9528 ‎(2)‎

• UCREFRP Catalog Number: 9528
• Author: Martinez, P. and N. P. Nibbelink.
• Title: Colorado Nonnative Fish Stocking Regulation Evaluation.
• USFW Year: 2004.
• USFW - Doc Type: Grand Junction, Laramie.
• Copyright Material: NO

densities are locations that fall in the top 10% of density for the species being considered). This '
analysis was performed by first measuring the nearest distance from high density backwater
locations to stocking locations (or high density pond locations), then measuring the nearest ,
distance from backwater locations with the species absent to the same potential source locations.
Theoretically, if the high density backwater locations are closer to pond or stocking locations
than locations where the species was absent, this lends evidence to stocking events or pond fish
populations being sources of a particular species to critical habitat (Figures 2 and 3). Of course,
if there is no difference between these two proximity measurements, this does not mean that the
ponds or stocking events are not potential sources. Rather, the ponds and stocking events simply
are not spatially related in the way we may expect, and therefore are not likely to be identifiable
sources using spatial pattern analysis. The proximity analyses we performed were as follows:
1. Comparison of stocking events vs. high backwater densities.
2. Comparison of high density pond locations vs. high backwater densiries.
Geostatistical analyses '
We used three primary geostatistical approaches to evaluate further potential spatial
patterns between potential sources of nonnative fish in the ISA and nonnative fish densities in
critical habitat. Two of these approaches expand on analyses described above by focusing more
formally on explicit spatial relationships between sources and backwater populations. The third
method concentrates explicitly within the backwater populations to describe how these
populations are distributed in the ISA, and how their distribution may influence the identification
of nonnative fish sources and direct nonnative fish control efforts.
All three of the techniques we describe below and employ here are based on the fact that
the correlation between fish densities at any two locations may vary depending on the distance i
between those locations. For example, we may expect a riverine location that is close to a
floodplain pond to have similar densities of similar species, IF that pond is a potential source to
the riverine location. Similarly, fish densities at locations that are very far apart will not, on 1
average, be correlated to one another. These geostatistical approaches can be more powerful
than visual and proximity approaches because they consider all densities of fish (not just high
ones) and many inter-site distances (not just the closest locations). ,
1. The modified h-scattergram
The h-scattergram is a plot of all pairs of measurements of the same attribute (i.e.
fish density) at locations separated by a given distance (h) in a particular direction
(Goovaerts 1998). We have modified this approach here by plotting the fish density at a
backwater location vs. the sum of the density in ponds (or number of ponds contaimng
that species) within a specified buffer distance from the backwater location. This
analysis was performed by first buffering the backwater locations, then summing the
ponds and species-specific fish density within 250, 500, 750, and 1000 m buffers
extending out from the backwater locations. If fish densities at these backwater locations
are related to densities of fish or ponds at particular scales, there should be relationships

10 ,