Biosystems Engineering (2007) 96 (4), 471–476 doi:10.1016/j.biosystemseng.2007.01.003 IT—Information Technology and the Human Interface A Model-based Approach to maximise Gross Income by Selection of Banana Planting Date P. Tixier 1 ; M. Dorel 2 ; E. Male´zieux 3 1 CIRAD, UPR 26, PRAM—BP 214–97285 Lamentin Cedex 2—Martinique, French West Indies, France; e-mail of corresponding author: tixier@cirad.fr 2 CIRAD, UMR SYSTEM, 97130 Capesterre Belle Eau—Guadeloupe, French West Indies, France; e-mail: dorel@cirad.fr 3 CIRAD, UMR SYSTEM, AGRO.M, Baˆt. 27, 2 place Viala, 34060 Montpellier, France; e-mail: malezieux@cirad.fr (Received 5 April 2006; accepted in revised form 17 January 2007; published online 12 March 2007) The harvest dynamic model (SIMBA-POP) was used to choose the best planting date of a banana field in order to optimise the gross income. The proposed method takes into account the seasonal variation of the selling price of bananas. The gross income was calculated weekly by multiplying the simulated number of bunches harvested weekly by the mean weight of bunches, and by the current selling price. The weekly gross income over the simulated period was then determined. The study was carried out in Martinique (French West Indies) for climatic conditions corresponding to 50, 350, and 650 m in altitude and for different field management strategies (number of cropping cycles before replanting). The result of our simulations showed that the optimal planting month enables a 10–36% increase in income over the worst month. r 2007 IAgrE. All rights reserved Published by Elsevier Ltd 1. Introduction Export bananas (Musa spp., AAA group, cv. Ca- vendish) currently covers nearly one million hectares worldwide (FAO, 2005). Bananas are rhizomatous herbs whose terminal bud produces the inflorescence. Each plant successively produces a series of bunches, each from a lateral shoot. The sequence can be repeated for one to fifty generations or more, which means that it can be considered as perennial (Turner, 1994). The main developmental stages of banana plants include sucker appearance (which the farmer may influence by selecting the sucker early or late), growth, flowering, and harvest. Banana crops represent a collection of individual plants derived from vegetative propagules. They develop at their own rhythm and do not follow a synchronous cycle. When the stages of new sucker, flowering, and harvest occur, the population structure spreads follow- ing lognormal functions (Tixier et al., 2004). Hence, at any given time, a banana crop consists of a population of individual plants at various developmental stages. The dynamics of banana bunch harvest follows peaks whose amplitudes tend to be wider over time, up to a continuous harvest after five to seven cropping cycles (Fig. 1). The development of bananas and the harvest dynamics depend on the climate, in particular the temperature (at least, in the case of appropriated water and mineral supply). The date of planting strongly influences the dates of harvest peaks. Furthermore, the selling price of banana varies over time with a regular pattern; the highest prices are generally during January–April while the lowest prices are during May–November (SNM, 2005). This pattern depends mostly on other fruits grown in the importing countries (Loeillet, 2005). It is regular over years, only its amplitude varies. Figure 2 shows the mean selling price (import price) of banana for Martinique for each month during 1999–2004 (FAO, 2005; ODM, 2005). A selling price dataset with a ‘minimal variation’ over the year was defined in order to test the sensitivity of the model to variations of this parameter. This ‘minimal variation’ data set was established by choosing for each month the value that is the closest to the annual mean value of the selling price and within the monthly variation range. ARTICLE IN PRESS 1537-5110/$32.00 471 r 2007 IAgrE. All rights reserved Published by Elsevier Ltd