Behaviour and landscape contexts determine the effects of artificial light on two crepuscular bird species

Context Artificial light at night (ALAN) is increasing worldwide, with many ecological effects. Aerial insectivores may benefit from foraging on insects congregating at light sources. However, ALAN could negatively impact them by increasing nest visibility and predation risk, especially for ground-nesting species like nightjars (Caprimulgidae). Objectives We tested predictions based on these two alternative hypotheses, potential foraging benefits vs potential predation costs of ALAN, for two nightjar species in British Columbia: Common Nighthawks (Chordeiles minor) and Common Poorwills (Phalaenoptilus nuttallii). Methods We modeled the relationship between ALAN and relative abundance using count data from the Canadian Nightjar Survey. We distinguished territorial from extra-territorial Common Nighthawks based on their wingboom behaviour. Results We found limited support for the foraging benefit hypothesis: there was an increase in relative abundance of extra-territorial Common Nighthawks in areas with higher ALAN but only in areas with little to no urban land cover. Common Nighthawks’ association with ALAN became negative in areas with 18% or more urban land cover. We found support for the nest predation hypothesis: the were strong negative associations with ALAN for both Common Poorwills and territorial Common Nighthawks. Conclusions The positive effects of ALAN on foraging nightjars may be limited to species that can forage outside their nesting territory and to non-urban areas, while the negative effects of ALAN on nesting nightjars may persist across species and landscape contexts. Reducing light pollution in breeding habitat may be important for nightjars and other bird species that nest on the ground. Supplementary Information The online version contains supplementary material available at 10.1007/s10980-024-01875-3.


Figure S2
Comparison of the Earth Observation Group's Version 1 (V1) and Version 2 V(2) annual composites for 2015.
Red areas show pixels assigned positive artificial light values by V1, but not by V2.Blue areas show pixels assigned positive artificial light values by V2, but not by V1.Visual inspection showed that blue areas (missed by V1) likely involved skyglow in the pixels surrounding artificial light sources.The red pixels (missed by V2) occurred in areas with low levels of human development where V1 identified light sources that V2 missed.V1 is available for only 2015 and 2016, and uses a combination of automated and manual processes for distinguishing artificial light from aurora in the Northern aurora zone (Elvidge et al. 2017).V2 is available for all years between 2012 and 2020, but applies an additional manual filter before manual editing in the Northern aurora zone, filtering out many dim lights in our study area that were found in V1.

Figure S3 Coefficient estimates from model fit with simulation data
We simulated relative abundance data using the coefficient values and scales selected by our model.These blue posterior density plots show the coefficient estimates from models fit using these simulated relative abundance values, the mean estimates (solid lines), and 95% credible interval (dashed lines).The pink lines show the coefficient value used for the simulation.To estimate the coefficients from the simulated data, we used the same process that we used for the real data.We used the BLISS procedure to select the scale for each covariate, then we refit the model with each covariate at its selected scale.For all models, the BLISS model correctly selected the ALAN or urban scale that was used to simulate the data, except in the case of Grassland of Territorial Common Nighthawks, where the model fit with simulated data selected 400 m instead of 1600 m.
Table S3 Scale(s) selected for each landscape covariate and the proportion of the posterior distribution that selected these scales.The solid lines show the mean coefficient estimate from model fit using each version of the annual composite and the dashed lines show the 95% CIs.V1 is the Earth Observation Group (EOG) Annual Composite V1 for 2016, V2 is the EOG Annual Composite V2 for the year in which the survey took place, and V1V2 is the average of these two composites.

Figure S5 Results from BLISS scale selection
The left side of the figure shows the proportion of the posterior selecting each scale.The right side of the figure shows boxplots of the coefficient estimates from the portions of the posterior distributions that selected each scale.

Figure S7 Posterior probability densities in the model for territorial Common Nighthawks excluding four influential survey stations near Victoria
Table S4 Estimated effect of increasing ALAN on relative abundance of nightjars at mean, median, and high levels of urban landcover Columns 1 and 2 show the coefficient estimates for ALAN and the interaction between urban landcover and ALAN in each model.The posterior distributions of the effect coefficients for other covariates are shown in Figure S5.Columns 3-5 show the expected change and the 95% CI for the change in the number of nightjars when ALAN increased from 0 to the 99 th percentile.This 99 th percentile ALAN value (shown in italics) was calculated for model and urban landcover level separately, using surveys with urban landcover less than or equal to the median, mean, and high (95 th percentile) urban landcover values within the buffer size selected for each model.Effect coefficient estimate (95% CI) for ALAN

Effect coefficient estimate (95% CI) for ALAN interaction with urban landcover
Change in nightjar counts between 0 nWcm -2 sr -1 and 99 th percentile ALAN value

Table S6 Coefficient estimates from detection probability models
Using the minute-by-minute detection data for each individual nightjar, we modeled the effect of artificial light and temporal covariates on the number of minutes (out of six) in which each individual was detected using a binomial GLM.Bolded cells contain coefficients whose 95% CIs do not overlap zero.In our model for extra-territorial Common Nighthawks, the effect of artificial light changed from positive to negative at when urban landcover at the 1600-m scale exceeded 18% (95% CI: 3%, 30%).These images are examples of 1600-m buffers around survey points where urban landcover is between 15% and 25%.These images from Esri World Imagery were generated in R using the leaflet package and the Esri World Imagery basemap.

Figure S4
Figure S4Posterior probability densities for ALAN and urban landcover covariates in the sensitivity analysis for the version of the annual composite used to estimate ALAN.

Figure S8
Figure S8 Cross-correlation among coefficient estimates for ALAN, urban, and their interaction term

Figure S9
Figure S9 Effects of (A) ALAN, (B) sun angle, and (C) Lunar illumination on detection probability in a 6-minute survey.We modeled the detection rate for each individual as a function of artificial light and the temporal covariates, and made predictions across ALAN, sun angle, and lunar illumination values, with ordinal day at its mean value (178.2).Values on the x-axes span the 5 th to 95 th percentiles of ALAN, sun angle, and lunar illumination values observed across all surveys.Dashed lines represent the 5 th and 95 th percentile values observed in surveys where each nightjar species/behaviour occurred.The 5 th percentile ALAN ad lunar illumination values for all species/behaviours were 0.

Figure S10
Figure S10Posterior probability densities for ALAN and urban landcover coefficients in our sensitivity analysis for including/excluding surveys with less than 90% detection probability

Table S1
Common Nighthawk associations with landscape features in previous studies

Table S2
DIC comparison for alternative model forms

Table S5
Marginal effects of landscape-scale ALAN on nesting nighthawks