Calculation of Upper Confidence Bounds on Proportion of Area containing Not-sampled Vegetation Types: An Application to Map Unit definition for Existing Vegetation Maps

Abstract

This paper explores the information forest inventory data can produce regarding forest types that were not sampled and develops the equations necessary to define the upper confidence bounds on not-sampled forest types. The problem is reduced to a Bernoulli variable. This simplification allows the upper confidence bounds to be calculated based on Cochran (1977). Examples are provided that demonstrate how the resultant equations are relevant to creating mid-level vegetation maps by assisting in the development of statistically defensible map units.