Fishes and other consumers excrete metabolic waste products, including dissolved nutrients rich in nitrogen, which is an essential nutrient for primary production. Relatively little is known about the magnitude and variability of nutrients excreted by fishes in kelp forest ecosystems and whether consumer-derived nutrients are important for supporting kelp productivity. In this study, the supply of ammonium (NH₄⁺) excreted by the dominant fishes (30 species representing ~85% of total fish biomass) was investigated on nearshore rocky reefs in California. Using rapid field incubations, the amount of excreted dissolved ammonium was measured as a function of body size (n = 460 individuals), and predictive models were developed relating mass to excretion rates at the family level. Mass-specific excretion rates ranged from 0.08 – 3.45 µmol·g^-1·hr^-1, and per capita ammonium excretion rates ranged from 5.9 – 2765 µmol·individual·hr^-1. Ammonium excretion scaled with fish body mass to the ¾ power, as predicted by the metabolic theory of ecology; mass-specific excretion rates were higher in smaller fishes, but larger fishes contributed more ammonium per individual. When controlling for body size, ammonium excretion rates were greatest among surfperch (Embiotocidae), damselfish (Pomacentridae), and wrasses (Labridae), and the general trophic groups of planktivores and micro-carnivores. When body size differences were considered, the greatest mean excretion rates per individual were observed in larger-bodied species, such as California Sheephead (Semicossyphus pulcher) and Lingcod (Ophiodon elongatus). Empirical estimates of nutrient excretion by fishes, among the first measured in temperate kelp forests, were consistent with those in other aquatic systems. Ultimately, empirically derived excretion rates are the first step to quantifying the relative importance of consumers to nutrient cycling in kelp forest ecosystems.

To quantify rates of ammonium excretion for the dominant fish species inhabiting nearshore rocky reefs and kelp beds in central and southern California, we utilized established techniques in the field for rapid empirical estimation of consumer-derived nutrient excretion in aquatic systems. We selected a subset of 30 species, from 13 families, to represent the dominant fish species comprising California kelp forests. The selected species represented ~85% of total fish biomass, based on estimates of density and biomass from extensive underwater visual surveys conducted by the Partnership for Interdisciplinary Studies of Coastal Oceans (PISCO) in central and southern California. Each fish species was assigned to one of five trophic guilds based on discrete, fixed trophic delineations (herbivore/omnivore, planktivore, micro-carnivore, macro-carnivore, and piscivore) as established in the literature. Herbivores/omnivores consist of species that consume a relatively large proportion of algae along with some animal matter, planktivores eat zooplankton, micro-carnivores eat small crustaceans (e.g., amphipods, isopods, shrimps, crabs, and copepods), macro-carnivores eat larger invertebrates (e.g., gastropods, crabs, polychaetes, shrimps, sea urchins) and may occasionally eat small fishes, and piscivores eat primarily fishes. Since many fishes within California kelp forests are opportunistic predators, species with a mixed diet were classified into a single trophic group based on the dominant type of prey taxa (>50% of the diet) consumed at the adult stage.

Fish collections occurred from May 2016 – September 2017. To quantify excretion rates for species representative of kelp forests along the entire coast of California, we sampled fish in both central and southern California. Species affiliated with colder water were collected from Stillwater Cove (36°33'42.8" N 121°56'48.5" W) and Point Lobos (36°31'18.4" N 121°56'33.9" W) in central California, whereas species affiliated with warmer water were collected off Santa Catalina Island (33°26'52.8" N 118°28'57.4 "W) in southern California.

For each species, we collected n = 5-45 individuals across a range of body sizes via a variety of methods, including hook-and-line and baited fish traps, as well as barrier nets and hand nets while SCUBA diving. Species with fewer than 5 individuals collected were pooled with similarly related species in the same family during the initial stage of species-level data analysis. Fish were collected during daylight hours between 0900 and 1500, and only individuals collected in good condition (i.e., not exhibiting signs of barotrauma) underwent experimental incubation to measure ammonium excretion rates. After capture, fish were either placed directly in a prepared incubation chamber or in a communal holding tank for <30 minutes, in cases where the fishing yield exceeded the rate of experimental processing. All containers with fish were covered and aerated to reduce stress. To reduce handling stress, we weighed each individual for wet mass (in grams) and measured total and standard length (in centimeters) following incubation. Fish were then released alive at the site of capture.

To calculate the ammonium excretion rate (NH₄⁺ per individual per gram body weight per hour) for each individual fish, we conducted experimental incubations to test the change in ammonium over a set time period. Each tank was lined with a 4 mm polyethylene bag, and the bag was changed between incubations. The tanks were filled with a known volume of seawater (1-75 L) which varied according to the size of the fish. During data analysis, nutrient excretion rates were calculated by standardizing the measured nutrient concentration by the water volume used. Control incubations without fish were conducted (n=4 per day) to quantify any changes in ammonium levels not due to excretion by the fish.

Samples for ammonium quantification underwent nutrient analysis on the day of collection using established fluorometric techniques. In the field, we added seawater samples to amber vials containing working reagents to begin the reaction process. The working reagent contained a mix of ortho-phthalaldehyde (OPA), sodium sulfite, and sodium tetraborate. Within 3-10 hours of collection, we analyzed the ammonium samples using a handheld Turner Designs AquaFluor Handheld Fluorometer (optical kit No. 8000-402). Standards with known concentrations of ammonium were used daily to create a standard curve for calculating ammonium concentrations and underwent the same treatment in the field.

To calculate the excretion rate of each individual, we plotted NH₄⁺ (µmol) as a function of experimental duration (minutes) and extracted the slope of the relationship using linear regression. Mass-specific excretion rates by species were calculated as total ammonium excreted per individual per hour, divided by the wet weight (g) of an individual. Per capita excretion rates were averaged over all sizes measured for that species, and reflect the total ammonium excreted, regardless of body size. Species with similar taxonomic identities and/or feeding habits were later grouped together and excretion was calculated at the level of a fish family or trophic guild, respectively.

To account for the effect of fish mass on excretion rates at the species-, family-, and trophic group levels, we used bivariate linear regressions to examine the strength of the association between NH₄⁺ excretion per individual per minute as a function of fish mass. Due to evidence of allometric scaling at the larger fish weights, we log-transformed excretion and body size to improve homoscedasticity and normality. To test the hypothesis that the concentration of ammonium excreted by kelp forest fishes differs based on phylogeny, we used an Analysis of Covariance (ANCOVA) with the factor of family identity, the covariate of body mass, and the interaction between family identity and body mass. The families Cottidae, Gobiidae, and Malacanthidae were excluded due to limited sample sizes and/or a limited range of body sizes sampled. To test the hypothesis that the concentration of ammonium excreted by kelp forest fishes differs among trophic guilds, we used an ANCOVA to test for statistically significant differences in mass-specific excretion rates with the factor of trophic guilds, the covariate of body size, and the interaction between the two.