Translate

Friday, April 17, 2015

Audubon Society: Climate Science or just Sticking Feathers on PIGs


September 2014 the Audubon Society launched their climate change campaign with a most remarkable assertion: 314 of 588 bird species are “on the brink” and “will lose 50 percent or more of their current ranges by 2080” due to rising CO2. Avid birder and renowned author Jonathan Franzen shared his resentment of these claims in a recent New Yorker article arguing that focusing on futuristic effects of climate change distracts conservationists from dealing with more immediate problems that can be more readily dealt with. Others were dubious of the hype because many of Audubon’s “climate endangered” species have been enjoying increasing population trends such as the recovering Bald Eagle. While Audubon contends that their provisional scenarios will help future conservation efforts, others have whispered that such an apocalyptic media campaign smacks of a crass fundraising gimmick that relies on dubious models and naive fears.



Are Audubon’s models so reliable they can justify hyping catastrophic conclusions? Will “Audubon science” promote better environmental stewardship?  Or, are their projections just another example of misplaced alarmism that has also obscured the critical issues facing butterflies, polar bears, emperor penguins, golden toad, pika or moose. Although we cannot ascertain Audubon’s intent, nor scientifically validate their projections for 2080, we can examine the skill of their models and the trustworthiness of their predictions. This essay illustrates the tremendous uncertainty of Audubon’s models and highlights some of the current research that presently contradicts Audubon’s predictions. Models that provide Pervasive Inadequate Generalizations are PIGs, and PIGs never provide reliable guidelines for wise environmental stewardship. The technical report behind Audubon’s apocalyptic media blitz simply merged Bioclimatic Envelope Models and downscaled Global Climate Models, and both models have been severely challenged within and without the scientific community.

1. Bioclimatic Envelope Models (BEMs) Uncertainty

Bioclimatic envelope models circumscribe the range of temperatures and precipitation that are deemed suitable for an individual species. BEMs are typically not determined by experimentally evaluating the species tolerances for any given range of temperatures and precipitation. BEMs simply correlate the temperatures and precipitation within a species’ current range. The major flaw in these models is the assumption that a species current range is limited solely by those climatic factors and the species is currently in equilibrium with those factors. But the availability of resources and competition with other species will also limit a species range. Landscape changes and overhunting have reduced many species’ range so that their current boundaries may only represent a fraction of what is climatically suitable.

For example, over a century ago the Greater Prairie Chicken ranged from southern Texas to North Dakota. Historically its climatic envelope encompassed a wide range of temperatures (both light and dark green in map below). However due to extensive hunting and habitat loss it was extirpated from most of its historic range (light green). A bioclimatic envelope based only on temperatures in its current range (dark green) would suggest the Greater Prairie Chicken depends on cooler temperatures of the northern Great Plains. But whether natural or man‑made factors raise temperatures 1-2 degrees by 2080, those temperatures would still be within the historical range experienced by Greater Prairie Chicken that once thrived in the southern end of its range.

Greater Prairie Chicken Historic Range


BEMs assume each species genetically conserves its reliance on a specified climate niche over millennia.  For that reason we believe species contracted their ranges towards the equator (or persisted in unique climate refugia) during the last ice age. Conversely, we must likewise assume birds expanded their ranges pole‑ward 6000 years ago during the Holocene Optimum when Northern hemisphere temperatures were 1° to 6°C warmer than today. Typical for most of the northern hemisphere, multi‑proxy evidence suggests the Great Plains were much warmer than today for most of the mid‑Holocene (Fig. 2). Whether man-made of natural, if future warming pushes species pole‑ward, would it be catastrophic as Audubon suggests? Or would species simply re‑colonize habitat that was lost due to the Little Ice Age between 1300 and 1850 AD?  The only reason species of the Great Plains would not re‑colonize the prolific grasslands of the Holocene Optimum, would have nothing to do with the current climate. It has everything to do with fire suppression and the loss of over 90% of the grasslands in most regions to agriculture and development.

 
Global warming good
Holocene Climate Change on Northern Great Plains 
BEMs are not determined by a species’ physiological limits. Indeed those limits have never been determined for most species. In addition, given that many species are confined to unique habitat and plant associations, a species’ current range is in large part limited by the climatic boundaries of its preferred vegetation (i.e. grasslands, forests, wetlands, etc.). This is why the consensus among conservation biologists is habitat loss has been the greatest threat to birds.  So it is highly likely that the ranges and abundance of bird species expanded and contracted in concert with expanding plant species during the Holocene. Just as Holocen warming benefitted grassland expansion, 9000 years ago tree line expanded to the shores of the Arctic Ocean, 100s of kilometers further north than observed today. In California researchers report that Sierra Nevada tree line was at higher elevations for most of the past 3500 years, but was pushed to lower elevations during the Little Ice Age. In some regions of Eurasia’s Ural Mountains, the cold of the Little Ice Age prevented any new trees from spouting for hundreds of years. The current warming that began 150 years ago has enabled a more productive forest ecosystem, so we can infer this warming has also been beneficial for bird species of the forest.  

Nonetheless, even if BEMs could fully determine the historical range of a species’ suitable macro‑climates over millennia, BEMs cannot predict how birds will exploit the varied habitats and micro-climates within that range. Paleontologists are increasingly finding “enclaves of benign environmental conditions within an inhospitable regional climate” that allowed species to persist during the Last Glacial Maximum. I have measured micro‑climates within just a 100‑meter radius. Temperatures along a gravelly roadside were 20° to 30°F higher relative to the forested area, and 10° to 15° F warmer than grassy and shrubby areas. In addition to those vegetation effects, varied topography creates a similar wide array of microclimates between north‑facing and south‑facing slopes. As daily temperatures fluctuate by 20° to 30°F, birds can easily exploit a wide variety of micro‑climates. It is likely this great variety of microclimates explains the complex range shifts that are not predicted by “Audubon science” and why so many species have not shifted their range at all over the past century. It also highlights a mechanism that will allow species to persist in their current habitat despite Audubon’s models.

Recent surveys of birds in montane California report that the elevation ranges of 223 breeding  species identified a century ago have not altered either their upper or lower range limits. For those species that did shift their range, just as many species moved down‑slope as up‑slope (Fig, 3). Again the difference appears to be more a function micro-habitats than an individual species response to global climate change. Of 53 species that were common to all 3 transects, (Lassen National Park, Yosemite and Southern California), only 5 species shifted their range in a similar manner. For 91% of the species, one population moved upslope in one region, another moved down‑slope or did not shift at all. Furthermore for those species that moved upslope, increased warmth was unlikely to have been the driving factor. Researchers reported that “although the northern (Lassen) region barely warmed on average over the last century, showing localized areas of marginal warming and cooling, the proportion of bird species shifting there was comparable to the other two regions that experienced substantial warming.”  Such results again argue that BEMs have very little skill predicting how species’ range will shift. It also suggests Audubon’s woeful predictions of 341 species “on the brink” are at best unsupported premature speculation.

birds moving downslope despite global warming
Proportion of California birds shifting breeding ranges upslope, downslope and no shift at all



2. Climate Model Uncertainty

To predict how BEMs would shift in the future, Audubon science employed IPCC global climate model predictions that have suggested uniformly and steadily increasing temperatures across North America (Fig. 4). But downscaled IPCC global climate models are notoriously bad at simulating local and regional climate. A 2015 study reported, “Examining the local performance of the [global] models at 55 points, we found that local projections do not correlate well with observed measurements. Furthermore, we found that the correlation at a large spatial scale, i.e. the contiguous USA, is worse than at the local scale.” This led the authors to ask if  “the most important question is not whether Global Climate Models can produce credible estimates of future climate, but whether climate is at all predictable in deterministic terms.”

Likewise Dr. Roger Pielke Sr. cited several peer-reviewed papers when he blogged, “regionally downscaled forecasts from global multi-decadal climate model predictions have no skill beyond whatever is in the parent global model.”…  “These global multi-decadal predictions are unable to skillfully simulate major atmospheric circulation features such the Pacific Decadal Oscillation [PDO], the North Atlantic Oscillation [NAO], El Niño and La Niña, and the South Asian monsoon.” Yet it those ocean oscillations that are the major drivers of climate change. Johnstone 2014 wrote that natural shifts in the Pacific Ocean’s circulation “account for more than 80% of the 1900–2012 linear warming in coastal NE Pacific SST [Sea Surface Temperatures] and US Pacific northwest (Washington, Oregon, and northern California) SAT [Surface Air Temperatures]. An ensemble of climate model simulations run under the same historical radiative [Greenhouse gasses and solar] forcings fails to reproduce the observed regional circulation trends. These results suggest that natural internally generated changes in atmospheric circulation were the primary cause of coastal NE Pacific warming from 1900 to 2012.”

In contrast to Audubon’s press releases touting modeled results predicting Maryland’s state bird, the Baltimore oriole, would soon be pushed northward and out of the state by global warming, instrumental data suggests no such change. Instrumental records highlight a century long cooling trend in the southeastern USA (Fig. 5), a region that climate researchers refer to as a “warming hole”. Additionally the brutal winters and record low temperatures for the past few years further stand in stark contrast to Audubon’s simulations that project increasing warmth and northward shifting wintering and breeding grounds. Those cooling trends do not refute the hypothesis of a warming contribution from rising CO2, but do demonstrate how greatly regional temperatures can depart from global climate projections due to natural dynamics. It also suggests Audubon’s climate science contributes precious little to bird conservation.

North America warming hole
IPCC warming prediction for  North America


Cooling temperatures in southeastern USA
North America "warming hole"


I have always argued that to be good environmental stewards, we must first understand local climate change, so I am heartened to see researchers are now realizing the Parmesan paradigm of a “coherent global climate fingerprint” does not explain changes in a species range or abundance. Echoing my sentiments, a 2014 research paper Beyond A Warming Fingerprint: Individualistic Biogeographic Responses To Heterogeneous Climate Change In California  the authors wrote, “populations respond to climate locally and local patterns of climate change often differ substantially from global patterns. As a result, we are unlikely to diagnose local climate change impacts using a global fingerprint.” 

I would add we are also unlikely to diagnose climate impacts using just regional average temperatures. The “average” temperatures in California have assuredly increased since the Little Ice Age, but the average has been driven by rising minimum temperatures, which are typically driven by land use changes and urbanization effects. While rising minimum temperatures may impact the rate of snowmelt, rising minimums do not significantly contribute to heat stress.

Only maximum temperatures exert heat stress on plants and wildlife, and compared to the 1900-1939 period, maximum temperatures have declined over most of California (Fig. 6). IPCC climate models failed to predict these regional cooling trends, in part, because IPCC models run hot and overestimate maximum temperatures. Although these cooling trends do not refute the warming potential of rising CO2, these cooling trends again demonstrate that natural climate variability can oppose CO2 warming and dominate surface temperature trends.

Maximum temperatures cooling in Californina
California's cooling maximum temperatures 




Allen’s hummingbird is an excellent example that demonstrates the failure of “Audubon science” when it combines a bad species BEM with an inadequate climate generalizations. Audubon science predicts the Allen’s hummingbird will lose 90 percent of its current breeding range as global warming shifts breeding habitat northward.  But as I watch these hummingbirds flit through my backyard, I know that such a loss will only happen when PIGs fly. In reality maximum temperatures have been cooling throughout most of their range. Second, there are 2 subspecies of Allen's hummingbird; one is migratory, and the other is a non-migratory permanent resident on the Channel Islands off southern California. As noted in Wikipedia, the non-migratory population dispersed to the mainland and colonized the Palos Verdes Peninsula of Los Angeles County in the 1960s. Since that time the subspecies has spread over much of Los Angeles and Orange Counties, spreading south through San Diego County. This southward expansion of breeding habitat towards the warmer regions is the exact opposite of Audubon’s behavioral predictions.

The migratory subspecies’ breeding habitat is generally restricted to California’s coastal fog belt and extends just into southern Oregon. Although there was a slight warming along California’s northern coast since the end of the Little Ice Age, there has been no warming trend its prime‑breeding habitat since 1950’s as observed in both instrumental data and isotope analyses of redwood tree rings (Fig. 7). Again there is no evidence to support Audubon’s dire predictions.

No warming along northern California coast.



So why is Audubon straying from habitat preservation issues to hype unlikely dire predictions that will more than likely give Audubon a black eye? As Franzen noted, climate change is a “ready-made meme, it’s usefully imponderable” and that is well understood by Audubon’s new president and CEO David Yarnold. Yarnold was not hired for his scientific expertise. He is a journalist. Before Audubon, he helped the Environmental Defense Fund double their revenues by pushing a climate change campaign, so it is no surprise that he is repeating those efforts for Audubon. But the new climate agenda seems more than a fundraising campaign. There is a definite political agenda. This week Audubon launched a new social media campaign #ClimateThing. The tactic appears to be less about protecting birds, but the perpetuating a meme that blames everything from the war in Syria to prostitution to stray kittens on rising CO2. We are constantly bombarded with media hype that everything we love is threatened by climate change. Hijacking the real conservation issues that face birds is just another example to be used as a political hammer.

Audubon wrote, “Clean drinking water is a #ClimateThing. Kids’ lungs are a #ClimateThing. Environmental justice, disease prevention, food security, economic mobility, family homes, beaches, ski slopes, vineyards, forests, whales, bats, butterflies, and salamanders–each of these is a #ClimateThing,    What is yours? “

So I suggest skeptics respond. Go on to twitter and tell Audubon your #ClimateThing demands better science, not fear mongering. Tweet David Yarnold (@david_yarnold) and tell him to get real and stop hijacking the sincere concerns of so many of its members. Reply to tweets and link to the analysis here or on my website and ask for explanations to why “Audubon science” diverges so far from reality. Ask how lowering CO2 concentrations will reclaim lost grasslands or restore watersheds that are so critical to birds. Ask how lowering CO2 will protect the truly endangered species on islands because humans introduced rats, cats and stoats against which these birds have no defense. Audubon and real conservation have become another scientific casualty inflicted by the politics of climate change.




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.

 
Diatoms 

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


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

Wednesday, February 4, 2015

Climate Horror Stories That Wont Die: The Case of the Pika (Stewart, 2015)


pika


Because most people can’t fathom how 0.8 degrees of warming over a century can be lethal compared to far greater changes on a daily and seasonal basis, advocates of CO2 warming have littered the media and scientific literature with apocryphal stories statistically linking cherry-picked data with that small temperature rise and suggest wide spread future extinctions (i.e. Polar Bears,  Walrus, Emperor Penguins, Edith Checkerspot, Moose, Golden Toad ). Pikas are another species that have been repeatedly targeted as an icon of impending climate doom. Pikas, or boulder bunnies, inhabit talus slopes (boulder fields) throughout western North America’s mountainous regions. Some suggest warming has been driving pikas up the mountain slopes, and they will soon be driven over the edge into the extinction abyss.

The doomsday stories of the pikas’ “impending extinction” began with a few contentious papers by Dr. Erik Beever. He re-surveyed a small subset of Nevada’s pika populations and reported 28% (7 of 25)  of pika territories, which had been occupied at the beginning of the 20th century, were now vacant. He suggested those 7 populations had gone extinct possibly due to climate change. That claim was then trumpeted by groups like the National Wildlife Federation with articles like “No Room at the Top.”

As they had done for polar bears and penguins, the Center of Biological Diversity argued climate change was threatening species with extinction and sued for pikas to be listed as federally endangered once, and as California endangered twice. The CBD alarmingly exaggerated Beever’s small survey to falsely report, “We've already lost almost half of the pikas that once inhabited the Great Basin.”  But to the credit of official wildlife experts, they rejected those lawsuits due to insufficient evidence. Dismayed that bad science had been rejected, the CBD called Obama a denier and Joe Romm bemoaned, “So long pika, we hardly knew ya.”

Now once again, dubious science is pushing pikas as another canary in the climate coal mine. Although the evidence has not supported the pika’s demise, Stewart (2015) constructed a model that would and published their projections in Revisiting The Past To Foretell The Future: Summer Temperature And Habitat Area Predict Pika Extirpations In California. These researchers predict “that by 2070 pikas will be extirpated from “39% to 88%” of California’s historical sites.  And once again the media is hyping that pikas are being pushed up the mountains to their doom.

In contrast to the hype, Dr. Andrew Smith, the International Union for the Conservation of Nature pika expert, has testified that pikas are thriving in California and should not be listed.  Although an avid defender of the Endangered Species Act, he argued that incorrectly listing the pika as endangered (see his letter here) would only subject the ESA to greater criticism and denigrate conservation science.

Due to possible climate change concerns, the US Forest Service was obligated to extensively survey pika habitat throughout the national forests of the Sierra Nevada and the Great Basin. Supporting Dr. Smith’s views, in 2010 they too reported thriving pikas. Overall, only 6% of observed pika territories were vacant. Due to the lack of connectivity with other pika territories, when a pika dies the smaller more isolated territories suffer longer periods of vacancy. Accordingly, the USFS reported that vacancy rates increased as surveys moved from the Sierra Nevada with its large interconnected talus slopes to more isolated habitat in the Great Basin. In the Sierra Nevada the vacancy rate was just 2%, in the southwestern Great Basin vacancies increased to 17%, and vacancies were highest, 50%, in more isolated habitat of the central Great Basin ranges. The larger percentage of unoccupied sites east of the Sierra Nevada crest was typically due to the greater difficulty of finding and re‑colonizing relatively small and isolated habitat.


USFS surveys provided more damning evidence that would lead to rejecting the CBD’s lawsuits. The benchmark for wildlife abundance and distribution in California had been Joseph Grinnell’s surveys from the early 1900s. Contrary to global warming theory, the USFS survey found many new active pika colonies several hundred meters lower than Grinnell had documented. In total, 19% of the currently known populations are at lower elevations than ever documented by any study during the cooler 1900s. Further north in the Columbia River Gorge, another independent researcher also found pikas at much lower elevations, surviving at temperatures much higher than the models had predicted.
Beever’s 2011 paper tried to counter those findings by arguing there was a nearly “five-fold increase in the rate of local extinctions and an 11-fold increase in the rate of upslope range retraction during the last ten years.” But Beever had badly manipulated his data. Surveying his 25 sites, he too had found 10 examples where pikas now inhabited lower elevations than previously documented. But he decided not to use those observations in his calculations. He, the editors and peer-reviewers unapologetically published his biased calculations to create his “11-fold increase in the rate of upslope range retraction”. Beever defended this statistical blasphemy by arguing pikas had likely always lived at those lower elevations, but had escaped detection by earlier observers (the equivalent of climate science infilling). Perhaps. It was possible. But by eliminating all new observations of pikas at warmer, lower elevations, he guaranteed their statistical upslope retreat.

Here’s an example of his calculations:   At Cougar Peak, a 1925 record documented the lowest elevation that pikas had inhabited was 2416 meters. Beever’s more recent surveys detected pikas living even lower on Cougar Peak at 2073 meters in the late 1990s, and at 2222 metes in  follow‑up surveys in 2003. Despite the fact that recent observations were all lower than 1925 by about 200 meters, Beever ignored the historical record. He simply subtracted the 1990s elevation from the 2003 elevation, to report climate had pushed pikas 149 meters higher. Furthermore, the Cougar Peak site was one of the sites Beever had initially reported as extinct.  Follow-up surveys found a robust population.

Vacant pika territories are natural and to be expected. Pika are very territorial and each year they drive their young away. Because pika live no longer than 7 years, (averaging 3 to 4 years in the wild), there is constant turnover at each site. A site remains vacant until a young pika, driven from another territory, randomly scampers into that vacancy and claims ownership. Without knowing how often a talus pile alternates between occupied and vacant, simply reporting observations of a vacant site tells us nothing about 1) why it is vacant, 2) when it was vacated, and 3) if it will soon be recolonized. Unfortunately vacancies have been misleadingly called extinctions. To illustrate, in the most recent paper by Stewart, his team initially found 15 vacancies, but a re‑survey the following year, found that 5 of those sites were now re‑colonized, a 33% reduction in “extinct locations” in just one year.

Re‑colonization has similarly undermined other classic doomsday stories. Parmesan’s iconic 1996 paper reported global warming had increased extinctions for the Edith’s Checkerspot butterfly, but most of those extinct colonies in the Sierra Nevada have now been re‑colonized. Unfortunately the re‑colonization information was never published. (read here and here).

The IUCN’s Dr. Andrew Smith is the only researcher with results from long term pika monitoring that actually provides insight into the natural frequency of “extinction” and re‑colonization.

Pika on Bodie ore piles
Pika on Bodie ore piles


In California’s abandoned desert mining town of Bodie, pika have colonized discarded ore piles. Dr. Smith tracked the vacancy rates of 76 ore piles from 1972 to 2009. As expected, during those 37 years Smith observed 107 local extinctions, balanced by 106 re-colonizations. Like pika habitat elsewhere in the Great Basin, on average 30% of the ore piles were unoccupied at any given time, but that vacancy rate was highly variable. Some years the vacancy rate was as high as 52%, and other years as low as 11% (see chart below). In his first survey in 1972, Smith found that 82.3% of the ore piles were occupied by pika. In 2009, pika again occupied 82.8% of their possible sites. Coincidentally Stewart (2015) found 85% (57/67) of his re-surveyed sites are now occupied. Without accounting for such a wide range of variability, the percentage of vacant territories tells us precious little about any climate effects. But in contrast to Smith’s analysis, Stewart presented vacant territories as evidence of global warming caused “extinctions”.

Pika Colonizaton and Extinctions at Bodie
Pika Colonizaton and Extinctions at Bodie


Although Smith’s research establishes a natural frequency of vacancy rates, it still doesn’t tell us why a site became vacant. In Beever’s 2003 paper, the seven “extinct” sites he attributed to climate change had other more plausible explanations. One site had half of the talus removed for road maintenance, another site had become a dump site, and a third site had scattered shotgun shells throughout the talus.

Like rabbits, and a truly endangered species of pika in China, pikas have been hunted and poisoned because they compete with livestock for vegetation. All of Beever’s extinct  sites were heavily grazed. Furthermore pikas do not hibernate. They create hay piles to sustain them through the winter. Any significant loss of vegetation will likely cause pikas to abandon their talus. Although studies have reported significant effects from grazing competition, Stewart (2015) did not include grazing as a variable in his climate change model.

Stewart (2015) created a model that only included 1) area of talus and 2) summer mean temperature as the determinants of local pika extinctions. Assuming that model represents reality, they then argued that according to projected warming from CO2 driven models, pika will become increasingly “extirpated from 39% to 88% of these historical sites”.

But talus area is the more critical variable, and the average summer temperature is highly questionable. Larger talus areas sustain more pika territories, and provide protection for dispersing young looking for vacancies. With more adjacent territories, there are more young pika who can immediately occupy any abandoned territory. In contrast the smallest talus areas, often sustaining just a single territory, are islands that lack connectivity to other territories. Vacant territories must wait to be randomly colonized by dispersing young from some distance talus. As the distance between isolated territories increases it is less likely that randomly dispersing young will re‑colonize a vacated territory. But the degree of connectivity was also never considered in Stewart’s model. As seen in his diagram below (I added the red lines for reference), the vacancies can be readily explained just by the talus area and random dispersal.

Stewart 2015 pika model
Stewart 2015 pika model


If the size of the talus area had been modeled as the only predictor of pika vacancies, any large talus area, (areas above the upper red line), would correctly predict full occupancy, accounting for 31% of the sites (20 of 67), regardless of temperature. Small talus area (areas below the lower red line) would correctly account for 70% of the vacancies (7 of 10 vacancies) regardless of temperature. In talus of intermediate areas, only 7% of the sites were vacant (3 of 39) which is close to the overall 6% finding of the USFS surveys. That 7% vacancy rate is easily accounted for by random extinction/colonization events, and the percentage is far better than vacancy rates Dr. Smith reported for Bodie’s ore piles. 

The higher temperatures reported at the 3 vacancies with intermediate talus areas may have been the result of a more barren dry landscape typical of the eastside of the Sierra Nevada. If so, lack of food, not higher temperatures may be the critical factor. Stewart never asks if the vacancies are due to higher temperatures, less reliable vegetation, or distance from other territories. Stewart’s model statistically linked higher temperatures to pika vacancies, but that link depends on what sites he includes or omits in his database.

Beever’s data had similarly suggested higher temperatures were killing pika, but his analysis excluded data from nearby populations thriving at warmer and lower elevations just 93 miles away from 71% of Beever’s extinct sites. At Lava Beds pika were flourishing at an average elevation 900 feet lower than the average elevation of three nearby extinct sites. Temperatures at Lava Beds also averaged an additional 3.6°F higher, and precipitation was 24% less. But Beever analyzed those sites separately. Likewise Stewart was clearly more interested in a connection to global warming. In his introduction he speculated, “climate change forces range contractions, species may effectively be ‘pushed off’ the tops of mountains by warming climate.” He also referenced Parmesan’s bad climate science connection for support. To create a link to global warming, Stewart needed to use average summer temperature as the other model variable.

During high temperatures, heat-sensitive pika will seek refuge beneath the cooler talus. However Stewart argues such behavior reduces critical foraging time and thus possibly reduces winter survival. Perhaps. During extreme warm days, pika are known to become crepuscular, restricting their foraging to the twilight hours. However if that is the key mechanism, then using the average temperature is simply wrong. The average temperature is amplified by minimum temperatures of the early morning when overheating is not a problem. If Stewart was sincerely concerned about induced heat stress, then the correct metric would be the afternoon maximum temperatures. But maximums were not even considered in Stewart’s choice of models.

Not considering maximum temperatures would seem shamefully negligent, but Stewart was aware that other studies had already determined no correlation with maximum temperatures. Stewart referenced Beever (2010) who wrote, ““Although pikas have been shown to perish quickly when experimentally subjected to high temperatures, our metric of acute heat stress was the poorest predictor of pika extirpations.” Because maximum temperatures had revealed no acute heat stress, Beever adopted the term “chronic heat stress” which was just a more alarming way to say the average temperature.  But even using average temperature,  Beever still concluded, “Climate change metrics were by far the poorest predictor of pika extirpation. Stewart’s own data supported the conclusion that climate metrics provided poor explanatory power. 

Stewart also cherry-picked a start date to argue, “documented 1°C increases in California-wide summer temperature over the past century, strongly suggest that pikas have experienced climate-mediated range contraction in California over the past century.” However if one examines the data Stewart links to for northeastern California, where most of their “extinctions” were observed, recent summer maximum temperatures have not exceeded the 1920s and 30s. If pika extinctions were truly “climate-mediated”, then the high temperatures of the 20s and 30s should have been the main driver. Furthermore during that 20s and 30s, pika experienced the most rapid temperature increases of about 2°C (4°F) in just 3 decades.

Northeast California Maximum Temperatures
Northeast California Maximum Temperatures


Stewart made one more feeble attempt to justify using average summer temperatures. He reported that a 2005 paper by Grayson revealed pika have been forced to move ever upwards as climate warmed throughout the Holocene. (See graph below) But Stewart seems unaware that he damaged is own argument. Several studies, using proxies and models, have shown the Great Basin was warmer during the Middle Holocene by 1 to 2.5°C.  Using Stewart’s logic, as global warming approaches temperatures seen in the mid Holocene, pika should descend to lower elevations.

Although summer temperature data has very little predictive power regards pika biology, it was Stewart’s only link to CO2 climate models. Using that dubious link to summer temperatures, he projects impending climate doom and widespread pika extinctions. But if Stewart was truly concerned about preserving pikas, instead of preserving CO2 theory, then all the data suggests small talus areas that are subjected to grazing are the relevant concern. To protect the pikas’ forage, simply fencing off livestock from the edge of those small talus slopes would be a simple affordable solution. Stewart’s own data also suggests, along with the USFS surveys,  that wherever there is large talus area, there has been nothing to suggest imminent extinctions. So why does the pikas’ climate change extinction story persist?

Grayson's depiction of elevations of pika habitat in the Holocene
Grayson's depiction of elevations of pika habitat in the Holocene