Data from: Artificial light at night and warming alter behaviour of the nocturnally active diadematid sea urchin Centrostephanus rodgersii
Description
Artificial light at night (ALAN) is a rapidly increasing global threat that can disrupt the behaviour of marine organisms through both top-down and bottom-up effects, impacting key processes such as predation and herbivory. At the same time, global ocean warming is having profound impacts on critical behaviours across a wide range of taxa. In urbanised coastal habitats, where these stressors are likely to co-occur, the combined impacts of light pollution and warming are predicted to significantly alter marine ecosystems through changes in the behaviour of key consumers, such as sea urchins. We investigated the impacts of light pollution and warming, over 9 weeks, on the behaviour of the diadematid sea urchin Centrostephanus rodgersii which normally shelters during the day and forages at night. At the start of the experiment (1 week), ALAN and warming increased nighttime sheltering behaviour. The effect of warming appeared to reduce over time while effects of ALAN persisted, particularly in ambient treatments. While we found no effects on daytime behaviour in the first week of experiment, by the end of the experiment, under ambient temperatures, we saw reduced daytime sheltering of urchins in ALAN treatments. Altered behaviour patterns due to ALAN and warming could impact the ecological role of urchins, with consequences for benthic marine communities.
Files
Steps to reproduce
We collected 40 individuals of C. rodgersii in Sydney Harbour, Australia. We conducted an aquarium experiment over 9 weeks with four treatments, ALAN warm, ALAN ambient, dark warm and dark ambient (n = 10 tanks for each) to examine sheltering behaviour. During weeks 1, 2, 5 and 9, urchins were photographed from above every 10 minutes for up to 24 hours to record urchin behaviour in each tank, with recording repeated for two nights in each sample week for each tank. We simultaneously used CammPro I826 Infrared Body cameras to photograph dark treatments at night, and GoPro Hero and OSMO Action cameras to photograph during the daytime and in ALAN treatments. Every 10 minutes, urchin activity was classified as ‘sheltering’ (urchin was inside the flowerpot, or between the flowerpot and the tank wall), or ‘outside of shelter’ (urchin was elsewhere in the tank). We initially categorised ‘outside of shelter’ and ‘feeding’ (urchin in contact with kelp) separately, however we were unable to reliably track kelp location in dark treatments. Therefore, these categories were combined into the ‘outside of shelter’ category. Statistical analysis was conducted in R. Due to gaps in recording caused by cameras overheating, we selected a subset of daytime and nighttime hours to maximise the number of replicates. Specifically, within each week, we filtered the data to include only tanks with complete footage from a ‘nighttime’ period (11 hours, 7pm – 5am) and a consecutive ‘daytime’ period (5 hours, 8am – 12pm) on the following day. Using the slice sample function in the dplyr package, we randomly selected one combined night/day period per tank, per week. As a result, different urchin individuals were selected in each week, however daytime and nighttime data always corresponded to the same individual urchin within a 24-hour period. Replicate numbers varied between weeks and treatments, ranging from n = 4 - 9 (Table S1). To avoid non-independence resulting from repeated measures, we only analysed data from the start and end of the experiment (week 1 and week 9), and day and night were analysed separately due to different number of hours in each. Data from all weeks are presented graphically to provide insights on changes in urchin behaviour over time. To assess the effects of light (fixed factor, two levels: ALAN and dark), temperature (fixed factor, two levels: ambient and warm) and their interaction on the total frequency of sheltering behaviour during day and night in week 1 and week 9, we fitted separate linear models using the glmmTMB package (Brooks et al. 2017) in RStudio (v 4.4.0, R Core Team 2022).