Some types of measurements about biodiversity that can only be collected with field-based forest sampling may supply information complementary to that acquired remotely. However, field-based forest sampling, especially over large tracts of land, is an expensive and labour intensive exercise. The application of spatial statistics in this context enables the ability to estimate attributes that vary in space at unsampled places from more or less sparse sample data. The importance of defining objectives, understanding the attribute to be modelled, choosing the appropriate interpolation method, assessing the results is discussed and some of the statistical methods available to address each of these issues is presented Tree-species composition of the coastal and tablelands native forest estate managed by State Forests of New South Wales is used a case study, using data from native forest inventory. Examples of how the datasets can be used operationally are also presented and associated uncertainty is discussed.