Your SlideShare is downloading. ×
Brazilian Cerrado geolinked data and qualitative models
Upcoming SlideShare
Loading in...5
×

Thanks for flagging this SlideShare!

Oops! An error has occurred.

×

Introducing the official SlideShare app

Stunning, full-screen experience for iPhone and Android

Text the download link to your phone

Standard text messaging rates apply

Brazilian Cerrado geolinked data and qualitative models

202
views

Published on

Published in: Education, Technology

0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total Views
202
On Slideshare
0
From Embeds
0
Number of Embeds
3
Actions
Shares
0
Downloads
1
Comments
0
Likes
0
Embeds 0
No embeds

Report content
Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
No notes for slide

Transcript

  • 1. Brazilian Cerrado ontology network and qualitative models: a case study application of geolinked data approach to Ecology Adriano Souza University of Brasilia, Brazil 2013
  • 2. This presentation • Introduction • Background and state of the art • Research question • Hypothesis • The proposal • Work in progress • What is next… 2 of 41
  • 3. Ecoinformatics A field of research and development focused on the interface between ecology, computer science, and information technology. “Ecologists have recognized the need for integrated data systems to support cross-disciplinary collaboration to understand the basic ecological principles that govern the biosphere.” Green et al. (2005) 3 of 41
  • 4. AI Technologies can be useful to theoretical development in Ecology: • To organize knowledge bases compatible with computers, including qualitative and quantitative knowledge; • To perform fast assessment of assumptions, hypotheses and other ideas in a theoretical context; • To determine the consequences and the logical consistency of long and complex paths of ecological reasoning. Maybe the most immediate impact of AI Technologies will be on the way ecologists organize, develop and implement models. Rykiel (1989) 4 of 41
  • 5. Why models are necessary? • To build and use models contribute to… • Understand the structure of systems; • Predict the behavior of systems; • Control variables in order to obtain specific results. • They are used in: • Scientific research; • Decision making and management; • Education and training. 5 of 41
  • 6. Ecological models To build ecological models is a complex task because... 1. Ecological models are heterogeneous, including both qualitative and quantitative knowledge; 2.It is hard to collect data and perform experiments; 3.The available data is incomplete, inaccurate, uncertain and many times expressed in qualitative terms; 4.The theoretical foundations and the laws (or first principles) are still under development. • New approaches to ecological modeling are required! Qualitative Reasoning 6 of 41
  • 7. Qualitative Reasoning (QR) It is an area of artificial intelligence that creates representations for continuous aspects of the world to support reasoning with little information The use of QR models can contribute to clarify many aspects and to improve the understanding of causal reasoning chains involving environmental factors and changes in populations and communities. Salles & Bredeweg (2006) 7 of 41
  • 8. Qualitative Ecological Models Are promising because • Allow to build and run simulations with incomplete knowledge; • Allow to create a rich vocabulary about a variety of systems; • Explicit representation of causality which gives support to explanation of systems from its structure; • Contribute to improve comprehension about complex systems and fosters the decision making process. Advantages over numerical models • Inaccurate prediction, but CORRECT • Easy exploitation of alternatives • Automatic interpretation 8 of 41
  • 9. Qualitative Reasoning 9 of 41
  • 10. Semantic Web Ecology QR • DataOne • OBOE • JournalMap • Domain Ontologys 10 of 41
  • 11. 11 of 41
  • 12. 12 of 41
  • 13. What all savannas have in common? “Savanna occurs over a vast range of conditions that have little in common except for their inability to support rapid tree growth.” (Hoffman et al., 2012) Source: Challenges and opportunities in remote sensing of global savannas. Colorado State University. Available in: http://www.nrel.colostate.edu/projects/srs/ 13 of 41
  • 14. 2 main reasons why do I care about Cerrado 1. Ecological Theory • Stability Biodiversity • Equilibrium x Non-equilibrium 2. Conservation Biology Hotspot 14 of 41
  • 15. Unstable Disturbance-driven savannas Disturbances such as fire, grazing and browsing are required to maintaing both trees and grasses in the system Disturbances such as fire and herbivore, although capable to modify tree to grass ratios, are not necessary for coexistence Stable Climatically determined savannas Sankaran et al. (2005) 15 of 41
  • 16. About 50% of Cerrado is Deforested Total Area: 2.047.146,35 Km2 Cerrado vegetation type (%) IBGE (2004) Savanna 61% Forest 32% Cerrado cover type (%) IBGE (2004) Natural 51% Grassland 7% Deforested 48% Water 1% 16 of 41
  • 17. Cerrado + Poor acid, aluminum toxicity soils + State-of-the-Art Technology + Farmers = Remaining areas Deforested area accumulated until 2008 17 of 41
  • 18. Amount of soybean production (ton) in 2005 per municipality 18 of 41
  • 19. Cerrado vegetation characteristics Panoramic view Physiognomies OM and Soil Fertility % General Structure of Cerrado Plants height Cover % Aerial view 19 of 41
  • 20. Research question How to use the GeoLinked data approach to integrate datasets along with qualitative reasoning ecological models, in order to improve the understanding of ecological mechanisms and facilitate access and management of environmental information? 20 of 41
  • 21. Exploring hypothesis The use of qualitative conceptual simulation models, associated to data sources made available by geolinked data semantic techniques, can improve the interpretative and predictive capacity over the data available about the dynamics of Cerrado vegetation. 21 of 41
  • 22. Qualitative Model Repository Export Recover Geosparql Qualitative Model OWL format QR Ontology Ad-hoc application GeoLinkedData Browser (map4RDF) Data and Metadata (RDF) Cerrado Ontology Recover OBOE Vegetation Dynamics Data repository Fire Ontology Meteor. Ontology Data cube Fire and Meteorological Data (INPE) 22 of 41
  • 23. Study A Species: Ouratea hexasperma Density: 73 Fire frequency: Fire Protected + geographical information meteorological data agriculture owl:sameAs owl:sameAs GeoLinked data to retrieve appropriate models Simulation owl:sameAs Study B Species: Ouratea hexasperma Density: 83 Fire frequency: Low + geographical information meteorological data agriculture 23 of 41
  • 24. Methodological aspects: the Life Cycle Model Specification Exploitation Modeling Publication RDF generation Links Generation Iterative incremental life cycle 24 of 41
  • 25. Specification • Ontology Requirements Specification Document Ontology Requirements Specification Document 1 Purpose The purpose of the Cerrado ontology network is to represent the scientific knowledge about the ecology and dynamics of Cerrado plants, it should express how the structure of the plant communities in Cerrado is and how they change over time. 2 Scope Because of the complexity and extend of the domain, the scope of the ontologies will focus to cover the following subdomains: plant community dynamics and fire. 3 Implementation Language The Ontologies will be developed using the Web Ontology Language OWL, once it is part of the W3C recommendation for the Semantic Web. 4 Intended End-Users 1. Researchers and scientists seeking to understand the functioning of savannah plant community. 2. Environment managers of conservation units and those responsible for making the public policies for environment and biodiversity conservation. 3. Ecological and environmental information and data about Brazilian Cerrado users. 5 Intended Uses 1. To store data and provide information about diversity, composition and dynamics of Cerrado wood plants. 2. To propose a standard and management practice of the data available about the Cerrado vegetation. 3. To propose a service in which the user can search for species, its location and assess the changes in populations over time. 6 Ontology Requirements a. Non-Functional Requirements NFR1: The ontology network must give support to a multilingual scenario for Portuguese and English. b. Functional Requirements: Groups of Competency Questions For the functional requirements it was used the competency questions technique (Gruninger and Fox, 1994) recommended by NeOn methodology. 25 of 41
  • 26. Specification • Competency Questions # Competency Questions for CCOn 1 What is a biome? 2 What is a savanna? 3 What characterizes a savanna? 4 What are the determining factors of savannas? 5 What is Cerrado? 6 What characterizes the Cerrado? 7 What is a population? 8 What is population growth? 9 Which processes determine the size of a population? 10 What is mortality? 11 What is natality? 12 What is an ecological community? 13 What are the types of ecological communities? 14 What is biodiversity? 15 What is the species richness of a community? 16 What factors determine the species richness of a community? 17 What is a plant community? 18 What are the types of plant communities? 19 What are wood plants? 20 What are herbaceous plants? What are the main measurements of biological diversity of a community 21 or ecosystem? # Competency Questions for Fire Ontology 65 Where does occur the wildfires in Cerrado? 66 How often Cerrado vegetation burns? 67 In what period of the year does wildfires occur in Cerrado? 68 What is wildfire? 69 Where are located the places with similar temperature range? Where are located the places with maximum temperature in a 70 certain time? 71 What are the types of wildfires? 72 What is the severity of each burn event in Cerrado? 73 What is the temperature of a location in a given time? 74 What is the relative humidity of a location in a given time? 75 What causes a burn event? 76 What are the effects of a burn event? 26 of 41
  • 27. Modeling: ontologies reused Reusability Usability CCOn Environmental Ontology (EnvO) Crop-Wild Relations FAO (CWR) Simple Knowledge Organization System (SKOS) Fire Ontology Environmental Ontology (EnvO) Extensible Observational Ontology (OBOE) 27 of 41
  • 28. Reused x New Terms Table 2. Reused Classes Number of terms Resource CCOn Fire Ontology OBOE 0 2 ENVo 3 1 CWR 24 0 Table 3. Reused Properties Property Origen Reused in exactMatch SKOS CCOn mappingRelation SKOS CCOn Table 4. New terms Ontology N of classes N of properties N of individuals CCOn 58 21 137 Fire Ontology 49 17 9 Total 107 38 146 28 of 41
  • 29. CCOn 29 of 41
  • 30. CCOn 30 of 41
  • 31. Fire Ontology 31 of 41
  • 32. Fire Ontology 32 of 41
  • 33. Use case example Dataset A Study area Fragment First inventory Last inventory Domain specific ontologies OBOE Gerais de Balsas Colonization Project, Maranhão, Brazil 1 2 1995 1995 2002 2002 Fire Mean Annual Mortality rate Mean Annual Recruitment rate Biennial Is a Community dynamics Measurement Is characterized by Biennial 2.73 Of-Characteristic Recruitment rate 4.88 3.25 5.86 Cerrado sensu stricto has-measurement Is part of Observation ofEntity Dataset B Location Tree Jatobá Biological Reserve, Bahia, Brazil 2004 Observation ofEntity Fire 1991 Last inventory Mortality rate Recruitment rate Wood plant affects First inventory Fire Is-a hasCharacteristic has-measurement Protected 1.93 3.72 Fire Characteristic Measurement Of-Characteristic Is-a Fire frequency 33 of 41
  • 34. Ontology Evaluation • Consistency: • Pellet reasoner • Pitfalls (OntolOgical Pitfalls Scanner - OOPS) • Conceptual Evaluation (experts) 34 of 41
  • 35. Ontology evaluation: Pitfalls 100 Table 3. Ontology Pitfalls found in Fire Ontology 70 Critical Important 60 Minor Usability-profiling Structural 50 Functional 90 80 Version P02 P04 P05 P07 P08 P10 P11 P13 P22 P29 P38 Total 70 0.3 2 11 0 0 54 ONT 7 7 ONT 0 - 81 60 0.4 0 1 1 0 49 ONT 0 4 ONT 2 - 57 50 0.5 0 1 0 0 53 0 0 2 ONT 0 - 56 40 0.6 0 0 0 0 51 0 0 0 ONT 0 - 51 30 0.7 0 0 0 1 24 0 5 0 0 0 ONT 30 20 0.8 0 0 0 1 24 0 0 0 0 0 ONT 25 10 40 30 20 10 0 0 0.3 Table 4. Ontology Pitfalls found in Ccon Ontology 0.4 0.5 0.6 0.7 0.8 100 Version P04 P05 P08 P11 P13 P22 P31 P35 P38 Total 0.8.1 2 0 83 8 8 ONT - - - 101 0.8.2 1 1 83 6 8 ONT - - - 0 0 88 2 0 ONT - - - 90 0.9.1 0 0 24 2 0 ONT 1 1 ONT 28 0.9.2 0 0 24 0 0 ONT 1 1 ONT 26 80 80 60 60 40 40 20 20 0 0 99 0.9.0 100 0.8.1 0.8.2 0.9.0 0.9.1 0.9.2 35 of 41
  • 36. Expert Evaluation • 2 Questionnaires were elaborated (G forms) • Quality • Completeness • Correctness • Likert scale and Open Questions Support material Evaluation On line documentation Bioportal visualization Competency questions Fire Ontology Questionnaire Expert Answers Questionnairies CCOn Questionnaire 36 of 41
  • 37. What is next RDF generation geometry2RDF Google Refine Links Generation owl:sameAs Publication Virtuoso, Pubby Exploitation Map4RDF SILK 37 of 41
  • 38. Data Sources for RDF generation Data sources for vegetation dynamics on scientific literature: • Souza, A. (2010). Estrutura e Dinâmica da Vegetação Lenhosa de Cerrado sensu stricto no período de 19 anos, na Reserva Ecológica do IBGE , Distrito Federal , Brasil. 68p. Dissertação de mestrado. Departamento de Ecologia. Universidade de Brasília. • Roitman, I.; Felfili, J.M.; Rezende, A.V. (2008). Tree dynamics of a fire-protected cerrado sensu stricto surrounded by forest plantations, over a 13-years period (1991-2004) in Bahia, Brazil. Plant Ecology. 197: 255-267. • Moreira, A.G. (1992). Fire protection and vegetation dynamics in the Brazilian Cerrado. Ph.D. thesis, Harvard University, Cambridge, MA, U.S.A. • Aquino, F. D. G., Walter, B. M. T., & Ribeiro, J. F. (2007). Woody community dynamics in two fragments of “cerrado” stricto sensu over a seven-year period (1995-2002), MA, Brazil. Revista Brasileira de Botânica, 30(1), 113–121. • Libano, A. M., & Felfili, J. M. (2006). Mudanças temporais na composição florística e na diversidade de um cerrado sensu stricto do Brasil Central em um período de 18 anos (1985-2003). Acta Botanica Brasilica, 20(4), 927–936. Meteorological Data Source Fire Occurrence and risk INMET INPE Maps and environmental data IBGE CSR IBAMA MMA LAPIG 38 of 41
  • 39. Outlook • This work presents a plan for a pilot study to be a test. • It involves linked geographical, meteorological, ecological and environmental open data provided by Brazilian government agencies. • A methodology based in a Geolinked data approach is adopted to create a case study aiming investigate the application of linked data principles to ecology. 39 of 41
  • 40. Outlook • A relevant question to be investigated in this preliminary pilot study is how to integrate qualitative reasoning models along with maps and other data, be able to reason with the data and make inferences and finally to show the results. • The topics addressed in this work have potential to boost both applications of geolinked data technologies to new areas, and to open new perspectives for research involving ecological data management, integration and use. 40 of 41
  • 41. Thank you! 41 of 41