Habitat modification experiment failed to find evidence for crested auklet population enhancement: A response to Divoky

Divoky (2018) commented negatively on our paper (Major et al. 2017) concerning our before-after control-impact (BACI) study to test whether habitat modification could be an effective means of restoring nesting habitat of crested auklets (Aethia cristatella) breeding on Gareloi Island, Alaska, USA. Our study arose from a possible requirement to replace crested auklets killed in the Selendang Ayu oil spill (National Oceanic and Atmospheric Administration 2015). Our criteria for assessing restoration of crested auklets were rigorous and experimental: we quantitatively measured crested auklet numbers via counts of birds visible in crevices, and calculated capture-mark-recapture estimates of numbers of individuals attending study plots and delivering chick meals. The salient conclusion of our paper was that, based on experiments at Gareloi during 2009–2013, we did not find evidence of a positive effect of vegetation modification on crested auklet numbers. Divoky (2018) states that he disagrees with this conclusion. Fortunately, our fieldwork was designed to test the hypothesis of population enhancement scientifically (Major et al. 2017) and is easily reviewable. We welcome the opportunity for constructive debate and are happy to discuss and elaborate upon specific concerns related to our methods, analyses, and conclusions. During 2009–2013, Major et al. (2017) undertook a BACI experiment designed to test whether habitat modification (i.e., devegetation) at a large crested auklet colony could be an effective means to enhance numbers of breeding crested auklets, allowing for population recovery. Using 3 quantitative techniques, Major et al. (2017) concluded that the experimental data obtained provided no evidence that habitat modification would be an effective means of restoring crested auklets. Divoky (2018) suggests that our analyses and the data used in those analyses were inadequate and flawed, and because of this our conclusions are unjustified. We disagree and will address each of Divoky’s (2018) concerns as they appear in his letter. Divoky (2018) alleges that in our paper’s title we stated the aim of the study to be the restoration of a colony at Gareloi Island. Our paper title refers laconically to habitat modification as a possible means of restoring crested auklet colonies in general, not one specific colony. The aims of our study were, as stated in the abstract and introduction, to experimentally test whether habitat modification could be an effective means of restoring nesting habitat of crested auklets (Major et al. 2017). The mixed crested and least (A. pusilla) auklet populations at Gareloi total approximately 2 million birds and are close to the largest in the Aleutians, with birds breeding in crevices on 2 inland blocky lava flows with encroaching vegetation, and in beach boulder habitat (perpetually renewed by wave action, 30% of the population) completely surrounding the island (Paragi 1996, Jones and Hart 2006). The island is obviously not in need of restoration to support auklets, but it was an ideal location to perform this experiment. Divoky (2018) continues by alleging that we failed to disclose or discuss additional data that demonstrate that vegetation modification “. . .might indeed be an effective method of restoration for crested auklets.” Our paper (Major et al. 2017) included all the demographic data collected that could be statistically analyzed. We excluded time-lapse digital photographs showing activity a priori because we judged them to be unreliable (see below for further explanation). Divoky (2018) outlines various difficulties associated with auklet monitoring that are well-reviewed in the literature; auklets’ underground nesting in rock crevices and their sporadic visibility on the surface of colony sites make their populations very difficult to count or monitor (Jones 1992, Renner et al. 2006, Sheffield et al. 2006). These issues led to us using 3 demographic methods in our experiment: direct counts of visible crevices, numbers of crested auklets attending the experimental plots (estimated from capturemark-resighting), and numbers delivering food to chicks Received: 22 May 2018; Accepted: 29 May 2018