Invader removal triggers competitive release in a threatened avian predator

Abstract

Changes in the distribution and abundance of invasive species can have far-reaching ecological consequences. Programs to control invaders are common but gauging the effectiveness of such programs using carefully controlled, large-scale field experiments is rare, especially at higher trophic levels. Experimental manipulations coupled with long-term demographic monitoring can reveal the mechanistic underpinnings of interspecific competition among apex predators and suggest mitigation options for invasive species. We used a large-scale before-after control-impact removal experiment to investigate the effects of an invasive competitor, the barred owl (Strix varia), on the population dynamics of an iconic old-forest native species, the northern spotted owl (Strix occidentalis caurina). Removal of barred owls had a strong, positive effect on survival of sympatric spotted owls and a weaker but positive effect on spotted owl dispersal and recruitment. After removals, the estimated mean annual rate of population change for spotted owls stabilized in areas with removals (0.2% decline per year), but continued to decline sharply in areas without removals (12.1% decline per year). The results demonstrated that the most substantial changes in population dynamics of northern spotted owls over the past two decades were associated with the invasion, population expansion, and subsequent removal of barred owls. Our study provides experimental evidence of the demographic consequences of competitive release, where a threatened avian predator was freed from restrictions imposed on its population dynamics with the removal of a competitively dominant invasive species.

Keywords: Strix owls; competition; invasive species; population dynamics; removal experiment.

Fig. 1.

Before–after control–impact experiment used to estimate the demographic response of threatened northern spotted owls to removal of invasive barred owls. (A) Adult northern spotted owl (S. occidentalis caurina) and (B) adult barred owl (S. varia). Owl images credit: Patrick Kolar (photographer). (C) Overlap between the geographic ranges of northern spotted owls (US range in orange) and barred owls (hatched-blue) in North America. Ranges were approximated from https://birdsoftheworld.org/bow. Barred owls were historically limited to eastern North America. (D) Locations of treatment (barred owls removed) and control (no barred owls removed) segments of five long-term experimental study areas within the range of the northern spotted owl in Washington, Oregon, and California.

Fig. 2.

Estimated effects of barred owl removal on survival and dispersal movements of northern spotted owls in each of five individual study areas in Washington, Oregon, and California. (A) Weighted mean estimates of apparent survival of northern spotted owls before and after barred owls were removed. Mean survival was estimated separately for treatment (T; orange) and control (C; blue) areas in each time period using the best multistate model that included an effect of barred owl removal; estimates were weighted by the inverse of the variance of annual estimates. (B) Mean change in apparent survival attributable to barred owl removal in each study area, calculated using Eq. 2 with estimates shown in A as the mean difference in survival between control and treatment areas before and after removals. Error bars in A and B are 95% CI. (C) Movement probability of resident northern spotted owls from territories in the control area to territories in the treatment (removal) area before versus after removals occurred. No movement was detected on the HUP-WC study area (i.e., movement probability = 0).

Fig. 3.

Meta-analysis of estimated effects of barred owl removal on long-term population dynamics of northern spotted owls in five study areas in Washington, Oregon, and California. We show annual estimates of apparent survival, recruitment, and the rate of population change for treatment (barred owls removed) and control (no removal) areas. Estimates are from the best meta-analysis model with all study areas combined. Shaded regions represent 95% CIs; solid vertical lines indicate the start-date of barred owl removal on treatment areas.

Fig. 4.

Mean change in vital rates and population trends of northern spotted owls that was attributable to barred owl removal based on a meta-analysis of five experimental study areas in Washington, Oregon, and California. We show estimates of mean effect size for (A) apparent survival, (B) recruitment, and (C) rate of population change of northern spotted owls. Mean effect size was calculated using Eq. 2 and weighted geometric means of before–after, control–treatment estimates from the best meta-analysis model. Red-dashed vertical lines indicate mean effect sizes across the five study areas. Positive values with a 95% CI that did not overlap zero (black-dashed vertical line) indicate strong evidence that removals increased a given vital rate in the treatment (barred owl removal) area relative to the control area.