Towards a new generation of agricultural system data, models and knowledge products: Information and communication technology

Sander J.c. Janssen, Cheryl H. Porter, Andrew D. Moore, Ioannis N. Athanasiadis, Ian Foster, James W. Jones, John M. AntleAgricultural Systems 2016 · OpenReview

TL;DR

This paper argues for a paradigm shift in agricultural systems modeling, moving from fragmented, researcher-centric models to a collaborative, ICT-driven ecosystem. It proposes leveraging modern information and communication technologies (ICT) to create a "NextGen" of agricultural models, data, and decision support tools that are more accessible, interoperable, and useful for a wider range of stakeholders, ultimately aiming to improve global food security and sustainability.

What Problem Does This Paper Address?

Agricultural modeling, while advancing scientifically, has lagged in practical application. Current models are often developed in isolation, leading to redundancy, poor integration, and difficulty in re-use. This fragmentation hinders the ability to generate real-world solutions for the agricultural sector. The authors identify a critical gap: the disconnect between sophisticated modeling capabilities and the needs of end-users like farmers, extension officers, and policymakers. They argue that existing ICT infrastructure and practices are insufficient to bridge this gap, leading to a lack of accessible, actionable knowledge for decision-making in agriculture.

Methods

The paper proposes a "NextGen" approach to agricultural modeling, driven by advancements in Information and Communication Technology (ICT). The core methodology involves:

  • Use Case Analysis: Five distinct use cases representing diverse stakeholders (farm extension officers, researchers, investment managers, precision agriculture advisors, supply chain consultants) were analyzed to identify specific needs and deficiencies in current systems.
  • Knowledge Chain and Application Chain Frameworks: These conceptual frameworks are used to structure the discussion, illustrating how data is transformed into wisdom and how models are integrated into a broader infrastructure for delivering decision support.
  • ICT Technology Assessment: The authors examine various ICT developments, including big data, crowdsourcing, remote sensing, semantic web technologies, linked data, cloud computing, and mobile technologies, assessing their maturity and potential to benefit agricultural modeling.
  • Qualitative Evaluation: The paper provides a qualitative assessment of how these technologies can address the identified problems, focusing on interoperability, modularity, data integration, and user-centric design.

Datasets

Not explicitly discussed in the paper as datasets for training or evaluation. The paper focuses on the types of data needed and the challenges in accessing and integrating them, such as:

  • Global soil maps
  • Historical weather data
  • Experimental data
  • Social network data
  • Citizen observatory data
  • Satellite data
  • Statistics
  • Private data from industries and farmers

The paper emphasizes the need for standardized data protocols and the use of Linked Open Data principles to facilitate data sharing and integration.

Key Insights and Ideas

  • The "NextGen" Vision: The paper advocates for a fundamental shift from isolated, researcher-focused models to a collaborative, open, and user-centric ecosystem.
  • Bridging the Application Gap: The primary challenge is not just developing better models, but making them accessible and actionable for a wide range of users beyond academia.
  • ICT as the Enabler: Modern ICT, from big data analytics to mobile interfaces, offers the tools to overcome fragmentation and create a more integrated and useful agricultural knowledge system.
  • Data Interoperability is Crucial: The ability to seamlessly integrate data from diverse sources (sensors, satellites, social media, traditional statistics) is paramount for advanced modeling and decision support.
  • Modular and Granular Models: Moving towards component-based, reusable model modules is essential for flexibility, maintainability, and easier integration into larger systems.
  • User-Centric Design is Non-Negotiable: Future agricultural modeling applications must be designed with the end-user's needs and context at the forefront, moving away from a "push" model of technology.
  • Openness and Collaboration: Fostering an open-source ethos for data and software, coupled with collaborative development practices, is key to accelerating innovation and adoption.

Key Results

The paper does not present quantitative results from a specific model or experiment. Instead, its key findings are qualitative and conceptual:

  • Deficiencies Identified: Current agricultural modeling systems are characterized by fragmentation, poor data accessibility, lack of interoperability, and a disconnect from end-user needs.
  • Technology Maturity: Several ICT technologies (e.g., mobile, cloud, big data) are mature enough to be leveraged for agricultural systems modeling, but their application in this domain is nascent.
  • Essential Requirements for NextGen: The paper identifies critical areas for development: interoperability of data sources, modular and granular models, standardized reference data sets, and robust user requirements analysis methodologies.
  • Benefits of Adoption: Adopting these ICT advancements and best practices will enable higher analytical capacities, integrated use of new data sources, and ultimately, more effective decision support for sustainable agriculture.

Limitations and Future Work

  • Limitations:
    • The paper acknowledges that the proposed solutions are conceptual and require significant effort to implement.
  • The qualitative evaluation of technologies is dependent on the specific use cases and draft application chains presented.
  • The maturity of some technologies, like the Semantic Web and Linked Data, is still developing, and their practical application in agricultural modeling is yet to be fully proven.
  • Challenges remain in achieving true interdisciplinary modeling and overcoming technical barriers between different modeling frameworks.
  • Future Work:
    • The authors call for a research agenda focused on addressing interoperability, modularity, reference data, and user requirements.
  • They emphasize the need for widespread adoption of best practices in coding, data management, and collaborative development.
  • Further work is needed to develop user-centered design methodologies for agricultural models and to foster a collaborative community that can drive the development and adoption of NextGen applications.
  • The paper suggests that a core group of adopters will be vital to invest in the medium-to-long term to achieve the critical mass needed for benefits to materialize.

Summary updated

BibTeX
@inproceedings{janssen2016towards,
  author = {Sander J.C. Janssen and Cheryl H. Porter and Andrew D. Moore and Ioannis N. Athanasiadis and Ian Foster and James W. Jones and John M. Antle},
  title = {{Towards a new generation of agricultural system data, models and knowledge products: Information and communication technology}},
  booktitle = {Agricultural Systems},
  year = {2016},
  url = {https://openreview.net/forum?id=YKUXntttBh},
}
Figure