The relationships between environmental and disturbance factors in temperate deciduous forest ecosystem (Amasya/Turkey)

Deciduous forests face many disturbance factors. Grazing and cutting are the leading factors in this disturbance.The study area's vegetation was analyzed using numerical methods to identify plant communities and determine the relationship between environmental gradients and disturbance factors. The species diversity was calculated using alpha and beta diversity indexes. As a result, four different communities were identified in the study area. One of the communities was under grazing pressure while the other community was under cutting pressure. No disturbance factors were found in the remaining two communities. Elevation and soil moisture were found to be important in the distribution of plant communities. pH, soil moisture, soil % N content and canopy factors were found to be important. The highest Shannon-Wienner diversity index values were found in non-cutting and non-grazing forest communities. The lowest Shannon-Wienner diversity index values were found in grazing and cutting forest communities. Unlike the Shannon-Wienner diversity index, the highest beta index values were found in grazing and cutting forest communities. The lowest beta index values were found in non-cutting and non-grazing forest communities.

In temperate deciduous forests, approximately the 90% of vegetation consists of vascular plant diversity (Whigham, 2004;Gilliam, 2007). The composition and diversity of the ground flora in temperate deciduous forests are affected by the composition of the canopy species and soil and climate characteristics (Hunter, 1999;Augusto et al., 2003;Gilliam, 2007;Barbier et al., 2008). While underground vegetation contributes significantly to total biodiversity in temperate forests, it contributes less to total forest biomass (Gilliam, 2007).
In this study, we examined relationships among disturbance (grazing and tree cutting), environment factors (soil pH, soil nitrogen, soil moisture and light availability) and biodiversity in the temperate deciduous forest.

Materials and Method
The study area is located in the Yeşilırmak basin in the central region of Turkey. The study area is located between 400 m and 1100 m in altitude (Fig. 1). The study area has between oceanic and continental climates. The mean annual temperature and the mean annual precipitation are 13.9°C and 397.5 mm, respectively. The maximum mean temperature is 31.7 °C (August), while the lowest mean temperature is − 0.6 °C (January). The vegetation consists of Irano-Turanian and Mediterranean species. Natural flora has been affected by grazing and tree cutting.
Taxonomic nomenclature followed was that of Davis (1965Davis ( -1985 and Davis et al. (1988), Tutin and Heywood (1964-1980), Güner et al. (2000 and Güner et al. (2012). Four plots were selected from floristically and structurally homogeneous places according to the goal of study. Ten relevés was established for each plot, and the size of plots was determined according to the minimal area method (Westhoff and van Der Maarel, 1978). A cover-abundance value for each species in each relevés was determined using the Braun-Blanquet (1964) scale.
Soil samples for each relevés were taken at a depth of 35 cm. Soil pH values were measured using deionized water (1:1) by pH meter (Kacar, 2012). Soil nitrogen was determined by the way of micro-Kjeldahl method (Bradstreet, 1954). Water content was determined by the gravimetric method (Bayrakli, 1987, Kutbay andOk, 2003). Light availability was determined using a Lutron Light Meter LX-1102 (Schuster and Diekmann, 2005). The Shannon -Wienner diversity indices of the plant communities were calculated using the following formula (Magurran, 2004).
"s" is the total number of recorded species, "pi" is the proportion of percentage cover of the "i"th species to the sum of the percentage cover of all species and ln is the natural logarithm.
Evenness was quantified using Shannon's indices. Indices of the plant communities were calculated using the following formula (Magurran, 2004).

= ′/
where Hmax is maximum species diversity and calculated as log2 Pi.
Beta diversity is defined as spatial heterogeneity or pattern diversity was calculated using the Whittaker formula (Whittaker, 1960;Gulsoy and Ozkan, 2008).
where S is the total number of species, α is the mean species richness.
Plant communities according to disturbance factors were separated by using TWINSPAN procedure. To determine what environmental factors were significant, we also treated our data with Detrended Correspondence Analysis (DCA) and Principal Component Analysis (PCA). Numerical methods were performed by using the "Community Analysis Package 4 version" software (Seaby and Henderson, 2007).
Statistical analysis was performed by using a SPSS (25.0 version) software. The differences among plant communities were investigated by one-way ANOVA. The biodiversity parameters were assessed by Tukey's significant difference (HSD) test to rank the means.  Detrended correspondence analysis (DCA) diagram showed the existence of the gradient considering the first axis (Eigenvalue of axis 1 is 0.96). It is an elevation gradient. Plant communities have spread depending on the elevation. Plant communities in non-grazing and noncutting areas grouped at the left of ordination plot, whereas plant communities in cutting and grazing areas grouped at the right of the ordination plot (Fig. 3). The first two axes explained 81.79 % total variance of the Principal Component Analysis (PCA). PCA revealed that pH, soil nitrogen content (%), soil water content were found to be significant in axis 1, while light availability was found to be significant in axis 2. Soil nitrogen (%) and soil water content were negative in axis 1, while soil pH and light availability were positive in axis 1 and 2, respectively (Table 1).

Results
Light availability was positively correlated with plant communities in cutting areas, While soil nitrogen content and water content were negatively correlated plant communities in grazing areas. pH was positively correlated with plant communities in non-grazing areas (Fig. 4). Species diversity indices (H and J) were high in noncutting and non-grazing plant communities compared to the other areas. Beta diversity was high in cutting and grazing plant communities as compared to the other areas. Statistically significant differences were found among the beta diversities with respect to plant communities (Table 2).

Discussions
The effects of environmental and disturbance factors on plant communities in terrestrial ecosystems are significant (Davies et al., 2007;Pausas and Austin, 2001). These factors affect the establishment, growth and reproduction of species (Pierce et al., 2007;Duru et al., 2010;Frenette-Dussault et al., 2012). According to TWINSPAN and DCA analysis, we found the main four plant communities: grazing, cutting, non-cutting and non-grazing. These communities are distributed according to altitude. Because topographic factors (altitude, geographical aspect, and slope) are primary factors of vegetation distribution (Mark et al., 2000) and affecting plant diversity (Vujnovic et al., 2002). PCA analysis showed that pH, soil nitrogen content (%), soil water content and pH were determining factors of vegetation distribution. In the cutting area, light availability is the main determining factor on species composition, while in the non-cutting area, soil pH and nitrogen content are determining factors (Chai et al., 2016;Tardella et al., 2016). Additionally, unpalatable and thorny species are dominant in grazing vegetation, and fast-growing species are dominant in non-grazing vegetation (Tardella et al., 2016;Kılıç et al., 2018). Overgrazing in meadows and pastures damages the ground flora and prevents the regeneration of dominant species (Malik et al., 2016). However, Pettit et al. (1995) stated that overgrazing increases the proportion of unrelated species.
It has been found that grazing has a significant effect on species richness and diversity and that the number of species and diversity indexes are lower in these areas (Lu et al., 2017;Tälle et al., 2016). Besides, it has been shown that overgrazing negatively affects bush and tree species and thus decreases species richness (Roder et al., 2002;Kumar and Shahabuddin, 2005).
In cutting vegetation, light availability is the main factor (Tardella et al., 2016). Cutting causes permanent grazing gaps, and grassland species are recolonized (Dzwonko and Loster, 1998). These areas are called wood-pastures. If regeneration fails, wood-pastures become permanent. (Bergmeier et al., 2010). In non-cutting vegetation, canopy species have an excellent availability to take light as compared to subcanopy species. Besides, soil pH and nitrogen content affect ground flora formations (Augusto et al., 2003;Chai et al., 2016).
Species diversity indices (H and J) were high in noncutting vegetation compared to the cutting vegetation. The ground flora diversity and composition is influenced by the species composition of the canopy species (Barbier et al., 2008;Gilliam, 2007;Hunter, 1999). Beta diversity was higher in cutting and grazing plant communities than the other plant communities.
Overgrazing harms the ground flora in meadows and pastures, preventing the regeneration of dominant species (Malik et al., 2016). Also, the disappearance of shrub and tree species causes the species richness to decrease gradually (Roder et al., 2002;Kumar and Shahabuddin, 2005). Considering the results obtained, it was consistent with previous studies (Faria et al., 2018). Tree cutting and forestry studies increase habitat heterogeneity (Bergmeier et al., 2010).
Paying attention to the protection of biological diversity in forestry activities should be the main goal of sustainable forest management. In this study, we revealed that disturbance and environmental factors affect vegetation types and species composition.