Identifying Priority Lakes

We identified a candidate set of priority lakes for restoration activities and surveys by analyzing data from the Minnesota Loon Watcher Survey (LWS) program. LWS data were collected by citizen scientists to provide insight into Minnesota loon biology, behavior, and population dynamics. Although record keeping has varied over the years, the LWS generally provides the number of adult loons, total chicks fledged, and number of ANPs for each lake. We inferred the number of territories on each lake on the basis of the number of adult loons. An important feature of the LWS data is that reproductive data from lakes with multiple territories represent annual lake averages. Thus, we did not have territory-specific data for lakes large enough to support multiple breeding loon territories.

The Monitoring and Adaptive Management Plan (Open Ocean Trustee Implementation Group, 2019) prescribes that activities conducted under this project occur on lakes that offer the greatest benefit to loon reproductive success. Therefore, through an interative process, we identified candidate lakes that were reproductively underperforming. First, we identified LWS lakes in an eight-county study area (Aitkin, Becker, Beltrami, Cass, Clearwater, Crow Wing, Hubbard, and Itasca Counties) of north-central Minnesota with mean hatch rate of ≤0.61 chick per pair or, if hatch data were missing, lakes with ≤0.5 chick per pair present in August to provide a preliminary assessment of lakes available for sampling (91 lakes). We specifically used 0.61 chick hatched per pair because this value corresponds to an August chick production of 0.5 chick per pair after accounting for mean estimated chick survival from hatch to 6 weeks of age (that is, 0.82; Meyer, 2006). Furthermore, 0.5 chick per pair in August was used because previous research demonstrated that ≥0.48 fledged chick per pair was required to support a sustained breeding population (Evers and others, 2020). We then eliminated 46 of 91 lakes that were associated with a different long-term loon monitoring program, the Minnesota Loon Monitoring Program, to yield 45 lakes with an estimated 75 territories without ANPs. Second, we iteratively increased our definition of an underperforming lake by 0.1 fledged chick per pair in August (and the equivalent value for chicks hatched per pair after accounting for survival from hatch to August) until our target of 100 candidate territories was achieved. This resulted in the criterion of ≤0.85 chick hatched per pair or ≤0.7 fledged chick in August, which yielded 64 lakes in 7 counties and an estimated 111 potential territories (table 1). No lakes were selected in Clearwater County.

Loon Surveys

We removed two lakes from the 64-lake candidate set before any surveys were conducted (table 1). Specifically, we removed a large candidate multiterritory lake (Gull Lake, Cass and Crow Wing Counties) because the geographic extent of LWS data was unknown for this lake, and the large lake area (4,051 hectares [ha]) precluded us from completely surveying the lake. In addition, we removed a small lake with a single candidate territory because we were unable to obtain access (Unnamed Lake, Aitkin County).

We surveyed lakes from May 3 to August 14, 2021, to collect pretreatment data before ANP deployment in 2023. Observers had a background in waterbird identification through field ornithology coursework or work experience and participated in training sessions provided by an experienced loon biologist, during which loon survey techniques, loon behaviors, nesting habitat, nest physical structure, and determination of nest fate were described. We defined territories through a post hoc evaluation of the field data. We defined a territory as an area with three observations of a pair of loons, of which at least one observation included a territorial behavior (such as territorial interaction with another loon, nest searching, nest building, or copulation), or one or more observations of a subsequent stage in the reproductive process (for example, incubation or chicks). Thus, we refer to areas surveyed with a breeding pair as “candidate territories” and areas that met our post hoc definition of a territory as simply “territories.” We divided survey efforts across the loon reproductive season into three time periods: (1) candidate territory occupancy surveys (May 3 to June 18, 2021), (2) nest-monitoring surveys (June 21 to July 16, 2021), and (3) chick survival surveys (August 2 to 13, 2021).

We surveyed selected lakes at least four times to identify loon territories for territory occupancy surveys from May 3 to June 18, 2021. We completely surveyed each lake by motorboat, kayak/canoe, or shoreline observation using public water access. We secured access through private property following USGS private landowner access protocols if a lake did not have public access. Observation time at each lake depended on variables that affected our ability to count loons and determine breeding status (for example, lake area and shape, environmental conditions, survey method, and loon behaviors). Larger lakes were circumnavigated by boat with frequent stops to scan for loons using binoculars and the naked eye, especially within small inlets and coves, near islands, and over large expanses of open water. Shoreline surveys were conducted on small lakes (<8 ha) that were entirely visible from terrestrial vantage point(s) using binoculars, a spotting scope, and the naked eye. We collected information on environmental variables that could affect loon or nest detection, including visibility and water surface conditions. Behavioral cues (such as parallel swimming, courtship behavior, copulation, and territorial interactions) were used to determine pairing status if more than one loon was located on a lake. Common loon behavioral cues were also used to denote candidate territory boundaries on larger multiterritory lakes.

We recorded all information from all surveys using an ArcGIS Survey123 application. We recorded the following information at each lake for each territory occupancy survey: (a) number of adult loons, (b) adult loon behavior, (c) number of candidate territories, (d) number of nests, and (e) number of subadult loons (in other words, loons in nonbreeding plumage). The location of each adult and subadult loon was documented by using mapping capabilities of ArcGIS Collector. For each breeding pair identified on a lake, observers noted evidence of nesting activity. If no breeding pairs were identified on a lake following four occupancy survey visits of the candidate territory, the lake was dropped from further consideration. Likewise, candidate territories with an existing ANP were dropped from the study because these candidate territories will likely have an ANP in 2023 and could not be randomly assigned to treatment (ANP) or control (no ANP) categories. Thus, only candidate territories that had documented nests without an ANP were further surveyed to monitor nest status.

We conducted nest-monitoring surveys every 3–14 days on candidate territories where nests were documented with a breeding pair. The goals of nest-monitoring surveys were to document nest survival and failure, hatch rates, and renest attempts through mid-July. Nest-monitoring surveys were conducted to minimally interrupt a nesting attempt. Observation time near each nest was situation dependent, with consideration taken to conduct research efficiently and allow loons to quickly resume normal nesting behavior. We recorded the following information for nest-monitoring surveys: (a) nest status, (b) potential disturbances affecting the nest, (c) loon behavior, (d) presence of Simulium annulus (black flies) at the nest, (e) nest type, (f) approximate distance of the nest to shore, (g) vegetative cover surrounding the nest (unobscured, partially obscured, or obscured), and (h) number of eggs in the nest (when possible). Nest status (a) was defined as one of five categories: (1) nest not located, (2) nest building, (3) adult incubating, (4) nest failed, or (5) nest successful. A candidate territory where observers failed to detect a nest during the early occupancy or nest-monitoring periods could represent either a breeding pair that did not nest or a false negative where the breeding pair nested, but observers failed to detect the nest.

A nest was considered successful if one or more eggs hatched. We considered eggs to have hatched if chicks were observed on or around the nest or accompanied by parents away from the nest. Additionally, we considered eggs to have hatched if egg membranes were present in the nest bowl, which was assumed to indicate a successful nest and subsequent chick depredation before the next survey. Loon chicks were observed on the back of adults within the first few days of hatching, beneath adult “tented” wings during inclement weather, and located in emergent vegetation during territorial disputes or clashes with intruder or floater loons. Nest locations were documented using the mapping capabilities of ArcGIS Collector, and more accurate Global Positioning System (GPS) coordinates were obtained for some nests using a hand-held GPS receiver after the nest attempt finished or failed. Eggs were counted when adults flushed from the nest or otherwise were not present. When eggs hatched, the following information was recorded: (a) the number of eggs hatched, (b) the number of unhatched eggs, (c) the number of chicks present, and (d) chick age in weeks.

We identified a nest as failed if eggs were present in the nest without an incubating adult for two consecutive surveys. A nest failure was also identified in cases where intact eggs were misplaced nearby on a shoreline or in shallow water, no intact eggs were present but eggshell fragments were in the nest or nearby prior to the expected hatch date range, or an absence of eggs was accompanied by evidence of predators in the immediate area. We estimated a range for expected hatch date based on the first date on which a nest was considered active with an incubating adult and an incubation period of 28 days (Sjölander and Ågren, 1972) We inspected the immediate area around failed nests for mammalian scat, fur, and tracks for evidence of the cause of nest failure. Environmental conditions (for example, inclement weather events, black fly infestation) and physical human impacts (for example, excessive boat traffic) that may have occurred prior to the current survey were also documented as potential causes of nest failure. Where nest abandonment was suspected, the temperature and odor of intact eggs were evaluated, and eggs were left for inspection during a subsequent survey. Photographs of nest sites and supportive evidence of nest failure were submitted with documentation via Survey123. Where a nest failed, the following information was recorded: (a) whether the nest was flooded, (b) whether the nest was abandoned, (c) whether nest material was disturbed or displaced, (d) the potential cause of disturbed or displaced nest material, (e) the number of intact eggs in the nest, (f) the number of intact eggs nearby, (g) egg odor, (h) whether the intact eggs sink or float, (i) whether the eggs were depredated, (j) whether eggshell fragments were present from the current year, (k) whether predator tracks, fur, or scat were present, (l) whether predator species were observed, and (m) the presence/absence and abundance of black flies at the nest site. We specifically recorded information on black fly abundance because previous research has demonstrated that outbreaks of black flies can increase nest abandonment and reduce reproductive success in years with severe outbreaks (Piper and others, 2018).

Loon chick survival surveys were conducted twice on all 62 lakes from August 2 to 13, 2021. Loon chick survival surveys served two purposes. First, these surveys provided information on survival of chicks that had been previously detected. If one or more chicks were present during the nest-monitoring period but were not detected during the first chick survival visit, we conducted 1–2 additional surveys. Second, these surveys provided information on survival of chicks that had not been previously detected. Loon chick survival surveys involved locating adult loons tending to chicks in open water. When adult behavior appeared normal in the absence of chicks, the search area was widened to allow for the possibility that older chicks had wandered away from adults.

Areas deemed unoccupied on the basis of occupancy surveys of candidate territories were not systematically surveyed during the nest-monitoring period. However, unoccupied areas were often traversed while we were in transit to survey a candidate territory with an active nest during the nest-monitoring period. Thus, previously unoccupied areas were deemed occupied following chick detections during nest-monitoring or chick survival surveys. Such a newly documented candidate territory likely represented a false negative during occupancy surveys of candidate territories. We included these territories in summary statistics below.