Abstract Changes in the spatial distribution of land cover and land use can have significant impacts on ecological processes at multiple scales; estimating these changes provides critical data for both monitoring and understanding land-use effects on these processes. One approach to mapping landcover changes, particularly useful over longer periods of time, is comparison of existing landcover maps, (post-classification change analysis). The accuracy of these maps is often unknown and varies depending on data sources and interpretation techniques; therefore, separating change on the ground from differences attributable to sensors and methods is both critical and problematic. Through a novel map comparison method applying major axis regression at multiple spatial grains of analysis, this study partitioned accuracy into components of bias and precision in comparing maps, which aided selection of an optimal analytical grain size. Comparisons between contemporaneous maps showed the magnitude and distribution of error alone, while between-period analyses indicated both cumulative map error and change on the ground. These methods enable exploration of the nature of error and identification of differences between maps, while accounting for the imprecision and bias inherent in the source documents. Mapping landcover change delineates landscapes under recent disturbance pressure, and these measures are more effective as performance indicators for broad-scale evaluation of natural heritage policies and habitat restoration initiatives when error in the data is identified and accounted for.