MODELLING AN INTENSIVE BANANA CROPPING SYSTEM IN ECUADOR USING A BAYESIAN NETWORK
The quality of many watercourses around the world has been seriously affected by intensive agricultural practices. Bananas are the second most important fruit crop in the world and its export depends on intensive monoculture production systems in tropical conditions. Ecuador is the biggest banana export producer in the world but its banana production suffers from the leaf fungal disease called black Sigatoka. Black Sigatoka is controlled by intensive (once every two to three weeks) aerial fungicide treatments thereby contaminating rivers bordering the plantations. Several studies have modelled the fate of pesticides in soil and water compartments, but few included the relationship with the banana production system. The development of a suitable model is difficult due to the unsynchronized perennial behaviour of banana plantations. In this article the development and use of a Bayesian Network model for banana production management is presented using a river basin in Ecuador as a case. The Bayesian Network analysis includes a graphical model with key variables in the production system and conditional and marginal probability distributions derived from expert knowledge and field data analysis. The resulting model has been validated and is used to estimate the probability of maintaining high banana production rates under several management scenarios.
Nolivos , I., Van Biesen, L. and Swennen, R.L. (2011). MODELLING AN INTENSIVE BANANA CROPPING SYSTEM IN ECUADOR USING A BAYESIAN NETWORK. Acta Hortic. 919, 89-98
black Sigatoka, canonical models, causal-effect relationships, uncertainty assessment, Chaguana river basin