Carbon stock in rubber tree plantations in Western Ghana and Mato Grosso (Brazil)

• Published in: Forest ecology and management 2008-04, Vol.255 (7), p.2347-2361
• Main Authors: Wauters, J.B., Coudert, S., Grallien, E., Jonard, M., Ponette, Q.
• Format: Article
• Language: English
• Subjects: Age effect; Allometric equation; Animal and plant ecology; Animal, plant and microbial ecology; Artificial regeneration. Forest nurseries. Planting; Biological and medical sciences; Carbon; Forestry; Fundamental and applied biological sciences. Psychology; Hevea brasiliensis; Site effect; Sowing and planting; Synecology; Terrestrial ecosystems
• ISSN: 0378-1127; 1872-7042
• DOI: 10.1016/j.foreco.2007.12.038

Cultivation of rubber trees on non-forested land could act as a carbon sink by sequestering carbon in biomass and indirectly in soils. International political and economical interests, following the Kyoto Protocol, require estimates of this carbon sequestration. The carbon stock of two rubber tree plantations ( Hevea brasiliensis (Willd.) Muell.-Arg.) was assessed in two contrasting climatic areas: western region in Ghana (WG; 2–14-year-old) and Mato Grosso (MG; 14–25-year-old) in Brazil. Trees (76 in WG and 210 in MG) spanning a range of stand ages, clone types and planting layouts were felled and partitioned into log, live lignified branches of selected diameter (large, medium and fine), dead branches, non-lignified fine branches, leaves, taproot and lateral roots. Allometric relationships (log-transformed power functions) based on trunk circumference at a height of 170 cm ( C 170) were used to predict the tree foliage, aboveground, belowground and total carbon content (kg C tree −1) ( r 2 = 0.86–0.99). The way in which carbon content varied as a function of C 170 depended both on components and sites. Sampling plots (25 in WG and 34 in MG) were established in stands of varying ages, clones and planting layouts, to quantify carbon pools. The tree carbon stocks (t C ha −1) estimated from the allometric relationships for total tree were fitted as a function of stand age using a non-linear sigmoid curve of Gompertz (total tree r 2 = 0.841 and 0.854 in WG and MG, respectively). Predicted tree carbon stock for 14-year-old stands was 83% higher in WG (76.3 t C ha −1) than in MG (41.7 t C ha −1), which was partially explained by a difference in tree height growth between the two locations resulting from the contrasting sites. In addition to tree components, the carbon stocks associated with dead wood on the ground, understorey vegetation, litter layer, fine roots (0–60 cm), soil organic carbon (0–60 cm) and charcoal (0–60 cm) were assessed using linear mixed models to evaluate the influence of clone, site, planting layout and age. Only age appeared to have a significant effect for some pools. For 14-year-old stands, the contribution of the soil organic carbon pool (0–60 cm) ranged between 39 and 69% of the total carbon stock, which amounted to 135 and 153 t C ha −1, in WG and MG, respectively. The cumulated contribution of all remaining soil and ground pools (dead wood on the ground, understorey vegetation, litter layer, fine roots and charcoal) was about 5% in