Fractal modelling of the remotely sensed two-dimensional net primary production pattern with annual cumulative AVHRR NDVI data

A major challenge facing ecologists and environmental managers studying the Earth as a global system is the quantitative description of broadscale net primary productivity (NPP) patterns. For instance, at broad spatial scales, direct estimations of NPP obviously cannot be considered and analysis of remotely sensed data appears to be the appropriate tool. In this letter, we show a method based on fractal statistics to measure the two-dimensional spatial pattern of broad-scale remotely sensed net primary production values. The results of applying this method on annual cumulative NDVI data (SigmaNDVI) obtained from the AVHRR sensor for a portion of western United States show that fractal analysis can be used to summarize the self-similar spatial pattern of broad-scale SigmaNDVI values over a large spatial range.