Allometric Equations for Estimating Grevillea Robusta Biomass in Farming Landscapes of Maragua Sub-County

Abstract

Grevillea robusta (Silk Oak) is widely interplanted with food crops in Maragua to enhance tree Biomass on farms. a practice that enhances biomass content on farms. However, models for estimating total biomass of G. robusta are lacking. This study sought to develop allometric equations for estimating G. robusta tree biomass using easily measurable predictor variables of bole diameter and height hypothesized as Biomass does not vary among tree components in different Agroecological Zones (AEZ), Tree component biomass does not differ with trees sizes G. robusta biomass stocks does not vary among AEZs. A stratified systematic sampling on Geographical Information System (GIS) platform was used to subdivide each of the four AEZs, Upper Midland 1 (UM1), Upper Midland 2 (UM2), Upper Midland 3 (UM3) and Upper Midland 4 (UM4) into three equal polygons. At the centre of each polygon, a one hectare sample plot was established and Diameter at Breast Height (DBH) for all G. robusta trees measured. Thirty three sample trees were randomly selected for destructive biomass measurements. They were felled, stumps uprooted and tree divided into different components. Samples for each component were weighed for fresh weights and oven dried at 1050C (woody components) and 700C (foliage). Biomass data for all sample trees was used to develop allometric equations. Fresh/dry weight ratios were computed and used to derive total biomass for each of the tree components and for the whole tree. The above ground and below ground biomass was used to calculate root/shoot biomass ratio (R/S) while root length and tree height were used to calculate root depth/tree height ratios. The linear, exponential, logarithmic, power and polynomial functions were used to estimate biomass from DBH and height data. The best fit equation was selected based on the lowest Standard Error of the Estimate (SEE), lowest Mean Residual Error (MRE) and Coefficient of determination (R2). Of the fitted functions the polynomial equations had the highest R2, lowest SEE and lower MRE values. The equation to estimate Total Tree Biomass (TTB) = 0.322DBH2+7.934DBH-19.26 (R2=0.99), Above Ground Biomass (AGB) = 0.248DBH2+6.243DBH-15.45 (R2=0.98) and Bellow Ground Biomass (BGB) = 0.074DBH2+1.688DBH-3.791 (R2 0.98). Use of height/or product of height and DBH as predictors resulted in a decrease in R2 and high SEE values. T-test for (AGB, BGB, TTB) indicated no difference between predicted and actual biomass (T=0.54,P=0.601,T=1.714,P= 0.117 and T = 0.422 ,P = 0.68 respectively). Developed equations were also compared with other existing equations for validation. The best fit equation estimated TTB in the AEZs was 13.926 tonha-1,13.109 tonha-1,10.869 tonha-1 and 11.827 tonha-1 in UM1, UM2, UM3 and UM4 respectively, showing uniformity of stocking across the landscape (F=2.87,P=0.675). DBH was found to be a reliable predictor of biomass (AGB, BGB and TTB) in farming landscapes of Maragua Biomass allocation to different tree components does not differ in the 4 AEZ implying that one allometric equation can be used to estimate the biomass of a specific tree component in all the AEZ of the study area but tree Biomass varies with tree sizes.. The developed equations will be useful in estimating G. robusta tree/component biomass in the farms in support of marketing for energy, timber and other wood uses in the area.