Analysis of land cover change is one of the major challenges in the remote sensing and GIS domain, especially when multi-temporal or multi-sensor analyses are conducted. One of the reasons is that errors and inaccuracies from multiple datasets (for instance caused by sensor bias or spatial misregistration) accumulate and can lead to a high amount of erroneous change. A promising approach to counter this challenge is to quantify and visualize uncertainty, i.e.
, to deal with imperfection instead of ignoring it. Currently, in GIS the incorporation of uncertainty into change analysis is not easily possible. We present a concept for uncertainty-aware change analysis using a geovisual analytics (GVA) approach. It is based on two main elements: first, closer integration of change detection and analysis steps; and second, visual communication of uncertainty during analysis. Potential benefits include better-informed change analysis, support for choosing change detection parameters and reduction of erroneous change by filtering. In a case study with a change scenario in an area near Hamburg, Germany, we demonstrate how erroneous change can be filtered out using uncertainty. For this, we implemented a software prototype according to the concept presented. We discuss the potential and limitations of the concept and provide recommendations for future work.