of near-natural beech stand management in Northern Germany
For. Ecol. Manage.
(2003) - et al.
Overstory succession in a mixed Quercus petraea-Fagus sylvatica old growth forest revealed through the spatial pattern of competition and mortality
For. Ecol. Manage.
(2014)