Abstract
Improved information about the forest will create new possibilities for efficient forest biomass production with due consideration to other ecosystem values. This project aims to develop data a new paradigm for keeping forest databases up to date by utilising data assimilation for incorporating the increasing flow of sensor data. In this data assimilation application a geographical model with forest data will be forecasted with growth functions. When new data from remote sensing, field inventories, harvesters, or field based sensors become available, data assimilation techniques such as Kalman filtering or Bayesian statistics will be used to update the forest model in proportion to the information value of the new data. In the near future, optical satellite data will be available every few weeks and canopy height models from digital photogrammetry will be available with a few years interval. These data can be used to continuously update forest information, as opposed to static maps. Development of the data assimilation paradigm will be carried out at test sites where time series are available with field reference data, airborne and terrestrial laser scanning, optical satellites, digital aerial photogrammetry, interferometric SAR, and harvester data. In addition, a case study where data assimilation will be used to update forest databases made from five year old airborne laser scanner data will be carried out in cooperation with forest companies and the relevant authorities.
Sweden