Sampling methodology tradeoffs: evaluating monitoring strategies for the endangered humpback chub (Gila cypha) in the Little Colorado River, Arizona
Pearson, Kristen Nicole, author
Kendall, William, advisor
Winkelman, Dana, committee member
Hess, Ann, committee member
Implementation of a reliable monitoring program is essential to informed population management. When recovering a sensitive species, priority should be on minimizing human induced negative effects, given already reduced population abundance. Thus, it is crucial to evaluate monitoring programs and make changes when more efficient techniques become available. To assess tradeoffs in sampling effort first necessitates obtaining accurate demographic parameter estimates. However, obtaining such estimates may be challenging especially when assessing a migratory species monitored on its spawning ground. Due to concerns regarding sampling availability, in such cases, it may be necessary to evaluate temporary emigration from the study site to avoid generating biased estimates of survival, detection and spawning probabilities. Evaluating temporary emigration is especially important when non-annual spawning is anticipated, as skipped spawners may be unavailable for detection during annual sampling events. Since the late 1980s, population monitoring for the potamodromous humpback chub (HBC) Gila cypha within the Lower Colorado River Basin (LCRB) has focused on hoop-net sampling within their primary spawning ground, the Little Colorado River (LCR). However, questions remain unanswered regarding their spawning strategy. Thus, due to the likely presence of both resident and migratory fish and suspected non-annual spawning, I evaluated temporary emigration from the LCR, which I equate to skipped spawning. Using, robust design mark-recapture methodologies, I was able to generate unbiased estimates of survival and skipped spawning probabilities as well as spawner abundance. Given concern for handling induced stress due to intensive hoop-net sampling and to gain additional insight into HBC life history strategies and population dynamics, in 2009, a passive detection system (i.e. full duplex PIT tag antenna array) was implemented in the LCR. With the addition of the array, this afforded an opportunity to evaluate sampling methodology tradeoffs between hoop-netting and array detections. Thus, using simulation analysis, and demographic parameter estimates generated from my skipped spawning analysis, I assessed the potential benefits and shortcomings of reducing hoop-net sampling effort and supplementing recapture data with passive array detections. From my analysis, I found considerable evidence for skipped spawning among both male and female HBC. Females on average had a higher probability of failing to spawn in a year subsequent to spawning (i.e. γ"male = 0.46 (95% credible interval [CRI]: 0.11, 0.81) and γ"female = 0.55 (95% CRI: 0.30, 0.75), although better sexing data is necessary to confirm this difference. Annual variability in skipped spawning probability was high (i.e. process variance (σ2) = 0.306) while survival probability remained stable throughout the study period (i.e. S = 0.75 (95% CRI: 0.66, 0.82), σ2 = 0.005). Based on my most reliable skipped spawning probability estimates, (i.e. probability a spawner transitions to a skipped spawner (γ") = 0.45 (95% CRI: 0.10, 0.80) and a skipped spawner remains a skipped spawner (γ') = 0.60 (95% CRI: 0.26, 0.83)) which exclude sex, I found HBC in the LCRB have an average breeding cycle of every 2.12 years, conditional on survival. By employing these estimates in simulation analysis, I found that hoop-net sampling can be reduced and supplemented with array detections without negatively affecting estimability of adult HBC survival and skipped spawning probabilities, given detection efficiency of the array remains sufficiently high. Because the array provides insight outside of traditional sampling periods and does not require repeated handling of this imperiled fish, it affords a viable means of reducing hoop-net sampling effort, thus, offering a potentially more efficient monitoring strategy.
Includes bibliographical references.
Includes bibliographical references.