Wednesday, March 25, 2015

Ocean Acidification: Natural Cycles and Ubiquitous Uncertainties

In 2002, Scripps’ esteemed oceanographer Walter Munk argued for the establishment of an Ocean Observation System reporting,  much of the twentieth century could be called a “century of undersampling” in which “physical charts of temperature, salinity, nutrients, and currents were so unrealistic that they could not possibly have been of any use to the biologists. Similarly, scientists could find experimental support for their favorite theory no matter what the theory they claimed.” Due to that undersampling MIT’s oceanographer Carl Wunsch (2006) likewise noted, “Among the more troublesome distortions now widely accepted, one must include the notion that the ocean circulation is a simple “conveyor belt” and that the Gulf Stream is in danger of ‘turning off’.”

Another such favorite theory, mistakenly offered as a fact, speculates we are now witnessing increasing anthropogenic ocean acidification, despite never determining if current pH trends lie within the bounds of natural variability. Claims of acidification are based on an “accepted scientific paradigm” that “anthropogenic CO2 is entering the ocean as a passive thermodynamic response to rising atmospheric CO2.” Granted when all else is equal, higher atmospheric CO2 concentrations result in more CO2 entering the oceans and declining pH. But the ever-changing conditions of surface waters exert far more powerful effects. Whether we examine seasonal, multi-decadal, millennial or glacial/interglacial time frames, ocean surfaces are rarely in equilibrium with atmospheric CO2. Relative to atmospheric CO2, seasonal surface water can range up to 60% oversaturated due to rising acidic deep water. Due to the biological pump, CO2 concentrations can be drawn down, leaving surface waters as much as 60% under‑saturated (Takahashi 2002). Thus we cannot simply attribute trends in surface water pH to equilibration with atmospheric CO2. We must first fully account for natural ocean cycles that raise acidic waters from deeper layers and the biological responses that pump CO2 back to ocean depths.

[note: in this essay I use “acidic” in a relative sense. For example, although the pH of ocean water is 7.8 at 250 meters depth and is technically alkaline, those waters are “more acidic” relative to the surface pH of 8.1.]

To appreciate the importance of pH altering dynamics, consider the fact that pure water has a neutral pH of 7.0. Rainfall quickly equilibrates with atmospheric CO2, and pH falls to ~5.5. Dark‑water rivers such as the Rio Negro drop to pH 5.1. In contrast, due to a combination of biological activities and geochemical buffering, the average pH of ocean surfaces (and some rivers) rises to ~8.1. In other words, after equilibration with atmospheric CO2, powerful factors combine to remove 99.8% of all acidifying hydrogen ions from rainwater. The balance between upwelled acidic waters versus carbon sequestration and export by the “biological pump” is the key factor maintaining high pH in oceanic surface waters, and the communities of plankton that operate that pump undergo changes on seasonal, multidecadal and millennial time scales; changes we are just beginning to understand.

In Bates 2014, A Time-Series View of Changing Surface Ocean Chemistry Due to Ocean Uptake of Anthropogenic CO2 and Ocean Acidification, they simplistically argued declining ocean pH is “consistent with rising atmospheric CO2. But a closer examination of each site used in their synthesis suggests their anthropogenic attribution is likely misplaced.  For example, at the Hawaiian oceanic station known as HOT, based on 10 samplings a year since 1988, researchers reported a declining pH trend. But that trend was not consistent with invasions from atmospheric CO2. An earlier paper (Dore 2009) had observed, “Air-sea CO2 fluxes, while variable, did not appear to exert an influence on surface pH variability. For example, low fluxes of CO2 into the sea from 1998–2002 corresponded with low pH and relatively high fluxes during 2003–2005 were coincident with high pH; the opposite pattern would be expected if variability in the atmospheric CO2 invasion was the primary driver of anomalous DIC accumulation. (DIC is the abbreviation for Dissolved Inorganic Carbon referring to the combined components derived from aqueous CO2, including bicarbonate and carbonate ions.)

Those higher fluxes of CO2 into the surface likely stimulated a more efficient biological pump resulting in higher pH. That rise in pH is consistent with experimental evidence demonstrating CO2 is often a limiting nutrient (Riebesell 2007), and adding CO2 stimulates photosynthesis. That most photosynthesizing plankton have CO2 concentrating mechanisms suggests CO2 is often in chronic short supply.

The greatest concentrations of CO2 upwell from depth to invade surface waters. As seen below in the illustration by Byrne 2010 from the northern Pacific, the ocean’s pH (thus the store of DIC) rapidly drops from 8.1 at the surface to 7.3 at 1000 meters depth. Dynamics such as upwelling bring deeper waters to the surface reducing pH, while dynamics such as the biological pump shunt carbon back to deeper depths and raise surface pH.  At the risk of oversimplifying a myriad of complex dynamics, oceans basically undergo a 4-phase cycle that determines the average annual surface pH. Any adjustments to this cycle will alter trends in pH over decadal to millennial time periods.

Vertical profile ocean pH

Phase 1: Varied rates of upwelling and winter mixing raises acidic water to the sunlit surface  
and lowers pH.

Phase 2: Specific plankton communities, largely diatoms respond quickly to the arrival of
nutrients in the surface waters, and rapidly sequester and export carbon back to depth. Phase-2 productivity also generates dissolved and suspended organic carbon that is transported laterally to other regions. When community photosynthesis absorbs CO2 faster than respiration releases it or upwelling injects it, surface pH rises.

Phase 3: As available nutrients are depleted, diatom populations dwindle and other plankton
communities dominate such as coccolithophores and photosynthesizing bacteria. Instead of rapidly exporting carbon, this plankton community is better at retaining and utilizing nutrients. The utilization of suspended and dissolved organic carbon and increased grazing by populations of zooplankton increase respiration rates relative to new photosynthesis, so pH declines.

Phase 4: A “regional equilibrium” is established as accumulated organic carbon from previous
phases is depleted and new, but lower, levels of productivity are balanced by community respiration. That balance raises pH. This equilibrium is fleeting and lasts until a new burst of nutrients reaches sunlit waters. The supply of nutrients rising to the surface cycles seasonally as well as over decades, millennia and glacial/interglacial intervals, so that short interval trends are embedded in much longer trends. This is one reason why computed pH trends by Bates 2014 statistically explained only a minor portion of pH variability even after removing seasonal trends.


First consider that oceans store 50 times more CO2 than the atmosphere. A small change in the rate by which deep acidic water reaches the surface is the major determinant of surface pH trends. Nutrients, acidity, and density increase with depth, but not all depths contain a balanced supply of nutrients critical for photosynthesis. To bring denser water to the surface requires a significant input of energy that is primarily provided by the winds or tides (Wunsch 2004). Stronger winds generate more upwelling and winter mixing.  Thus cycles of oceanic and atmospheric circulation that strength and weaken winds, raise varied combinations and concentrations of nutrients to the surface, which accordingly affects the biological pump and pH.

For example in temperate oceans, winter cooling of surface waters allows winds and storms to mix surface waters with CO2 rich waters from as deep as 500 meters. This lowers surface pH, so that relatively insignificant inputs from atmospheric CO2 are undetectable. (Takahashi 2002, 1993). Several researchers have observed significant correlations between winter mixing and the North Atlantic Oscillation (Ullman 2009, Steinberg 2012). A positive NAO is associated with stronger westerly winds and also correlates with a stronger subpolar gyre. Counter-clockwise gyres in the northern hemisphere increase regional upwelling when they strengthen. So changes in NAO-driven upwelling cause multi-decadal oscillations in the plankton communities and pH.

In the Pacific, El Nino years strengthen the Aleutian Low and the Pacific subpolar gyre, similarly increasing regional upwelling. In contrast during La Nina years, gyre upwelling decreases but trade winds speed up and intensify coastal and equatorial upwelling. The frequency of El Niño’s vs La Niña’s varies over 40 to 60 year cycles of the Pacific Decadal Oscillation. Although periods of increased upwelling decreases pH, due to undersampling it is not clear how this extrapolates across the whole Pacific Basin during the 20th century.

Upwelling also varies on millennial scales. During the Roman Warm Period, Medieval Warm Period and the Current Warm Period, La Nina-like conditions with stronger trade winds dominated (Salvatteci 2014) causing above average upwelling and higher productivity. During cooler periods like the Dark Ages and Little Ice Age, the Pacific was dominated by El Nino-like conditions with less upwelling and lower productivity. Claims that oceans have acidified since the Little Ice Age due to anthropogenic CO2 (Caldeira 2003) may be true, but the uncertainties are huge. It is just as likely increased upwelling caused more acidic modern oceans, or it is equally possible that modern oceans are less acidic if increased upwelling stimulated a biological pump that sequestered and exported enough carbon to offset acidic upwelling.

Global ocean acidification is determined by averaging sink regions with out‑gassing source regions. Opposing regional trends add significant uncertainty when determining global calculations. As illustrated by the yellows and reds in the Martinez-Boti (2015) illustration below, there are vast regions where so much DIC is upwelled, on average the ocean is exhaling CO2. Regions that are net sources of out-gassing CO2 experience lower pH solely due to upwelling of ancient waters, and the pH is lower than predicted from simple equilibration with the atmosphere.

Oceanic regions of outgassing CO2 sources and CO2 sinks

Paradoxically, oceans also experience acidification if weakening winds reduce upwelling. For example due to changing locations and strength of the InterTropical Convergence Zone (ITCZ), trade winds over northern Venezuela’s Cariaco Basin undergo decadal and centennial shifts in strength. When the ITCZ moved south during the Little Ice Age, upwelling and productivity in the Cariaco Basin declined. At the end of the LIA, the ITCZ began moving northward and upwelling and productivity increased (Gutierrez 2009). Recently the ITCZ moved further northward due to more La Niña’s and the negative Pacific Decadal Oscillation, and regional winds declined. Consequently researchers reported anomalously shallow seasonal upwelling that brought more DIC to the surface but fewer critical nutrients that reside at lower depths. This resulted in decreased productivity and a decrease in diatom populations. Less productivity and less carbon export did not offset upwelled DIC, so the regional pH declined (Astor 2013). Despite Astor serving as a co-author, Bates 2014 oddly failed to mention this pH altering dynamic, choosing to attribute Cariaco’s declining pH trend to rising anthropogenic CO2.

In contrast to the Cariaco Basin, a negative Pacific Decadal Oscillation increases upwelling along the Americas west coast, stimulating the highly productive/high carbon-export community of phase-2. Upwelled DIC is quickly sequestered and exported by large single-celled diatoms. With their relatively heavy siliceous shells, dead diatoms rapidly sink carrying carbon to the sea floor. Larger zooplankton graze on diatoms and their large fecal pellets and carcasses also carry carbon rapidly to depth.  Diatom blooms along California and Oregon spark increased krill and anchovy populations, which attract feeding humpback whales from Costa Rica and seabirds like the Sooty Shearwater from New Zealand, confounding any attempts accurately measure the carbon budget.

As illustrated in the Evans et al graph below, coastal upwelling can over‑saturate the surface waters to 1000 matm, 2.5 times above atmospheric pCO2 (represented by dashed horizontal line). Within weeks the biological response sequesters and exports that carbon so that concentrations of surface water CO2 fall as low as 200 matm; a concentration that would still be under-saturated relative to the Little Ice Age’s atmosphere. Relative to these rapid seasonal changes in pH, fears that marine organisms cannot adapt quickly enough to the relatively slower changes wrought by anthropogenic CO2 seem overblown.

Upwelling and the Biological pump along the Oregon Coast

Still such fears filter researchers’ interpretations. Along the west coast of North America, planktonic sea snails called pteropods, begin life feeding on algal blooms ignited by seasonal coastal upwelling. As illustrated in scanning electron micrograph “a”, shown below from (Bednarsek 2014), pteropod shells are heavily dissolved during the first few weeks of life due to acidic upwelled water. Picture  “b” shows a larger more mature shell with the outer part of the shell experiencing no dissolution. As the snails matured, either upwelled acidic waters subsided or the snail was transported seaward to less acidic waters by the same currents that promoted upwelling. The result is pteropod shell dissolution is a very localized, short duration phenomenon.

Nonetheless in a study sponsored by NOAA’s Ocean Acidification Program Bednarsek 2014 argued those examples of shell dissolution were caused by anthropogenic carbon writing, “We estimate that the incidence of severe pteropod shell dissolution owing to anthropogenic OA has doubled in near shore habitats since pre-industrial conditions across this region and is on track to triple by 2050.” But such “conclusions” are unsupported speculation at best. The study failed to determine if upwelled waters were any more acidic now than during any other seasonal or La Nina upwelling event. Most studies suggest upwelling declined during the Little Ice Age, and the resumption of stronger upwelling is the result of a natural cycle. But Bednarsek (2014) simply used a formula equilibrating past and present atmospheric CO2 to compute surface water pH. But such methodology is meaningless. No net CO2 diffusion from the atmosphere to surface waters occurs when upwelling has oversaturated surface pCO2, and as shown in the Evans et al graph, due to the biological pump surface waters remained undersaturated relative to both current and LIA atmospheric CO2. Shame on those NOAA scientists for such biased interpretations.

Dissolution of pteropod shells from Bednarsek 2014

On all time frames, when upwelling subsides and nutrients and carbon become scarce, diatom populations dwindle and oceans transition to Phase 3. Coccolithophore and bacterial communities that were relatively minor constituents, begin to dominate. Smaller bacteria remain suspended in the surface layers and export much less carbon. Grazing on increasingly abundant bacteria and accumulated organic carbon, promotes greater zooplankton populations.  As a result, community respiration rates increase, and higher CO2 concentrations lower surface pH.


Coccolithophores are large single-celled alga encased by several ornate calcium-carbonate “coccoliths”, so that sinking dead individuals do export carbon relatively quickly. However the construction of coccoliths metabolically increases surface pCO2, lowers pH and counteracts the “biological pump”. When calcium combines with carbonate ions to form coccoliths, the reaction releases acidifying CO2. Likewise the growth of pteropods’ calcium carbonate shells also increases CO2. It seems paradoxical that one of the greatest fears of ocean acidification is the dissolution of carbonate shells, yet the very process of creating those shells increases acidification and lowers surface alkalinity.

Several researchers suggest coccolith formation evolved to provide much needed CO2 for photosynthesis in under-saturated waters. Experimental evidence reveals higher concentrations of CO2 results in lower rates of coccolith formation but proponents of worrisome acidification argue this is evidence of acidification’s deleterious effects. However the same response would be expected if the rate of coccolith formation responds to the available supply of CO2 required for photosynthesis. Furthermore if they are so vulnerable to acidification, how did coccolithophores evolve and survive over 200 million years ago, when atmospheric CO2 was at least 2 to 3 times higher than today?

Without copious supplies of nutrients from upwelling, productivity in subtropical gyres is much lower and diatoms constitute a minor component of that plankton community. But they still undergo cyclic changes. In the Atlantic, Steinberg (2012) describes a 113% decrease in diatoms between 1990 and 2007 in contrast to stable coccolithophore populations and a rapidly increasing community of photosynthesizing bacteria. In turn rapidly increasing communities of small zooplankton graze on the bacteria resulting in increased community respiration rates. Three sites from Bates 2014 are located in subtropical gyres: HOT near Hawaii, BATS near the Bermuda and ESTOC near the Canary Islands. And all three are exhibiting these classic phase-3 patterns with increasing respiration rates (Lomas 2010, Gonzalez-Davila 2003, Peligri 2005, Steinberg 2012), which accounts for declining pH trends. As shown by Steinberg 2012, those trends are significantly correlated with multi-decadal climate indices – the North Atlantic Oscillation plus three different Pacific Ocean climate indices”.

Global pH decreased when oceans transitioned from the Last Glacial Maximum (LGM) to the current interglacial Teleconnections between the Atlantic and Pacific have been confirmed  as warm periods in the Greenland ice core correlate with periods of extended periods of upwelling along the California coast (Ortiz 2005). Recent research also links simultaneous multi‑millennial cycles of upwelling and higher productivity in the sub‑Antarctic Atlantic and equatorial Pacific. Most research suggests that at the end of the LGM, Antarctic began to warm followed by a rise in atmospheric CO2. Although the precise mechanism of CO2 out‑gassing during the deglacial period has been under debate, there is a growing consensus that circulation changes caused aged waters rich in nutrients to upwell in subpolar Antarctic waters. Via oceanic tunneling, those deep Antarctic waters also upwelled in the equatorial eastern Pacific. Using foraminifera proxy data, the graphic below from Martinez-Boti (2015) shows periodic upwelling of subpolar Antarctic waters (on the left in blue) caused regional pH to decline from the LGM maximum of 8.4 to about 8.1 at the beginning of the Holocene. Due to the biological pump and/or reduced upwelling during the early and mid Holocene, pH rises and bounces between 8.25 and 8.15.

Based on CO2 concentrations determined from Antarctic ice cores, Martinez‑Boti also constructed a green “Equilibrium pH” trend indicating the surface pH if it had simply equilibrated with atmospheric CO2.  For most of the past 20,000 years, surface waters were not in atmospheric equilibrium and more acidic, so those regional oceans were typically a source of out‑gassing CO2. The graphs on the right (in red) show the same pattern for the equatorial eastern Pacific but with data that extends further into the LGM.

Ocane pH variability over pst 20,000 years from Martinez‑Boti 2015

Calvo 2011 examined ocean sediments to determine the strength of upwelling versus the biological pump plus the relationship between diatoms and coccolithophores over the past 40,000 years. Their research found lower productivity during the LGM and lower diatom abundance relative to coccolithopheres. As upwelling increased around 20,000 years ago so did ocean productivity and the proportion of diatoms. They concluded upwelling enhanced the biological pump but it was “not sufficient to counteract the return to the atmosphere of large amounts of CO2 delivered by the oceans through an enhanced ventilation of deep water.”

Ratio of Diatoms to Coccolithophores from Calvo 2011

Finally examining sediments in the eastern equatorial Pacific, Carbacos 2014 found “a clear prevalence of dominant La Niña-like conditions during the early Holocene, with an intense upwelling and high primary productivity.” High levels of productivity persisted through the Holocene Optimum until productivity dramatically declined around 5,500 years ago. Since that time Carbacos 2014 reports, “An alternation between El Niño-like and La Niña-like dominant conditions occurred during the late Holocene, characterized by a clear trend toward prevailing El Niño-like conditions, with a low primary productivity.” During the past 5,000 years, that lower productivity coincided with increased dominance of coccolithophores and declining proportions of diatoms. That suggests the oceans have been in a phase-3 multi-millennial decline in pH superimposed on multidecadal cycles driven by the Pacific Decadal and Atlantic Multidecadal Oscillations.

It is also worth noting, as seen in the graph below, throughout the Holocene changes in atmospheric CO2 did not correlate with temperature. However atmospheric CO2 did track changing plankton communities. During the early and late Holocene, atmospheric CO2 concentrations were relatively low and stable during periods of high productivity with higher ratios of diatoms. When ocean productivity crashed overall around 5,000 years ago, the proportion of CO2 producing coccolithophores increased and atmospheric CO2 likewise increased by about 20 matm. A similar annual increase in CO2 has been observed in modern oceans and similarly attributed to increased proportions of coccolithophores (Bates 1996). 

So where are the oceans headed? If history repeats itself, declining solar insolation will result in less upwelling, lower productivity, a reduced biological pump and higher pH. Or perhaps higher levels of atmospheric CO2 will increase productivity as observed in several experiments, or perhaps rising CO2 will cause a deleterious decline in pH?  The ubiquitous uncertainties from the current undersampling of oceans allows anyone to “find experimental support for their favorite theory no matter what the theory they claimed.” But I can say for sure, I would not trust any predictions that failed to account for changes in upwelling and the various responses of the biological pump.

Contrasting Holocene Temperatures and Atmospheric CO2