Partitioning Spatial Model Uncertainty Based on Joint Spatial Simulation

Abstract

In this study, an uncertainty analysis procedure for joint sequential simulation of multiple attributes of spatially explicit models used in geographical informational systems was developed based on regression analysis. This procedure utilizes information obtained from joint sequential simulation to establish the relationship between model uncertainty and variation of model inputs. Using this procedure, model variance can be partitioned by model input parameters on a cell by cell basis. In the partitioning, the correlation of neighboring cells is accounted for. With traditional uncertainty analysis methods, this is not possible. In a case study, spatial variation of soil erodibility from a joint sequential simulation of soil properties was analyzed. The results showed that the regression approach is a very effective method in the analysis of the relationship between variation of the model output and model input parameters. It was also shown for the case study that: (1) the uncertainty of soil erodibility of a cell is mainly propagated from its own soil properties; (2) the interactions of soil properties of neighboring cells could reduce uncertainty of soil erodibility; (3) it is sufficient for uncertainty analysis to include the nearest three neighboring cell groups; and (4) the largest uncertainty contributors vary by soil properties and location.

Sensitivity analysis of a decision tree classification to input data errors using a general Monte Carlo error sensitivity model

“…Consequently, sensitivity and uncertainty analyses have been the subject of much attention in spatial and environmental sciences (Crosetto et al 2000, Crosetto and Tarantola 2001, Jager and King 2004. Sensitivity and uncertainty analyses have been used to assess model parameters (Hamby 1994, Hwang et al 1998, McKenney 1999, and input data, both continuous (Davis and Keller 1997, Wang et al 2000, Goovaerts 2001, Canters et al 2002, Gertner et al 2004 and categorical (Goovaerts 1996, Finke et al 1999, Canters et al 2002, Hines et al 2005.…”

“…Many sensitivity analysis methods have been proposed (see Hamby 1994 for a good review), some of which have been used by GIS researchers. They include the one-at-a-time method (McKenney 1999), variance based method (Finke et al 1999, automatic differentiation (Hwang et al 1998), and the regression method (Gertner et al 2004). Most of the existing sensitivity methods, however, are designed for outputs that are interval or ratio data types and are not suitable for categorical (nominal) data.…”

“…The other common approach is to use geostatistical methods to directly generate an effective input data surface from samples (e.g. Davis and Keller 1997, Atkinson 1999, Goovaerts 2001, Canters et al 2002, Gertner et al 2002, Gertner et al 2004. The advantage of the geostatistical approach is that it can take into account the spatial auto-correlation of a single attribute as well as spatial cross-correlation between different attributes.…”