Forecasting weekly Canary tomato exports from annual surface data

Sea shipping is the main transport mode used by Canary farmers to export tomatoes to the European markets. Provincial associations of Canary growers negotiate charter fees with the shipping companies for the whole exporting period and, therefore, provide a unified sea transport service. When such a negotiation takes place each year, the individual growers’ decisions about planting surface are usually known. However, the forecasting of the distribution of tomato exports over the whole harvesting period would help Canary associations make more timely and effective decisions. In this paper, a model is proposed to forecast weekly Canary tomato exports conditioned on a given total planting surface. A seasonal model is formulated to deal with the weekly seasonal pattern of Canary tomato yields per hectare by means of evolving splines. Such a model is a useful tool to forecast weekly yields. From these forecasts, weekly tomato exports beyond the end of the sample are also forecast by taking the total planting surface into account. To illustrate the aptness of this framework, the proposed methodology is applied to a weekly series of tomatoes exported to the European markets from 1995/1996 to 2010/2011 harvests