An integrated agroforestry-bioenergy system for enhanced energy and food security in rural sub-Saharan Africa

Most people in rural sub-Saharan Africa lack access to electricity and rely on traditional, inefficient, and polluting cooking solutions that have adverse impacts on both human health and the environment. Here, we propose a novel integrated agroforestry-bioenergy system that combines sustainable biomass production in sequential agroforestry systems with biomass-based cleaner cooking solutions and rural electricity production in small-scale combined heat and power plants and estimate the biophysical system outcomes. Despite conservative assumptions, we demonstrate that on-farm biomass production can cover the household’s fuelwood demand for cooking and still generate a surplus of woody biomass for electricity production via gasification. Agroforestry and biochar soil amendments should increase agricultural productivity and food security. In addition to enhanced energy security, the proposed system should also contribute to improving cooking conditions and health, enhancing soil fertility and food security, climate change mitigation, gender equality, and rural poverty reduction. Supplementary Information The online version contains supplementary material available at 10.1007/s13280-024-02037-0.

The 40 selected trees were harvested in February 2020, i.e., 34 months after the woodlots were established.Harvested trees were divided into three fractions: Leaves, pods and twigs below 5 mm in diameter; branches (between 5 and 20 mm in diameter); and logs (above 20 mm in diameter).For each harvested tree, we recorded the total fresh mass of each fraction and collected two sub-samples from each fraction (n = 240 sub-samples in total) to determine the moisture content.The sub-samples were dried at 105 °C until no significant decrease in mass was detected, and the final dry mass was recorded.The moisture content for each sub-sample was calculated as a percentage of fresh mass and averaged per tree and fraction.Finally, the dry mass for each harvested tree and fraction was estimated based on its total fresh mass and mean moisture content.
Allometric models were developed to estimate the tree biomass for each fraction in the ten improved fallows based on the DBH (cm) and dry mass (kg) data from the 40 harvested trees.
Due to non-linearities in the data, we performed a log-log transformation of the DBH and dry mass data.We then fitted least-square linear regression models to the transformed datasets (ln (dry biomass) = β0 + β1 ln (DBH)), one model for each fraction, using the lm function within the stats package in R. The estimates of the regression coefficients, their associated standard error and p-value, and the R 2, for each model, are presented in Table S2.2.Plots of residuals were used to visually assess that model assumptions of normality, linearity, and homoscedasticity were not violated.We then used these regression models to predict the dry biomass of the different fractions for all individual trees in the 10 improved fallows based on their DBH.Model predictions computed on the logarithmic scale must be transformed to the original scale (kg).However, a simple exponential-based transformation generates bias (Finney 1941), and applying correction factors is often recommended to correct this bias (Clifford et al. 2013).Here, we used the MM correction factor (Shen and Zhu 2008), as recommended by Clifford et al. (2013) when predicting the biomass of new trees.The total predicted biomass per fraction for each of the ten improved fallows is shown in Table S2.3.

Appendix S3. Compiled data on biomass production in improved fallows
Table S3.1.Tree biomass production from selected species in improved fallows across sub-Saharan Africa and median, Q1, and Q3 values.Mass is expressed on a dry basis.The tree species, location, Köppen-Geiger climatic zone (Kottek et al. 2006), mean annual precipitation, dominant soil type (order) according to the USDA soil classification system (USDA-NRCS 1999), fallow period, tree density, and source are indicated .

A
tree inventory was conducted in December 2019 in the ten improved fallows.All trees above 10 mm Diameter at Breast Height (DBH) were labelled and measured for DBH using a caliper.Based on the DBH distribution (Figure S2.1), we established four DBH classes (10 -25 mm; 25 -50 mm; 50 -75 mm and >75 mm).For each class, 10 trees were randomly selected for harvest, totaling 40 trees.

Figure
Figure S2.1.Diameter at breast-heigh (DBH) distribution of Sesbania sesban trees in the ten study improved fallows in Kitalale and Siaya sites, western Kenya (fallow period of 34 months).The vertical lines indicate the DBH class thresholds.

Table S2 .
1. Location and area of the 10 Sesbania sesban improved fallows in western Kenya where we estimated tree biomass production.

Table S2
Table S2.3.Predicted total dry biomass (kg) per tree fraction for each improved fallow.
.2. Regression model summaries for the three models, one for each of the defined tree fractions(leaves,  twigs and pods; branches; and logs).