The performance of local generic allometries in estimating biomass carbon of trees in agriculturally dominated landscapes of Nyando basin, Kenya

Abstract

Climate change is a real phenomenon whose effects ought to be mitigated, based on assessments that are informed by scientific evidence. Trees in agricultural landscapes play a significant role in mitigating climate change hence a robust inventory of tree biomass in these landscapes is essential for the accurate estimation of their potential in sequestering carbon. Therefore, applicable allometric equations that yield accurate biomass estimates of trees in agricultural landscapes are essential. This study sought to develop and test the performance of local generic allometries (multi-species equations) for estimating biomass carbon. The study was undertaken in the Nyando Basin and it involved obtaining non-destructive measurements of trees (e.g., DBH, tree height, crown diameter) and destructive measurements (trunk, branch, leaf and root biomass). Regression analysis was then run on various predictive parameters (e.g., DBH, height and crown diameter) to derive allometric equations and to test their accuracy. The obtained biomass was used to determine the biomass carbon by an adequate conversion factor of 0.47.The performance of local generic allometries and globally recommended equations for tropical trees was tested on 33 destructively sampled trees from two sites in the Nyando Basin whose diameters at breast height ranged from 7.1 cm to 70.6 cm. Eighteen trees of several species were sampled on the site in the lower block of the basin. The site sampled in the upper block of the basin was a plantation with pure stand of Eucalyptus grandis. Local equations estimated aboveground biomass (AGB) with lower error compared to published global equations for tropical ecosystems, which overly estimated or underestimated AGB. The local equation by Kuyah et al. (2012a) conservatively estimated AGBwith a relative error (RE) of -2.7% while the site-generated equation had a RE of 5.5%. The local equationsalso underestimate biomass for the smaller trees while overestimating the biomass of the larger trees but within an acceptable range of error. Both the local and global equationsperformed with varying accuracies in estimating belowground biomass (BGB) and their resultant RE were quite high. The RE of Kuyah 2 equation and the site-generated equation (Equation 3) were 53.0% and 8.9% respectively while the RE for the global equations by Mokany et al. (2006) and Cairns et al. (1997) were 64.9%and 13.0%respectively. The study concludes that local generic allometries developed in a specific agricultural landscape perform better than global ones in estimating biomass in the sampled agricultural landscapes. The study therefore recommends the use of local allometries in biomass carbon assessment for agricultural landscapes.