A new journal puts every step of modelling research on the record

Modelling has quietly become the engine room of modern agricultural, ecological and environmental science. Agent-based simulations track pollinator populations across fragmented landscapes, coupled food-system models inform national policy, and machine-learning pipelines forecast crop yields and environmental risk. Yet the people who build, calibrate and document these models have long watched their most consequential work vanish into a publishing system that was never designed for it. Formal model descriptions, calibration studies, implementation documentation and software workflows have rarely found a natural home in traditional journals, leaving the foundations of applied modelling uncited, unreviewed and unrewarded.

That gap is what a newly relaunched journal, Agricultural and Environmental Modelling, or AEM, published by Pensoft, exists to close. The title is the evolution of the Food and Ecological Systems Modelling Journal and its earliest predecessor, the Food Modelling Journal, and it has now broadened its scope, deepened its Open Science ambition and redesigned itself around the full lifecycle of modelling research. Its founding principle is stated bluntly: credit, where credit is due. Every step in the modelling lifecycle, the editors argue, deserves a citation.

The technical case for this approach rests on how modelling science is actually done. A complex model is not born in a single paper. It emerges through cycles of formal specification, software implementation, calibration, testing and application, often carried out by different teams across different project phases. Each of those steps represents genuine scientific work, and each deserves recognition. Contemporary research in environmental and social science highlights the concept of invisible labour, where the hidden efforts that underpin a modelling project’s success remain largely overlooked and go unrewarded in academia. In parallel, citation bias means researchers often use a model’s foundational influence but cite only their own prior implementations, so the foundational reasoning and assumptions defining a model’s parameters remain uncited.

The consequences are more than academic. As modelling research grows in complexity, the underlying assumptions and reasoning become black boxes that are rarely scrutinised by policy-makers and stakeholders, producing a lack of transparency across the modelling research lifecycle. When a coupled food-system model informs a regulatory decision, the provenance of its parameters matters enormously. AEM’s answer is a portfolio of unique article types that make each research object citable, discoverable and reusable: a Formal Model article provides citable recognition for foundational mathematical and conceptual design, while Model Testing and Calibration and Model Implementation and Documentation articles put the later stages of a model’s development on the peer-reviewed record. Standard Research Articles, Review Articles, Software Descriptions, Data Papers, Short Communications and a Modelling Methods type for standalone descriptions of smaller models complete the structured portfolio.

The journal’s technical standards are correspondingly strict. AEM commits to the FAIR principles, meaning Findability, Accessibility, Interoperability and Reusability, applied to diverse scholarly research outputs, and it goes beyond disseminating narrative and data to incorporate models and code into its peer-reviewed content. All data and source code underlying published work must be made openly available under an Open Source Initiative-approved licence. Computational workflows should be registered in WorkflowHub.eu, and machine-learning-based methods must comply with the DOME-ML guidelines for reproducibility and open benchmarking, which some journals have already successfully implemented in their publishing process.

Reproducibility requirements extend to the deployment of machine-learning models themselves. Models must be reproducible and presented in an easily deployable, containerised format, such as Code Ocean or a Jupyter Notebook, so that other researchers can rerun and verify published results rather than trust a narrative description. Mandatory requirements for publication also include the addition of provenance and versions of code, archived and cited according to the Software Citation Principles. In practice, this means a reader can trace exactly which version of which codebase produced which figure, a level of traceability that traditional modelling publications have almost never provided.

The relaunch arrives at a moment when academic publishing must evolve to keep pace with the integrity challenges presented by artificial intelligence. Pensoft addresses this by establishing transparency standards through its ARPHA Publishing Platform. On submission, authors are now required to declare any use of Large Language Models within manuscripts, to ensure accountability. In parallel, Pensoft is optimising the peer review process, with the ARPHA platform harnessing the power of AI to identify potential high-quality peer reviewers. The journal also inherits a proven record of innovation from its preceding titles, including an experimental living article publication and the publication of Formal Model articles that have already gained a reputation within scientific circles and across bibliographic platforms.

AEM is archived in four databases, namely CLOCKSS, Zenodo, Portico and Zendy, and indexed in over 40 services, including AGRICOLA, CABI, ChronosHub, CNKI, FAO AGRIS, ResearchGate and Swisscovery, amongst others. Its scope spans the full breadth of the domains it serves, from agriculture, covering management systems, production dynamics and agroecological modelling, to food systems, including food safety, production, quality, resilience and supply-chain control, ecological systems, the bioeconomy, social-ecological systems with coupled human-environment interactions and governance, and environmental science and natural resource management. Submissions may describe model development, methodological innovation, application of models to substantive research questions, validation and benchmarking studies, or reproducible computational workflows that advance understanding or decision-making in coupled human-environment systems.

The journal is written for and by the modelling community, and its primary audiences reflect that. Academic researchers developing or applying computational and mathematical models in the agricultural, socio-ecological, food and environmental sciences sit alongside regulatory and applied scientists using models in environmental risk assessment, food safety evaluation and ecosystem management. Software developers and data engineers who produce the tools and workflows on which modelling science depends are equally central, as are funding agencies and policy bodies requiring formal, peer-reviewed documentation of model foundations, and interdisciplinary teams working across the natural sciences, social sciences, computer science and agronomy. Because distributed teams across different project stages frequently produce modelling research, AEM enables appropriate attribution and citation of contributions from model developers, data providers, software engineers and analysts, and it integrates the Contributor Role Taxonomy, CRediT, in its submission interface and published content to recognise individual contributions within large collaborative projects.

The relaunch of the journal as Agricultural and Environmental Modelling marks both an evolution and a commitment. The editors retain the founding mission of FESMJ and FMJ to make modelling outputs publishable, findable and reusable, and they extend it, with a broadened scope to match where the science is going, article types that reflect how modelling research is actually done, and built-in standards for open code, open data and FAIR compliance that the field now demands. Submissions are invited from across all disciplines within the journal’s scope, and modellers are encouraged to contact the editors to begin a conversation about specific needs in promoting sharing, use and reuse. Editor applications are also welcome from experts around the globe who share the journal’s vision and values. Whether describing a formal mathematical framework for the first time, documenting the calibration of a long-standing model, sharing data or presenting a new computational workflow that others can build upon, modellers now have a place to put that work on the record, and credit is the founding principle of the enterprise.

Subject of Research: Agricultural and Environmental Modelling journal relaunch

Article Title: A journal by modellers, for modellers: Introducing Agricultural and Environmental Modelling

Article References: Topping, C. J., Nogoy, N., Boyadzhieva, I., Stoev, P., & Penev, L. (2026). A journal by modellers, for modellers: Introducing Agricultural and Environmental Modelling. Agricultural and Environmental Modelling, 8, Article e194160. https://doi.org/10.3897/aem.8.194160

Image Credits: AI Generated

DOI: 10.3897/aem.8.194160

Keywords: AEM, modelling, FAIR principles, open science, reproducibility, agriculture, environmental science, journal, modellers, Introducing, Agricultural, Environmental