Locally driven interannual variability of near‐surface pH and Ω_A in the Strait of Georgia

Abstract

Declines in mean ocean pH and aragonite saturation state (ΩA) driven by anthropogenic CO2 emissions have raised concerns regarding the trends of pH and ΩA in estuaries. Low pH and ΩA can be harmful to a variety of marine organisms, especially those with calcium carbonate shells, and so may threaten the productive ecosystems and commercial fisheries found in many estuarine environments. The Strait of Georgia is a large, temperate, productive estuarine system with numerous wild and aquaculture shellfish and finfish populations. We determine the seasonality and variability of near‐surface pH and ΩA in the Strait using a one‐dimensional, biophysical, mixing layer model. We further evaluate the sensitivity of these quantities to local wind, freshwater, and cloud forcing by running the model over a wide range of scenarios using 12 years of observations. Near‐surface pH and ΩA demonstrate strong seasonal cycles characterized by low pH, aragonite‐undersaturated waters in winter and high pH, aragonite‐supersaturated waters in summer. The aragonite saturation horizon generally lies at ∼20 m depth except in winter and during strong Fraser River freshets when it shoals to the surface. Periods of strong interannual variability in pH and aragonite saturation horizon depth arise in spring and summer. We determine that at different times of year, each of wind speed, freshwater flux, and cloud fraction are the dominant drivers of this variability. These results establish the mechanisms behind the emerging observations of highly variable near‐surface carbonate chemistry in the Strait.

The sensitivity of estuarine aragonite saturation state and pH to the carbonate chemistry of a freshet-dominated river

“…These same coastal and topographic features create long residence times, causing carbon to accumulate and making the strait DIC-rich relative to the open ocean ) despite a strong, seasonal DIC upwelling signal over the outer shelf (Bianucci et al, 2011). The large freshwater footprint, together with the abundance of previous circulation (e.g., LeBlond, 1983;Pawlowicz et al, 2007), ecology (e.g., Masson and Peña, 2009;Allen and Wolfe, 2013), and acidification studies (e.g., Moore-Maley et al, 2016;Ianson et al, 2016) make the strait an ideal system for investigating the response of pH and A to freshwater carbonate chemistry in a complicated estuarine setting.…”

“…The dynamics have been well studied (e.g., Pawlowicz et al, 2007), giving us the information to effectively parameterize higherdimensional processes with a one-dimensional (1-D) model (Collins et al, 2009;Allen and Wolfe, 2013). The benefits of a 1-D model are the quick run times that allow us to simulate many parameter variations over multiple annual cycles ( Moore-Maley et al, 2016). Accurately simulating mixing of the stratified plume with the waters below is required to reproduce the biology and chemistry of the plume.…”

“…For each permutation of TA f and DIC f : TA f , we ran the model over the same 12-year period (2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012) used in previous studies involving this model (e.g., Moore-Maley et al, 2016) to maintain consistency with previous work and to ensure a wide range of climatological forcing regimes. With 18 possible combinations of TA f and DIC f : TA f , we ran the model a total of 216 times.…”

“…We use profiles of temperature, salinity, fluorescence, chlorophyll a, nitrate, and dissolved silica (1999 through 2012) measured near the model site (Pawlowicz et al, 2007;Masson, 2006;Masson and Peña, 2009;Peña et al, 2016) (Diane Masson, personal communication, 2014) to initialize the model (see Moore-Maley et al, 2016). Model runs are initialized in fall and run through a full year and then beyond until the end of the following December.…”

“…Initial phytoplankton, zooplankton, and detritus concentrations are determined according to Moore-Maley et al (2016). Since few DIC and TA data are available, we use a representative fall profile (11 September 2011 to initialize model DIC and TA.…”