Population structure and spatial point-pattern analysis of a mono stand of Acacia polyacantha along a catena in a savanna ecosystem
Introduction
Many savanna landscapes are shaped by catena-forming processes resulting from the interaction of rainfall and terrain morphology (Davies et al., 2016). Soil catenas facilitate the processes of illuviation and eluviation in soil profiles (Khomo et al., 2011). One important feature of catenas is the mobilisation and redistribution of solutes, colloids, and particles in the upper sections, leading to soil differentiation across the slope (Khomo et al., 2011, Schimel et al., 1985). Indeed, this process makes soils found on upper catenas to be sandy, well-drained, and nutrient poor, whereas soils in low-lying areas are moist and relatively rich in clay, organic matter and nutrients (Khomo et al., 2011, Schimel et al., 1985, Scholes and Walker, 1993). Although many woody savanna species have been observed to be strongly associated with catenal positions (Baldeck et al., 2014), little is known on the effect of catena on structure and spatial distribution of woody species that are adapted to all catena positions. Topographical changes of forest structure in tropical rainforests has been explored (Homeier et al., 2010, Takyu et al., 2002, Webb et al., 1999), but savanna woodlands are still poorly understood, more so for mono stands of natural woodlands. Tree to tree interaction together with catena formation have the potential to influence the size class distribution and spatial arrangement of individual trees. Furthermore, spatial pattern analysis enhances our understanding of changes in tree structure across the catena, since it can pick processes such as competition and facilitation that have an impact on vegetation structure (Getzin et al., 2008).
The spatial distribution of plants can be explored using mapped plant locations (Getzin and Wiegand, 2007). The small-scale positive or negative interactions of plants can be interpreted through spatial point pattern produced by spatial statistics such as Ripley’s K function (Ripley, 1981, Ripley, 1976), the pair-correlation function (Law et al., 2009, Stoyan and Stoyan, 1994), or the distribution function of nearest-neighbour (NN) distances (Diggle, 2003). One advantage of point pattern analysis is its ability to separate neighbourhood processes (i.e. first order effects resulting from the interaction of organisms under investigation) from environmental heterogeneity, second order effects such as favourable and unfavourable patches in the environment (Getzin et al., 2008, Wiegand and Moloney, 2004). Once site specific heterogeneities have been controlled, it is possible to directly compare second-order effects of sites that differ in environmental heterogeneity (Getzin et al., 2008).
One prominent hypothesis in tree point pattern analysis suggests that rare species are at an advantage because trees have lower survival in areas of high conspecific density due to increased attack by natural enemies, a process known as negative density dependence (NDD) (Bagchi et al., 2011). Therefore, NDD is expected to be more pronounced in vegetation mono stands. Individual plants are not only influenced by population density, but rather by the proximity, size, and activity of its immediate neighbours (Mack and Harper, 1977). In the absence of disturbances such as fire, NDD competition is considered to be responsible for shaping stand-level structure (Kneeshaw and Bergeron, 1998). Furthermore, biological interactions, seed dispersal and germination success during early stages of stand establishment are also important (Martínez et al., 2010).
In cases of strong density dependence, trees that survive the effects of competition are expected to be more regular with clustering expected to decline with increase in size classes (Barot et al., 1999). Therefore, understanding competition is critical for predicting its effects on forest stand structure. Nevertheless, one major challenge in spatial point pattern analysis is that, different mechanisms can produce the same spatial pattern (random, regular or aggregated) (Pielou, 1960). For example, the generation of an initial random template via germination and an uneven age distribution (Kenkel, 1998). However, because trees are sessile organisms, several studies have managed to infer mechanisms from the observed spatial point pattern (Getzin et al., 2006, Gray and He, 2009). For example, processes such as dispersal, competition, disturbance and spatially structured environmental predictors can be evident in the point pattern analyses, leading to great inferential gain (McIntire and Fajardo, 2009).
In order to understand the effect of catena on the structure and spatial distribution of a mono stand of Acacia polyacantha, we sampled trees in plots across the catena. The specific objectives were to: (i) determine the effect of catena on the structure of A. polyacantha, and (ii) determine the spatial distribution of A. polyacantha across the catena. We hypothesized that trees in the upper section of the catena are of highest density and smallest structure (height, canopy cover and basal area) due to shallow soils, lack of adequate moisture and nutrients (Grubb, 1977, Homeier et al., 2010, Khomo et al., 2011, Takyu et al., 2002). Based on the increase in size of the trees down the catena (Homeier et al., 2010), we predicted the spatial distribution of the trees to change from clustering in the upper section to random in the middle section and overdispersion in the bottom section of the catena, with negative density depended thinning being the most prevalent mechanism shaping distribution at the bottom. Furthermore, higher environmental harshness in the upper section of the catena was expected to increase the importance of facilitation as a structuring mechanism (Callaway et al., 2002). Moreover, general increase in moisture and nutrients down the catena (Schimel et al., 1985) were expected to produce intense competition in the bottom section leading to deviation from CSR to overdispersion.
Section snippets
Study site description
The study site is located at (latitude: 17°18′–17°19′S and longitude: 31°11′–31°12′E) in a farming area about 10 km west of Bindura town, Zimbabwe. The average rainfall is 700–1000 mm per annum, and mean annual temperature is 26.5 °C. The slope of the study area faces north and receives solar radiation all year round. The catena is generally gentle from the upper section to the bottom (slope angle = 2.6%). The slope angle did not vary much in each catena section from upper to bottom catena (
Population density and structural variables
The number of trees recorded in the upper, middle and bottom sections of the catena were 71, 96 and 33, respectively. Tree density was highest in the middle section of the catena followed by the upper catena and least in the bottom section of the catena. There was significant difference (F2,197 = 115.14, p < .0001) in the mean basal area per tree across the catena (Table 1). Tree mean basal area for bottom catena was significantly higher than for both upper and middle sections by factors of 3.3
Discussion
Trees had larger canopy cover, basal area and taller in the bottom section of the catena than in the middle and upper catena sections. The abundance of trees was inversely related to the structural variables down the catena. Although we expected all tree structural variables to be larger in the middle section compared with the upper section of the catena, only height was greater. Lack of difference in tree basal area and canopy cover between upper and middle sections of the catena could have
Conclusions
We show that tree structure (height, basal area and canopy cover) for A. polyacantha is highest in the bottom section of the catena, an observation that corroborates previous studies that recorded valley bottoms to contain large trees (Baldeck et al., 2014, Homeier et al., 2010, Takyu et al., 2002, Webb et al., 1999). Because of the large tree size in the bottom section of the catena, competition has led to increased nearest neighbour distances, confirming negative density dependent. Although
Acknowledgements
We thank the two anonymous reviewers of this article who helped us increase the clarity and quality of our work.
References (54)
- et al.
Dynamics of some forests in KwaZulu-Natal, South Africa, based on ordinations and size-class distributions
South African J. Bot.
(1995) - et al.
Asymmetric tree growth at the stand level: random crown patterns and the response to slope
For. Ecol. Manage.
(2007) - et al.
Spatial point-pattern analysis for detecting density-dependent competition in a boreal chronosequence of Alberta
For. Ecol. Manage.
(2009) - et al.
Characterising wide spatial variation in population size structure of a keystone African savanna tree
For. Ecol. Manage.
(2012) - et al.
Impact of rainfall and topography on the distribution of clays and major cations in granitic catenas of southern Africa
Catena
(2011) - et al.
Spatial associations among tree species in a temperate forest community in North-western Spain
For. Ecol. Manage.
(2010) - et al.
Changes of woody plant interaction and spatial distribution between rocky and sandy soil areas in a semi-arid savanna, South Africa
J. Arid Environ.
(2011) - et al.
On tests of spatial pattern based on simulation envelopes
Ecol. Monogr.
(2014) - Baddeley, A., Rubak, E., Turner, R., 2015. Spatial Point Patterns Methodology and Applications with R. Chapman and...
- et al.
and Habitat associations in tropical trees
Ecology
(2011)