ABSTRACT
ABSTRACT Once established, spatial databases to support forest management need to be maintained to support current operations, to provide a record of past activities and to evaluate the possible outcomes of management decisions. This chapter describes the simulation of longterm spatial error propagation in a database covering a forest area in Austria. Various scenarios of error propagation were produced using Monte Carlo simulation. Arcs in each of the five states of the database were perturbed with different levels of error, depending on both the type of feature and the date of data acquisition. In this way, the simulation modelling not only reflected the degree of uncertainty associated with different linear features but also accounted for the likelihood that surveys on separate occasions may be of different quality. Results of this simulation modelling indicate that an initially high quality database can become significantly degraded if it is amended with low quality data. The amendment of an initially low quality database with high quality data, however, provides a comparatively small improvement in the overall quality of the database. The chapter concludes by demonstrating how the benefits obtained from improving data quality may be assessed in relation to the cost of database maintenance.
