Allometric equations for biomass estimations in Cameroon and pan moist tropical equations including biomass data from Africa

Moist tropical forests in Africa and elsewhere store large amounts of carbon and need accurate allometric regressions for their estimation. In Africa the absence of species-specific or mixed-species allometric equations has lead to broad use of pan moist tropical equations to estimate tree biomass. This lack of information has raised many discussions on the accuracy of these data, since equations were derived from biomass collected outside Africa. Mixed-species regression equations with 71 sample trees using different input variables such as diameter, diameter and height, product of diameter and height, and wood density were developed to estimate total aboveground biomass and biomass of leaves and branches for a Cameroon forest. Our biomass data was added to 372 biomass data collected across different moist tropical forests in Asia and South America to develop new pan moist tropical allometric regressions. Species-specific and mixed-species height diameter regression models were also developed to estimate heights using 3833 trees. Using only diameter as input variable, the mixed-species regression model estimates the aboveground biomass of the study site with an average error of 7.4%. Adding height or wood density did not improve significantly the estimations. Using the three variables together improved the precision with an average error of 3.4%. For general allometric equations tree height was a good predictor variable. The best pan moist tropical equation was obtained when the three variables were added together followed by the one which includes diameter and height. This study provides height diameter relationships and wood density of 31 species. The pan moist tropical equation developed by , estimates total aboveground biomass across different sites with an average error of 20.3% followed by equations developed in the present study with an average error of 29.5%.