Estimation Of Carbon Stocks In Acacia-Commiphora Woodlands In Kiteto District, Tanzania

ABSTRACT

Carbon stocks of Acacia-Commiphora woodlands was estimated using site specific biomass allometric models developed in Kimana Village, Kiteto District, Tanzania. Fifty sample trees covering a range of diameter at breast height (DBH) from 5.9 cm to 79.2 cm from 12 different tree species were destructively sampled to develop biomass and volume models. The sample trees included three dominant species namely Commiphora africana, Balanites aegyptiaca, and Acacia tortilis. Candidate non-linear models were fitted where DBH or combination of DBH and tree height (Ht) were used as predictor variables against separate tree components. Evaluation of model performance at the beginning was based on statistical significance of parameters: small root mean square error, higher R2, and low mean prediction error. Final model selection was based on Akaike’s Information Criterion, since the selected functions had different number of parameters. The developed models gave an option of using either DBH alone or in combination with Ht, depending on availability of inventory data and financial capability of the user since an additional variable has implication on inventory costs. The average basic density for tree species sampled ranged from 0.32 to 0.75 with an average of 0.59 g cm-3. Forest inventory was carried out with 164 plots of 15m radius systematically laid out at sampling intensity of 0.75%. The number of tree stems per hectare was 971, basal area 5.96 m2 ha-1 and volume 38.8 m3 ha-1 with a carbon content of 21.96 t ha-1. Carbon stock estimates computed from the developed biomass allometric equation with that derived from volume and basic density shows no significant difference for small to medium tree sizes. The developed models can be used in other woodlands with similar conditions and within DBH range of the sample trees.