A weighted normalized likelihood procedure for empirical land change modeling
- 13 Downloads
A critical foundation for empirical land change modeling is the mapping of transition potentials—quantitative evaluations of the readiness or suitability of land to go through a transition. This paper presents a procedure based on empirically determined normalized likelihoods of transition. It shows that these normalized likelihoods equate to posterior probabilities if case–control sampling is carried out among historical instances of change and persistence. The posterior probabilities can then be aggregated at the pixel level across multiple covariates using linear opinion pooling where the pixel-specific weight for each covariate is determined locally by its ability to distinguish between the alternatives of change or persistence. Thus, covariates with spatially varying diagnostic ability can be productively incorporated. The resulting algorithm is shown to have a skill comparable to that of a multi-layer perceptron approach with the advantage of high efficiency and amenability to distributed processing in a cloud environment.
KeywordsLand-use change Land-cover change Model development Model evaluation Empirical modeling Machine learning
This research was conducted as part of a grant from Esri Inc. to develop a cloud-based implementation of Clark Labs’ Land Change Modeler. We are very grateful to Esri for this support.
- Angelsen A, Boucher D, Brown S, Valerie M, Streck C, Zarin D (2011) Modalities for REDD+ reference levels: technical and procedural issues. The Meridian InstituteGoogle Scholar
- Eastman JR (2006) IDRISI 15.0: The, Andes edn. Clark University, WorcesterGoogle Scholar
- Eastman JR (2014) The TerrSet geospatial monitoring and modeling system. Clark University, WorcesterGoogle Scholar
- Eastman JR, Solorzano LA, Van Fossen ME (2005) Transition potential modeling for land-cover change. In: Maguire DJ, Batty M, Goodchild MF (eds) GIS, spatial analysis, and modeling. ESRI Press, California, pp 357–385Google Scholar
- Heidke P (1926) Berechnung des erfolges und der gute der windstarkvorhersagen im sturmwarnungsdienst. Geogr Ann 8:301–349Google Scholar
- Jordan MI (1995) Why the logistic function? A tutorial discussion on probabilities and neural networks, MIT computational cognitive science, MIT Computational Science Report 9503Google Scholar
- Platt J (2000) Probabilistic outputs for support vector machines and comparison to regularized likelihood methods. In: Smola A, Bartlett P, Scholkopf B, Schuurmans D (eds) Advances in large margin classifiers. MIT Press, CambridgeGoogle Scholar
- Pontius RG Jr (2000) Quantification error versus location error in the comparison of Tso, B., and Mather, P.M., 394 (2011) categorical maps. Photogram Eng Remote Sens 66(8):1011–1016Google Scholar
- Pontius RG Jr, Cornell JD, Hall CAS (2001) Modeling the spatial pattern of land use change with GEOMOD2: application and validation for Costa Rica. Agric Ecosyst Environ 1775:1–13Google Scholar
- Rish I (2001) An empirical study of the naïve Bayes classifier. IBM Research Report RC 22230Google Scholar
- Verified Carbon Standard (2013) VCS module VMD0007 REDD methodological module: estimation of baseline carbon stock changes and greenhouse gas emissions from unplanned deforestation (BL-UP) version 3.1. http://database.v-c-s.org/sites/vcs.benfredaconsulting.com/files/VMD0007%20BL-UP%20v3.2.pdf. Accessed 8 Dec 2018