Skip to main content

Sustainable Historic Environments hoListic reconstruction through Technological Enhancement and community based Resilience

Periodic Reporting for period 1 - SHELTER (Sustainable Historic Environments hoListic reconstruction through Technological Enhancement and community based Resilience)

Reporting period: 2019-06-01 to 2020-11-30

Historic Areas are one of our most valuable achievement as a society due to their cultural and natural values and their impact on peoples well-being. But current disaster statistics and studies usually do not consider heritage as a sensitive and valuable element. The higher vulnerability of materials and structures, difficult accessibility and density of the urban fabric makes the improvement of the resilience of historic areas a challenge. The overall objective of SHELTER is to develop data-driven and community-based approaches, tools and solutions to improve resilience in the historic areas. SHELTER proposes a change of paradigm, where the disturbance is not an unexpected event anymore, instead it is foreseen, accepted and addressed for transformation. SHELTER aims to generate and demonstrate a way to improve resilience that is the result of the interplay of two processes collaborating to feature individual solutions for each case: a data-driven platform that supports the diagnosis, decision making, implementation and monitoring based on existing knowledge and heterogeneous data, and an Open Lab approach that provides a continuous framework for local knowledge extraction, citizen´s engagement, co-creation, capacity building and innovation. SHELTER is organized to develop and demonstrate a highly adaptable and replicable systemic approach based on the identification and activation of the inherent resilient characteristics of historic environments (heritage-led resilience) and the preservation of its values (conservation-friendly resilience). SHELTER will adopt a case studies based approach with two objectives: to generate the required knowledge regarding the impact of different impacts in diverse typologies of Natural and Cultural Heritage and to co-validate the suitability, adaptability and replicability of the SHELTER framework, methodologies and ICT tools to different contexts. The five case studies (Ravenna-Italy, Seferihizar-Turkey, Dordrecht-Netherlands, Natural Park of Baixa Limia-Serra Do Xurés – Spain and Sava River Basin) have been selected considering their natural and cultural value and diversity, their especial exposure to diverse hazards, their geographical representability and climate conditions and their scale and typology.
•The knowledge base and the operationalization of the data and knowledge has finished with four deliverables where existing data sources have been mapped and operationalised, next and best practices have been characterized, an ontology has been developed and implemented in a wiki, the Data lake has been defined and implemented and the Multiscale data model for georeferenced data from the Open labs has been defined.
•The quantification and assessment of resilience have set its basis in this first period. The conceptual framework for assessment was defined and the framework of indicators to monitor resilience and the methodology for characterization of cultural and natural heritage has been delivered.
•The second draft of the existing solutions portfolio for the whole DRM cycle is ready.
•The architecture of the data-driven platform has been defined and agreed and some of the common components have been fully designed, implemented, and deployed.
•A procedure for organization and management of the Open Labs has been defined and the bottom-up approach established in the first cycle. To complete the GLOCAL approach an International Stakeholder Workshop was organized by UNESCO. During this reporting period, two cycles of workshops where fully completed and the third one is on-going, allowing the validation of the preliminary results of SHELTER.
•Existing technologies in OLs for the preparedness phase have been identified and analysed as a baseline for the development of the early warning model and the chatbot for crowdsourced data collection and social media data extraction service.
•Developments regarding co-creation approaches and adaptive governance have included the definition of the strategic blueprints for each Open Lab, the definition of the community-ICT interaction rulebook and the methodology for local knowledge extraction. The mapping of adaptive governance schemes is an on-going task. The Open Labs have begun to ‘self-diagnose’ weakness and opportunities in the DRM as a result of the work.
•The Project Web Site was established and the dissemination and communication plan was submitted and approved.
The foundation of the SHELTER framework has been defined, setting natural, tangible and intangible heritage as a central piece of the complex interactions between the resilience and the social, technical, ecological and economic forces. The definition of a multiscale data model will allow the spatial identification and analysis of the dynamics among these components and the designing of innovative strategies for resilience enhancement. The challenge lies in how to transform a large amount of existing information in knowledge. SHELTER has classified and evaluated data sources according to their compatibility with standards and the use of metadata (INSPIRE). Besides, an in-cloud Big Data storage solution is proposed with a data catalog for the efficient search of stored information (CKAN Data Catalog) and the multiscale data model is defined based on international standards, that facilitate the compatibility and durability of the proposed solution. SHELTER has developed a method for collaborative and multiscale characterisation of CH assets. It the next periods, it will be completed by the definition of the impact of natural hazards, the establishment of the constraints for the strategies applicability and identification of the assets contributing to inherent resilience. A cross-scale, multidimensional assessment framework based on indicators has been established. The innovation of this approach lies in linking physical vulnerability and risk management concepts and approaches with a more general and multidimensional resilience approach focused on the singularities of historic environments. The most common approach for rapid damage assessment in general and flood delineation is based on the use satellite imagery combined with supervised machine learning classifiers. The proposal in SHELTER is based on applying a traditional Machine Learning algorithm (Random Forest) and an adapted Deep Learning model (U-Net). Landsat data is mostly used for delineating burned areas, and the approaches combine expertly defined thresholds with classic machine learning algorithms. SHELTER exploits a subset of deep neural networks - convolutional neural networks (CNNs) - to find burned regions. Social networks and the analysis of the information generated allow obtaining useful information on natural disasters. Existing solutions are usually specific and tend not to generalize well and perform poorly when applied to data coming from different sources. SHELTER aims to study the performance of recent deep learning models for automatic analysis of images from multiple sources and classification applicable to multiple languages. SHELTER is implementing an innovative Open Labs approach based on a more horizontal, cross-scale resilience model through adaptive governance and collaborative thinking that places stakeholders in the centre of the process. The designed process allows the Open Labs the co-identification of problems and priorities, co-creation, test and pilot of solutions, co-monitoring and awareness-raising.
SHELTER framework
Crowdsourcing solutions for citizens engagement in preparedness and response
SHELTER cultural and natural macrocategories