An Ontology Design Pattern for Modeling Pollution? Saad Ahmad[0000−0002−3040−508X] , Sudhir Attri[0000−0002−3667−9110] , and Raghava Mutharaju[0000−0003−2421−3935] Knowledgeable Computing and Reasoning (KRaCR) Lab, IIIT-Delhi, New Delhi, India. {saad18409,sudhir18267,raghava.mutharaju}@iiitd.ac.in Abstract. Pollution has been identified as a significant risk to global ecosystems and living beings. However, the information about pollution is fragmented and there is no meaningful organization of information despite several ongoing efforts to monitor it. Organizing the informa- tion about the pollution, such as the pollutants, their observations at different spatio-temporal points and the carriers in the form of an on- tology will be very helpful to the applications that work with the dif- ferent heterogeneous pollution data sources. We propose an ontology design pattern (ODP) for pollution that captures its general characteris- tics and can be used as a building block for modeling specific categories of pollution such as air, water and soil. The Pollution ODP is available on the ODP portal at http://ontologydesignpatterns.org/wiki/ Submissions:Pollution. It is also available for public comments at https://github.com/kracr/aq-structured-platform/blob/main/On tology/PollutionODP/PollutionODP.owl with an Apache License 2.0. Keywords: Pollution · Ontology Design Pattern · Pollutants · Air Pol- lution. 1 Introduction The introduction of substances into the environment that are harmful to the liv- ing organisms is defined as pollution [11]. These harmful substances can be solids, liquids, or gases produced in higher than usual concentrations. These substances are referred to as pollutants. Pollution could affect many natural resources, such as the air, water sources and soil. The need to study these pollution types is of utmost importance. Air pollution is regarded as a global health emergency and the effect of bad air quality is deadly. It leads to asthma, other respiratory ill- nesses and heart disease. Air pollution is responsible for more deaths than many other risk factors, including malnutrition, alcohol use and physical inactivity [6]. Similarly, there is a pressing need to study water and soil pollution [1]. Different ? Copyright © 2021 for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0). 2 S. Ahmad, S. Attri, R. Mutharaju heterogeneous data sources are made use of to build pollution monitoring appli- cations [2,3]. An ontology design pattern(ODP) [8] that captures the abstract details of the pollution will be beneficial to these applications. 2 Related Work Attempts to model pollution have focused on air pollution, probably because of the readily available data of pollutant concentrations. Claudine et al. [12] built an air quality ontology and used that along with the 3D models of a city. The ontology proposed in [13] links air pollutants with meteorological factors. EnvO1 is a very broad ontology that describes the environment by focusing on biomes. Dalia et al. [4] built an air pollution ontology by focusing on species, sensors, pollutants and meteorological factors. They have designed one ontology for each of these factors. We studied these ontologies along with several data sources such as pollutant concentration2 , weather data3 and wind trajectory4 to design the Pollution ODP. We also capture the spatio-temporal aspect of pollution in the ODP, which has not been considered in some of the pollution ontologies. The Pollution ODP has been annotated to indicate the ODPs that were used and submitted for a review on the ODP community portal5 . 3 Pollution ODP Description The pollution ODP makes use of the trajectory ODP [9], observation ODP [5] and the stub meta-pattern [10]. We use cpannotationschema6 to describe the intent, scenarios, consequences and components of the ODP. The schema of the pollution ODP is given in Figure 1. We describe the concepts and the properties of the pollution ODP in the subsequent sections. 3.1 Concepts – Pollution. It is the core concept in the ODP to represent pollution and is linked to the contributors of pollution. Some of the instances of Pollution could be air pollution, water pollution, soil pollution, space pollution, and sound pollution. – Observation. The Observation concept is modeled from the Observation ODP. It represents a spatio-temporal observation. We use it to capture the concentration and prescribed standards for a particular pollutant. With re- spect to the Observation ODP, Pollutant and PrescribedStandardForPo- llutant is the situation and TimeEntity, PlaceEntity are the parameters of the observation. 1 http://environmentontology.org/ 2 https://cpcb.nic.in/ 3 https://weatherstack.com/ 4 https://www.ready.noaa.gov/HYSPLIT disp.php 5 http://ontologydesignpatterns.org/wiki/Main Page 6 http://www.ontologydesignpatterns.org/schemas/cpannotationschema.owl An Ontology Design Pattern for Modeling Pollution 3 Fig. 1. Core concepts and properties of the Pollution ODP – DirectContributor. The DirectContributor concept represents the con- tributors that directly affect the pollution. Pollutant is a subclass of Direct- Contributor. Some examples of pollutants include biological pollutants (viruses, pathogens, bacteria, etc.), chemical pollutants (toxic metal, radionuclides, organophosphorus compounds, gases, etc.) and physical pollutants (sound, thermal energy, space debris, etc.) present in a particular environment. – IndirectContributor. The IndirectContributor concept represents con- cepts that indirectly contribute to the pollution at a particular spatio-temporal point. These include environmental factors like temperature, the air or water streams flowing into or out of a particular place, the socio-economic factors such as policies and demographics. We have modeled only the Carrier con- cept as the subclass of IndirectContributor since other indirect contribu- tors are specific to certain kinds of pollution. – PlaceEntity and TimeEntity. These concepts denote the place and time for a spatio-temporal observation. They are linked to the Observation and TrajectoryPoint concepts by the atPlace and atTime properties. – PrescribedStandardForPollutants. This is a stub meta-pattern [10] that can be used to describe the prescribed ranges for pollutants. For a partic- ular location, pollutants have a defined range of permissible concentrations that are specified by the various global authorities7 . A standard for a pollu- tant may change with time and place. Hence the PrescribedStandardFor- Pollutants concept is linked to the Observation concept by the has- Observation property. 7 2005 WHO guidelines prescribe the range for air pollutants such as particulate mat- ter (PM), ozone (O3), nitrogen dioxide (NO2), etc., available at http://whqlibdo c.who.int/hq/2006/WHO SDE PHE OEH 06.02 eng.pdf?ua=1. 4 S. Ahmad, S. Attri, R. Mutharaju – Carrier. This concept is a subclass of the IndirectContributor concept and represents the air, water, or other kinds of streams flowing into or out of a particular place. It is linked to the Trajectory concept through the hasTrajectory property. Carriers are generally observed to affect the con- centration of pollutants at a particular place and are important in modeling pollutants. To represent the pollutants that might be carried through a tra- jectory, the Carrier concept is linked to the Pollutant concept by the carriesPollutant property. To specify the location of the source of pol- lutants in a carrier trajectory, nearby property can be used. This links the pollutant sources such as factories in the case of wind stream carrier or drains in the case of water stream carrier to the TrajectoryPoint. The applica- tions that have such a requirement can make use of the nearby property, but we excluded it from the ODP because it is not sufficiently general. – Trajectory. The Trajectory concept represents a set of ordered spatio- temporal points and has been directly adopted from [9]. It is linked to the TrajectoryPoint and TrajectorySegment through the hasPoint and has- Segment properties. The nextPoint property links trajectory points in the appropriate order within a trajectory. The segments in the trajectory are defined by a starting trajectory point {xi , yi , ti } and an ending trajectory point {xj , yj , tj } where ti , tj denote time points such that ti < tj . The TrajectorySegment concept represents this notion of a segment which is connected to two fixes through startsFrom and endsAt properties. 3.2 Properties – atPlace, atTime. They connect Observation and TrajectoryPoint con- cepts to the PlaceEntity and TimeEntity respectively. Since an instance of Observation or TrajectoryPoint can be associated with at most one timestamp, the atPlace and atTime properties are functional. – startsFrom, endsAt. These functional properties connect a Trajectory- Segment to starting and ending TrajectoryPoint representing the starting and ending point of a segment. – carriesPollutant. This property represents the pollutants that a Carrier can carry. It connects Carrier to the Pollutant concept and represents the pollutants being carried away by a carrier. The trajectory of the carrier dictates the displacement of pollutants. – hasTrajectory. This property connects the Carrier concept to the Trajectory. – hasContributor. This property connects the Pollution concept to the Contributor concept. – hasPoint. This property connects the Trajectory with the TrajectoryPoint. – hasPrescribedStandards. This property connects the Pollutant concept to the stub PrescribedStandardForPollutant concept. – hasObservation. This property connects Pollutant and PrescribedStand- ardForPollutant concept to the Observation concept. It can be used to capture the concentration of pollutants or a prescribed standard for a par- ticular pollutant with the TimeEntity and PlaceEntity as the parameters of the observation. An Ontology Design Pattern for Modeling Pollution 5 – hasSegment. This property connects the Trajectory to the Trajectory- Segment. – nextPoint. This property links each point represented by the TrajectoryPoint to the next point forming a chain of ordered points. 3.3 Axioms of the Pollution ODP The axioms that are part of the Pollution ODP are discussed here. Carrier v ∀carriesPollutant.Pollutant (1) Carrier v ∃hasTrajectory.Trajectory (2) Pollution v ∃hasContributor.Contributor (3) Observation v ∃atPlace.PlaceEntity (4) Observation v ∃atTime.TimeEntity (5) − TrajectoryPoint v ∃hasPoint .Trajectory (6) Trajectory v ∃hasSegment.TrajectorySegment (7) hasSegment ◦ startsFrom v hasPoint (8) hasSegment ◦ endsAt v hasPoint (9) Axioms 6, 7, 8 and 9 capture the relation between the Trajectory, Trajectory- Point and the TrajectorySegment concepts. 3.4 Competency Questions The Pollution ODP answers the following competency questions. 1. What are the contributors of the pollution? 2. What is the pollutant concentration at a particular time and place? 3. What are the carriers that contributed to the pollution? 4. What are the pollutants carried by a carrier? 5. What are the prescribed standards for a particular pollutant? 6. What is the trajectory of a carrier for a pollutant? 4 Use Cases The Pollution ODP can be used as a building block to model various pollu- tion types such as air, water and soil. They can be added as subclasses of the Pollution concept. An example use case of air pollution can describe the con- centration of pollutants in the air at a particular spatio-temporal point through the Observation and Pollutant concepts. Similarly, the pollutants carried by the air stream can be captured by the Carrier concept. The Weather concept can be added as a subclass of IndirectContributor concept in the concrete implementation of this ODP to represent the weather related factors that con- tribute to air pollution. These concepts can be extrapolated to water, soil and other types of pollution as well. 6 S. Ahmad, S. Attri, R. Mutharaju 5 Conclusion and Future Work Pollution exists in many forms in the environment and affects individuals as well as ecosystems. An ontology design pattern for modeling various sources and characteristics of pollution can be used effectively as a building block by multiple applications that work with pollution data. We introduce an ODP for modeling pollution and discuss its competency questions, concepts and properties. We also describe some use cases of the ODP. The ODP and its documentation are publicly available at http://ontologydesignpatterns.org/wiki/Submissions: Pollution and at https://github.com/kracr/aq-structured-platform/bl ob/main/Ontology/PollutionODP/PollutionODP.owl with an Apache License 2.0. We plan to use the Pollution ODP to model air pollution by considering the pollutant concentration, weather and wind trajectory data sources. The air pollution ontology will be populated using these data sources by converting the semi-structured data (csv, json, etc.) into an RDF8 graph using YARRRML [7]. Acknowledgements. This work is partially supported by the Infosys Center for Artificial Intelligence (CAI) at IIIT-Delhi, India. References 1. Air and Water Pollution: An Important Nexus of Transportation and Health. In: Meyer, M., Elrahman, O. (eds.) Transportation and Public Health, pp. 65–106. Elsevier (2019). https://doi.org/10.1016/B978-0-12-816774-8.00004-9 2. Bellas, J., Hylland, K., Burgeot, T.: Editorial: New Challenges in Ma- rine Pollution Monitoring. Frontiers in Marine Science 6, 820 (2020). https://doi.org/10.3389/fmars.2019.00820 3. Cromar, K.R., Duncan, B.N., Bartonova, A., Benedict, K., Brauer, M., Habre, R., Hagler, G.S.W., Haynes, J.A., Khan, S., Kilaru, V., et al.: Air Pollution Monitoring for Health Research and Patient Care. An Official American Thoracic Society Workshop Report. Annals of the American Thoracic Society 16(10), 1207–1214 (Oct 2019) 4. Dalia Dahleh, M.S.F.: An Environmental Ontology for Global City Indicators (ISO 37120) (09 2016) 5. Eva Blomqvist: Submissions:Observation - Odp. http://ontologydesignpatter ns.org/wiki/Submissions:Observation, (Accessed on 15/07/2021) 6. Goldsmith, J.: Effects of Air Pollution on Human Health. In: STERN, A.C. (ed.) Air Pollution and its Effects (Second Edition), pp. 547–615. Academic Press, second edition edn. (1968). https://doi.org/10.1016/B978-0-12-666551-2.50021-2 7. Heyvaert, P., De Meester, B., Dimou, A., Verborgh, R.: Declarative Rules for Linked Data Generation at your Fingertips! In: Proceedings of the 15th ESWC: Posters and Demos (2018) 8. Hitzler, P., Gangemi, A., Janowicz, K., Krisnadhi, A.A., Presutti, V.: Ontology Engineering with Ontology Design Patterns: Foundations and Applications. IOS Press, NLD (2016) 8 https://www.w3.org/TR/rdf11-primer/ An Ontology Design Pattern for Modeling Pollution 7 9. Hu, Y., Janowicz, K., Carral, D., Scheider, S., Kuhn, W., Berg-Cross, G., Hitzler, P., Dean, M., Kolas, D.: A Geo-ontology Design Pattern for Semantic Trajectories. In: Tenbrink, T., Stell, J., Galton, A., Wood, Z. (eds.) Spatial Information Theory. pp. 438–456. Springer International Publishing, Cham (2013) 10. Krisnadhi, A., Hitzler, P.: The stub metapattern, pp. 39–45 (12 2017) 11. Manisalidis, I., Stavropoulou, E., Stavropoulos, A., Bezirtzoglou, E.: Environmen- tal and Health Impacts of Air Pollution: A Review. Frontiers in Public Health 8, 14 (2020). https://doi.org/10.3389/fpubh.2020.00014 12. Métral, C., Falquet, G., Karatzas, K.: Ontologies for the Integration of Air Quality Models and 3D City Models (2012) 13. Oprea, M.M.: AIR POLLUTION Onto: an Ontology for Air Pollution Analysis and Control. In: Iliadis, Maglogiann, Tsoumakasis, Vlahavas, Bramer (eds.) Artificial Intelligence Applications and Innovations III. pp. 135–143. Springer US, Boston, MA (2009)