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

1
~J

Visual pattern detection
Visual pattern detection involves primarily the use of maps and symbology that carefully
represents point locations and values. By using, for example, symbols whose size are related to
the fish density at a particular location, we can detect whether or not there are clusters of
locations that have high values for a particular species. Sometimes the visual patterns observed
can lead directly to spatial anomalies that are very useful in identifying problem locations, or
developing further testable hypotheses (Nibbelink and Rahel, in review). Furthermore, by
combining point pattern maps of riverine densities with those offish stocking densities or
floodplain pond fish densities, we can potentially see whether high densities from particular
sources maybe related to high densities of nonnative fishes in critical habitat. This approach
looked for clusters that might indicate potential sources of nonnative fishes to backwaters within
the ISA. Furthermore, we looked at whether treated ponds related to lower densities of
nonnative fishes in backwaters. The analyses conducted were as follows:
1. Comparison of stocking (all known stocking events) densities vs. backwater fish densities.
2. Comparison of treated ponds to backwater fish densities.
3. Comparison of nonnative fish densities in ponds vs. backwaters.
4. Nonnative fish densities in backwaters, pre- and post-regulation periods.
It was not possible to simply pool and compare all of the fish data from backwater
sampling due to different sampling protocols (Table 1), so we had to choose carefully when
using data to reflect species composition and density in the ISA. Most of these data were
collected using river miles or kilometers as indications of location. Therefore, data for species
and number were linked in the GIS to the nearest 0.1-river mile designation. To get a robust
picture of the density of species in the riverine backwaters, we used only the data sets from
Bundy and Bestgen (2001) and Trammel et al. (2002) as their methodologies were the most
thorough and similar. To further facilitate comparisons among these data sets, only the first-pass
removals were utilized due to the different levels of effort expended by each project.
We mapped nonnative fish densities for each of these sampling programs to visualize
hotspots for nonnative species in critical habitat, and to describe the relative densities of each for
select species of interest. Three nonnative species were of particular interest in backwaters, due
to their prevalence. These included largemouth bass and fathead minnow, which can be stocked
under the new regulations, and one species that is not stocked, green sunfish. The inclusion of
green sunfish was due to several factors: 1) these fish are ubiquitous and appear to be self-
sustaining in a variety of stream sizes and habitats, and 2) there has been recent interest in the
concept of concentration areas for centrarchids, referred to as "hot spots", within the Grand
Valley reach of the Colorado River (Kesler 2002), and 3) there is some evidence that green
sunfish share the top trophic position in backwater food webs in the Grand Valley along with
largemouth bass and black crappie (Martinez et al. 2001).
Proximity analyses
We next looked more formally at whether high densities of nonnative fishes in
backwaters were closer than expected to stocking locations or high densities in ponds ("high"