Conference,
Irrigation Modelling Language for decision support
(2009)cited By (since 1996) 1.
Abstract
Irrigation Decision Support Systems (DSS) have seen poor uptake in Australia for a number of reasons, one of which is the lack of their flexibility in allowing users to choose data sources they perceive as most relevant to them to generate the decision support they receive. The number of local and remote data sources available to DSS users for use in irrigation scheduling is increasing due to technological progress in both the agricultural engineering and information technology sectors. A standardisation of the outputs of data sources used by irrigation DSS and their internal processes would help to solve this. While there are already large international projects, such as Sensor Web Enablement (SWE) and WaterML, aimed at standardising sensor communication, internet data transfer and natural resource data sources, irrigation-specific standardisation work has not been undertaken. This paper makes a start on this by providing some background information on irrigation data sources and other standardisation approaches affecting the irrigation sector, defining requirements for irrigation standardisation and providing details of tests of standardised data source output. Surveys in the 2007/2008 and the 2008/2009 irrigation seasons listed data sources that irrigators currently use for their management decisions. A large range was found especially when non-biophysical factors, such as irrigators' personal calendars, were classified as data sources. Non-biophysical factors affecting irrigation management have long been known to cause a 'gap' between the support irrigation DSS can provide and industry practice, therefore this classification of non-biophysical factors as data sources is a first step towards codifying their effects so that future DSS may incorporate them and bridge this gap. Apart from the data source range, this survey work showed there is currently no consensus among irrigators as to the total set of useful data sources with irrigators changing data source use over time and accepting new data sources. This further highlights the need for data source flexibility if irrigation DSS are to remain relevant to irrigators. The many data sources used, or potentially used in irrigation decision making, are often heterogeneous in form (units, structure, range timestep etc) and therefore Informatics - the science of information use - needs to be considered when working to combine them. Sub disciplines of informatics include the study of data formats, standards and semantics as well as information fusion, comparison and presentation. Informatics work within these disciplines aims to allow heterogeneous data sources to be meaningfully accessed and used through systems such as DSS. Much standards work, such as SWE and WaterML, results from informatics. In order to allow irrigation DSS to offer users choice with respect to the data sources used through them and to allow that set of data sources to be expanded as technological progress generates new ones, not only must the technical standardisation of data source outputs occur but a conceptual informatics framework that describes the data required, and the techniques used, to generate irrigation decision support advice is also required. One way in which DSS may be able to help bridge the 'gap' between science and industry is to offer truly effective flexibility that will allow both biophysical and non-biophysical data sources to be used. We present summarised results from our surveys that indicate the range and type of current data sources used for irrigation decision making and give examples that show where other standardisation projects have catered for them and where they have not. We determine the position and scope of an Irrigation Modelling Language (IML) that can be used to further add to the other projects' standardisation efforts and fill gaps between them. After detailing these general requirements of an IML, we present a start to the formalisation of the conceptual model that is needed to underpin an IML by defining decision support processes and their requisite data sources. We show such formalisation with data mark-up allows DSS more data source choice and conclude that the wide adoption of such an IML framework would lead to more flexible irrigation DSS.
