Search

Crop models are frequently used in agronomy for simulating crop variables at a discrete time step. This paper describes the application of an artificial neural network in developing a model for yield forecasts in durum wheat, using back-propagation algorithms based on the mechanistic model AFRCWHEAT2 (Porter, 1993; Porter et al., 1993). Given the relevant number of inputs (16) required to operate AFRCWHEAT2, we have tried to develop a simpler model based on a neural network, in order to match AFRCWHEAT2 performance while substantially reducing the need of inputs.

Crop models are frequently used in agronomy for simulating crop variables at a discrete time step. This paper describes the application of an artificial neural network in developing a model for yield forecasts in durum wheat, using back-propagation algorithms based on the mechanistic model AFRCWHEAT2 (Porter, 1993; Porter et al., 1993). Given the relevant number of inputs (16) required to operate AFRCWHEAT2, we have tried to develop a simpler model based on a neural network, in order to match AFRCWHEAT2 performance while substantially reducing the need of inputs.

Although many Decision Support Systems (DSSs) have been developed for crop management, DSSs have contributed little to practical agriculture because of the so-called ‘problem of implementation’; underutilisation has been ascribed to both technical limitations of the DSSs and to farmer attitude towards DSSs. A new DSS, named vite.net , was developed for sustainable management of vineyards and is intended for the vineyard manager (the person who makes decisions about the vineyard management or suggests the proper actions to the grape-grower). The DSS has two main parts: (i) an integrated system for real-time monitoring of the vineyard components (air, soil, plants, pests, and diseases) and (ii) a webbased tool that analyses these data by using advanced modelling techniques and then provides up-todate information for managing the vineyard in the form of alerts and decision supports. The information is tailored to a vineyard, or part of a vineyard, or a number of vineyards that are uniformly managed throughout the season. In the design and development of vite.net , the implementation problem was specifically addressed by: (i) focusing on the important vineyard problems with a holistic approach (the DSS incorporates overall management solutions for growers); (ii) using automation and integration in data collection, and supporting flexible input efforts by the user; (iii) developing and validating fit-topurpose, mechanistic, dynamic models; (iv) designing a user-friendly interface and providing complete and easy-to-understand information; (v) delivering the DSS through the Web and thereby enabling both continuous updating by the provider and flexible access by the user; (vi) designing the DSS with the goal of assisting the decision maker (by providing necessary information) rather than replacing the decision maker; (vii) involving potential users during vite.net development and testing so as to obtain insight into how users make decisions; (viii) communicating the benefit