Agricultural modeling has long suffered from fragmentation in model implementation. Many models are developed, there is much redundancy, models are often poorly coupled, model component re-use is rare, and it is frequently difficult to apply models to generate real solutions for the agricultural sector. To improve this situation, we argue that an open, self-sustained, and committed community is required to co-develop agricultural models and associated data and tools as a common resource. Such a community can benefit from recent developments in information and communications technology (ICT). We examine how such developments can be leveraged to design and implement the next generation of data, models, and decision support tools for agricultural production systems. Our objective is to assess relevant technologies for their maturity, expected development, and potential to benefit the agricultural modeling community. The technologies considered encompass methods for collaborative development and for involving stakeholders and users in development in a transdisciplinary manner.

Summary: Enabling data reuse and knowledge discovery is increasingly critical in modern science, and requires an effort towards standardising data publication practices. This is particularly challenging in the plant phenotyping domain, due to its complexity and heterogeneity. We have produced the MIAPPE 1.1 release, which enhances the existing MIAPPE standard in coverage, to support perennial plants, in structure, through an explicit data model, and in clarity, through definitions and examples. We evaluated MIAPPE 1.1 by using it to express several heterogeneous phenotyping experiments in a range of different formats, to demonstrate its applicability and the interoperability between the various implementations. Furthermore, the extended coverage is demonstrated by the fact that one of the datasets could not have been described under MIAPPE 1.0. MIAPPE 1.1 marks a major step towards enabling plant phenotyping data reusability, thanks to its extended coverage, and especially the formalisation of its data model, which facilitates its implementation in different formats. Community feedback has been critical to this development, and will be a key part of ensuring adoption of the standard.

The global effort toward sustainable development has initiated a transition in water management. Water utility companies use water-pricing policies as an instrument for controlling residential water demand. To support policy makers in their decisions, the authors have developed DAWN, a hybrid model for evaluating water-pricing policies. DAWN integrates an agent-based social model for the consumer with conventional econometric models and simulates the residential water demand-supply chain, enabling the evaluation of different scenarios for policy making. An agent community is assigned to behave as water consumers, while econometric and social models are incorporated into them for estimating water consumption. DAWN’s main advantage is that it supports social interaction between consumers, through an influence diffusion mechanism, implemented via inter-agent communication. Parameters affecting water consumption and associated with consumers’ social behavior can be simulated with DAWN. Real-world results of DAWN’s application for the evaluation of five water-pricing policies in Thessaloniki, Greece, are presented.

Machine learning (ML) expands traditional data analysis and presents a range of opportunities in ecosystem service (ES) research, offering rapid processing of 'big data' and enabling significant advances in data description and predictive modelling. Descriptive ML techniques group data with little or no prior domain specific assumptions; they can generate hypotheses and automatically sort data prior to other analyses. Predictive ML techniques allow for the predictive modelling of highly non-linear systems where casual mechanisms are poorly understood, as is often the case for ES. We conducted a review to explore how ML is used in ES research and to identify and quantify trends in the different ML approaches that are used. We reviewed 308 peer-reviewed publications and identified that ES studies implemented machine learning techniques in data description (64%; n = 308) and predictive modelling (44%), with some papers containing both categories. Classification and Regression Trees were the most popular techniques (60%), but unsupervised learning techniques were also used for descriptive tasks such as clustering to group or split data without prior assumptions (19%). Whilst there are examples of ES publications that apply ML with rigour, many studies do not have robust or repeatable methods. Some studies fail to report model settings (43%) or software used (28%), and many studies do not report carrying out any form of model hyperparameter tuning (67%) or test model generalisability (59%). Whilst studies use ML to analyse very large and complex datasets, ES research is generally not taking full advantage of the capacity of ML to model big data (1138 medium number of data points; 13 median quantity of variables). There is great further opportunity to utilise ML in ES research, to make better use of big data and to develop detailed modelling of spatial-temporal dynamics that meet stakeholder demands.