ABSTRACT
This chapter discusses the effects of data aggregation on modeling, in an attempt to contribute to the research of scaling up environmental models that use both remotely sensed and GIS data. The effects of aggregation methods, the multiple scale nature of spatial data, and the role of semivariance-based fractals in forecasting scale effects are then discussed. Differences can be considerable among these methods if applied to scaling up environmental models. The multiscale nature of the biomass image reflects an assemblage of environmental processes operating simultaneously but at different scales (operational scale). This complex, multiple scale pattern characterizes spatial data and particularly remotely sensed images. To scale up environmental models, spatial variation of data values is the primary concern; a semi variance-based fractal dimension is thus the most appropriate because it measures how data variation changes with sampling interval.
