Environmental Data Science: a new open access venue for the transformative potential of AI and data science in addressing environmental challenges

It’s my pleasure to announce the launch of Environmental Data Science, a new peer-reviewed, open access journal dedicated to the potential of artificial intelligence and data science to enhance our understanding of the environment and to address climate change. 

As of 10 December 2020 – coinciding with both NeurIPS 2020 and the AGU Fall Meeting – the journal is open for submissions. 

As Editor-in-Chief, I am delighted to be joined by nine Editors who have been working at this interface of data science and environmental challenges for some time. Supported by our publisher Cambridge University Press – a not-for-profit department of the University of Cambridge – we plan to develop Environmental Data Science into the definitive venue for the use of data-driven methods to tackle environmental problems.

I have been working for over a decade on the use of machine learning to shed light on climate change. I co-founded the International Conference on Climate Informatics, which turned 10 years old this year. Climate Informatics was inspired in part by the success of Bioinformatics in harnessing data science to accelerate scientific discovery. Through its annual conferences, Climate Informatics has built an interdisciplinary community committed to realising the potential of new data science methods and tools for leveraging insights from the massive amounts of climate data, both observed and simulated through physics-driven climate models. 

Environmental Data Science aims to build upon the momentum created by Climate Informatics, along with other workshops, conferences, and institutes that have emerged in this space, in order to serve all researchers using data science approaches for environmental problems. Often their work is simultaneously viewed as too applied for computer science journals and too methods-focused for disciplinary journals in their particular domain. 

We have deliberately defined both the data and environmental aspects of Environmental Data Science broadly, in order to reflect the fact that we are only at the start of realising the potential of data-driven approaches to environmental challenges.

Major data and methodological themes of the new journal include:

• Artificial intelligence (AI) and data science algorithms and architectures to make sense of large, complex and heterogeneous data, such as machine learning, computer vision, data mining, image analysis, contemporary statistical modelling and econometrics. These methods have greater potential to learn from data, assess uncertainty and improve estimates and predictions than many conventional models in the literature.
• Spatiotemporal data: Beyond time-series analysis, AI and data science have provided new methods to deal with fields and streams of spatiotemporal data with complex, non-linear and non-stationary relationships over space and time.
• Reasoning under uncertainty is a central consideration for these new approaches. They should seek to provide applications that help stakeholders make difficult decisions, for example in relation to climate change mitigation and adaptation.
• Data assimilation techniques allow AI and data science methods to learn from observational data, which can be noisy, sparse, and incomplete.
• Data collection, reproducibility, and computing infrastructure have improved, including remote sensing and other data-driven technologies, as well as citizen science and collaborative approaches. Cloud environments enable data storage and processing and can host services providing data to different scientific communities. However, these advances also raise new questions for research methodology

Environmental themes will encompass the following, and more: 

• Water cycle, atmospheric science, (including air quality, climatology, meteorology, atmospheric chemistry / physics, paleoclimatology)
• Climate change (including carbon cycle, transportation, energy, and policy)
• Sustainability and renewable energy (the interaction between human processes and ecosystems, including resource management, land use, agriculture and food)
• Biosphere (including ecology, hydrology, oceanography, glaciology, soil science)
• Societal impacts (including forecasting, mitigation, and adaptation, for environmental extremes and hazards)
• Environmental Policy and Economics

Why submit to EDS?

In Environmental Data Science, we not only plan to create a trusted, peer-reviewed venue overseen by Editors with a track record in this emerging area, but also to use Open Access to promote wider interest in and understanding of the potential of data-driven approaches for the environment. All articles will be published under Creative Commons licensing, enabling them to be re-distributed and re-used with proper attribution.  

We will publish application papers that tackle problems in an environmental field, enabled by data science, as well as methods papers that use novel data science approaches demonstrated in one or more environmental applications. We will also publish review papers that provide a thorough overview of relevant fields and subfields, or that survey the uptake of data science within existing environmental disciplines. We are also interested in receiving data papers that describe, in a structured way, important and re-usable environmental data sets that reside in publicly accessible repositories, and position papers and perspectives that provide expert opinions on issues related to environmental data science. 

We recognise that this is a burgeoning area, and in addition to Climate Informatics there are other workshops and conferences that explore the interface between data science and environmental science – and welcome contact from groups interested in partnering with the journal. We’ve outlined a way to submit proposals for special collections of articles arising from relevant events.  

Open Access is just the tip of a deeper Open Research movement with which we will engage in a nuanced way. In our Transparency and Openness Promotion policy, we encourage authors to make data and code related to their research openly available, but recognise that this is not always possible because of the methods involved or the nature of the data. We know that computational notebooks, such as Jupyter Notebooks, are increasingly important in this and other fields. In terms of peer review, we will launch with a standard single-blind policy, but we have enabled post-publication commenting and, as the journal develops, we will consider more open forms of peer review when this is feasible and appropriate.

So why submit to Environmental Data Science? To recap, here are five key reasons:  

1. Gain quality peer review feedback on your work from editors and reviewers who have expertise in the use of data science in environmental disciplines. 
2. Publish, if accepted, your work under open license to make it freely available to read, distribute and re-use in a venue that also provides you with options for making pre-prints, data and code openly available. 
3. Reach a wider audience through impact statements published with articles, conveying the significance of your work.
4. Align your conference or workshop with Environmental Data Science, as a venue that can make peer-reviewed outputs open and discoverable.
5. Help us build a community of authors, reviewers and editors advocating for the transformative potential of data science for a better understanding of the environment. 

You can read our Instructions for Authors here, follow us @envdatascience on Twitter, and sign up here to receive alerts about our first publications. We also encourage you to contact us on eds@cambridge.org with your ideas.