%0 Research article %T Multispecies allometric equations for shrubs and trees biomass prediction in a Guinean savanna (West Africa) %A Kouamé, Yao Anicet Gervais %A Millan, Mathieu %A N'Dri, Aya Brigitte %A Charles-Dominique, Tristan %A Konan, Marcel %A Bakayoko, Adama %A Gignoux, Jacques %D 2022 %J Silva Fennica %V 56 %N 2 %R doi:10.14214/sf.10617 %U https://silvafennica.fi/article/10617 %X

Currently, tools to predict the aboveground and belowground biomass (AGB and BGB) of woody species in Guinean savannas (and the data to calibrate them) are still lacking. Multispecies allometric equations calibrated from direct measurements can provide accurate estimates of plant biomass in local ecosystems and can be used to extrapolate local estimates of carbon stocks to the biome scale. We developed multispecies models to estimate AGB and BGB of trees and multi-stemmed shrubs in a Guinean savanna of Côte d’Ivoire. The five dominant species of the area were included in the study. We sampled a total of 100 trees and 90 shrubs destructively by harvesting their biometric data (basal stem diameter Db, total stem height H, stump area SS, as well as total number of stems n for shrubs), and then measured their dry AGB and BGB. We fitted log-log linear models to predict AGB and BGB from the biometric measurements. The most relevant model for predicting AGB in trees was fitted as follows: AGB = 0.0471 (ρDb2H)0.915 (with AGB in kg and ρDb2H in g cm–1 m). This model had a bias of 19%, while a reference model for comparison (fitted from tree measurements in a similar savanna ecosystem, Ifo et al. 2018) overestimated the AGB of trees of our test savannas by 132%. The BGB of trees was also better predicted from ρDb2H as follows: BGB = 0.0125 (ρDb2H)0.6899 (BGB in kg and ρDb2H in g cm–1 m), with 6% bias, while the reference model had about 3% bias. In shrubs, AGB and BGB were better predicted from ρDb2H together with the total number of stems (n). The best fitted allometric equation for predicting AGB in shrubs was as follows: AGB = 0.0191 (ρDb2H)0.6227 n0.9271. This model had about 1.5% bias, while the reference model overestimated the AGB of shrubs of Lamto savannas by about 79%. The equation for predicting BGB of shrubs is: BGB = 0.0228 (ρDb2H)0.7205 n0.992 that overestimated the BGB of the shrubs of Lamto savannas with about 3% bias, while the reference model underestimated the BGB by about 14%. The reference model misses an important feature of fire-prone savannas, namely the strong imbalance of the BGB/AGB ratio between trees and multi-stemmed shrubs, which our models predict. The allometric equations we developed here are therefore relevant for C stocks inventories in trees and shrubs communities of Guinean savannas.