Vincent de Leijster

127 Chronosequence analysis of economic performance of agroforestry coffee farms in Colombia 6 collinearity, we did not include correlated factors and chose the factor that was most strongly related to economic performance and we did not include multiple ‘factor variables’ of one ‘factor group’ in the same model. We ranked the models with different factor variable combinations using AIC, considering the best models those with a deltaAIC ≥2. We visually assessed whether the residuals of the final model met the assumptions of homoscedasticity and normal distribution by inspecting density plots and residuals vs. fitted values plots. The analyses were performed using the R packages ‘lme4’ and ‘stats’. 6.3 Results We found that 79% of the farms (112 out of 142) were profitable in terms of coffee net revenues, and this fraction increased to 82% (117 out of 142) when considering the actual net revenues. Average farm characteristics are presented in Table 6-2. Table 6-2. Descriptive statistics of farm and management characteristics, vegetation structure and coffee yield and price. The ‘*’ indicates whether the values only refer to agroforestry farms, in other cases all farms are included. Mean Minimum Maximum n Productive coffee area (ha) 7.5 ± 13 0.3 97 154 Elevation (m) 1580 ± 186 1110 2020 153 Coffee density (shrubs ha-1) 5140 ± 1290 1000 9090 153 Time since pruning of coffee shrubs (y) 6.8 ± 6.6 1.0 30 111 Tree density (tree ha-1)* 98 ± 85 7.5 655 118 Tree species richness (species farm-1)* 4.9 ± 2.6 1 13 118 Musa density (plants ha-1) 138 ± 164 0 940 143 Coffee yield (kg ha-1) 1640 ± 2340 73 26800 142 Coffee price (€ kg-1) 1.56 ± 0.15 1.09 2.63 151 6.3.1 Development of economic performance and costs and benefits We found that ‘time since transition to agroforestry’ did not significantly explain variation in ‘gross coffee revenues’, ‘net coffee revenues’ and ‘total net revenues’ (Figure 6-1). We found an overall decreasing trend for ‘coffee gross revenues’ for the entire time-series (psR 2 =0.0015, F=3.15, P=0.08; Figure 6-1 and Table 6-3), but not for the first 20 years (psR 2 =0.0013 F=2.6, P=0.11; Figure 6-1). More specifically, we found that coffee yield significantly decreased, and coffee price did not vary over time (appendix Figure A6-2).