Tying It All Together
Learning Objectives
After completing this module, you will be able to:
- Identify the three primary “products” that come out of synthesis groups.
- Understand the metadata and other features that make published datasets useful.
- Evaluate the reach and reproducibility of an ecological synthesis project’s outputs.
- Create a plan for your synthesis team’s research products that applies contribution, publishing, and citation practices that will benefit the team.
Introduction
So far, we’ve made the point that ecological synthesis research is collaborative and inclusive, and that it integrates a wide range of data. Synthesis research is also intended to be influential and useful. There are many definitions of “influential and useful” to consider here, but successful synthesis research tends to expand the boundaries of knowledge and aims to improve human lives or the environment. 
The ability to accomplish this in synthesis research frequently depends on what knowledge or products are created, and how the synthesis team disseminates and communicates them to the outside world.
There are three interconnected, publishable products that are the most common outputs from a synthesis project (or potentially any research project, really): the data, analytical workflows (code for data cleaning or statistics, for example), and research results. Each of these elements is a valuable product of synthesis science, and each one should reference the others. In this module we’ll discuss the mechanics of publishing each one, and then how they can be connected and made accessible for the long-term.
Designing and Publishing Datasets
Estimated time: 12 min
In Module 2 we discussed some considerations for creating and formatting harmonized data files useful for synthesis research. We also introduced the importance of metadata for describing data and making it more usable. Publishing harmonized data files and descriptive metadata together as a dataset ensures that the data products produced by a synthesis team are findable, accessible, interoperable, and reusable (FAIR). FAIR data are an important output for almost any ecological synthesis project.
The FAIR principles, standing for Findability, Accessibility, Interoperability, and Reusability, are a community-standard set of guidelines for evaluating the quality and utility of published research data. Making an effort to meet the FAIR criteria promotes both human and machine usability of data, and is a worthy objective when preparing to publish data from a synthesis research project.
The FAIR principles were first defined in the paper by Wilkinson et al (2018)1. Since this time, many resources have arisen to guide the implementation the FAIR principles23, and to quantify FAIR data successes and failures in the research and publishing communities45.
Activity 1: Evaluate published datasets
Lets start our journey to publishing datasets by looking at some that are already published. Form breakout groups and course instructors will assign each group a dataset (a DOI) for evaluation. With your group, answer these questions about the dataset:
- Where were the data collected?
- What variables were measured and in what units?
- What is the origin of the data and how have they been altered since collection?
- Were the first three questions easy to answer? Why or why not?
Example dataset: Jarzyna, M.A., K.E. Norman, J.M. LaMontagne, M.R. Helmus, D. Li, S.M. Parker, M. Perez Rocha, S. Record, E.R. Sokol, P. Zarnetske, and T.D. Surasinghe. 2021. temporalNEON: Repository containing raw and cleaned-up organismal data from the National Ecological Observatory Network (NEON) useful for evaluating the links between change in biodiversity and ecosystem stability ver 1. Environmental Data Initiative. https://doi.org/10.6073/pasta/7f0e0598132e3fea1bfd36a4257af643.
Example dataset: Craine, Joseph M. et al. (2019). Data from: Isotopic evidence for oligotrophication of terrestrial ecosystems [Dataset]. Dryad. https://doi.org/10.5061/dryad.v2k2607
Example dataset: Wieder, W.R., D. Pierson, S.R. Earl, K. Lajtha, S. Baer, F. Ballantyne, A.A. Berhe, S. Billings, L.M. Brigham, S.S. Chacon, J. Fraterrigo, S.D. Frey, K. Georgiou, M. de Graaff, A.S. Grandy, M.D. Hartman, S.E. Hobbie, C. Johnson, J. Kaye, E. Snowman, M.E. Litvak, M.C. Mack, A. Malhotra, J.A.M. Moore, K. Nadelhoffer, C. Rasmussen, W.L. Silver, B.N. Sulman, X. Walker, and S. Weintraub. 2020. SOils DAta Harmonization database (SoDaH): an open-source synthesis of soil data from research networks ver 1. Environmental Data Initiative. https://doi.org/10.6073/pasta/9733f6b6d2ffd12bf126dc36a763e0b4
Example dataset: Woods, B., Trebilco, R., Walters, A., Hindell, M., Duhamel, G., Flores, H., Moteki, M., Pruvost, P., Reiss, C., Saunders, R., Sutton, C., & Van de Putte, A. (2021). Myctobase (1.1) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.6131579
Example dataset: Ross, C.W., L. Prihodko, J.Y. Anchang, S.S. Kumar, W. Ji, and N.P. Hanan. 2018. Global Hydrologic Soil Groups (HYSOGs250m) for Curve Number-Based Runoff Modeling. ORNL DAAC, Oak Ridge, Tennessee, USA. https://doi.org/10.3334/ORNLDAAC/1566 (2018)
Metadata
One thing that Activity 1 introduces is the importance of metadata. Metadata are data about the data. As a general rule, metadata should describe
- Who collected the data
- What was observed or measured
- When the data were collected
- Where the data were collected
- How the data were collected (methods, instruments, etc.)
- Sometimes, stating why the data were collected can help future users understand data context evaluate fitness for use.
Including metadata of this nature makes data more usable, and helps prevent the deterioration of information about data over time, as illustrated in the figure below (from Michener et al. 19976).
Data Provenance Metadata
Provenance metadata deserves special attention for ecological data synthesis projects. Data provenance refers to information detailing the origin of the values in a dataset, which is particularly important for synthesis projects that bring together data from many different sources. Synthesis activities typically produce new data products that are derived from the original source data after they have been cleaned, harmonized, and analyzed. Provenance metadata should be included with the derived products to point back to the original source data, similar to the way bibliographic references point to the source material for a book or scholarly article.
A few other notes on provenance:
- At its simplest, documenting data sources as you collect and analyze the source data is a great start on provenance metadata.
- Many data repositories provide guidelines, tools, and features for data provenance metadata7.
- Provenance metadata can become very detailed if the software and computing environment is also taken into account. This is an active area of study 89.
Licensing
Published datasets should include a license in every copy of the metadata that defines who has what rights to use, reproduce, or distribute the data. Licensing decisions should be made in consultation with the synthesis team after considering the nature of the data (does it contain human subject data, for instance?), its origin (including restrictions on source data, if applicable), and the requirements of the funders and institutions associated with the project. For publicly-funded environmental research data, it is generally appropriate to use open licenses, and the Creative Commons CC-BY attribution, and CC0 public domain, licenses are probably a good choice for most ecological synthesis data. This is not legal advice and your mileage may vary.
Metadata Creation and Management
Assembling metadata should be an integral part of the data synthesis activities discussed in Module 2, and can even be built-in to the workflow and project management practices of a project. Make sure to plan for and start creating metadata early in a synthesis project. Below are a few ways to do that.
- Keep a detailed project log and populate it with metadata for the project, including information like:
- what source data the team is using and where they came from.
- how data are being analyzed and methods used to create derived products.
- who is doing what.
- Start creating distinct publishable datasets (data plus metadata) as data are processed and analyzed. The team can do this:
- locally, using a labeled directory for the cleaned, harmonized, of derived data, along with related code and metadata files. Metadata files may be plain text, or use a metadata template.
- with a repository-based metadata editor, such as ezEML from the Environmental Data Initiative (EDI) repository.
- Get a professional data manager or data curator involved with the synthesis project. For example, the LTER Network has a community of “Information Managers” 10 trained in data management, metadata creation, and data publishing. Research data repositories11 and associated data curators12 may also be a good resource.
Reproducibility and the creation of metadata are closely related. Your team’s detailed documentation of the research process allows for reproducible science, and can be mined as a source of metadata during data publication.
Deciding What to Publish
The overall design of the dataset to be published is often difficult to imagine, particularly for people new to using or creating datasets. One of the most common questions data managers hear is “What should we publish?” This is usually a question about what files to include in the published dataset, or what data will be useful as a published dataset.
What should be included in a published dataset?
As we learned in Activity 1, every dataset is different, but the answer to “What should we publish?” usually comes down to:
- Publish any data used to generate research results.
- Publish any data that will be used by others (scientists, managers, public stakeholders), including raw data.
- If reproducibility is of interest or concern, publish the workflow.
- Usually this means publishing code, such as scripts written in R, python, or a shell language.
- What code? Any scripts used to process or analyze the data, or to generate research results like figures, are fair game.
- Sometimes code, especially detailed, reusable workflows like an R package, can stand alone as an independent publication. We discuss that in a later section.
And of course… always publish descriptive metadata about any of the above.
These are general rules, but you can also look at advice from a repositories like EDI and BCO-DMO, or from a research network like NEON. Asking a data manager, especially one involved with the synthesis group’s work, can also be helpful, as will discussion among the full synthesis team.
Choosing and Publishing to a Repository
There are many, many research data repositories available to researchers now13, making the choice of where to publish data fairly challenging. A few basic data repository features are essential when publishing a synthesis dataset. First, the repository should issue persistent, internet-resolveable, unique identifiers for every dataset published. Generally this will be a Digital Object Identifier, or DOI, that can be cited every time the dataset is used after publication. Second, repositories should require, and provide the means to create/publish, metadata describing each dataset. Without requiring at least minimal metadata, no repository can ensure that published data are FAIR. Finally, research data repositories should be stable and well supported so that data remain available and usable in perpetuity. Choosing a repository from the CoreTrustSeal certified repository list is one way to assess this. Beyond this, asking a few questions about the dataset will help with repository selection:
- Who are the likely users for this data? Will they belong to a specific scientific discipline, research network, or community of stakeholders?
- How specialized are your data? Do they fall into a common data type or follow a special formatting standard?
- Will the data be updated regularly?
- Does the repository charge for publication?
- Will the dataset benefit from some level of peer review?
More specialized repositories tend to offer enhanced documentation, custom software tools, and data curation staff that will review submitted data and assist users with data publication. Selecting a data repository with metadata requirements or standards, and a review and curation process for submissions, will help ensure that you are publishing a more FAIR data product. Consulting a project data manager if one is available to the synthesis team will also help with repository selection. After making a choice, the process of publishing data varies from repository to repository.
Additional Data Publishing Resources
Communicating Research Results
One of the primary goals of synthesis research is to find useful, generalizable research results about the system under study. Most often this means writing scientific journal articles. While we aren’t going to go into full detail about what constitutes, or how to write, a manuscript for a journal, there are some unique features of writing articles for synthesis projects. First, data papers are often an important product for synthesis groups, and these are somewhat different than standard research journal articles. Second, given, the large size and cooperative nature of most synthesis teams, a collaborative writing process is called for. An appropriate collaborative writing method, and some team norms and contribution guidelines, should be in place to reduce the potential for conflict or mistakes.
Data papers
A data descriptor article, usually known as a data paper, is a peer-reviewed journal article written to introduce and describe a (usually) new dataset. For synthesis teams, who are often producing a harmonized dataset as their first major research product, writing a data paper to accompany the dataset makes sense as a way to introduce the data, demonstrate their utility, and get the word out about the dataset. Data papers also lay the groundwork for any future papers that will answer the science questions of interest to the synthesis team.
Data papers may be simpler and shorter than research articles (not always though), but there are still a few gotchas that can arise. Below are some recommendations, and the rationale behind them.
- Publish the dataset described by the data paper in a reputable data repository.
- Although some data paper publishers host data themselves, they are usually published only as supplementary material for article, or are only held for review. Most data-focused journals require that accepted data papers should describe and reference a dataset published in a research data repository. Follow the guidance above to select a repository and prepare the dataset for publication.
- Be sure to cite the data paper and the dataset properly.
- The existence of a data paper and a dataset, each describing the same data and each with its own DOI, can create confusion about what to cite in related works. If the novelty and utility of the dataset, or the methods used to assemble it, are being referenced by a related work, then it may be most appropriate to cite the data paper. If the actual data are being used (analyzed, interpreted, etc.) in a related work, then cite the published dataset. In many cases it is expected to cite both.
- Don’t shortchange the metadata in the published dataset just because there is also a data paper.
- Consider the future usability of the data the data paper describes, and ensure that the associated published dataset contains detailed, community-standard metadata. Not all users will see the data paper, and data paper publishers may have incomplete or quirky requirements for metadata.
Some examples of data papers related to synthesis projects:
- Komatsu, Kimberly J., et al. “CoRRE Trait Data: A dataset of 17 categorical and continuous traits for 4079 grassland species worldwide.” Scientific Data 11.1 (2024): 795. https://doi.org/10.1038/s41597-024-03637-x
- …
A few suggested venues for publishing data papers:
- Scientific Data (Nature Publishing Group)
- Data (MDPI)
- PLOS ONE (usually termed “database papers”)
- The ESA journal Ecology, and quite a few other disciplinary journals, now publish data papers.
GBIF also maintains a helpful list of data paper journals.
Writing collaboratively
Writing a paper with a large team can be a challenge. It is important to encourage team members to contribute in a way they are comfortable with, but there is the potential for technical, editorial, and personal conflict without some prior planning. Practically, there are two models for writing a manuscript with a bunch of contributors.
In this model manuscripts live mainly in web-based writing platforms managed by a cloud service provider (e.g. Google Docs) and all contributors write and edit the document within that platform. Contributions may be asynchronous or synchronous since version control and conflict resolution is generally built into the platform. Most platforms have additional collaboration features, such as user account management, suggested edits, and commenting systems.
Software platform: Google Docs, Microsoft 365 Online, Overleaf (LaTeX)
Pros: Strong collaboration features (user/permission management, contribution tracking, comments and suggestions). No need to distribute copies and then merge contributions.
Cons: Can be unfamiliar to senior contributors. Easy to lose track of links. Limited formatting features compared to local word processors. Privacy/tracking concerns.
This model relies on word processing software installed on contributors’ local machines. Copies of the manuscript are distributed to contributors for asynchronous writing and editing assignments, and contributions are then merged together into a synchronized version of the manuscript. In large teams, it may be best to have one person managing the copy/merge process.
Software platform: Microsoft Word (usually), email
Pros: Familiar to most. Integrates with local data management practices. Most word processors have powerful collaboration and versioning features now. Advanced formatting and editing. Less reliance on cloud providers.
Cons: License pricing and institutional availability may be limited. Multiple versions in use, and the copy/merge workflow can easily generate conflicts or become unmanageable in large groups.
In addition to these practical considerations, there are some team considerations as well
- Make the expectations for contributing to a manuscript clear.
- How, when, and where should contributions be made
- Authorship expectations discussed in advance
- Make space for new, or early-career team members to contribute.
- Efficiency and experience level aren’t good reasons to exclude contributors
- Synthesis papers are a great learning experience and career opportunity
- Team discussions are preferable to unilateral editorial decisions.
- This can help avoid hurt feelings during the editing process.
- It can be beneficial to have a manuscript coordinator.
- The coordinator can help split up writing and editing tasks equitably
- Someone needs to manage conflicts, check for consistency, etc.
- Often this is the lead author
Connecting the Pieces
We’ve now covered how a synthesis team should approach creating and publishing its main research outputs (data, code, results). Now we’ll discuss how to begin making these useful to the world, which starts with making sure the products of synthesis research point to each other. Lets begin with an activity.
Activity 2: Synthesis project detective
Estimated time: 12 min
Form breakout groups and course instructors will assign each one a link to a product from a synthesis project (the code, a paper, a dataset, etc.). Using any means necessary (metadata, web search, etc.) figure out what other products are related (other publications, source/derived data, etc.) and who is involved in the synthesis team. Answer these questions as a group:
- If your group received a link to a paper, were you able to find datasets and a code repository (for an analytical workflow)?
- If your group received a link to a code repository, were you able to find papers and datasets?
- If your group received a link to a dataset, were you able to find papers and a code repository?
- Who was involved in the synthesis project?
- Could you understand the overall scope and impact of the synthesis project? Why or why not?
More ways to synthesize
We’ve talked about the three most common products of synthesis: papers, datasets, and workflows. But, we’ve also seen that there are plenty of other ways to share synthesis research! Education and outreach can become an important goal in for some synthesis teams, and providing access to data and actionable research results, such as forecasts, can be very useful to stakeholders. As time goes, on synthesis teams may produce many things that meet these goals and needs, moving well beyond the three kinds of products we’ve already talked about. See below for a few ideas and examples.
Synthesis research produces new scientific knowledge that other researchers, students, or stakeholders can learn and build on. Synthesis can also generate applied-science tools and methods that others need to learn how to use for themselves. Teaching modules are an important way of sharing both of these outcomes, and of broadening the reach of a synthesis project.
Examples:
- The EDDIE project is a clearinghouse of contributed teaching materials for the earth and environmental sciences.
- This website is an example of teaching materials produced by a synthesis team.
Interactive web applications can provide users with easy access to scientific datasets, especially large ones, analytical results, visualizations, interpretation, and many, many other things. Creating web apps is not necessarily an easy task, but if your synthesis team has the expertise, or access to web developers, web apps may be useful for outreach, or as tools the synthesis team itself can use. Frameworks like Shiny (for R), Streamlit, or Flask (both for python), and services like Shinyapps.io and Plotly, can make creation of apps relatively painless.
Examples
- An app for finding and exploring ecocomDP data in the NEON and EDI repositories.
- A dashboard app for the NEON ecological forecasting challenge.
- The Jornada LTER interactive viewer for weather station data.
Some research efforts have developed automation systems for research data processing, analytics, and publishing. These often fall into the “continuous integration/continuous deployment” class of web-enabled software and data pipelines, in which one software processes (data processing, analytics, publication, etc.) may be automatically triggered by events that occur in another, connected software service (such as adding new data to a GitHub repository). These technologies enable researchers to build software pipelines that can be useful for quality control of new data, updating forecasts, and rapid deployment of data or analysis products.
Examples:
- The Portal Project in southeast Arizona has developed a well-described near-term ecological forecasting pipeline.17
- Automated quality control of dendrometer band data.18
- Forecasting Lake and Reservoir Ecosystems (FLARE) project.
At a certain point, the outputs of a synthesis project can become numerous and challenging to present to the public in an organized way. Project websites can serve as a gateway to an entire synthesis project by providing comprehensive listings of project outputs (papers, datasets, GitHub repositories, etc), a narrative for the research, appealing images or graphics for outreach, and links to related projects, funders, or institutions. GitHub Pages sites are a common solution for creating simple, cost-effective (free, usually) project websites nowadays, but there are other options. A good project website can become a cohesive, engaging clearinghouse for information about a synthesis project, but they can be laborious to create and keep up-to-date.
Examples:
Linking synthesis products together
Reflecting on all the information above, we can see one common feature of the many different products of a synthesis team: they exist primarily as digital objects on the internet. The internet may seem fluid, but fortunately there are ways to identify and connect these digital objects in a stable way.
Persistent identifiers
Persistent identifiers, or PIDs, are references to digital objects that are intended to last a long time. For objects on the internet, they are intended to be unique, i.e. having a 1:1 relationship between the PID and the digital object, and machine actionable, meaning they can be understood by software like web browsers. There are many different types of PIDs, but the most useful ones in the context of publishing research products are:
- Digital Object Identifiers (DOI), used to identify digital publications like journal articles, datasets, or governement reports.
- Open Researcher and Contributor ID (ORCID), used to identify individuals, usually in the context of research or publishing activities.
- Research Organization Registry (ROR), used to identify organizations, also in the context of research and publishing, primarily.
These identifiers can and should be associated with all journal articles and published datasets resulting from synthesis projects. DOIs and ORCIDs can easily be associated with GitHub and other code repositories as well.
Citing synthesis products
The best way to ensure that use of a research product is recognized is through proper citation. This is already common practice for journal articles, but is only recently being adopted for published datasets. The most logical place in an article to cite a published dataset is in the Methods section and in the Data Availability Statement, which most reputable journals now require. Be sure to check journal data sharing requirements well in advance so that data publication preparation can begin early enough. When citing datasets, be sure that the full bibliographic entry is correctly included in the article’s References list. Citation of code is not as widely practiced, but some journals require it and it is a best practice.
From Currier and Sala 202219. Note that source datasets are properly cited in the Data Availability Statement, meaning an in-text citation is given and the full bibliographic entry is provided in the article reference list (not shown). The DOIs included here are helpful for quickly finding the data.
All original and derived phenology data produced by the authors, and R scripts for data processing, statistical analyses, and figure production are publicly available in the Environmental Data Initiative (EDI) repository. EDI package knb-lter-jrn.210574001.2 (Currier & Sala, 2022a) contains daily phenocam image data, derived timeseries and associated scripts for processing and is available at https://doi.org/10.6073/pasta/836360dce9311130383c9672e836d640. EDI package knb-lter-jrn.210574002.2 (Currier & Sala, 2022b) contains observed phenological indicators and environmental drivers as well as associated scripts for final analyses and figure construction presented in this manuscript and these data are available at https://doi.org/10.6073/pasta/d327a77f6474131db8aa589011e29c29. No novel code was generated by the authors of this manuscript. The precipitation data used in all analyses are derived from G-BASN data in EDI package knb-lter-jrn.210520001 (Yao et al., 2020) available at https://doi.org/10.6073/pasta/cf3c45e5480551453f1f9041d664a28f. Daily air temperature summaries from 4 June 1914 to the present for the Jornada Experimental Range Headquarters (NOAA station GHCND:USC00294426) are freely available upon request via the National Ocean and Atmospheric Administration (https://www.ncdc.noaa.gov/cdo-web/datasets/GHCND/stations/GHCND:USC00294426/detail).
Data used in the figures are included in the supplementary material. The full dataset will be provided upon reasonable request to the corresponding author.
Maintaining Momentum
As we discussed in Module 1, starting a synthesis project benefits from motivating scientific questions, a well-planned foundation for team science, and significant activation energy from the team. When successful, synthesis projects gather enough momentum to be productive for many years. Below are a few ideas on how to maintain this momentum.
Give everyone credit
Everyone deserves credit for the work they do, and in academic environments this is too often overlooked. Synthesis working groups commonly begin without any dedicated personnel support, which means that some participants, usually early-career scientists, will be contributing unpaid time to the project. In the absence of pay, leaders of a synthesis team should take the initiative to make sure everyone receives appropriate credit and opportunities for career advancement when they contribute to the project. Below are a few thoughts on how to do that.
- Discuss and define in advance some of the contributions team members will make.
- This is particularly important for deciding authorship of journal articles.
- The CRediT framework is a good starting point.
- More detail on this is in Module 1.
- Be willing to credit participants for a wide variety of contributions.
- This includes writing code, cleaning data, taking meeting notes, and more.
- Make sure all contributors have an ORCID. They are easy to obtain and widely used.
- Use ORCIDs to associate contributors with a research product whenever possible.
- List contributors on websites, GitHub repositories, and other public-facing team materials.
- Its nice to include affiliations, bios, links to profile pages, and other information too.
- Don’t rely on any one metric for valuing contributions to the team.
- Code commits in GitHub, for example, may reflect the input of many people besides the one that actually wrote and committed the code.
- Don’t forget students, technicians, early-career scientists, and others.
- Don’t forget to put your name on your work!
What are we missing here?
Encourage new contributions
Interests and commitment to synthesis projects change over time. To sustain active research contributions by the team, and continued use of the data, make sure new people can find a way to participate.
- Provide a path for new data contributions.
- This follows from making the data preparation/harmonization workflow reproducible.
- Have open meetings when possible.
- This helps bring in new team members that are interested and willing to contribute.
- Give all team members the freedom and support to lead analyses, papers, and other valuable project activities.
Find support
Maintaining momentum for a synthesis project over the long term is highly dependent on the ability to keep scientists engaged and find support for dedicated personnel time. Usually this means getting monetary support in the form of grants.
- Explore and apply to the funding sources presented in Module 1.
- Personnel support may need to come from larger grants since working group funding often provides only meeting support.
- Think creatively about how to get students and postdocs participating in synthesis projects.
- If student/postdoc research interests & plans overlap, dedicating some time to synthesis group work can lead to career-building opportunities (networking, high-impact papers).
- Promote the synthesis team’s work!
- It is difficult to attract interest from new participants and new resources for a project without doing this.
HAVE FUN!
When done correctly, ecological synthesis research means having lots of fun doing science with a great team.
Footnotes
Wilkinson, M., Dumontier, M., Aalbersberg, I. et al. The FAIR Guiding Principles for scientific data management and stewardship. Sci Data 3, 160018 (2016). https://doi.org/10.1038/sdata.2016.18↩︎
Bahim, C., Casorrán-Amilburu, C., Dekkers, M., Herczog, E., Loozen, N., Repanas, K., Russell, K. and Stall, S. (2020) ‘The FAIR Data Maturity Model: An Approach to Harmonise FAIR Assessments’, Data Science Journal, 19(1), p. 41. Available at: https://doi.org/10.5334/dsj-2020-041.↩︎
Gries, Corinna, et al. “The environmental data Initiative: Connecting the past to the future through data reuse.” Ecology and Evolution 13.1 (2023): e9592. https://doi.org/10.1002/ece3.9592↩︎
Michener, W.K., Brunt, J.W., Helly, J.J., Kirchner, T.B. and Stafford, S.G. (1997), NONGEOSPATIAL METADATA FOR THE ECOLOGICAL SCIENCES. Ecological Applications, 7: 330-342. https://doi.org/10.1890/1051-0761(1997)007[0330:NMFTES]2.0.CO;2↩︎
Lerner, et al., “Making Provenance Work for You”, The R Journal, 2023. https://journal.r-project.org/articles/RJ-2023-003/↩︎
White EP, Yenni GM, Taylor SD, et al. Developing an automated iterative near-term forecasting system for an ecological study. Methods Ecol Evol. 2019; 10: 332–344. https://doi.org/10.1111/2041-210X.13104↩︎
O’Brien, Margaret, et al. “ecocomDP: a flexible data design pattern for ecological community survey data.” Ecological Informatics 64 (2021): 101374.↩︎
Kim, A. Y., Herrmann, V., Barreto, R., Calkins, B., Gonzalez-Akre, E., Johnson, D. J., Jordan, J. A., Magee, L., McGregor, I. R., Montero, N., Novak, K., Rogers, T., Shue, J., & Anderson-Teixeira, K. J. (2022). Implementing GitHub Actions continuous integration to reduce error rates in ecological data collection. Methods in Ecology and Evolution, 13, 2572–2585. https://doi.org/10.1111/2041-210X.13982↩︎
Currier, Courtney M., and Osvaldo E. Sala. 2022. “ Precipitation versus Temperature as Phenology Controls in Drylands.” Ecology 103(11): e3793. https://doi.org/10.1002/ecy.3793↩︎