Disturbance from human activity can cause reduced productivity of coastal birds that nest on sandy beaches. A common method to protect coastal birds from human disturbance is the use of signs and fencing to close off a section of beach used for breeding. This management action requires public compliance and might require enforcement, such as in the use of volunteer stewards stationed at protected colonies that provide education and enforcement. I assessed the effectiveness of active stewardship as a conservation measure to protect nesting Least Terns (Sternula antillarum) in coastal Mississippi by determining if colony-level productivity (fledglings produced per nest) was correlated with stewardship effort (hours that stewards protected each colony), the rate of disturbance from human and natural sources, and additional factors. Observers surveyed 24 Least Tern colonies in Harrison County, Mississippi, twice weekly during the 2017–2019 breeding seasons to record number of nests present, fate of marked nests, predator tracks within the colony, and number of fledglings produced per colony. Concurrently, stewards recorded during their work shifts all sources and durations of disturbance events that caused the terns to flush or respond with defense. Least Tern daily nest survival increased with colony size and stewardship effort, and was lower during intervals that included weekends and evidence of owl presence. Total productivity was negatively associated with avian predator disturbance rate and total time adults spent flushed, but was not associated with stewardship effort. The results of this study demonstrate that active stewardship can have a positive effect on Least Tern productivity by increasing nest success, whereas current stewardship practices were not sufficient to increase chick survival, even in a system with a paucity of ground predators.

Colony Survey Data:

Audubon Mississippi staff conducted surveys at tern colonies throughout the breeding seasons, visiting at least once per week until all adults and fledglings had dispersed. Staff counted the number of active nests, flightless chicks, and fledglings using either an exterior or interior survey protocol. Staff conducted exterior surveys by standing outside the symbolic fencing at one corner of the colony and scanning the interior with binoculars (8x42) and scope (60x power). Birds seen in incubation posture provided a nest count. Fledglings often congregated on the shoreline and were easily counted from outside the colony. At large colonies (>250 pairs) or colonies that contained visual obstructions such as dunes, staff conducted the exterior survey by visiting multiple viewpoints. Staff conducted interior surveys by walking systematically through the entire colony, keeping a count of all nests, chicks, and fledglings observed, and by noting and identifying any predator or human tracks within the colony. All colonies were surveyed using a combination of interior and exterior protocols throughout the season. The exterior protocol was employed whenever an interior survey would cause excessive disturbance (though staff would enter these colonies at other times for the purpose of checking nest plots), such as in the case of large colonies, colonies with a large number of flightless but mobile chicks, days with high winds or other inclement weather, hot and sunny days, or when time and logistics otherwise necessitated an exterior survey.

Disturbance Data:

The stewards recorded all interactions they had with the public and conducted continuous disturbance and behavior observations throughout their work shifts. When recording public interactions, stewards classified the interaction as an intervention if they prevented or interrupted a disturbance (such as somebody entering the colony, approaching with an off-leash dog, or setting off fireworks near the colony boundary), or as education if the interaction only involved answering questions or teaching the public about the birds. Interventions were often accompanied by education efforts that were not tallied under that category. In addition, stewards assessed the reaction of the person to the intervention or education, and classified the reaction as positive if the person appeared receptive, interested, or apologetic, or negative if the person appeared dismissive, disinterested, or combative. For the disturbance observations, anytime a portion of the tern colony flushed or reacted defensively (e.g. by dive-bombing a potential threat), the stewards recorded the source of disturbance, duration of the disturbance event to the nearest minute (using a watch or cell phone), and the maximum proportions of the colony that reacted by flushing, engaging in defensive behavior, or failing to react. Proportions for these reactions totaled one. The disturbance event was considered over once all terns had returned to pre-disturbance behavior. Any subsequent flush or defensive behavior was recorded as a new disturbance event. Simultaneous and sequential disturbance events were rare, but in cases of simultaneous disturbance events, stewards categorized the source as which source was present during the majority of the time that terns spent reacting. For sequential disturbance events caused by different sources, events were assigned separate durations and sources accordingly. Stewards categorized disturbance sources as pedestrian (any human approaching or entering the colony on foot, including staff entering for research purposes), dog (off-leash or on-leash), avian predator of eggs and chicks (gull, crow, Gull-billed Tern), avian non-predator (primarily Osprey [Pandion haliaetus] but also herons, egrets, shorebirds, passerines, and near-passerines), vehicle, aircraft, other (miscellaneous and infrequent sources such as kites flown nearby and beach balls or trash blowing into the colony), and unknown (source could not be determined).

Nest Survival Data:

Each year, I selected up to 6 colonies where individual nests were marked for nest survival estimation. I selected colonies systematically in an attempt to include a range of colony sizes and geographic locations. I attempted to mark all nests in selected colonies in 2017 but used nest plots in subsequent years. Nest plots (3–6 per colony) were 10- x 10-m in extent and placed randomly throughout each colony. I placed a 20-cm wooden stick marked with the nest number ~5–20 cm to the north of each nest; densely-spaced nests at many colonies prevented me from placing the stick any farther from a nest. Audubon staff visited each nest plot 2–3 times per week, marking all new nests that appeared within the plot boundaries and checking the status of existing nests. Staff estimated the nest initiation date based on laying sequence for nests found prior to clutch completion and assumed a 21-day incubation period (Thompson et al. 1997) ±3 days to determine the range of expected hatch dates. For nests discovered after clutch completion, staff estimated a range of possible nest initiation dates based on the date of nest discovery within the nest plot and date of the previous visit to the plot. Expected hatch dates were not calculated for nests that had complete clutches on the day that plots were established. During each visit, staff assigned each nest one of the following fates: active with eggs, hatched, failed, or unknown. Nests were assigned as hatched if at least one downy chick was present, or if the nest scrape was empty and an appropriately-aged chick was < 1 m from the nest that could not have come from any other nest within that radius. Nests were assigned as failed if there were signs of depredation (yolk or punctured egg), the scrape was empty but there were fresh predator tracks leading to the scrape, or if the scrape was empty and the date did not fall within the expected hatch window. If eggs were present but appeared abandoned as evidenced by partial burial or bleaching of the shell, then staff assigned the fate as unknown but checked the nest again during subsequent visits. If the nest continued to appear untended during the next visit, it was assigned as failed. Staff also assigned nests as failed if there was evidence of tidal overwash or washout due to heavy rain (nest and scrape gone and sand freshly washed over). Staff assigned fate for nests as unknown if the nest was empty but evidence did not meet any of the previous criteria, or if the nest was empty with no evidence of hatch or failure and staff had not been able to determine an expected hatch window.

Data files are unprocessed except for QA/QC that occurred at the ends of each season. I have included the R script used to process the data for analysis.