When a cumulative number-size distribution of data follows a power law, the
data set is often considered fractal since both power laws and fractals ar
e scale invariant. Cumulative number-size distributions for data sets of ma
ny natural phenomena exhibit a "fall-off" from a power law as the measured
object size increases. We demonstrate that this fall-off is expected when a
cumulative data set is truncated at large object size. We provide a genera
lized equation, herein called the General Fitting Function (GFF), that desc
ribes an upper-truncated cumulative number-size distribution based on a pow
er law. Fitting the GFF to a cumulative number-size distribution yields the
coefficient and exponent of the underlying power law and a parameter that
characterizes the upper truncation. Possible causes of upper truncation inc
lude data sampling limitations (spatial or temporal) and changes in the phy
sics controlling the object sizes. We use the GFF method to analyze four na
tural systems that have been studied by other approaches: forest fire area
in the Australian Capital Territory; fault offsets in the Vernejoul coal fi
eld; hydrocarbon volumes in the Frio Strand Plain exploration play; and fau
lt lengths on Venus. We demonstrate that a traditional approach of fitting
a power law directly to the cumulative number-size distribution estimates t
oo negative an exponent for the power law and overestimates the fractal dim
ension of the data set. The four systems we consider are well fit by the GF
F method, suggesting they have properties characterized by upper-truncated
power laws.