The family of Kappa indices of agreement claim to compare a map's observed classification accuracy relative to the expected accuracy of baseline maps that can have two types of randomness: (1) random distribution of the quantity of each category and (2) random spatial allocation of the categories. Use of the Kappa indices has become part of the culture in remote sensing and other fields. This article examines five different Kappa indices, some of which were derived by the first author in 2000. We expose the indices' properties mathematically and illustrate their limitations graphically, with emphasis on Kappa's use of randomness as a baseline, and the often-ignored conversion from an observed sample matrix to the estimated population matrix. This article concludes that these Kappa indices are useless, misleading and/or flawed for the practical applications in remote sensing that we have seen. After more than a decade of working with these indices, we recommend that the profession abandon the use of Kappa indices for purposes of accuracy assessment and map comparison, and instead summarize the cross-tabulation matrix with two much simpler summary parameters: quantity disagreement and allocation disagreement. This article shows how to compute these two parameters using examples taken from peer-reviewed literature.

This paper compares two land change models in terms of appropriateness for various applications and predictive power. Cellular Automata Markov (CA_Markov) and Geomod are the two models, which have similar options to allow for specification of the predicted quantity and location of land categories. The most important structural difference is that CA_Markov has the ability to predict any transition among any number of categories, while Geomod predicts only a one‐way transition from one category to one alternative category. To assess the predictive power, each model is run several times to predict land change in central Massachusetts, USA. The models are calibrated with information from 1971 to 1985, and then the models predict the change from 1985 to 1999. The method to measure the predictive power: 1) separates the calibration process from the validation process, 2) assesses the accuracy at multiple resolutions, and 3) compares the predictive model vis‐à‐vis a null model that predicts pure persistence. Among 24 model runs, the predictive models are more accurate than the null model at resolutions coarser than two kilometres, but not at resolutions finer than one kilometre. The choice of the options account for more variation in accuracy of runs than the choice of the model per se. The most accurate model runs are those that did not use spatial contiguity explicitly. For this particular study area, the added complexity of CA_Markov is of no benefit.

This paper addresses two grand challenges in the development of methods for Geographic Information Science (GIS). First, this paper presents techniques to compute a cross‐tabulation matrix for soft‐classified pixels. Second, it shows how to compute the cross‐tabulation matrix at multiple scales. The traditional approach to construct the cross‐tabulation matrix uses a Boolean operator to analyse pixels that are hard‐classified. For soft‐classified pixels, the contemporary approach uses a Multiplication operator; the fuzzy approach uses a Minimum operator; whereas this paper proposes a multiple‐resolution approach that uses a Composite operator. There are difficulties with the traditional, contemporary, and fuzzy methods of computing the cross‐tabulation matrix. The proposed multiple‐resolution method resolves those difficulties. Furthermore, the proposed method facilitates multiple‐resolution analysis, so it can examine how results change as a function of scale. The paper derives the equations to compute cross‐tabulation matrices at multiple resolutions and connects those equations to ontological foundations of GIS.

We use remote sensing and GIS to map changes in land cover and to identify systematic land-cover transitions in Southwestern Ghana. Landsat Thematic Mapper satellite imagery of 1990 and 2000 is used to create two land-cover classifications, and the two maps are then compared to produce transition matrices both for protected and for unprotected areas. These matrices are analyzed according to their various components to identify systematic landscape transitions based on deviations between the transitions observed and the transitions expected owing to random processes of change. The results show that closed forest regions inside the protected area transition systematically to bare ground or bush fire, but closed forest outside the protected area transitions systematically to open cultivated woodland. These results are consistent with the hypothesis that logging is the main cause of the loss of closed forest inside the protected areas whereas farming is the main cause of the loss of closed forest outside the protected areas. The research highlights the need for the implementation of this methodological approach to landscape change. Identification of strong signals of forest transformation is particularly important in the light of efforts by policy makers to curb deforestation in Ghana.