Delivering Web based Spatial Decision Support Tools with Open Source Technology

02
May
05.02.2016 |
stephenb
Stephen Bathgate's picture

Forest Research has been developing a new version of its decision support platform using a suite of open source geospatial web technologies to support a broad range of users across the forestry sector, primarily in Great Britain. Since 2007 web based services have been used to deliver decision support tools to a range of stakeholders including the private and public forestry sector, researchers and educational users.

The need to replace previous versions of decision support tools arose from the requirement to provide spatial context alongside stand level models.1 This was particularly important with Ecological Site Classification (ESC), a Forest Research decision support tool that advises on tree species choice according to climatic and edaphic(soil) variables. The legacy stand level tool did not explicitly indicate national or local topographic influences upon ESC outputs, due to the technology used.2

With the emergence of spatial requirements, a new solution was required. Given the small scale of the project, control of costs was a major driver, but also was the need to maintain delivery to a broad range of users with different computing infrastructure. Open source software offers developers the ability to minimise licensing costs associated with the delivery of deployment while providing industry strength solutions and compliance with open data standards. The use of open geospatial data standards, such as web map services (WMS), facilitates delivery of information not only to web based users (Figure 1), but also those using desktop geospatial information systems (GIS).

Douglas Fir climatic suitability according to Ecological Site Classification

 

 

 

 

Among the open source technologies used are Openlayers, Geoserver, Geotools and PostGIS.3 Openlayers provides the interactive map elements, PostGIS manages spatial data while Geoserver delivers raster and vector maps via web based services. The Geotools libraries provide a means to query and manipulate spatial data alongside model (e.g. ESC) codes.  In the preparation of spatial datasets the R statistics packages provides integration with geospatial libraries (e.g. GDAL) enabling automation of data preparation. QGIS, an open source desktop GIS package, is also used where simple geoprocessing tasks are necessary. All the development and deployment is undertaken using open source operating systems and development tools.

Although the costs of licensing have been eliminated, there have been costs associated with establishing cloud based hosting capabilities, and developing the necessary skills to use the open source GIS toolkits. There are some areas where open source licensing requirements prevented certain tools being used, for instance where the code was free for development but not deployment.

Currently the legacy and successor versions of the decision support portal are running in parallel as some functions have yet to be migrated. This has provided a means to quantify relative use of an aspatial versus spatial system. From September 2015 to April 2016 the new version provided around 12000 map requests and over 8000 site analyses - in contrast the legacy version provided just over 3000 site analyses. Direct feedback also favoured the open source spatial system, and several forest management organisations have integrated the web map services in their business systems, owing to the interoperable services provided by the tools.

[1] Ecological Site Classification version 4, http://www.forestdss.org.uk/geoforestdss

[2] Ecological Site Classification version 3, http://www.eforestry.gov.uk/forestdss

[3] OSGeo, http://www.osgeo.org