Rising global temperature will have profound impacts on species and ecosystem functioning. Species existing near their thermal thresholds will be particularly vulnerable to these changes, and those species that rely on, or preferentially use, artificial structures may face pronounced effects. Gaining insights into the anticipated climate changes, both present and future, is crucial for informing conservation practices and the utilisation of artificial structures in conservation efforts. Using three years of data, we quantified and compared the temperature of artificial nest boxes installed between 1986 and 2006 and natural nest burrows of a fringing population of Little Penguins existing at the north-western limit of their range. Nest boxes were ineffective at replicating conditions of natural nests, exhibiting consistently higher daily maximum temperature (~2°C) and exceeding upper thermoneutral limits for longer than natural nests. Fine-scale biotic and abiotic nest characteristics influenced maximum nest temperature and exposure duration. Simulated temperature increase of 2°C predicted an increase in the number of days exceeding hyperthermic conditions (≥35°C) by up to 49%. Such increases will expose penguins to potentially fatal thermal conditions, particularly during the late breeding and moulting phases of their annual cycle. This study revealed that current and future thermal environments of Little Penguin terrestrial habitat on Penguin Island can exceed physiological limits for this species. Intervention to improve artificial nests and better quantify consequences is urgently needed given recent estimates of a declining population and increasing risk of local extinction.

Our objective was to describe and quantify temperature in artificial nests and how they differ from natural nest burrows, examine the influence of climate and nest attributes (location and vegetation cover) on nest microclimate. To do this, we measured air temperature within artificial nest boxes and natural nests for 3.5 years from July 2013 – January 2017.

Nest Types

Nest Boxes: Timber** nest boxes are broadly similar in design, with minor differences depending on the year of installation. The boxes are either square or rectangular in shape, with slight variations in timber thickness and entrance design (hole versus tunnel). Additionally, some boxes include ventilation holes on the sides, while others do not.

Natural Nests: A total of 20 natural vegetation nests were located in 2013 with an additional 33 nests identified and monitored in 2014, 2015 and 2016

Nest characteristics data

For each nest, a suite of nest habitat variables thought to influence nest temperature was recorded. These were stratified to represent (1) characteristics describing the position of the nest within the landscape, and (2) characteristics directly associated with the nest. Landscape position characteristics included topographical measurements (slope, aspect and elevation). Nest site measurements included physical attributes (relating to the dimensions or physical structure of a nest) and vegetation characteristics. 

Field measurements of vegetation characteristics were measured between March 2013 to December 2016 capturing seasonal and annual dynamics. To determine the spatial scale at which vegetation most strongly influences microclimate, cover was quantified for: (1) nest box lid cover (box cover; for nest boxes only), (2) within one meter of the nest (quadrat cover), and (3) in the broader surrounding habitat, within approximately four meters of the nest (vegetation cover). These measurements included a combination of visual cover estimates and quantitative estimations of digital images.

Description of nest characteristic measurements recorded for both Natural nests and nest boxes

Slope – Slope of the ground on which the nest is located (degrees)      

Elevation – Elevation of the of the position where the nest is located (m)    

Aspect – Aspect of the hill face on which the nest is located (degrees)

Species Composition – Species of plant that dominated the vegetation surrounding or covering the nest (measured once annually during winter)

Quadrat cover – Visual assessment of vegetation cover (%) within a circular plot (1 m diameter) centred over the nest, categorised into: 1 = <5% cover, 2 = 5- 24% cover, 3 = 25 – 49% cover, 4 = 50 – 74% cover and 5 = 75 – 100% cover). (Measured 4 times annually during winter, spring, summer, autumn)

Vegetation cover – Quantitative assessment of proportion green vegetation cover (%) within a rectangular quadrat centred over the nest site. Measured annually at the beginning of autumn for all boxes in 2013, 2014, 2015, and 2016. Natural burrows were included in 2015 and 2016.

Entrance bearing – The bearing (degrees) recorded once only for boxes and annually in winter for natural nests

Description of nest characteristic measurements for Natural nests only

Bush wall – Thickness of vegetation measured from the nest cavity ‘ceiling’ to the outside edge of the vegetation (mm).  (Annual measurement during winter)

Discrete bush – Natural nest is either an isolated bush or part of a larger vegetation patch. (Annual measurement during winter)

Nest bush height – Maximum height of the vegetation directly over the nest (mm). (Annual measurement during winter)

Nest bush length – Length of nest shrub from the edge containing the entrance to the opposite edge of bush (mm). (Annual measurement during winter)

Nest bush width – Width of nest shrub (mm). (Annual measurement during winter)

Cavity cover - Percentage vegetation directly over the nest cavity measured using a Gopro HERO4 camera positioned in the centre of the nest bowl facing upwards. (Bi-annual measurement in summer and winter)

Description of nest characteristic measurements recorded for nest boxes only

Box cover - Visual assessment of percentage of box lid covered by vegetation, categorised into: None = <5% cover, Moderate = 5 – 74% cover and Full = 75 – 100% cover). Measured seasonally (winter, spring, summer, autumn)

Vents - Presence or absence of ventilation holes

Local weather data

To account for the effect of local weather conditions (wind, ambient temperature, humidity) on nest temperature, half-hourly measurements of ambient temperature, relative humidity, wind speed and direction were sourced from the Bureau of Meteorology (BoM) at Garden Island meteorological station. From these we extracted the daily maximum temperature (°C), relative humidity (%), and the average wind speed (kph) and wind direction at four time points (06:00, 09:00, 12:00 and 15:00) for each sampling day. Wind direction was categorised as either ‘onshore’ or ‘offshore’. Wind speed between each timepoint was highly correlated and consequently only the windspeed and wind direction at 12:00 was retained to evaluate the effect of wind conditions on nest temperatures during the observed peak heat periods.

Nest microclimate data

Temperature (°C) and relative humidity (%) inside both natural nests and nest boxes at were recorded continuously at 30-minute intervals. As there were fewer loggers than sampled nests, loggers were rotated between nests over two and a half years with most nests containing a logger for approximately 12 months.

To quantify nest temperatures, half-hourly temperature measurements were first averaged by hour for each nest site and sampling date. Half-hourly measurements of ambient temperature recorded at the BOM Garden Island weather station were also averaged by hour for each day. For each sampling date (24-hour period), we calculated the daily maximum, daily minimum, daily mean and temperature range for ambient and sample nests. Departure values were calculated for each sampling date and nest by subtracting the daily ambient measurement from the corresponding daily nest measurement. For nests that recorded temperatures ≥30°C or ≥35°C, we also calculated the total number of hours per day a nest recorded temperatures of 30°C or higher and 35°C or higher (exposure hours).

Exploratory data analysis revealed a slight non-linear relationship between maximum nest temperature and maximum ambient temperature, with a minor change in the effect occurring around 25°C. To account for this, we introduced a categorical variable, hereafter referred to as ‘temperature category’, where sample dates were categorised according to whether the maximum ambient temperature exceeded 25°C or not.