Climate Change TimeLine: An Ontology to Tell the Story so Far

In this paper we present an ontological approach to build a knowledge-base on climate change related facts. Our knowledge space, developed in Ontology Web Language (OWL), enables knowledge within the Semantic Web. It allows a dynamic classification and a semantic characterization of the different events that can be related to each other, as well as to external sources of information. The resulting ontology is structured as a timeline which aims to describe the climate change story from multiple perspectives, including scientific, social, political and technological ones. We have created an instance of the ontology as an attempt to tell the climate change story so far. Such a population is based on the collection of factual information and critical literature review, with a focus on relevant theories, happenings, social and political initiatives. In this paper we present an ontological approach to build a knowledge-base on climate change related facts. Our knowledge space, developed in Ontology Web Language (OWL), enables knowledge within the Semantic Web. It allows a dynamic classification and a semantic characterization of the different events that can be related to each other, as well as to external sources of information. The resulting ontology is structured as a timeline which aims to describe the climate change story from multiple perspectives, including scientific, social, political and technological ones. We have created an instance of the ontology as an attempt to tell the climate change story so far. Such a population is based on the collection of factual information and critical literature review, with a focus on relevant theories, happenings, social and political initiatives.


I. INTRODUCTION
Climate change is a threat to all lives on Earth, yet very little is being done to address the issue at hand. Despite anthropogenic climate change being a fact from a scientific perspective, the science behind climate change is still disputed [1].
In simplistic terms, the concept ''climate change'' indicates, according to Climate Change in Australia, a prevailing change in climate statistics such as temperatures, greenhouse gases and sea levels [2]. This can occur due to natural phenomena, as well as due to human activities [3].
More recently, climate change has often been associated with changes in the amounts of greenhouse gases due to human activities that have resulted in a steady increase in temperatures [4]. This rise in temperatures has set off a series of ecosystem alterations resulting, for instance, in the melting of icebergs in the Arctic and Antarctic, rising sea levels, an increase in storms and calamities, and the destruction of The associate editor coordinating the review of this manuscript and approving it for publication was Malik Jahan Khan . natural habitats leading to wildlife extinctions [5]. Anthropogenic climate change is on the verge of disrupting our lives as we know it [6].
Climate change emergency is considered to be the no.1 threat to humanity at the moment. However, the actions taken in response are considered far away from the needed to assure sustainability. Indeed, the science of climate change itself is often disputed, and there are people who disbelieve its validity [7]. Establishing policies to address climate change implies the need for global agreements, as well as a strong commitment by local governments. This, evidently, is a major challenge for humanity, while global warming continues to break records [8].
More recently, climate change has been gaining attention in the public debate because of Greta Thunberg, 1 a young Swedish environmental activist who has been promoting worldwide demonstrations to ask governments for concrete action. Greta's activity has generated an enormous consensus but, unfortunately, also some criticisms and denigrating attacks.
The climate change story has already been making headlines, the next chapters of which will probably define the future of our planet in the context of anthropogene [9]. However, the climate change story is defined by the progressive convergence of a number of theories rather than by a linear collection of simple facts. Although in many cases, such theories have been supported by clear scientific evidences, climate change has never been considered an actual priority despite their objective relevance and practical impact in our lives. However, a social awareness around climate change is quickly rising. We believe that at this stage it is extremely important to tell the story so far allowing multiple interpretations and perspectives. Indeed, the climate change story cannot be limited to scientific perspectives, but needs to be understood in the light of social and political contexts as well; it has to be analysed taking into account other aspects of life, such as inequities and inequalities.
In this paper we present an ontological approach to build a knowledge-base on climate change related facts. Our knowledge space, developed upon Semantic Web technology, enables knowledge within the Semantic Web [10], assuming, therefore, a Linked Data philosophy [11]. It allows a dynamic classification and a semantic characterisation of the different events that can be related to each other, as well as to external sources of information. The resulting ontology (CCTL) is structured as a timeline which aims to describe the climate change story from multiple perspectives, including scientific, social, political and technological ones. We have created an instance of the ontology based on critical literature review and on the collection of factual information from relevant sources as an attempt the tell the story so far with a focus on most relevant theories, happenings, social and political initiatives. Such a data infrastructure is significantly more consistent than a simple specification of ordered facts. Last but not least, the proposed ontological approach focuses on the backbone of the semantic structure, namely on the systematic definition and integration of climate change related facts as part of a unique knowledge base. Such approach is agnostic with respect to the domain vocabulary. It can be an external vocabulary, as well as it can be collaboratively defined by users or inferred by the semantic analysis of referred documents.
As the climate change story is complex and diversely articulated, we do not pretend to provide a comprehensive set of facts and events. The focus of this research is on the ontological model to describe climate change facts which can be semantically enriched and defined in a given context through internal and external linking.

A. METHODOLOGY AND APPROACH
The ontology described in the paper has been developed by following an iterative process composed of the following steps: • Development of an ontological structure to define a timeline of potentially correlated events. This very first stage aimed to provide a simple agile data structure to arrange events, facts, data and theories along the time dimension. At this stage, correlations are completely generic as more specific relationships are expected to be introduced by users within the different application domains.
• Primary classification of climate change related facts and happenings. In order to better organise the target knowledge, we have provided a primary classification for climate change related facts and, additionally, some more fine grained concepts. As usual in knowledge engineering, this base taxonomy is expected to evolve according to the users' needs. We have also provided a generic class that identifies an element of the timeline. It allows to skip the use of the taxonomy in case such classification doesn't match needs in a given application context.
• Population of the ontology based on literature review and the analysis of recent facts. We have generated an instance of the ontology from a literature review and recent facts analysis. Such a population doesn't pretend to include all relevant facts, happenings and theories. It rather aims to provide a base dataset to evolve by integrating contributions by users from different areas of expertise.
• Semantic enrichment by identifying correlated concepts. The instance of the timeline defined in the previous step has been enriched by identifying correlations among the different timeline elements. Such correlations have been described by adopting the provided vocabulary.
• Validation and consistency checking within well-known development environments. Finally, we have checked our implementation within well-known ontology development environments with the support of reasoners, query wrappers and visualization tools. Details are provided later on in the paper.

B. STRUCTURE OF THE PAPER
The paper follows with a related work section. The core part is composed of two sections that deal with the description of the ontology (sec. III) and of climate change timeline (sec. IV) respectively. The former section focuses on the vocabulary by providing an overview of main concepts and relationships existing among them, as well as some details about the ontology implementation. The latter presents the ontology population by describing some key facts related to climate change. Moreover, it includes examples of specifications by adopting the ontological format provided, and examples of complex query on the resulting semantic structure. The paper ends with a conclusion section, which also addresses potential future work.

II. RELATED WORK
In this section we discuss the importance of narratives and public perceptions in the climate change issue, both with the VOLUME 8, 2020 adoption of ontologies to describe the different aspects of climate change.

A. NARRATIVES AND PUBLIC PERCEPTIONS
The climate change story is definitely a controversial one, making headlines and raising consciousness at different social levels, including elites (e.g. politicians and decision makers) and common people [12]. In this process of progressive awareness, narratives play a relevant role [12] and may influence public perceptions [13]- [15]. A well recognised influence can be celebrity endorsements. Indeed, public figures have a wide audience reach and, in general, they can potentially act as influencers among followers. On the negative side, associating the climate change issue with a public figure can cause a perception of climate change as part of the figure's brand [16]. Thus, the narrative is swayed by what the figure portrays rather than being based on indisputable facts and on a real awareness of the problem.
Timelines are effective to summarize narratives and are extensively used in different works, for example in The Discovery of Global Warming 2 [17]. In [18] authors highlight some notable developments along time in the creation and use of emissions, as well as other timelines (e.g. on globalization [19]) may indirectly deal with climate change.
Historically, news outlets largely communicated climate change topics as one of scientific significance [20]. However, the media often prioritize news that may easily catch the attention of people, i.e. sensationalist news. Indeed, news that may have an impact on people's lives in a relatively long term are often not even perceived as ''real'' news. That has an impact on how people perceive the reality of climate change.
Survey and poll data normally help to understand the impact of such influences on public perceptions, and to measure the effective level of awareness among people. Studies such as the one by [21] show that public knowledge gained traction around the late 1980s where up to 70% of respondents were aware of climate change; this was a marked improvement from early 1980s where only about 30% of respondents were aware of climate change. However, according to those studies, the actual knowledge was limited. Through the 1990s, awareness and understanding of impacts of climate change improved, but a lack of knowledge of what causes climate change persisted through to the millennium. To this day, the human contribution to climate change is disputed to the point that even the reality of climate change is disbelieved by many.
An effective awareness of climate change and its effects on human life is progressively building and demands clear response (e.g. for public health [22]) by governments [23].
We model our dataset and, therefore, our knowledge space as a timeline, meaning we consider the time dimension to organize the target knowledge. However, as it will be extensively explained later on in the paper, the ontological approach intrinsically supports multi-dimensional data and complex query from different perspectives.

B. ONTOLOGIES FOR CLIMATE CHANGE
Because of its complexity, climate change is unlikely to be entirely described by one single ontology [24]. Indeed, an ontology that fully addresses climate change should include concepts from various disciplines, implicitly assuming pluralism [24]. For instance, the semi-automatic method proposed by [25] allows the progressive extension of the concept hierarchy from a seed ontology by mining textual data from the Web sites. More in general, a collaborative approach assures effectiveness in building large scale knowledge bases [26]. In [27] the authors deal with technoscientific ontologies of climate change. In [28] authors adopt an ontological approach to evaluate the impact of a changing climate on food and waterborne diseases. Ontology has been adopted to model the impacts of agriculture and climate changes on water resources [29]. An ontology based on nature-society mutuality is adopted to deal with adaptive living with climate change in the rural areas [30]. An interesting application of ontology is to track the provenance of information [31].
As far as we know, there is no ontology to support the description of climate change related facts and their integration with contextual information. The ontology proposed in this paper aims to build a climate change timeline by adopting an approach which is agnostic with respect to the domain vocabulary. Indeed, the ontology primarily addresses the definition of the timeline backbone, namely the specification and the correlation of climate change related facts. Such facts and events are dynamically classified, meaning the underlining taxonomy provided is expected to evolve as per users' need. Similarly, the domain vocabulary, which is not object of this paper, may be created by merging external vocabularies from different disciplines, as well as it can result from the analysis of referred documents. We have chosen this open approach because of the intrinsic transdisciplinarity that characterises climate change. This heterogeneity makes hard a centralised definition of a domain vocabulary, and advises a ''local'' approach closely related to the various disciplines. Moreover, the characterization and semantic enrichment of described facts is based on external linking over the Semantic Web infrastructure. It allows the understanding of climate change related facts within a rich context as a result of the association with content on the Web. Additionally, the adoption of Semantic Web technology facilitates information re-use and interoperability. Therefore, the specifications provided may be easily integrated within generic purpose knowledge bases (e.g. DBpedia [32]).

III. CCTL ONTOLOGY: DESCRIBING CLIMATE CHANGE RELATED FACTS
The Climate Change TimeLine (CCTL) is an OWL ontology 3 to describe climate change related facts. It has been developed with the support of Protege 4 [33] and validated within that same environment by adopting common reasoners, such as PELLET [34] and HermiT [35]. Main classes are reported in Table 1, while properties are listed in Table 2.
An overview of the ontology is depicted in Figure 1. It is structured in three ideal layers, including a main categorization, a secondary categorization and a set of supporting concepts. These three layer are separately described below.

A. MAIN CLASSIFICATION
Because of its specific purpose, CCTL is designed around the central concept of climate change related fact (CCTL:ClimateChange_Fact). Such a class represents the most possible generic concept and it is used to classify a generic fact.
Primary classification assumes five different sub-classes ( fig. 1) as follows: • Scientific Evidence. It is some scientific evidence, normally documented in literature.
• Social Awareness. It differs from the previous one because it is related to an evidence or event directly associated with social awareness about climate change.
• Political Initiative. Any kind of political initiative directly or indirectly related to climate change.
• Theory. Theory on climate change, i.e. a prediction of a future happening or event due to climate change. As a theory may or may not be supported in the real world by some factual evidence, we prefer to differentiate such concept from the previously mentioned scientific evidence.
• Technology. A technology that has an impact (CCTL:Impact), positive, negative or neutral, on climate change. As we are implicitly focusing on climate change caused by human activity, the progress and, therefore, the underpinning technology plays a central role.
• Happening. An event related to climate change (e.g. a glacier disappearing). In many cases, it can be considered equivalent of the generic fact. It normally refers to events.
Facts can be related to each other with the generic relationship CCTL:relatedTo. This is similar to the omonymous property in PERSWADE-CORE [36], but it is stated as a symmetric property as per OWL specification: This generic relationship provides a simple, yet effective, way to navigate the resulting knowledge graph from an event correlation point of view. Sub-properties to define more specific relationships among facts will be object of future work.

B. SECONDARY CLASSIFICATION
Ontologies are formal descriptive tools [37] that are extensible by definition. Thus, a core taxonomy can evolve according to users ' needs.
Our core taxonomy has been extended to include two very common concepts, demonstration (CCTL:Demonstration) and policy (Policy) as in a common meaning. The former is understood as a sub-class of both CCTL:SocialAwareness and CCTL:PoliticalInitiative, as we consider significant demonstrations to have an impact at both a social and political level. The latter is assumed to be a kind of political initiative. A given policy is normally associated with a scope (CCTL:PolicyScope) -i.e. national, trans-national, worldwide -as well as with the organization which proposes or adopts it.

C. SUPPORTING CONCEPTS
Regardless of its generic or fine-grained classification, each fact assumes a description in a natural language (CCTL:decription) and a label (rdfs:label).
Additionally, the time dimension is modelled as a relationship to a single year (to be specified through the property CCTL:year) or to a year range defined by a starting and ending year (CCTL:startYear and CCTL:endYear respectively). The association of a given event with the time can also be specified by using the corresponding data properties (CCTL:year_s, CCTL:startYear_s and CCTL:endYear_s). Generally speaking, that level of detail is considered appropriate to build the target timeline. A more sophisticated association with time may rely on specific ontologies (e.g. Time Ontology in OWL 5 by W3C).
In order to build a consistent knowledge space, facts should be always associated with proper references, typically scientific literature, newspaper articles, reports, datasets or Web content. Because of the characteristics of ontological data models, this kind of link becomes a critical factor to build effective knowledge graphs. Normally a reference is identified by a URL or identifier (e.g. the DOI 6 ). Further info, the full citation (CCTL:citation) for instance, may be associated with references.

IV. THE CLIMATE CHANGE TIMELINE
The current version of the timeline includes the events reported in table 3, 4, 5, 6, 7, 8, 9, 10 and 11. Each table refers to one of the main categories for events classifica-tion as previously discussed. Events related to technology (Table 11) are limited to ICT in the current version of the ontology. The ontology is represented as a knowledge graph in figure 2.
In the following subsections, we focus on examples of use. We first provide an example of project descriptor by using an external vocabulary. In general, a project descriptor allows to specify metadata and generic information about a given project and, eventually, to register it within a target system. The descriptor normally depends on the target system. Our example is based on the PERSWADE-CORE model [36]. Additionally, we provide an example of formal specification of events by adopting the ontology. Finally, we propose some examples of complex query.

A. PROJECT DESCRIPTOR
A project descriptor is provided by adopting PERSWADE-CORE ontology [36]. The specification of the most relevant aspects, including project provenance, method adopted, main deliverables, aim and scope and similar concepts is depicted in fig. 3.
Such a descriptor is defined by the following OWL code (turtle syntax):

B. EXAMPLES OF SPECIFICATION ADOPTING CCTL
An example of specification involving 3 different events is proposed in fig. 4.
The first fact is a theory proposed in 1968 by Mercer. He predicted that a relatively slight rising of temperature could lead to the melting of the West Antarctic Ice Sheet which would in turn raise sea levels by about 5 metres [38].
The formal specification in OWL is as follows: A second fact is related to Okjökull (Ok glacier) that was declared dead in 2014 7 . The fact is considered to be correlated to the previous one as an old theory has somehow predicted a happening. Finally, we consider a recent (2019) demonstration in Iceland where a number of activists celebrated the funeral of that same glacier. Such a demonstration has been reported by media worldwide, for instance by SBS News. 8 The fact is intrinsically correlated to the previous one.
The formal specification is proposed below:

C. EXAMPLES OF COMPLEX QUERY
Apart from automatic reasoning by inference, developing upon Semantic Web technology assures full support to complex query through standard languages, i.e. SPARQL. 9 Few simple examples of SPARQL query on the proposed ontology are: • All correlated facts. By executing the SPARQL code reported below, all facts that are related to others are retrieved.
SELECT ?x ?y WHERE { ?x CCTL:relatedTo ?y } • All facts related to a fact that happened in a given year. The query reported below is an elaboration of the previous one that retrieves facts related to a fact that happened in a given year.   range. This kind of query may become important in understanding the different phases of climate change story and the progressive consolidation of the different theories.

V. APPLICATIONS
Because of the ontological approach adopted, our contribution aims to provide a generic semantic description of a  climate change timeline. Such a data infrastructure is agnostic with regards to applications and, in general terms, enables knowledge within the Semantic Web in which data may be inter-linked and semantically enriched via (semantic) interoperability.
In our opinion, the benefits of using such a data infrastructure could be especially effective, among others, in the context of the following applications: • Communication framework. One of the most empathised issues around climate change is related to the VOLUME 8, 2020 communication between scientists and other stakeholders, especially decision making and citizens. Focusing on people perceptions, it is commonly accepted that, for a long time, climate change has not been considered like an actual and immediate threat. It is well known that history may help to not redo the same mistakes. Climate change may be making alarming headlines, however, its most critical chapters are yet to be written. The support of the systematic definition of climate change related facts can be useful to tell the story so far in context and to progressively enrich semantic specifications, including the relationships among the different happenings. Such kind of application may be useful to further raise awareness on the topic.
• Specific-purpose knowledge base. The same reasons that suggest a more effective and consistent communication among the different stakeholders point out the need for actual knowledge around climate change. Knowledge bases could contribute significantly in this sense as they are easily accessible though the Web, and, if properly designed and structured, may provide a valuable convergence point for informed discussion. We believe that the organization of the knowledge in a timeline can contribute to reach a wide audience focusing on the causality relationship between theories and happenings.
• Information systems. An ontological structure implemented upon Semantic Web technologies intrinsically provides some facilitations to information integration. Indeed, as the specifications are provided in a standard language, information can be accessed and integrated in any context, including systems whose purpose is not specifically related to climate change.
• Knowledge specification and re-use within multidisciplinary environments. Climate change related knowledge results from multi-disciplinary studies and it is intrinsically complex. This intrinsic interdisciplinary requires data, information and knowledge sharing among systems in different domains. The ontological approach fulfills requirements in terms of formal specification, interoperability and re-use.
We believe that a simple specification of events along the time dimension would not have provided a consistent support in the context of the above mentioned applications. On the contrary, a full ontological approach enables the progressive extension of the base taxonomy and the specification of a shared domain vocabulary resulting from the different disciplines and applications.

VI. CONCLUSION AND FUTURE WORK
We have provided an ontology (CCTL) to effectively describe climate change related events that can be related to each other, as well as they can be linked to external sources of information. CCTL is structured as a timeline which aims to describe the climate change story from multiple perspectives, including scientific, social, political and technological ones. We have created an instance of the ontology based on critical literature review and on the collection of factual information from relevant sources as an attempt the tell the story so far with a focus on most relevant theories, happenings, social and political initiatives.
Developed upon Semantic Web technology, such a knowledge space is enabled within the Semantic Web in which concepts are uniquely defined to enhance the interoperability model (Semantic Interoperability [40]). By definition, ontological structures are extensible and provide re-usable formal specifications for knowledge and data, which can be easily linked with each other and interchanged through the Web infrastructure.
Our approach allows the progressive definition of a knowledge base to tell the climate change story. The different key events, facts, theories and scientific evidences may be formally specified and automatically integrated as part of a unique knowledge base. Moreover, the specification of the relationships existing among the different climate change related facts provides further capabilities in terms of analysis supported by standard query languages. Semantic enrichments enable the specification of the context as linkage to external sources of information. Finally, by adopting an ontological format in a standard language, data may be easily re-used and interchanged across different systems.
Future work will aim to the application of CCTL in different contexts. CCTL is expected to evolve as per users' needs. For instance, a breakdown for the category CCTL:Technology could be provided according to some technology classification schema. The domain vocabulary is not object of this paper and will be obtained by analysing the references associated with the different events. Simple techniques based on keywords extraction will allow a systematic and fully automatic association of the key terminology with each event. The evolution of such domain vocabulary along the time is intrinsically supported because of the organization of the information along the time dimension. Nevertheless, the hierarchical organisation of such concepts will require a supervised approach involving experts from the different disciplines.

APPENDIX. ONTOLOGY POPULATION
We report in this appendix the current population of the ontology organised by category.