Animal consumers excrete dissolved inorganic nitrogen, an essential nutrient for regulating primary production and macroalgal growth in the ocean. Often overlooked in attempts to explain kelp forest productivity, relatively little is known about the role of consumers in nutrient cycling in temperate systems, such as the magnitude of nutrients excreted and the factors that influence spatial and temporal variability in consumer-derived nutrients. To investigate the supply of ammonium (NH₄⁺) excreted by the dominant members of the nearshore rocky reef fish community, we combined empirically-measured relationships between excretion rate and body mass with data on fish density and size structure from visual SCUBA surveys conducted from 2005-2018 in the northern Channel Islands, California, USA. The fish community excreted a substantial amount of ammonium to the kelp forest (mean: 95.5-131.3 µmol·m^-2·hr^-1), however excretion varied substantially among sites and over time (range: 23.1–247.9 µmol·m^-2·hr^-1). The ammonium supply was influenced by the establishment of marine protected areas (MPAs) that restricted fishing activities and environmental characteristics that influenced geographic and temporal shifts in the overarching fish community structure. Fish-derived ammonium excretion rates were 50-80% greater inside MPAs compared to areas open to fishing, while environmental and habitat characteristics (e.g., habitat vertical relief, kelp biomass, wave exposure, chlorophyll a) explained 85% of the spatial variation in community excretion rates. In contrast, large-scale oceanographic phenomena (e.g., North Pacific Gyre Oscillation index) and past patterns of fish recruitment explained 57% of the temporal variation in nutrient excretion over the 14-year time series. Results suggest that fish-derived ammonium may provide an important and underappreciated nutrient source to kelp beds, particularly during low-nutrient periods (e.g., seasonal reductions in upwelling or El Niño-Southern Oscillation events), and that fishing disrupts these nutrient cycling pathways. Fishes likely play a critical role in supporting the resiliency of kelp forest ecosystems by supplying a relatively constant source of reduced nitrogen that can be utilized by giant kelp and other macroalgae to fuel primary production of biogenic habitat.

Study system

The northern Channel Islands, California, USA, are located in the middle of a dynamic oceanographic boundary formed by the cold California Current to the west and the warmer Southern California Countercurrent to the east. The convergence and mixing of the currents result in substantial variation in productivity and the formation of strong thermal gradients from west to east that influence benthic community structure across four islands, covering a scale of approximately 100 km. These kelp forests support diverse assemblages with many invertebrates and fishes that rely on the modification of the habitat by the canopy forming giant kelp (Macrocystis pyrifera). In addition, spatial variation in fish biomass throughout the wider Southern California Bight, including the northern Channel Islands, is strongly associated with physical environmental and habitat characteristics. Fish abundance and biomass has also been observed to vary across the Channel Islands in response to the establishment of a network of MPAs, with biomass nearly doubling for species targeted by fishing activities since 2003.

Fish Community Survey Methodology

Visual surveys of kelp forest assemblages conducted by the Partnership for Interdisciplinary Studies of Coastal Oceans (PISCO) during summer/fall were used to quantify fish abundance and biomass, in order to estimate nutrient excretion by the fish community. Sampling effort has changed over time, but included n=46 sites in the northern Channel Islands at the peak of effort, with sites inside and outside MPAs. At each site, PISCO divers conducted n=8 to 12 fish transects (30 m long × 2 m wide × 2 m high) at multiple levels in the water column: benthic, midwater, and canopy in depths from 0-20 m. Transects were laid out in a stratified-random design, with multiple non-permanent transects located in fixed strata (i.e., outer, middle, and inner edges of the reef). At each level in the water column, one diver counted and sized the total length (in centimeters) of all the fishes per transect (excluding small cryptic species). Experienced underwater observers are accurate to 1 cm in their visual estimates. Although PISCO surveys started in 1999, we included only the year 2005 onwards in order to encompass when the survey effort expanded to include sufficient sites for analysis over time. To focus only on resident kelp forest species, we excluded elasmobranchs and highly migratory species. Although over 125 species associate with rocky reefs in the Southern California Bight, we focused on the 50 resident species comprising 95% of the bony fish biomass in the northern Channel Islands. 

Estimating fish community excretion

To estimate nitrogen excretion by the fish community on an areal scale, we calculated the mass of ammonium excreted by each individual observed on a transect (NH₄⁺ µmol·m^-2·hr^-1) using the family-level mass-specific equations determined in Shrestha et al. (2024). The formulas estimate ammonium excretion by fish weight (both log transformed); thus, we converted the estimated lengths of each fish observed on a transect to weight (in grams, wet mass) using published weight-length formulas. We then summed the excretion of each individual observed on a transect over the three levels of the water column (bottom, midwater, canopy) to estimate total ammonium excreted by all fish per transect per area of benthos. The total rate of excretion per unit area at each site was then averaged across transects each year, with the mean excretion in a site-year as the lowest level of replication.

Testing spatial and temporal variability in fish community excretion

We first evaluated spatial differences in the supply of ammonium by fishes among islands due to MPA protection, using data from n=33 sites (15 MPA and 18 non-MPA) surveyed for at least 3 years (all sites). We also explored spatial and temporal patterns in excretion using data from a subset of n=19 sites (10 MPA and 9 non-MPA) surveyed concurrently from 2005-2018 (long-term sites).