Forests play a significant role in offsetting anthropogenic CO2 emissions and hence in climate change mitigation (Brown et al. 1996, IPCC 2013, Brienen et al. 2015). With the widespread concern about human activities increasing level of atmospheric CO2 there is a need to assess the potential of native forests in carbon sequestration and storage (C) (Johnson and Kern 2002, Sharma et al. 2010, Borah and Garkoti 2011, Gandhi and Sundarapandian 2017). Forest biomass determines the potential amount of C that can be added to the atmosphere or sequestered in terrestrial ecosystems when they are managed for meeting emission targets (Brown et al. 1999, Brienen et al. 2015). Estimation of the existing C stocks in different forest ecosystems in regional and local level helps in making decisions about C management within the forest ecosystems (Sahu et al. 2016). Forest C stock is also a useful measure for comparing structural and functional attributes of forest ecosystems across a wide range of environmental conditions (Brown 2002, Gandhi and Sundarapandian 2017).
Tropical forests are considered as the most diverse terrestrial ecosystems and largest repository of aboveground C stock covering only about 6% of the earth surface (Beer et al. 2010, Pan et al. 2013, Brienen et al. 2015). Due to rapid deforestation, agricultural expansion, urbanization and industrialization since last few decades, they are being degraded and losing significant portion of biodiversity (LaFrankie et al. 2006, Brienen et al. 2015). Anthropogenic disturbances in forests result in changes in community structure, which in turn influence ecosystem function and processes (Collins 1987, Pawar et al. 2014, Sicard and Dalstein-Richier 2015). In view of the growing threats to tropical forests it would be interesting to understand how natural forests and their phytosociological attributes are affected by the progressive degradation due to anthropogenic activities; and functional relationship between such attributes and C storage (Chapin 2000, Tilman 1988, Srivastava and Vellend 2005, Kirby and Potvin 2007).
Being part of Indo-Burma biodiversity hotspots, tropical forests of southern Assam are rich from biodiversity standpoint (Borah and Garkoti 2011). The tropical forests of southern Assam still retain significant proportion of its phytodiversity, possibly due to long years of isolation and in some cases due to difficult terrains (Borah et al. 2014). However, the forests closer to the human settlements are facing high level of anthropogenic pressure such as extraction of wood for timber unplanned and unsustainable extraction of non-timber forest products, etc., by the local people (Borah 2012, Borah et al. 2014). These anthropogenic activities not only accelerated the biodiversity loss in the forests of the region but also other ecological functions such as productivity, carbon stocks, regeneration and forest hydrology (Borah 2012, Borah et al. 2014, Athokpam et al. 2014, Borogayary et al. 2018) . In this paper we estimate the above ground C stocks of the tree species in tropical forests of Assam along a disturbance gradient and establish relationships with different phytosociological attributes.
6.2 Materials and Methods
2.1 Study site
The present study was conducted in the three districts viz., Cachar, Hailakandi and Karimganj of southern Assam (24°08′-25°05’N, 92°15?-93°15?E.). The study area is popularly known as Barak Valley (Fig. 6.1). The valley is characterized by hot and humid climate. The three districts cover 6920 km2 geographical area and of which 55% was covered by forests (FSI 2017). According to Champion and Seth (1968), vegetation of the region is dominated by Cachar tropical evergreen forests and Cachar tropical semi evergreen forests. Cynometra polyandra, Mesua ferrea, Stereospermum personatum, Artocarpus chama, Palaquium polyanthum, Mesua floribunda, Dysoxylum binectariferum, Trewia nudiflora and Pterygota alata are the dominant tree species in these forests (Boarh et al. 2014).
6.2.2 Vegetation sampling
Based on reconnaissance survey on the forest type, canopy density and disturbances, 26 forest sites were selected for the present study (Fig. 6.1). The vegetation of each forest site was analyzed through a belt transect of 500 m × 10 m in each forest following Ganesh et al. (1996), Borah and Garkoti (2011). Each belt transect was again subdivided into 50 quadrats of 10 m × 10 m size along its length. In each quadrat, all the woody plants (excluding lianas) with > 10 cm circumference at breast height (CBH) were considered as tree and recorded with their CBH (Singh and Dadhwal 2009, Borah et al. 2014). The cut stumps and lopped trees in each quadrat were also counted and their girths were measured at ground height (i.e. 10 cm from the ground) for estimating the disturbance index. Specimens of each species were brought to the laboratory and identified with the help of “Flora of Assam” (Kanjilal 1934-1940), “Assam’s Flora” (Chowdhury et al. 2005) and the herbarium of the Botanical Survey of India, Shillong.
6.2.3 Disturbance index
A disturbance index for each forest site was calculated following Kanzaki and Kyoji (1986), Rao (1990) and Borah (2012). The disturbance index (DI) was calculated as the basal area of “cut trees measured at the ground level expressed as fraction of total basal area of all trees
DI (%)= (Basal area of cut stumps × 100)/(Total basal area (cut stumps basal area+Standing tree basal area))
Based on disturbance index, the forest sites were classified into (i) undisturbed forest (disturbance index 0%), (ii) mildly disturbed forest (disturbance index upto 20%), (ii) moderately disturbed forest (disturbance index 20-40%) and (iv) highly disturbed forest (disturbance index above 40%). Of the 26 forest sites, five sites were recorded as undisturbed forests, six forest sites were mildly disturbed, seven sites were moderately disturbed and eight sites were highly disturbed forests. Generally undisturbed and mildly disturbed forests were more than 6 kms and moderately and highly disturbed forests were less than 6 km away from the human habitation. Fuel wood and bamboo collection, tree felling for timber, cattle grazing and wild edible collections were the common practices in moderately and highly disturbed forests, but such practices were rare in the undisturbed and mildly disturbed forests.
6.2.4 Phytosociological attributes
The phytosociological data were quantitatively analysed for frequency, density, basal area, relative density, relative frequency and relative dominance (Curtis and McIntosh 1950). Distribution of tree density and basal area in different DBH classes were estimated by following Mueller- Dombois and Ellenberg (1974). Trees were categorized into ten DBH classes starting with 90 cm class and tree density and basal area each DBH class was estimated.
Diversity index was calculated following Shannon ; Wiener (1963) as follows
Where pi is the proportion of individuals of ith species and total number of individuals of all species.
The concentration of dominance was calculated following Simpson (1949) as follows-
Where pi is the proportion of individuals of ith species and total number of individuals of all species.
6.2.5 Aboveground biomass (AGB) and aboveground carbon (AGC)
Because of high species richness (132 tree species) in the studied forests, species-specific regression models were not used. Though, there are several regression equations for estimating aboveground biomass (AGB), we used the following equation developed by Brown (1997) for simplicity, less prediction error and higher R2 value. The regression model is
Y= 21.297-6.953 (D) + 0.740 (D2) with R2 value 0.87.
Where Y is the aboveground biomass (AGB) and D is diameter of the tree.
We assumed vegetation carbon equal to 50% of biomass for all the tree components. The estimates are based on assumption of common carbon content per biomass unit as in many other similar studies (Brown and Lugo 1982, Montagnini and Porras 1998, Borah et al. 2013).
6.2.6 Statistical Analysis
ANOVA was performed to compare the average tree density, basal area, diversity index, AGB and AGC among the forest categories with the help of SPSS 16.0. Pearson correlation coefficient was calculated to determine the functional relationships among the phytosociological attributes of the forests.
6.3. Results and Discussion
6.3.1 Forest Structure
A total of 132 tree species was recorded in 26 forest stands. The highest number of species (92) was recorded in mildly disturbed forests and the lowest (47) in highly disturbed forests (Table 6.1). The Shannon diversity index of undisturbed forests ranged from 1.72 to 2.14 while it was recorded 1.78 to 2.19, 1.29 to 1.71 and 1.11 to 1.35 respectively in mildly, moderately and highly disturbed forests. The average Shannon diversity index was highest (1.98) in mildly disturbed forests followed by undisturbed (1.94), moderately (1.51) and highly disturbed (1.23) forests (Table 6.1). The average tree density was highest in mildly disturbed forests (693 ± 12 trees ha-1) followed by undisturbed (676 ± 10 trees ha-1), moderately (675 ± 17 trees ha-1) and highly disturbed (328 ± 6trees ha-1) forests. The average basal area was highest (42.9 ± 1.6 m2 ha-1) in undisturbed forests followed by mildly disturbed forests (39.6 ±1.8m2 ha-1), moderately disturbed forests (20.8 ± 1.0m2 ha-1) and highly disturbed forests (14.4 ± 0.9m2ha-1). The Simpson dominance index was highest (0.075) in highly disturbed forests and lowest (0.047) in mildly disturbed forests (Table 6.1). All the phytosociological attributes (species number, tree density, diversity and basal area) were significantly different in different forest categories (Table 6.1). The forest structural analysis reveals that the species richness (47-92 species), Shannon index (1.11-2.19), tree density (328-693 trees ha-1) and tree basal area (14.37-42.91 m2 ha-1) recorded in different forest categories are comparable with earlier studies (Table 6.2) in tropical forests of north-east India by Nath et al. (2005), Deb and Sundriyal (2008), Deb et al. (2009), Borah and Garkoti (2011), Thapa et al. (2011), Borah et al. (2013, 2014), Nandy and Das (2013).
The highest species richness, diversity and density of tree species in the mildly disturbed forests were due to the favourable conditions (such as less competition for available resources, more sunlight penetration to the forest floor through canopy gaps etc.) for the growth and regeneration of tree species (Boarh and Garkoti 2011). Pressures such as relentless extraction of fuel wood, tree felling, and non timber forest products including bamboo collection were found responsible for lower species richness, density, diversity and basal area in moderately and highly disturbed forests (Boarh and Garkoti 2011, Borah et al. 2014).
6.3.2 Distribution of density and basal area in different DBH classes
Density and basal area distribution in different DBH (Diameter at Breast Height) classes can be used as indicators of changes in population structure and species composition (Newbery and Gartlan 1996). The distribution of tree density and basal area in different DBH classes of undisturbed and mildly disturbed forests revealed a reversed J-shaped and J-shaped curve, respectively. In undisturbed and mildly disturbed forests, tree density decreases and basal area increased with increasing DBH classes (Fig. 6.2). In both undisturbed and mildly disturbed forests, tree density was high in 90 cm DBH class. High tree density in lower size DBH class indicates continuous regeneration of the forest ecosystems and higher total basal area in higher DBH class indicates that the forests are undisturbed and old growth. Similar trends were reported for different tropical forests of north east India by Khamyong et al. (2004), Nath et al. (2005), Deb et al. (2009), Borah ; Garkoti (2011), Borogayary et al. (2018). In moderately disturbed and highly disturbed forests both tree density and basal area showed fluctuating distribution curves. In moderately disturbed forests, the density of younger trees (DBH