Search USGSSearch

Geology topics

Lucas K. Kowalewski

Publications and source records attributed to Lucas K. Kowalewski.

2 recordsLinked to original sources

Detection probabilities of Flathead Catfish in small Kansas impoundments

A primary challenge of Flathead Catfish Pylodictis olivaris management is uncertainty associated with sampling strategies and resulting ambiguity in population-level information. Assessment of impoundment and environmental conditions that affect detection probability may aid in reducing sample variance and benefit inferences regarding changes to Flathead Catfish populations. We sampled eight small impoundments in Kansas (37–114 surface ha) using low-frequency electrofishing in summer, 2021. We revisited sites nine times over three months using an occupancy modeling framework to estimate the influence of impoundment and environmental conditions on detection probability of Flathead Catfish. We employed an information theoretic approach and ranked models built with impoundment as a random effect and three environmental variables predicted to influence detection of Flathead Catfish in small impoundments. Detection probability across all populations was 0.526 (SE = 0.020) and was influenced by water temperature, mean depth of the impoundment, and proportion of impoundment sampled. Generally, detection probability increased with all measured variables. The inclusion of detection probability in assessments of Flathead Catfish in small impoundments can inform interpretation of catch-related metrics. Further, variable detection suggests collection of multiple samples during a defined sampling period might be more suitable for characterizing populations than a single sample.

Kansas

Accuracy or precision: Implications of sample design and methodology on abundance estimation

Sampling by spatially replicated counts (point-count) is an increasingly popular method of estimating population size of organisms. Challenges exist when sampling by point-count method, and it is often impractical to sample entire area of interest and impossible to detect every individual present. Ecologists encounter logistical limitations that force them to sample either few large-sample units or many small sample-units, introducing biases to sample counts. We generated a computer environment and simulated sampling scenarios to test the role of number of samples, sample unit area, number of organisms, and distribution of organisms in the estimation of population sizes using N -mixture models. Many sample units of small area provided estimates that were consistently closer to true abundance than sample scenarios with few sample units of large area. However, sample scenarios with few sample units of large area provided more precise abundance estimates than abundance estimates derived from sample scenarios with many sample units of small area. It is important to consider accuracy and precision of abundance estimates during the sample design process with study goals and objectives fully recognized, although and with consequence, consideration of accuracy and precision of abundance estimates is often an afterthought that occurs during the data analysis process.

Ecological Modelling