The tropicalization of temperate marine ecosystems can lead to increased herbivory rates, reducing the standing stock of seaweeds and potentially causing increases in detritus production. However, long-term studies analyzing these processes associated with the persistence of tropical herbivores in temperate reefs are lacking. We assessed the seasonal variation in abundances, macrophyte consumption, feeding modes and defecation rates of the range-extending tropical rabbitfish Siganus fuscescens and the temperate silver drummer Kyphosus sydneyanus and herring cale Olisthops cyanomelas on tropicalized reefs of Western Australia. Rabbitfish overwintered in temperate reefs, consumed more kelp and other macrophytes in all feeding modes and defecated more during both summer and winter than the temperate herbivores. Herbivory and defecation increased with rabbitfish abundance but this was dependent on temperature, with higher rates attained by big schools during summer and lower rates in winter. Still, rabbitfish surpassed temperate herbivores, leading to a five-fold acceleration in the transformation of macrophyte standing stock to detritus, a function usually attributed to sea urchins in kelp forests. Our results suggest that further warming and tropicalization will not only increase primary consumption and affect the habitat structure of temperate reefs but also increase detritus production, with the potential to modify energy pathways.

Herbivorous Fish Abundance and Rates of Herbivory

Fish abundance was surveyed using Diver Operated Stereo-Video (S-DOV) before deploying bioassays of kelp filmed with Remote Underwater Videos (RUV). Three or four 25 × 5 m S-DOV transects were sampled along the ecotone between reef and seagrass, separating each transect by a minimum of ~10 m to ensure independence of replicates and ignored fish appearing from the back to avoid double counting (Goetze et al. 2019). Herbivory on kelp was assessed with tethered bioassays consisted on a cluster of at least nine individual ~15 cm long lateral blades of Ecklonia radiata attached to 0.5 m rods simulating kelp canopy, one cluster was deployed per reef in each sampling day, representing one independent sampling unit. Tethers were deployed within the typical feeding timeframe of diurnal herbivorous fish, from the morning until the afternoon (8:00 – 16:00 hrs) (Fox et al. 2009) and filmed for 3-4 hours with GoPro cameras to identify the species responsible for the consumption of kelp, their relative abundances (MaxN), their bite rates on kelp (bites h^-1) and defecation rates (feces h^-1). In addition to herbivory on tethered kelps, bites on other macrophytes attached to the substratum and drifting in the water column were also registered during RUV analyses (e.g. kelp, seagrass, Sargassum spp, Ulva spp., Hypnea spp.). We classified the herbivory modes as: browsing (substrate attached macrophytes, including kelp tethers), kelp browsing, drifting (non-attached macrophytes) and total herbivory (browsing + drifting consumption).

Kelp lateral blades of bioassays were pressed between a white background and a perspex glass and photographed before and after deployment. The photographs of kelp were analyzed using the software ImageJ (rsb.info.nih.gov/ij/) to calculate the area consumed per time (cm² hr^–1) (Vanderklift et al. 2009; Wernberg et al. 2006; Zarco-Perello et al. 2017). This was transformed to biomass (g hr^–1) using a linear area-weight regression. Herbivory and defecation rates were standardized for each fish species; bite rates estimates were multiplied by weight-specific bite sizes (cm²), calculated from bite size-weight regressions using bite measures from specimens of both species, while defecation rates were multiplied by the mean weight of each species calculated from the corresponding S-DOV carried out immediately before each herbivory assay (Shantz et al. 2015). Underwater temperature data for each reef was recorded in situ during the fish and herbivory surveys using hobo data loggers.