Causes of the apparent scale independence of fractal indices associated with forest fragmentation in Bolivia

Urban influence on changes in linear forest edge structure

“…Fractal dimension also increased in forest patches adjacent to afforesting old fields (Narumalani et al, 2004). Several other studies have used area-based fractal dimension to quantify deforestation (Ewers and Laurance, 2006), forest fragmentation (Kojima et al, 2006), and forest structure (Imre and Bogaert, 2004;Kostylev et al, 2005;Motisi et al, 2004).…”

Scaling functions for landscape pattern metrics derived from remotely sensed data: Are their subpixel estimates really accurate?

“…Area-weighed mean patch fractal dimension (AWMFD), computed as an area-weighted mean of the fractal dimension index for each patch in the analysed class, where the patch fractal dimension equals two times the logarithm of patch perimeter (where the perimeter is divided by four to correct for the raster bias in perimeter) divided by the logarithm of patch area (McGarigal et al, 2002). AWMFD attains its minimum value (AWMFD = 1) for compact shapes and, as with the AWMSI, increases for more complex or elongated shapes (Kojima et al, 2006;Saura and Carballal, 2004). …”

“…Quantifying the downscaling accuracy through AI (instead of other simpler measures such as the standard deviation between Y act and Y est ) is particularly important and necessary in this context to take into account the different ranges of variation that each metric has in reality. These can be quite different depending on the analysed metric and often much narrower than its theoretical range of variation (Kojima et al, 2006;Saura and Martínez-Millán, 2001;Saura, 2002). For example, the subpixel estimate for a metric with a low variability in actual landscape patterns may, by being close to the actual target value (low relative error), provide a false impression of accuracy.…”

“…Numerous studies have analysed the effect of spatial resolution on landscape pattern metrics (Benson and MacKenzie, 1995;Frohn, 1998;Frohn and Hao, 2006;Frohn et al, 1996;García-Gigorro and Saura, 2005;Hlavka and Livingston, 1997;Kojima et al, 2006;Millington et al, 2003;Saura, 2001Saura, , 2004Turner et al, 1989;Wickham and Riitters, 1995;Wu et al, 2000Wu et al, , 2002Wu, 2004). It is well known that there are large differences in the values of the fragmentation indices derived from satellite images with different spatial resolutions, and the general trends of the variations of these metrics with spatial resolution have been characterised.…”

Riparian responses to extreme climate and land-use change scenarios

“…AWMPFD is a shape-category descriptor that measures the complexity of riparian patches by calculating a perimeter-area ratio. As a fractal dimension, it ranges between 1 and 2 and is scale-independent (Turner et al, 1989;Kojima et al, 2006). Low values are obtained when a patch has a compact and simple form with a small perimeter relative to the area.…”