The study area is located in Tankawati natural hill forest, which is under the administrative control of Padua forest range of Chittagong (South) Forest Division, Bangladesh. It lies between 21°57'08''N and 22 ° 09'13''N latitude and between 92 ° 07'32''E and 92 ° 02'22''E longitude, and area of the forest is 1,123.72 ha. The elevation of the study area ranges be- tween 14 m and 87 m above m a.s.l. . The study area has the moist tropical maritime climate with high rainfall. The mean minimum and maximum temperature is 21.97°C and 30.51°C, respectively. The highest concentration of precipitation is found from May to September, pre- and post-monsoon periods of rain during April, May and October. The mean monthly relative humidity is high through- out the year. Soil is brown sandy loams, somewhat excessively drained, Barkal soil series and classified according to the USD a Taxonomy by Alam et al. as Udic Utochrept. The study was based on field data collection through physical measurement, field observation and laboratory analysis, and it was conducted from January to December 2009. a systematic sampling method was used for identification of each intersection point in the field. The geo-position of the study area was mapped out, and then one-minute intervals were inserted in the map and finally, a total of 72 intersection points was located in the map. Each point was identified using GPS in the field. Primarily land use of each intersection point was identified in the field. after that a total of 20 intersection points was taken in the natural forest, following that four sampling plots of 20 m × 20 m sizes were selected for tree species around each intersection point. a sampling plot of 2 m × 2 m in size was selected in the centre of the intersection point for shrubs, herbs and grass species, and litter fall collection. Finally, a total of 80 plots for tree species, 20 plots for shrubs, herbs and grass species, and 20 plots for litter fall collection were considered. In the fixed grid lines, to estimate carbon stock, the number of stems was counted, Spiegel Relascope and diameter tape were used to measure tree height and diameter, respectively. Tree increment core sample from each tree was collected with wood borer at breast height (1.3 m). In each sampling plot of shrubs, herbs and grass species, the number of each species was counted and their samples were uprooted for laboratory analysis. The litter fall included leaves, fruits, seeds, barks and twigs, etc. In the litter fall sampling plot, all weeds and brushes were cleared, and fresh weight of the collected litter fall was measured in the field using a field balance. Fallen litter was collected after six months and the average litter fall for six months was converted to annual litter fall per ha to estimate the biomass of the litter fall in the forest. The carbon stock in soil was estimated to a depth of 1 m using an earth augur from selected geo-position (plot size of 2 m × 2 m) and a total of 20 plots were sampled at five different depths: 0–1 cm, 1–3 cm, 3–14 cm, 14–30 cm and 30–100 cm. about 774 individuals of 56 tree species were measured in the sampling plots. a model of Brown et al. was used to determine aboveground biomass of each tree; so far literature showed that this method is one of the most suitable methods for biomass carbon stock estimation in tropical forests. The model: Y = Exp.{ − 2.4090 + 0.9522 ln (D 2 × H × S)} where: Y – aboveground biomass (kg), H – height of tree in meters, D – diameter (cm) at breast height (1.3 m), S – wood density (t·m −3 ) for specific species.
Below ground biomass was calculated consider - ing 15% of the aboveground biomass. Collected undergrowth samples, in total of 817 individuals, were divided into above- and below-ground identifying the collar region and then fresh weight was taken. after that, the weight value was multiplied by the number of individuals of each species in all sampling plots. The loss on ignition method was used to estimate biomass car - bon stock. In this method, initially fresh weights of vegetative samples were taken, then dried at 65°C in the oven for 48 hours to take dry weight. Oven- dried grind samples were taken (1 g) in pre-weight - ed crucibles, after that put in the furnace and that was followed by ignition for one hour. after cool- ing, the crucibles with ash were weighted and then the percentage of carbon was calculated according to Allen et al.. ash (%) = (W3 − W1 )/(W2 − W1 ) × 100, C (%) = (100 − ash) × 0.58 (considering 58% carbon in ash-free litter material) where: C – biomass carbon stock, W1 – weight of crucible, W2 – weight of the oven-dried grind sample and crucible, W3 – weight of ash and crucible. During field work, soil from each depth was collected to determine organic carbon, and soil core was used to calculate bulk density for different depths. Field’s moist soil cores were dried in an oven at 105°C for eight hours, and re-weighted to determine moisture content and dry bulk density. To estimate the percentage of carbon in the soil, samples were analyzed by the wet oxidation method. Duncan’s Multiple Range Test (DMRT) was done for data of each of the parameters using SPSS package to determine levels of significance.