Departmental Research Projects
British National Space Centre (BNSC) Customer Partner Project
A research project of the Department of Geography.
Monitoring of woodland establishment and growth for the UK forest industry.
The British National Space Centre uses Customer Partnership Projects to encourage the business sector to make use of Earth observation (satellite) data by part funding the development of a demonstration system to help address a specific commercial problem. Projects are short term, normally six months to one year, and are awarded on a competitive basis after peer review and government scrutiny. This project arose because the UK forest sector makes little use of satellite data for resource management purposes compared to many other countries. Completed in 2002, the partnerships developed in the BNSC project feed into the ForestSAFE project running at present.
The aim of this project was to demonstrate how satellite imagery can improve a discrete but important part of the forest industry's ongoing business needs; the process of monitoring the establishment and growth of new plantation forestry. The project integrated innovative techniques that combine satellite imagery, ground survey and GIS-data to address an important aspect of commercial forestry operations. The CPP aimed to deliver a prototype software system that was fully compatible with the Forestry Commission's (FC) own GIS to complement its forest survey and monitoring work.
This Customer Partnership Project has demonstrated the potential of medium resolution (20-30m pixel size) optical satellite imagery to monitor the establishment and growth of Sitka spruce (Picea sitchensis) plantations in Britain. The project has shown that tree growth can be monitored remotely in order to identify stands that fail to establish as commercially viable crops. The mechanism for monitoring establishment involves predicting key forest growth parameters from satellite imagery reflectance and comparing this with field observations. The first stage in this process is to model the relationship between forest structure and reflectance data from the Landsat and SPOT satellites. This was investigated for 24 stands of Sitka spruce ranging in age from 2-17 years. Regression models showed that average tree height and basal area are most strongly correlated with reflectance data (R2=0.86, P<0.01 and R2=0.61, P<0.01). Stem density could not be predicted satisfactorily. The models derived for Landsat 7 ETM+ and SPOT 4 HRV data were almost identical.
Independent field validation of these models indicate that the Standard Error (SE) of crop height predictions is 1.5m. The project has shown that tree growth can be monitored, and that an appropriate level of ground survey data required to calibrate measurements derived from satellite image data is one ground survey plot per 200 hectares of forest. Furthermore, the satellite derived data can easily be integrated into a GIS (Geographical Information System) allowing foresters to interrogate and visualise the information over large areas.
Four demonstration products were developed for the Forestry Commission to evaluate. Two of these, tree height and tree height variability prediction and a Forest change data layer, are based upon satellite image data. Feedback from potential end users indicated that these products directly addressed business needs and could be used operationally with little modification. The other two products that were demonstrated included a digital field based forest inventory system and a GIS user interface. Feedback on these products indicated that more work would be required to tailor these to work with existing computer systems and business practices.
For further information, please contact Professor Danny Donoghue.