Allometric relationships between stem diameter, height and crown area of associated trees of cocoa agroforests of Ghana

Allometric models which are used to describe the structure of trees in agroforestry systems are usually extrapolated from models developed for trees in forest ecosystems. This makes quantitative assessment of the functions of shade trees in agroforestry systems challenging since increased availability of light and space in these systems may induce structural differences from those growing under forest conditions. We addressed this issue by providing species-specific allometric information on the structural characteristics of associated shade trees on cocoa agroforestry systems and assessed if allometries conformed to theoretical predictions. At the plot level, stand and soil characteristics affecting tree structural characteristics were assessed. The study was conducted in cocoa agroforestry systems at Suhum, Ghana. The height-diameter at breast height (H-DBH) allometry had the best fits (R2 = 53–89%), followed by the crown area (CA)-DBH allometry (R2 = 27–87%) and then the CA-H allometry (R2 = 22–73%). In general, the scaling exponents of the CA-DBH, H-CA and H-DBH allometries conformed to the metabolic scaling theory (MST). However, both the CA-DBH and H-DBH allometries diverged from the geometric similarity model. Though forest tree species had similar crown areas as fruit trees, they were slenderer than fruit trees. Tree slenderness coefficients were positively correlated with soil P, Ca, Cu and the ratios (Ca + Mg):K, (Ca + Mg):(K + Na) and Ca:Mg, but not C:N while DBH and H were correlated with soil P and C:N ratio. Our results show that critical soil nutrients and their ratios affects shade tree structural attributes (e.g. slenderness and CA), which possibly restrict variations in species-specific allometries to a narrow range on cocoa systems. Furthermore, shade tree species richness and density are better predictors of relative canopy projection area (a proxy for shade intensity) than tree species diversity. In conclusion, the results have implications for shade tree species selection, monitoring of woody biomass and maintenance of biodiversity.


Supplementary Information 1: Soil data collection, processing and chemical analysis methods
The soil samples were collected from two depths, namely 0-15cm and 15-30cm, for the purpose of conducting a chemical analysis and soil bulk density determination.A 5cm-bladed Eijkelkamp soil auger was utilized to extract five soil samples from each depth in each plot.
The auger was marked at the corresponding depths, and manually driven into the soil to extract the samples.The collected samples from each layer in each plot were combined, thoroughly mixed, and subsampled for further analysis.The soil subsamples were dried in an oven at 105°C for 48 hours, sieved through a 2mm mesh, and ground in an agate ball mill at 290rpm for 15 minutes prior to chemical analysis.To determine soil bulk density, two soil samples per depth per plot were collected using 139cm 3 bulk density cylinders.These samples were obtained from the soil wall after digging.All samples were collected on the same day to ensure consistency.
Soil water content (%WC) was determined using the formula WC= [(M1 -M2)/M2] × 100, where M1 represents the fresh weight of soil and M2 is the weight of soil that was oven-dried at 105°C for 48 hours.Soil bulk density was estimated as BD (g cm -3 ) = [(W1 -W2)/V] × (100 -%CF)/100, where BD represents soil bulk density, W1 and W2 are the weights of empty trays and oven-dried soils in trays, respectively, and V represents the volume of the bulk density cylinder.The coarse soil fraction (CF) was used to adjust for the effect of stone fragments on bulk density.Soil particle size distribution was determined using laser ablation (Bechman Coulter LS 200) and classified into textural classes based on the USDA soil triangle (Soil Survey Division Staff, 1993).
Soil pH was measured in a 1:2.5 soil:solution slurry using a pH meter (pH 209, Hanna Instruments), which was calibrated with pH 4.01 and 7.00 buffer solutions.Electrical conductivity (EC, mS cm -1 ) was also measured in the same soil slurry using a portable electrical conductivity meter (Combo pH and EC, Hanna Instruments), which was calibrated with 1413 μS cm -1 standard solution.To determine Olsen extractable P (hereafter P), soil samples (2 g) were extracted with 30 ml of 0.5 M sodium bicarbonate, which was thoroughly mixed with 0.05% w/v polyacrylamide, shaken, and centrifuged at 3500 rpm for 15 min.Phosphorus in the extract was estimated at 880 nm in a 1 cm cell using a spectrophotometer and the blue phospho-molybdate method with ascorbic acid as a reducing agent.
Total soil organic carbon (SOC) and total nitrogen (N) were analyzed using a CN analyzer [Thermo Scientific™ Flash™ 2000 Organic Elemental Analyzer (OEA)].Extractable Ca, Mg, K, and Na, as well as the concentrations of Mn, Al, Cu, and Zn, were determined using ICP-MS (Thermo Scientific™ iCAP™ TQ) after extracting 2 g of each soil sample with 20 ml of 1 M NH4NO3, centrifuging at 3500 rpm for 30 min, filtering, and diluting 1 ml of the supernatant with 9 ml of 2% HNO3.Effective cation exchange capacity (ECEC) was determined as the sum of exchangeable bases and exchangeable acidity.Nutrient stocks (Mg or kg ha -1 ) were estimated as the product of nutrient concentration, bulk density, depth, and unit conversion factor.The total nutrient stocks for the 0-30 cm depth were used for the correlation analysis involving soil nutrients.