In this chapter we will discuss issues related to the development of a general UML model that covers a large class of similar models, the class of water-balance and irrigation-scheduling models [PSH04]. Many irrigation-scheduling and water-balance models have been developed and published in the past. These models have been used for both research purposes and as management tools. Models used for research purposes generally represent the system and underlying processes in greater detail than do management models. [AW85] distinguished between (i) mechanistic and functional, and (ii) rate and capacity models. Mechanistic models are based on fundamental processes, whereas functional models simplify the representation of processes. Rate models are driven by time and define rates of change within a system; capacity models are driven by input amounts and define amounts of change. However, even within these broad categories, models differ in their assumptions and representation of water-balance processes. [MLB98] make a detailed analysis of the assumptions and the representations of the waterbalance models. Water-balance models have been used as stand-alone applications and as components of larger agricultural-system models. For example, a water-balance model developed by Ritchie has been integrated into numerous simulation models, including the cotton simulation model OZCOT [Hea94], CERES-Wheat [R085], and is used by the Decision-Support System for Agrotechnology Transfer (DSSAT) [Rit98] which includes crop simulation models for a number of agronomic crops. Irrigation-scheduling models are generally standalone applications that have been designed as management decision-support tools.

Similar models may differ in their input data requirements and their use. [OEKOl] used THESEUS developed by [WegOO], a modeling system containing a number of sub-models, for water-balance and crop simulation, representing the soil, plant, and atmosphere, which can be combined to create simulation models. The system contains a number of water-balance models; users can select one that meets their complexity and data requirements. They summarized and distinguished between models according to their output, the equations used in the model and the input data requirements.

Although water-balance and irrigation-scheduling models may have been developed for different purposes and vary widely in their input requirements and representation of processes, they do share a number of commonalities. For example, they all typically require some soil and weather data. There is also often overlap in the processes represented, although they may be calculated by different methods. For example, most models include water removal by évapotranspiration. This may be calculated by the model: A historical value or an input requirement. The process of water movement is simulated by these types of models either as amounts moving into the soil profile and stored within it, or by rates of change in soil water content. In order to identify common elements and relationships, a number of waterbalance and irrigation-scheduling models were compared.

In the case of soil water-balance and irrigation-scheduling models, the common system elements are the soil, plant, and weather. The behavior of these elements is model-specific and is defined by the processes accounted for by the model.

0 0

Post a comment