Effective stewardship of ecosystems to sustain current ecological status or mitigate impacts requires nuanced understanding of how conditions have changed over time in response to anthropogenic pressures and natural variability. Detecting and appropriately characterizing changes requires accurate and flexible trend assessment methods that can be readily applied to environmental monitoring datasets. A key requirement is complete propagation of uncertainty through the analysis. However, this is difficult when there are mismatches between sampling frequency, period of record, and trends of interest. Here, we propose a novel application of generalized additive models (GAMs) for characterizing multi-decadal changes in water quality indicators and demonstrate its utility by analyzing a 30-year record of biweekly-to-monthly chlorophyll-a concentrations in the San Francisco Estuary. GAMs have shown promise in water quality trend analysis to separate long-term (i.e., annual or decadal) trends from seasonal variation. Our proposed methods estimate seasonal averages in a response variable with GAMs, extract uncertainty measures for the seasonal estimates, and then use the uncertainty measures with mixed-effects meta-analysis regression to quantify inter-annual trends that account for full propagation of error across methods. We first demonstrate that nearly identical descriptions of temporal changes can be obtained using different smoothing spline formulations of the original time series. We then extract seasonal averages and their standard errors for an a priori time period within each year from the GAM results. Finally, we demonstrate how across-year trends in seasonal averages can be modeled with mixed-effects meta-analysis regression that propagates uncertainties from the GAM fits to the across-year analysis. Overall, this approach leverages GAMs to smooth data with missing observations or varying sample effort across years to estimate seasonal averages and meta-analysis to estimate trends across years. Methods are provided in the wqtrends R package.
The ecosystem response to anthropogenic nitrogen (N) loading in estuarine systems is determined by hydrodynamics, biogeochemical transformation rates, and other system-specific characteristics. Historically, San Francisco (SF) Bay has been an outlier from other estuaries with an unusual resistance to eutrophication, despite having extremely high rates of nitrogen loading. Recent increases in phytoplankton biomass and an unprecedented harmful algal bloom, however, have increased the urgency to understand rates and drivers of nitrogen removal in the system. To assess benthic N cycling rates, we conducted seasonal measurements across nine sites in South and Lower South SF Bay, the two sub-embayments with the highest rates of area-normalized N loading, to determine the rates and potential drivers of denitrification and dissimilatory nitrate reduction to ammonium (DNRA). Denitrification rates averaged 60.6 ± 8.1 µmol m−2 h−1 and were primarily coupled to nitrification. Denitrification rates were positively correlated with DNRA rates and % clay. DNRA rates ranged from 0 to 20 µmol m−2 h−1 and on an annual basis averaged ~ 10% of total benthic nitrate reduction, with a negative correlation to % clay content in the surface sediment. The measured denitrification rates account for the removal of, on average, 14% of N loaded annually to South SF Bay, leaving a sizeable portion for alternate fates (e.g., recycling, export, or burial) and potential for substantial temporal and spatial variability (1–79%). This identifies the relative importance of sediment denitrification in ecosystems characterized by high nutrients and low productivity.
San Francisco Bay, California, typically has chlorophyll a (chl- a ) concentrations below 10 µg L −1 , despite nutrient loadings exceeding those in many estuaries with recurring harmful algal blooms (HABs). However, in August 2022 there was a Heterosigma akashiwo (raphidophyte) bloom with chl- a concentrations exceeding 450 µg L −1 , resulting in widespread hypoxia and fish die-off. We used protist community (18S ribosomal ribonucleic acid gene metabarcoding) and environmental data collected 2–3 times per month from 2015–2023 to assess differences between 2022 and other years, and capture dynamics pre-, during, and post-bloom. The 2022 protist community clustered separately from other years in non-metric multidimensional scaling and hierarchical cluster analyses. A diatom bloom in June generated the highest median June chl- a concentration (15 µg L −1 ) recorded between 1977–2023 and drew down nutrients (nitrogen, silicate), leading to the establishment of a heterotrophic community in July with low dinoflagellate abundance (0.9% of protists versus 5.9% average in other years). The absence of dinoflagellates, potentially including H. akashiwo predators , plus suitable habitat (13–14 daylight hours, temperatures 17–21˚C, nitrate concentrations 17–24 µmol L −1 , neap tides) likely provided opportunity for H. akashiwo to proliferate from 0.03% to 48% of protists between July 20 and August 25. As the bloom collapsed, the genus Gyrodinium (dinoflagellate predator of H. akashiwo ) increased from 3.0% to 52% of protists between August 25 and August 31, likely exerting grazing pressure. Results show that antecedent protist community dynamics play important roles in shaping HABs, and interactions between biotic and abiotic factors govern the development of blooms.