Combining Inventories of Land Cover and Forest Resources with Prediction Models and Remotely Sensed Data
It is difficult to design systems for national and global resource inventory and analysis that efficiently satisfy changing, and increasingly complex objectives. It is proposed that individual inventory, monitoring, modeling, and remote sensing systems be specialized to achieve portions of the objectives. These separate systems can be statistically linked to accomplish all objectives, which might simplify statistical designs and improve efficiency and timeliness. This might be accomplished using an estimation technique known as the Kalman filter, which combines a theoretical or empirical prediction model with several multivariate time series of statistical estimates and remotely sensed data. An example is used to demonstrate one application that monitors land cover.
Czaplewski, Raymond L. 1989. Combining inventories of land cover and forest resources with prediction models and remotely sensed data. In: Lund, H. Gyde; Preto, Giovanni, technical coordinators. Volume 3, Global Natural Resource Monitoring and Assessment: Preparing or the 21st Century. Proceedings of the International Conference and Workshop; September 24-30; Venice, Italy. Bethesda, MD: American Society for Photogrammetry and Remote Sensing. p. 1079-1089.