Decision support tools for forestry using open source software
4 likes1,027 views
The document discusses decision support tools for forestry utilizing open-source software, emphasizing the importance of ecological site classification (ESC) in species selection and forest management. It highlights various technological frameworks and tools that assist in analyzing site properties and making informed decisions while noting challenges faced during implementation and the shift toward sustainable practices. The ongoing development of these systems aims to enhance user experience and expand functionality while accommodating a diverse user base in forestry.
Decision support tools for forestry using open source software
- 1. Decision Support Tools for Forestry using Open Source Software Stephen Bathgate, Andy Kennedy, Duncay Ray, Louise Sing and Phil Taylor 03 November 2016
- 2. 03/11/162 Introduction 1. A little bit of history on selecting tree species 2. Decision Support Systems for Forestry 3. Ecological Site Classification (ESC) 4. ESC4 Decision Support System 5. Conclusions and Questions
- 3. 03/11/163 FC Bulletin 30 (1957) • Climatic maps covering temperature, wind speed and moisture (rainfall and evapo-transpiration). • Site prescriptions according to bioclimatic zones.
- 4. 03/11/164 Some of the issues • During 60s and 70s the principle of matching species to site was turned around, and some sites were matched to species with ploughing and fertiliser. • The advent of Sustainable Forest Management, coupled with increasing costs, led to a reduction in fertiliser use in the 90s. Objectives changed away from production. • Skills and staff were lost. • Land available for afforestation was often degraded, poor choices about species potential arose. • The range of site types, tree species and potential future climates a forester might need to analyse required digital data and computer processing.
- 5. 03/11/165 ESC Bulletin 124 • Following research and development in the 1990s national scale guidance was published in 2001. • Was complemented by a computer based decision support system, field survey pack and training course. • ArcView extension. • Now embedded in many aspects of GB Forestry.
- 6. 03/11/166 Decision Support Systems 1.Tools/models that simplify a complex problem so that users can evaluate the relative merits of different management actions. 2. Intended to complement, not replace, expert/local knowledge. 3. Rubbish in, rubbish out. 4. Uncertainty/error/assumptions in models/data.
- 7. 03/11/167 What does a DSS do? • In this context 1. query the properties of a site (climate/soil) 2. analyse those properties with a scientific model 3. return the results to the user. • The climatic data is usually in raster format (250mx250m pixels). • Some models have spatial interactions (e.g. insects move according to their preferences for certain sites).
- 8. 03/11/168 Audience • Education/outreach • Researchers • Private sector forestry • Public sector forestry (operational/policy) With such a broad spectrum of users ESC tools have to be relatively agnostic to technology – web based tools and paper based systems.
- 10. 03/11/1610 Ecological Site Classification Climatic Factors -Temperature (AT5) -Continentality (CT) -Exposure (DAMS) -Moisture deficit (MD) Soil Factors - Soil Moisture Regime - Soil Nutrient Regime
- 11. 03/11/1611 Ardgarten Forest/Cobbler • Illustrate topographic limits on planting.
- 13. 03/11/1613 ESC Model AT CT DAMS MD SMR SNR Climatic factors (raster maps) Soil factors Species Model Yield Class Suitability.
- 14. 03/11/1614 Soil Map – New Forest
- 15. 03/11/1615 Pedunculate oak – New Forest
- 16. 03/11/1616 Guiding principles • When disseminating data/services: - avoid imposing technical requirements (and costs) upon end users (e.g. Windows only desktop tools). - ensure interoperability of geospatial data and services across platforms. • In ESC4 we use web map services (WMS), GeoJSON, CSV and other open data formats.
- 18. 03/11/1618 Technology Stacks Decision Tools Oracle App. Server Oracle Database (aspatial) ArcGIS + extensions Decision Tools PostGIS Geoserver R QGIS Geotools Tomcat
- 19. 03/11/1619 Open Source GIS R with raster plugins to generate national scale maps for species potential QGIS for visualisation, with qgis2threejs, sld4raster, quickmapservices Geotools provides set of libraries to work with spatial data (geotiffs/geojson) in Java OpenLayers3 for integration of spatial operations on the client side Geoserver to provide web map services to desktop and web based clients. Geomoose? Could this simplify the construction of this type of tool? PostGIS to store and manage spatial and aspatial data.
- 20. 03/11/1620 Design/Open Source Tools Linux Server Geoserver Geotools ESC Tomcat Geospatial Database PostGIS Web browser (HTML/JSON) Desktop GIS (WMS)
- 21. 03/11/1621 Use Cases • Stand (Site) scale (e.g. <15ha), point and click • Strategic scale, WMS maps, precalculated results from R scripts etc. • Operational scale with specific site data (e.g. soil survey)? Need to batch process large number of individual sites, e.g. to assess a whole forest block/design plan area. - Upload a geojson file based on a soil survey, process and send back to user so they can integrate with other design plan data.
- 22. 03/11/1622 ESC 4
- 24. 03/11/1624 ESC Site Analysis • To create an individual site analysis. - query 6-20 raster layers for point of interest. • Execute models based on site data. • Send results to user (web browser, file download depending on use case). • In use cases involving multiple sites this has to be repeated many times (including zonal stats/centroid derivation of mean attribute values for polygons). • PostGIS raster queries slower than Geotools, so far.. • Need to look at alternate approaches, e.g. rather than rasters create bioclimatic zones as vector data with attributes (loss of precision)? • Other ideas?
- 25. 03/11/1625 Suitability for Sitka spruce
- 26. 03/11/1626 Silver birch at Loch Ard
- 27. 03/11/1627 Douglas fir at Loch Ard
- 28. 03/11/1628 Strathyre – Wind risk
- 29. 03/11/1629 Next Steps • All very much a work in progress. • Refining user interface and working with user groups. • Adding other decision tools. • Optimisation of operational applications. • Looking at options for open source survey data collection e.g. QGIS extension. http://www.forestdss.org.uk/geoforestdss/esc4.jsp stephen.bathgate@forestry.gsi.gov.uk
- 31. 03/11/1631 Challenges (some historical) • Forestry Commission technical strategy based on Oracle technology and proprietary GIS. • Learning curve and range of open source GIS tools e.g. GRASS, Gvsig, QGIS + many others, Mapserver/Geoserver. What to choose? • Licensing constraints around certain tools and data. • Developing capacity/capability to support an open source infrastructure. • Range of user communities and their expectations. • Resistance to open source(or change). Security constraints, one tool per business requirement, concerns over integration and support. The creation of parallel open source environments was necessary to workaround certain barriers and prove the merits of open source.
- 32. 03/11/1632 Exposure (DAMS) • Wind exposure index. • Based upon topographic exposure (topex), elevation, valley funnelling effects and aspect. • FR conducted tatter flag surveys to build this data.
- 33. 03/11/1633 Decision Support Systems Ecological Site Classification Local Knowledge Professional experience Other evidence ESC is part of a decision making process, it is NOT the decision. This principle should apply across all decision support systems.