“Nature-based solutions” are actions inspired by nature which use the features and complex system processes of nature, such as its ability to store carbon and regulate water flows, in order to help societies address a variety of environmental, social and economic challenges in sustainable ways. Nature based solutions examples are green roofs, green walls, urban farms, etc.

Knowing the capacity of a city to host these types of nature based solutions allows defining what adaptation actions are available in the municipality, which needs to be modified, new areas that could accommodate these solutions and finally identify what their potential for climate change adaptation with such solutions.

Tecnalia Research & Innovation, with experience in the specific subject of climate adaptation, has developed a methodology to identify and map NBS in municipalities through the call KLIMATEK I + B + G: Projects R & D, Innovation and demonstration in climate change adaptation 2016.

In order to give a demonstrative character to the project, the methodology has been developed applying it to the municipality of Donostia/San Sebastián. Most of the NBS mapped in Donostia have been identified by the land uses established by the General Plan for Urban Planning and cartography available at the municipal and Basque Government level.

However, a major challenge in this type of project is to categorize, catalog and characterize the resources and capacities available in each municipality. Usually, this information has to be manually extracted by expert technicians and, in some cases; the temporal variation of the resources makes impossible to monitor their actual status. In conjunction with the Energy and Environment division, Tecnalia’s Computer Vision team has developed algorithms for the identification, classification and temporal monitoring of these assets based on information provided by aerial and satellite imagery.

One of the use cases analyzed corresponds to the identification and classification of building roofs. Green roofs provide multiple benefits as reducing the risk of flooding by collecting rainwater. This can reduce the ambient temperature, improves energy efficiency in buildings and the social benefits associated with urban agriculture. In addition, these types of interventions on roofs are easy to implement by the placement of trees in pots, tables, etc.

To accurately quantify the urban assets capabilities, it is crucial to identify which buildings have flat roofs. The information available for this identification is the own cartography of buildings of the municipality, which does not contain information on the height or inclination of the roof, and satellite images of ortho-photos or free tools. This information only allows the identification of flat roofs visually, making it a tedious, probably imprecise and unfeasible task in time and resources.

The developed algorithm performs a cataloging of each of the buildings based on two criteria:

  • Real slope of the roof.

  • Classification of the albedo based on the color of the roof.

First, each cadastral unit in the ortho-photo image is segmented. For each unit (building) located, an analysis of the slope is performed on the basis of the LIDAR signal by measuring the actual slope at each point of the roof. The slope map of the building is analyzed using a statistical algorithm that gives us the estimation of the real slope and the flat area available. Buildings with a suitable flat surface are labeled as flat roof by noting their flat zone percentage. Additionally, reflectance parameters of each roof are analysed to catalog them according to their energy efficiency (roof tile, dark roof, clear roof). The classification error, in the pilot case of the city of Donostia / San Sebastián, is less than 2% in the case of roof slope and 4% in the color roof characterization.

The LIDAR information is available for the whole Basque Autonomous Community, so this methodology of flat roofs identification could be applied in any municipality in a simple and fast way.

The use of available GIS and image ( Public Open Data) together with the use of image and data analysis algorithms allows, not only the automation of the cataloging processes but also the generation of automated services that were not affordable such as:

  • Automated Services on request.

  • Automatic resources classification.

  • Evaluation of the resource quality / deterioration.

  • Monitoring and immediate detection of anomalies.

In this sense, TECNALIA is developing different algorithms to properly identify and catalog different land uses, vegetation types, building material segmentation, as well as the temporal monitoring of changes.

Identification of flat surface (pixels) in urban areas with typical structure of gable roofs (reddish: sloped surgace, greenish/bluish: flat roof)

Identification of flat surface (pixels) in industrial zone (yellow-green: flat surface, reddish: sloped surface)