The Journal of Electronic Publishing publishes research about contemporary scholarly publishing issues and practices. The argument we put forward in this article for the journal is that scholarly electronic publishing, or the publication process itself as currently practiced, generates climate injustice. This publication process is far from equitable as it is determined by a small number of mega-publishers and some mega-societies who exert significant control and influence over digital publishing infrastructures and access to knowledge (Chen et al. 2019). We suggest that this process is actually fundamentally unjust, particularly for the Global South. As Márton Demeter and Ronina Istratii argue, “Extensive research has demonstrated that the global academy is characterised by a significant Global North–Global South imbalance. Editorial boards, selection committees, funding agencies, publishing houses are predominantly located in Northern high-income societies” (2020, 506).

For the purpose of this article, we have identified three aspects of the digital scholarly publishing process of research articles that contribute most to both epistemic and climate injustice (as we will also outline more in-depth later). The first is fee-paying open access policies. The second aspect is deficiencies in “digital publishing infrastructures” to support open research. Especially problematic are the use of PDF-oriented system architectures and the associated lack of application of technically open methodologies, such as the FAIR principles or the 5-star deployment scheme for data, to all parts of a publication. Third is the slow progress toward ethical regulatory control of publishing, including open access and open science policy adoption and compliance, to ensure widespread access to knowledge. Our research in this article is concerned predominantly with the second aspect. In other words, we will be looking at “digital publishing infrastructures” and how these can be improved by the adoption of digital good practice and the alternative publishing model proposed in this article based on semantic publishing.

Similar to electronic publishing, open access publishing is increasingly controlled by mega-publishers (Butler et al. 2023; Sundell 2021). As Lin Zhang et al. state, “APC [author payment charge] fees collected by DOAJ [Directory of Open Access Journals] journals probably exceeded 1.27 billion US dollars in 2020” (2022, 7654). Here open access publishing is not done as a public good but to extract profit for publishing corporations. The costs of fee-based commercial OA models are huge for the Global South, rendering mainstream scholarly publishing fundamentally unjust (Raju and Badrudeen 2022). Diamond open access is recognized by the open access movement as a more egalitarian open access model, as it charges no APCs and is both “free-to-publish and free-to-read.” The UNESCO Recommendation on Open Science recommends member states support “not-for-profit, academic and scientific community-driven publishing models as a common good” and further encourages “multilingualism in the practice of science, in scientific publications and in academic communications” (UNESCO 2021).

Looking specifically at climate research, what these recommendations reflect is that the Global South currently is not an equal contributor to discussions on climate justice given these fee and language barriers (among others), and they have to accept the agenda and the publishing models, infrastructures, standards, and processes that the Global Northern mega-publishers create and control. This agenda is English-language centric (Chowdhary 2024), which means that many of the countries in the Global South who already suffer disproportionally from the effects of climate change miss out on having their voices heard, both in how knowledge is reported about the climate problems they face and in developing potential mitigations. We believe that this reflects a major aspect of climate injustice in the current publishing system.

This climate injustice is further exacerbated by the fact that the process of publication is not aimed at knowledge transfer but at revenue generation for publishing corporations. This is also reflected in the lack of innovation and investment that publishers have shown in relation to publication technologies, where the predominant technology used in scholarly publishing for the last 25 years has been the portable document format (PDF). PDF is an inefficient way of transmitting knowledge as it is used as a mono-directional form of information transfer, created by publishers and forced upon readers. Readers thus have little say in how they consume climate research, and it is extremely difficult to transform PDFs into any other format, such as a format that does more than only support reading with human eyes. Converting PDFs into semantic form—that is, making them searchable, translating them into other languages (those who cannot easily read PDFs in English face access barriers to scholarly climate knowledge), and adapting them for different types of readers—is almost impossible.

In general, PDFs are not structured for ease of reading. They do not take account of the fact that in many countries most research consumption is on mobile devices rather than on laptops or desktops (Sobral 2020). On mobile devices PDFs do not display well or at all, which becomes an issue when displaying, for example, large tables that need to be rotated by 90 degrees, as well as tables and figures that are split over more than one page, which again are extremely difficult to read on a mobile device. Modern software allows research to be consumed in a much more efficient manner than PDFs.

To provide some quick practical examples of the issues the PDF format poses, in the open research project #semanticClimate, we address a number of non-academic constituencies both in India and Europe, including high school students (often without English as their first language) and city planners who need climate research to shape regional climate plans (when this research is not available in the kind of form that they can easily consume and poses difficulties for people with limited eyesight, etc.) (Nganji 2015).¹ In relation to the first set of non-academic constituencies, at a visit to Indian schools for the India-UK BBSRC–funded TIGR2ESS program, a project #semanticClimate is involved in, our aim was to meet the young students and encourage them (especially girls) to undertake science education.² The goal of the visit was also to explore possibilities for rural students to undertake short summer internships in Delhi. The students were all engaged in STEAM, including in climate change learning programs, but IPCC reports had not been translated or were not used as part of these programs because they are mostly only available in PDF format (Riley, n.d.). For our second set of non-academic constituencies—city planners developing climate plans who want to use climate literature—their issue mainly concerns locating and sharing derivative data and publications. The PDF format prevents both of these operations. To provide an alternative, we contributed to two prototypes as part of our research: (1) City - Open Climate Reader (#semanticClimate 2023), which demonstrated how derivative publications of the IPCC report could be collated and republished, and (2) Semantic Glossar (Worthington et al. 2024), which allowed city municipal employees to collaborate on building their own climate glossaries (this prototype was developed as part of the TH Köln Co-Site project and used #semanticClimate extracted IPCC Glossary and technology).

The research we present in what follows seeks to illustrate the functionalities and possibilities of semantic publishing and is the result of a semi-automated literature search on the topic of “climate justice.” Our research asked the following questions: What does the open access scholarly corpus know about the topic of “climate justice,” and what is the shape of the discourse as it exists in this literature? In which papers and journals has the topic appeared, and, reaching back several decades, how often has it been mentioned? The research we present here also functions as a demonstration of the open-source tools that will be used for future work to create the Climate Knowledge Graph (ClimateKG), which aims to make the IPCC report globally accessible. We will first discuss the importance of the IPCC reports in the climate science discourse and then outline how the proposed new model allows for improved syntactical and semantic structuring and indexing of publications (including the IPCC reports), which offers various improvements to the Global South’s access to and participation in the climate science discourse. This article should be seen as an indicative scoping exercise and conceptual modeling to show what the #semanticClimate software tooling is capable of. The tooling includes the following modular operations as examples: named entity recognition (NÉR), corpus creation, text and data mining (TDM), and Wikibase import/export.

The Intergovernmental Panel on Climate Change (IPCC) is the United Nations body for assessing the science related to climate change. The IPCC has a mandate to “provide governments at all levels with scientific information that can be used in climate policies” (IPCC 1988) and thus is charged with assessing pathways for mitigating climate change (Hughes 2024). The pathways represent likely futures for the earth’s climate and the state of climate change as well as the effects of mitigating actions (IPCC, n.d.-b). The actions to achieve viable pathways, such as the 1.5°C emission pathways, are then taken into the United Nations Framework Convention on Climate Change (UNFCCC), a process in which UN member states agree on how to limit climate change.³ The conjoined agencies of the IPCC and UNFCCC operate under a model of Shared Socioeconomic Pathways (SSPs) (Wei et al. 2018), which inherently focuses on climate justice as a global social justice issue between rich and poor nations of the Global North and Global South. The approach of SSPs has its opponents in climate change denial (Jylhä and Hellmer 2020; Milfont et al. 2018). In 2017, the first Trump administration in the United States announced its departure from the Paris Agreement and the decision to take a different approach to climate change (Wikipedia 2024e), which is to abandon the idea of a shared future for humanity, as succinctly hypothesized in Bruno Latour’s 2018 book Down to Earth. The IPCC reports are thus much more than a report on climate research or the science of climatology and the current status of the planet’s climate. Rather, they are a knowledge and policy instrument that forms a vital organ for bringing climate justice into legal frameworks and justice systems (UNDP 2023).

While repeating the assessment cycle over the 32 years since the first Assessment Report was published, the IPCC has evolved its process (IPCC 1992). The open access book The IPCC and the Politics of Writing Climate Change by Hannah Hughes (2024) covers in detail the working of the IPCC as a knowledge endeavor of herculean scale that is dealing with a dynamic field of study and with how the interface of scientific knowledge and society creates political tensions. The study examines the IPCC’s editorial workings and sheds light on the IPCC knowledge creation process, providing a useful mapping of the steps involved in the report authoring, the functioning of the report cycles, and how decision-making has changed over time. ClimateKG will incorporate some sample workflow mapping of IPCC report chapters into the knowledge graph creation.

The IPCC as a scientific working group has had to change their authoring process (how they create content) in relationship with and response to climate justice questions. For example, in 1996, the Second Assessment Report, Working Group III (WG3), with an emphasis on economic thinking, lacked a protocol to ensure proper consultation with the UN Observer Organisations and other stakeholders (Hughes 2024). As a result, the following calculation to assess the “social costs” of climate change was submitted in its report: “suggesting a cash value of $1.5 million to a human life in the OECD against $150,000 in developing countries” (Pearce et al. 1996, 177). The chapter was not approved by the subsequent plenary session. The procedures and rules for subsequent assessment reports were then tightened and codified, such as by assigning chapter lead authors from developing countries (IPCC 2013).

The IPCC reports themselves can be seen as part of a climate justice system of sorts, and the mechanisms of its functioning have evolved over the years. There are also many other agreements and treaties that play significant roles in the process of moving science knowledge into policy and legal frameworks, a few of which merit highlighting here. The initial IPCC Assessment Report in 1992 triggered the UN member states to take action on climate change. The treaty body was formed (the UNFCCC), and a key task of this body was to establish a framework for how nations would organize themselves to respond to the production of greenhouse gases (GHG), the cause of climate change. This framework was the Kyoto Protocol (UNFCCC 1997, n.d.-b). In 2015, UN countries negotiated the Paris Agreement “to limit the temperature increase to 1.5°C above pre-industrial levels” (UNFCCC 2016). Ongoing treaty negotiations take place in the decision-making body called Conference of the Parties (COP) (UNFCCC, n.d.-a). These multilateral agreements are then enacted in member state legislation. An example COP treaties clause illustrates how the COP acknowledges climate justice: “Also acknowledging that climate change is a common concern of humankind, Parties should, when taking action to address climate change, respect, promote and consider their respective obligations on human rights, the right to health, the rights of indigenous peoples, local communities, migrants, children, persons with disabilities and people in vulnerable situations and the right to development, as well as gender equality, empowerment of women and intergenerational equity” (UNFCCC 2022, 2). The IPCC authors, including the chair Jim Skea, have further made a submission to the International Court of Justice (ICJ) on an “advisory opinion” tasked from the court by the UN General Assembly Resolution 77/276 on “Obligations of States in Respect of Climate Change” (ICJ 2024a). It is in these contexts that climate justice has been codified into law: “China submits that developed countries have an obligation to bear their historical responsibilities. IPCC reports reveal that historical emissions from developed countries are the primary cause of the current climate crisis and injustice” (ICJ 2024b).

As we put forward here, the IPCC reports, or the way in which they are made available through digital publishing, are influenced by two key factors: the academic publishing industry and governmental grey literature practices (GreyNet International 2024). First, IPCC’s digital publishing implicitly follows the norms dictated by the academic publishing industry. The scientific work, research, and data science carried out by the IPCC make up the most authoritative climate science available and are carried out to the highest standards. The data science work of the IPCC has been modernized and applies modern open science methods using FAIR data principles, and data are openly stored on GitHub (GO FAIR, n.d.; IPCC-WG1, n.d.). Second, the IPCC’s remit is to “provide governments at all levels with scientific information that can be used in climate policies” (IPCC, n.d.-a). For the IPCC to achieve this, it has added clauses to its Summary for Policymakers (SPM), which are IPCC Assessment Report (AR) summaries that have to be agreed on unanimously by all signatories, nation-states, and parties (AR6 included 195 members).

The IPCC in their Sixth Assessment Report has a specific emphasis on climate justice while addressing climate change. The different volumes of AR6 describe the importance of equitable climate action, integrating local and Indigenous knowledge with scientific approaches for effective climate adaptations and mitigations (IPCC 2023a, 2023b, 2023c).⁴ The IPCC reports have raised the concern of climate justice because of the disproportionate adverse impacts of climatic hazards on vulnerable populations who have contributed less to climate change. Climate vulnerability has also increased due to gender inequalities, so gender considerations are crucial for climate justice (IPCC 2023b, chap. 08). The different findings from the reports have shown the need for climate justice while making adaptation strategies. The Paris Agreement and the 2030 Agenda for Sustainable Development have widened the scope of adaptation governance by linking adaptation to development and climate justice (IPCC 2023b, chap. 18; UNFCCC 2016; United Nations 2015). Many environmental and climate justice activists have drawn attention to considerations of economic and environmental inequalities to increase awareness and also advocate for stronger climate mitigation efforts (WG3, chap. 05). The Civil Society and Social Movements section (IPCC 2023c, chap. 14) focuses on protecting rights, adopting responsibility-based approaches to climate finance, adaptation, and mitigation. IPCC Observer Organizations have also raised questions about the equitable allocations of future climate budgets to developed and developing countries. The Climate Change 2023: Synthesis Report (SYR) has mentioned the disproportionate risks from climate hazards such as heat waves, droughts, and flooding to poor and marginalized communities who are less responsible for this adverse climate change (Calvin et al. 2023). The SYR has urged the organization and government to design and structure adaptation and mitigation strategies that should consider equity and justice. Climate justice recommendations mentioned in the volumes of the IPCC report assist in the development of policies that will consider each and every section of society.

To further contribute to the climate justice endeavors proposed within the IPCC reports, we propose a new publishing model that could be applied to the publication process of the IPCC reports. This model is informed by the semantic web, hypermedia, and the work of earlier technological visionaries who built and promoted global access to knowledge through granular indexing and linking—from Tim Berners-Lee, Alan Kay, Ted Nelson, Vannevar Bush, J. C. R. Licklider, Donald Knuth, to Paul Otlet—a century before the semantic web, and reaching even further back to annotation and hermeneutics (Wikipedia 2024d; Kay 1972; Nelson 1974; Bush 1945; Knuth 1992; Licklider 1965; Wright 2014; Krajewski 2011). Even with such a long history, semantic publishing is still waiting in the wings and has not reached the mainstream practice of scholarly publishing. The publishing model we put forth therefore represents a new type of publishing that technically enables multilingualism, better global discoverability, more accurate scientific categorization and indexing, and reuse and republishing. In addition, the focus here is on the knowledge being available in academic repositories or corpora as a whole as opposed to prioritizing individual papers or journals. To test this model, we conducted a semi-automated literature review on the topic of climate justice in two corpora of climate scholarship (the Europe PubMed Central [PMC] repository and IPCC’s Assessment Round 6 reports) and asked the following questions: What does the scholarly corpus know about the topic, and what is the shape of the discourse as it exists in this literature?

The literature search on “climate justice” is an example use case of employing the open-source tools that we then used on a larger scale to create the Climate Knowledge Graph (ClimateKG)—and some of the specifics of knowledge graph capabilities (Worthington 2025). Knowledge graphs are a database that has nodes and properties, and the term “graph” is a mathematical term for “relation” (Wikipedia 2024b). ClimateKG is an online graph database of a literature corpus that can be searched, and the search results can be saved online and collated into a multi-format publication. ClimateKG could be used for frictionless dissemination of the IPCC reports and would include cataloging the literature corpora, word search, content publishing, and reuse.

We searched and retrieved papers and data computationally from the open access repository Europe PMC (2024) and its 6.5 million open access papers, as well as all 70 chapters of the IPCC Sixth Assessment Report held on GitHub as HTML with IDs (Calvin et al. 2023). The output of this search showed where the term “climate justice” (in which papers and journals) is being used and creates tables using jQuery DataTables software to browse the collected results that are stored in a mini-corpus.⁵ Data analysis was then carried out to find associated terms, which we used to create the Climate Justice Dictionary, a word list that includes all the associated terms from Europe PMC and the IPCC and annotations from Wikipedia, Wikidata, and Wiktionary stored as a linked open data document.⁶

Why are knowledge graphs important? And why are they important to improve access to climate justice literature? The scale of scholarly literature and its organization is too complex for a person, group, or institution to easily navigate: to drill down, filter, retrieve, collaboratively review, reuse, or even cite. The ClimateKG project’s primary mission is to make the IPCC reports accessible globally. As argued earlier, the IPCC reports are authoritative climate science and policy literature that regularly report on the status of climate change, including ways that the dangers from climate change can be addressed. Currently the IPCC reports are published as PDF or with unstructured web versions, which makes them virtually unusable in modern information systems such as the web, and as a consequence, their promulgation, reach, and efficacy are limited. The most recent IPCC AR6 report is a substantial corpus itself, and it is the sixth assessment report since the IPCC was created 36 years ago, in 1988. A knowledge graph effectively creates multiple layers of indexing on top of the untouched “version of record” publication. Such a layered approach was adopted in geospatial cartography the first decade of the 21st century by the OpenStreetMap (OSM) project (Wikipedia 2024c). This kind of indexing is technically called semantic annotation and uses ontologies, taxonomies, and controlled vocabularies. Concepts within the publication can be tagged, for example, “climatology”⁷ and “climate justice.”⁸ When tagged using Wikidata, a host of trusted source IDs and language translation are available: 97 language translations and 50 identifiers for “climatology” and 40 translations and 8 identifiers for “climate justice.” Many other semantic and syntactic questions can be added as indexed and linked layers to the knowledge graph: IPCC pathways frameworks, use of climate models, data and supporting software used, use of other syntheses papers and their data, and the list goes on.⁹

Currently this syntactical and semantic structuring and indexing has to be done retrospectively until the new publishing model described here and by many others is carried out at the time of authoring and publishing (Stocker et al. 2024; Capadisli 2019; Papers with Code, n.d.).

Knowledge graphs are part of the modern implementation of the Semantic Web, which enables automated knowledge management using data about definitions and relationships. In the context of the web, “semantic” refers to defining meaning in a machine-readable way. But the web itself has a design fault—it has no cataloging or indexing system. The web’s creator, Tim Berners-Lee, proposed a solution to this, which he coined the Semantic Web. But his intervention came too late, after the web had already developed in another direction. Since then, those interested in the opportunities that the Semantic Web has to offer have had to try and continue implementing his vision under difficult circumstances. The tech giants of web search and social media (Meta, Google, Microsoft, Meta, and Twitter [now X]) exploited the lack of a central cataloging or indexing system for the web and created much of their wealth by making their own private knowledge graphs. An example of the Semantic Web is Google Knowledge Graph (Wikipedia 2024a), which has 500 billion facts (in 2020) and generates the info boxes on search results.

Our work on the ClimateKG involves building an open knowledge graph for the open access parts of the IPCC Assessment Report AR6. The knowledge graph would be used for dissemination of the reports and would include various functionalities—from cataloging the literature corpora using ontologies, taxonomies, and controlled vocabularies to creating word search interfaces as used on academic repositories—and opportunities for content publishing and reuse with the ability to save the search results collated into new full text editions and output them as multi-format (e.g., web, ePUB, PDF).¹⁴

As outlined earlier, the IPCC reports are the gold standard of climate science, but they are rudimentarily published as PDFs while some are also published as basic web pages. IPCC reports are ripe for semantification as many pre-existing parts just need connecting—data, citations, author list, and glossaries, and so forth. The ClimateKG goal would be to have the AR6 available online with the added functionality afforded by semantic publishing. The created knowledge graph then allows for queries that can be collated and exported as multi-format publications.

The #semanticClimate community has been developing software tooling since 2019 for searching and semantifying climate literature. In 2023 and 2024, numerous hackathons and further developments have resulted in several forms of semantification of the IPCC Sixth Assessment Report, including text and data mining conversion to HTML of all 70 chapters,¹⁵ 2,000 climate terms aligned with Wikidata,¹⁶ and the extraction of a glossary of 600 terms.¹⁷ In 2025, #semanticClimate will be working with TIB and the National Institute of Plant Genome Research (NIPGR) to build the ClimateKG as a web service. Being a library web service means that there is an institutional guarantee that ensures that material is authoritative, that availability is maintained, and that there is support for the different public and user groups.

The ClimateKG will be a modern catalog and index of climate science literature with a subject-focus on UN Climate and UN IPCC corpora. It will be open source and based on open science principles. The ClimateKG technology allows for search, semantification (adding meaning), data analysis, and machine learning and AI functions of all types (scientifically validated, open science compliant, and ethically based in the SDGs, which is necessary for climate justice), as well as for republishing. Republishing here means supporting frictionless and authoritative reuse of content; as explained above, our first planned product will be a search function that collates the search result sections into a reader that can be accessed online and outputted as a multi-format publication. An example is “IPCC Reports and City Climate Change Plans: Proof of Concept Prototype—Open Climate Reader” (#semanticClimate 2023).

The knowledge graph creates a map of the literature included in the corpora as a data set that can be used for a granular search and can retrieve results down to a sentence level. It also assigns meaning to content by using a combination of “definitions” and “relationships.” Definitions use controlled vocabularies, such as, type of thing “Organization”¹⁸ (e.g., the IPCC). The IPCC has expertise in the following, so has the property of “knowsAbout,”¹⁹ “climatology,”²⁰ and “climate justice,”²¹ which reflects its relationships.

With only these few bits of data (the organization IPCC and its expertise), we can tap into Wikidata linked open data on “climatology” and retrieve 97 language translations and 50 identifiers and on “climate justice” 40 language translations and 8 identifiers. The identifiers mentioned here are authority control systems and are used to disambiguate, validate, or carry out queries on terms, such as Stanford Encyclopedia of Philosophy ID, OpenAlex ID, JSTOR topic ID (archived), and so on.