7733 - Laserfiche WebLink

Home Browse Search

1 EXAMPLE A EXAMPLE B Mean BA Total Mean BA Total CPE Area Fish CPE Area Fish 5 100 500 5 200 1,000 5 500 2,500 10 100 1,000 5 1000 5,000 15 67 1,000 These examples depict two very different potential problems. Example A illustrates a situation where CPE does not identify a trend in fish abundance when one does exist. In contrast, CPE in example B shows a trend in abundance where such a trend does not exist. A plot of total fish vs mean CPE in the above examples would result in straight lines with slopes of either 1 or 0. Clearly these examples are over simplifications of a complex problem. They assume that all backwaters are equal and that Colorado squawfish do not discriminate among backwaters by water depth, water quality or clarity, or any other parameter that we can or can not measure. Further, they assume that Colorado squawfish are randomly distributed among all backwaters. River habitat and Colorado squawfish behavior are more complicated than these examples. More than likely, if mean CPE alone was not a valid index of total number of fish (and this estimate truly represents the actual number of Colorado squawfish), plots of estimated number of fish vs mean CPE would simply be scatters of points with no meaningful relationship. Plots of Gmean CPE and estimated number of fish in this study showed positive relationships between the two variables in most of the cases examined, although correlation coefficients and significance levels varied widely among years and reaches. The data plots were not random (which would indicate a potential problem); but they were not good linear relationships in all cases either. Because we do not have independent estimates of the total number of small fish in the study reaches, there is no way to determine whether one variable is more accurate than the other, or whether both are inaccurate. However, because backwater area varied during data collection and because there were apparent differences among investigators in identifying backwaters, we can assume that multiplying Total CPE by backwater area actually adds error to abundance indices in an unknown number of cases. Nonetheless, the two parameters do increase and decrease in relation with each other in many instances. This analysis suggests that variations in backwater area among years or sampling reaches is not a serious problem for ISMP because there were no dramatic shifts in 17