Birds exhibit an assortment of behavioral strategies to cope with variable environmental conditions during reproduction, including altering nest construction behaviors. In species building enclosed domed nests, the microclimate within nests is influenced not only by its structure and the surrounding vegetation but also by the orientation of the nest opening. Many grassland-dependent birds build dome-shaped nests with clear directionality of openings. We studied two species in northeastern Kansas, United States that typically orient their nests east to northeast in this region. However, in a drought year, both Grasshopper Sparrows (Ammodramus savannarum) and Eastern Meadowlarks (Sturnella manga) shifted orientations of their nests southward toward prevailing winds. We hypothesized that this shift reduced the deleterious effects of heat stress on parents and developing young by diminishing morning solar radiation and increasing cooling due to the prevailing southerly winds. To test this hypothesis, we measured temperature, humidity, and wind speed at pairs of unoccupied, field-collected sparrow and meadowlark nests, experimentally placed to face south or east (control) in a non-drought year. Nest orientation affected the daily microclimate patterns, with south-facing nests warming later in the day relative to east-facing nests. The temperature differences depended upon humidity, with south-facing nests being relatively cooler under more humid conditions. This work provides the first experimental evidence of the benefits of plasticity in nest construction under challenging thermoregulatory conditions and shows how ground-nesting birds may reduce thermoregulatory demands during incubation under climate variation.

We conducted the experiment at the Konza Prairie Biological Station (KPBS), a 3487-ha native tallgrass prairie located 15 km south of Manhattan, Kansas (Knapp et al. 1998). The KBPS is managed experimentally as part of the Long-Term Ecological Research program in units subject to multiple combinations of prescribed fire and grazing. The site has a mid-continental climate with mean daily temperatures in recent years of ~13 °C (Macpherson et al. 2008) and the mean maximum daily temperatures during July range from 29.4 °–35.6 °C (Craine et al. 2012). At KPBS, the sun rises 58.63 degrees from the north and sets 301.36 degrees from the north during peak summer, defined in this study as June–August (NOAA Solar Position Calculator, https://gml.noaa.gov/grad/solcalc/). Prevailing winds blow from south to north (Lysenko et al. 1994). There is high interannual variability in summer (May–September) precipitation with a 30-yr rainfall average of 65.2 cm (CV of 29.8%; Knapp et al. 2015). During May–July 2019, a total of 52.9 cm of precipitation fell at KPBS; in 2018, a drought year, 20.5 cm of precipitation fell over the same time period (Nippert 2022).

Grasshopper Sparrows nest in grazed native prairies of moderate height, little-to-no wooded areas, and patchy bare ground mixed with clumps of vegetation (Shaffer et al. 2021). Eastern Meadowlarks build nests away from habitat edges and select somewhat taller, denser vegetation than Grasshopper Sparrows (Hubbard et al. 2006). Both species construct enclosed, domed nests from dead grasses on the ground (Roseberry and Klimstra 1970; Slater 2004), often placing nests beneath clumps of vegetation (i.e., Amorpha canescens, Baptisia australis, Psoralidium tenuiflorum, and Andropogon gerardii). In this experiment, we manipulated naturally constructed nests, as opposed to artificial nests, to approximate realistic conditions to the extent possible. Since neither species reuses nests, we carefully collected completed (i.e., fledged or failed) nests in 2017 and 2019 from both species. We dug and cut nests out of the surrounding vegetation using scissors and trowels, placing individual nests in small cardboard boxes, which we stored in plastic tubs until experimental initiation. We paired nests by mass, a subjective measure of light penetration, and general appearance to minimize differences in microclimate due to slight differences in nest construction.

We conducted the experiment from 12 June–3 August 2019, spanning the hottest and driest parts of the breeding season and capturing times when the thermoregulatory costs of overheating presumably peak at our site. We compared nest microclimate in eight experimental trials; each trial lasted four–six days. We compared microclimate in four–six nest pairs during each trial for a total of 18 Eastern Meadowlark and 16 Grasshopper Sparrow comparisons (n = 34). Distributing trials throughout the season allowed us to detect consequences of nest orientation associated with specific weather conditions and to explore the magnitude of microclimatic differences between differently oriented nests temporally.

We performed the experiment on a single, flat region of our study area to control for potential microclimate and habitat differences within the study site (Fig. 1). We placed 34 pairs of nests in suitable habitat along parallel ~50-m transects, spaced 10 m apart using a handheld GPS unit (GPSmap 60CSx; Garmin, Olathe, Kansas). At the start of each transect or 10-m mark, we selected and flagged the nearest plant species commonly used by these species under which to place nests. If there were no plants of those species within a meter of the point, we continued to a new point another 10 m northward. Under the selected plant, we parted the grass and fastened two similar conspecific nests to the ground with a nail, with the backs of the nests no more than six cm apart. Within each pair, the two nest cups faced a different direction based on the most common orientation from drought and non-drought years: east (90°; Grasshopper Sparrow) or northeast (45°; Eastern Meadowlarks) to represent typical years, and south (180°; both species) to represent the drought year orientation (Smith et al. 2023).

We assessed the consequences of orientation on microclimatic variation by measuring the temperature, humidity, and wind speed at each nest. We attached a small datalogger (DS1923-F5# Hygrochron; iButton®) using Blu-Tack Reusable Adhesive (Bostic) to the underside of small plastic pizza savers and embedded the legs securely in the middle of each nest cup, suspending the dataloggers a couple of centimeters above the ground to measure conditions a nestling or parent would experience inside the nest. The shade from the pizza savers was unlikely to influence the recorded microclimate because the inside of these enclosed nests was already shaded from above. We programmed the dataloggers to record temperature (°C) and relative humidity (%RH) in 10-minute intervals throughout the day and night. We calculated mean hourly temperature and humidity values by averaging the six measurements per hour. We recorded wind speed (m/s) once per day, taking measurements during the hottest part of the day, between 10:30 and 13:00, using a handheld anemometer (Benetech, GM8908) at two locations relative to the nest: (1) immediately in front (at the height) of the nest entrance and facing the direction of the opening, parallel to the ground and perpendicular to the opening, and (2) a single set of measurements at 1 m above the nest pairs. We averaged eight measurements for each set of wind measurements, recorded every 10 seconds, to minimize the effects of gusts on mean wind speed comparisons.

Our response variables represented the differences between paired nests in temperature, humidity, and wind speed, calculated by subtracting the values of south-facing nests from their east-facing counterparts. We presented the difference values as opposed to the raw microclimatic variables to account for variation in the raw data unrelated to our experimental treatment. Thus, response variables reflect how birds mitigated expected costs that could occur if birds adopted standard orientations under drought conditions. Negative values indicated that south-facing nests were warmer, more humid, or windier than east-facing nests, whereas positive values indicated that east-facing nests were warmer, more humid, or windier than south-facing nests.

We used an information theoretical approach to evaluate four sets of models explaining variation in the following response variables: differences between nest pairs in (1) temperature (°C) and (2) humidity (% relative humidity; %RH), both measured from 09:00–17:00 which we anticipated would encompass peak daytime temperatures relevant to bird thermoregulation. To evaluate how humidity might modulate temperature differences, we again modeled (3) temperature differences, this time including raw humidity values measured in the south-facing nest as a covariate. We modeled differences in (4) wind speed, measured directly in front of nest cups. Because we only collected a single, mid-day set of wind measurements, we calculated mean differences in temperature and humidity between 11:00 and 13:00 to match the temporal resolution of the wind speed in analysis 4. Finally, we evaluated the consequence of wind speed (measured at 1 m above the nest pair) on temperature differences (5a) and humidity differences (5b) in the absence of temporal variables.