Fluctuations in perennial crop yields are a major source of instability for farmer livelihoods. Theory suggests fluctuations may be linked to trade-offs in resource allocation to growth and reproduction. Resource budget models (RBMs) have been developed to explain temporal patterns of resource allocation, though few applications of the model exist. Farm management practices are also likely to interact with parameters of intrinsic variability to ultimately determine agroecosystem stability. Despite the social, ecological and agricultural importance of reproductive variability, we know little about the underlying mechanisms nor how feedbacks with farm management practices affect tradeoffs. We use coffee (Coffea arabica) as a system and RBMs as a framework to characterize growth-reproduction trade-offs and the role of resource availability. On a farm in Costa Rica’s Central Valley, high and low fertilization treatments (n=40 plants each) were established in the early stages of fruit development. Two thinning treatments were imposed on a subset of plants where 75% or 100% of the initial fruit load was removed (n=5 plants per thinning-fertilizer combination). Measurements were taken at fruit initiation (June) and fruit maturity (November-January) to quantify the within-year effects of fertilization and initial fruit load on growth-reproduction tradeoffs.
Results/Conclusions
We observed a strong negative relationship between shoot growth and initial fruit load across all treatments. Thinned plants displayed greater shooter extension and fruit retention than unthinned plants. Plants under high fertilization tended to elongate their shoots more than those under low fertilization, but the trend was not statistically significant. These results provide evidence for a within-year growth-reproduction trade-off and the potential for resource availability to mitigate the trade-off, but not eliminate it. Our findings are consistent with predictions of the RBM that fruit filling is resource-limited. Surprisingly, there were also significantly more leaves on the branches of thinned plants when sampled in December than on branches of unthinned plants. Again, there was no effect of fertilization on the number of leaves. Further research will address the potential feedback between leaf retention and future resource gain for reproduction, an important parameter of the RBM, and consequences for growth-reproduction trade-offs. We discuss implications for long-term yield stability in coffee agroecosystems and the development of a coffee-specific resource allocation model.