“The Sea’s Missing Salt”: Austin and Humphreys (1990) propose a dilemma
“The known and conjectured processes which deliver and remove dissolved sodium (Na+) to and from the ocean are inventoried. Only 27% of the present Na+ delivered to the ocean can be accounted for by known removal processes. This indicates that the Na+ concentration of the ocean is not today in “steady state” as supposed by evolutionists, but is increasing with time. The present rate of increase (about 3 × 1011 kg/yr) cannot be accommodated into evolutionary models assuming cyclic or episodic removal of input Na+ and a 3-billion-year-old ocean. The enormous imbalance shows that the sea should contain much more salt than it does today if the evolutionary model were true. A differential equation containing minimum input rates and maximum output rates allows a maximum age of the ocean of 62 million years to be calculated. The data can be accommodated well into a creationist model.” -Excerpt from the abstract, Austin and Humphreys (1990)
The methodology by Austin and Humphreys is as straightforward as balancing your own bank account: subtract your total number of monthly expenses from your total monthly incomes, and you can calculate the net monthly change to the account. Their conclusion is likewise as simple as the following logic: last month, I added $100 to my account, so I currently have $1,100 in the account; therefore, my account could not have been opened more than 11 months ago.
Imagine this describes your bank account, which you actually opened some 20 years ago. You might be quick to respond in several ways: 1) the net change to my account is not always positive, because sometimes I spend more than I earn; 2) the net change to my account has not been $100 every month, but has been more or less in the past; or 3) if there is an error in accounting, I didn’t actually add $100 to my account. As it turns out, all three responses can be given to Austin and Humphreys, who—despite more than 30 years of new research on the Earth’s oceans and geochemistry—have not updated their ‘accounting’.
|Table 1 from Austin and Humphreys (1990), summarizing
model inputs of Na to the oceans. Units are in 1010kg/yr.
1. Rivers: Sea-spray component
The first item in Table 1 of Austin and Humphreys (1990) indicates that 50–55 million tons of sodium are added to the oceans via droplets of water containing sea salt, which fell into rivers draining into ocean basins. The origin of this sodium, however, is the ocean itself. As waves crash over the ocean, tiny droplets of salty water are carried off by the wind and deposited over the continents. Since this mass of sodium moves directly from the oceans to the rivers and then back again, it should not be included in the table of inputs. If you draw $20 from your account, only to deposit it back into the account, the net change is zero. Therefore, the real influx of sodium via sea-spray input to rivers is 0 tons/year.
2. Rivers: silicate weathering
Austin and Humphreys cite Meybeck (1987), who estimated that ~62 million tons of sodium are dissolved through chemical weathering of silicate minerals (e.g. feldspar) and delivered to the oceans via rivers. This estimate is based on modern analyses of rivers and major watersheds, however, and Meybeck notes that precise masses are very difficult to assess, due to a lack of direct measurements. Assuming the accuracy of their figure, in any case, we should also note how this number (62 million tons) can vary through time. Nobody expects that it would remain constant over hundreds of millions of years.
First, sodium delivery via silicate weathering depends on the global weathering rate, which itself depends on climate, sea level, and global tectonics. Glacial conditions enhance silicate weathering by crushing millions of tons of silicate minerals into fine powder, which gets washed downstream to the oceans. Therefore, sodium delivery should be less for a majority of Earth history, during which glaciers were absent. Higher sea level limits the amount of land (particularly sodium-rich coastal sediments) exposed to chemical weathering and erosion. Therefore, sodium delivery should be less for a majority of Earth history, during which sea level was higher and less land area was exposed. Finally, the formation of large mountain ranges, particularly where annual precipitation is high, contributes substantially to modern silicate weathering. Relatively recent mountain belts like the Himalayan and Sierra Nevadan ranges expose more silicate minerals to chemical weathering and erosion. They also promote strong precipitation (rain/snow) over the continents, by forcing air masses upward. Therefore, sodium delivery should be less for periods of Earth history when massive orogenic belts did not exist.
In any case, more recent work by Holland (2005) provides a better estimate of sodium from silicate weathering. Therefore, the total influx of sodium via silicate weathering should be ≤55 million tons/year.
3. Rivers: chloride solution
In the modern geological setting, a small percentage of the land area (<2%) is comprised of some very salty rocks. These small outcrops of halite, gypsum, and ancient marine clays contribute a relatively huge proportion of sodium to rivers draining into the oceans (today, as much as 75 million tons/year, including agricultural runoff). Quite simply, rock salt is far more soluble than minerals like feldspar, so any exposures of rock salt at the Earth’s surface will erode thousands of times faster than, say, granite and other silicate rocks.
Before we consider “chloride solution” to be a long-term Na input to the oceans, however, we need to ask: what is the source of sodium in these rocks? Geologists agree unanimously that these Na-rich minerals were precipitated largely from seawater, either as ocean basins became isolated (e.g. the Mediterranean Sea) when sea level was much lower, or as warmer climates evaporated more water from shallow seas. Whatever the mechanism, this source of sodium to the oceans ultimately derived from the same oceans! That being the case, Austin and Humphreys are wrong to include this flux in their table without adding it directly to the other side, because in the long-term, no more sodium can be dissolved from marine salt deposits than was removed at some point in the past.
In terms of our analogy from accounting, imagine that you sporadically withdrew money from your account and hid $20 bills around your house. Whenever these bills resurfaced (say, during ‘Spring Cleaning’), however, you took the money back to the bank and re-deposited it into the same account. The amount of money going back into your account cannot be more than the amount originally withdrawn (wouldn’t that be nice!). But according to the accounting by Austin and Humphreys (1990), an average of 75 million tons of sodium were added to the oceans every year, despite that less than 40 thousand tons (see Table 2) were ‘withdrawn’, on average, each year. Austin and Humphreys have failed miserably in this simple test of accounting.
So we have determined that Na from “chloride solution” should not be included in the table of Na inputs, so the actual number should be 0 million tons/year. I will include it here, because I will also consider Na removed by halite deposition, as Austin and Humphreys have done. However, I use a more reasonable estimate of long-term “chloride solution” using the data of Hay et al. (2006), who estimated halite burial and erosion over the course of the entire Phanerozoic. These data take into account the fact that at various points in Earth history, more or less halite has been exposed at the Earth’s surface. As it turns out, the total area of ‘salty’ outcrops (~1.3%) is much higher today than for the bulk of Earth history, because most of these outcrops are only Miocene in age. Prior to the Miocene, (>23 million years ago), these salt deposits didn’t exist and therefore could not have been dissolving back into the oceans. The amount of salt being dissolved from evaporite minerals and added back to the ocean has fluctuated substantially over time:
|Estimated influx of Cl– to the oceans over the Phanerozoic, according to
Hay et al. (2006). Each atom of Cl– should be accompanied by one Na+.
The average influx of sodium via “chloride solution”, according to data from Hay et al. (2006), was about 17.0–18.3 million tons/year, much less than the figure cited by Austin and Humphreys.
4. Ocean floor sediments
As marine sediments accumulate on the ocean floor, the uppermost centimeters of sediment tend to release sodium into the ocean while absorbing both Mg and Ca. This phenomenon was quantified for Atlantic Ocean sediments by Sayles (1979), cited by Austin and Humphreys. A later review of the topic by Drever, Li, and Maynard (1988) also cited Sayles (1979), whose estimate appears in Table 1.4 of their paper. This is the figure used by Austin and Humphreys (1990), who conclude that 5.0 x 1012 moles/yr of sodium (1.15 x 1011 kg/yr) are added to the oceans every year by this process. It is the largest single input of sodium used in the model by Austin and Humphreys (Table 1).
Although nobody questions that the diagenesis of ocean sediments (i.e. their chemical modification after burial) releases Na into the oceans, the calculated magnitude is very much in question. Drever, Li, and Maynard (1988) also include a previous estimate by Maynard (1976), which is 6 times smaller than the figure by Sayles (1979). Even the more comprehensive data from Sayles (1979) indicate substantial variation in this flux from one location to the next, and by no means have all the world’s oceans been studied in this manner. Drever, Li, and Maynard (1988) conclude:
“…it seems likely that the relative changes in [Na+] are correct but the absolute magnitudes are too high by a factor of at least 2.” (emphasis mine)
If we take the advice of Drever, Li, and Maynard (1988), whom Austin and Humphreys (1990) cite to obtain their figure, then the actual flux of sodium from ocean-floor sediments should be ~52.5 million tons/year or less. The associated error bars are high, however, and we can expect this flux to have varied over Earth history, since it depends strongly on the amount and composition of sediments delivered to the oceans.
5. Glacial silicates
Austin and Humphreys include the sodium input from “finely pulverized glacial silicates”, which they estimate crudely from the volume of rock being eroded by the Antarctic Ice Sheet. This process is important today, because most of Antarctica is covered by active glaciers. These massive ice sheets are missing, however, from the majority of Earth history. In fact, tropical plant fossils are common among sedimentary layers from Antarctica. Therefore, the estimated 39 million tons/year of sodium from “glacial silicates” is not applicable to a long-term model of the sodium cycle.
In addition, there is no direct evidence for how much sodium is shed from the Antarctic continent and dissolved in seawater, and the estimate by Austin and Humphreys is certainly way too high. The only study they cite is from 1964 and did not address sodium dissolution directly, let alone in Antarctic waters. Nonetheless, they assume that 64% of all glacially eroded rocks dissolve completely in seawater rather than accumulate as sediments. Is this realistic? Not at all.
The actual long-term influx of sodium from glacially pulverized silicates is slightly more than 0 million tons/year, but far less than the 39 million tons estimated by Austin and Humphreys. Even if we use their figure, we should multiply it by the small fraction of Earth history during which large continental glaciers existed, which yields ~1 million tons/year.
6. Atmospheric and Volcanic Dust, and 7. Marine Coastal Erosion
Austin and Humphreys once more make gratuitous assumptions about how much silicate dust/sediment completely dissolves in seawater. Their estimates of Na influx from these two processes are so low, however, that ignoring them completely would not change the total estimate of sodium inputs. Therefore, I will include their estimates to be generous/conservative.
8. Glacier ice
Yet again, Austin and Humphreys include a relatively insignificant process that is not applicable to the majority of Earth history. They estimate that ~1.2 million tons/year of sodium are added from glacial ice containing salt trapped from the atmosphere. If the glaciers were absent, however, this tiny amount of halite dust would either be washed back to the oceans through rivers or buried in surface sediments. Once again, I will include their estimates to be generous/conservative, but I want to highlight the unscientific nature of their methods, which they employ under the guise of being thorough. If we have no reason to expect that large glaciers were present for the past 62 million years, then why include this flux in a model that supposedly characterizes the last 62 million years of Earth history? Austin and Humphreys most certainly know better, so the fact that glacial ice is included as a sodium input reveals the dishonest tactics behind their work.
9. Volcanic Aerosols
This flux depends, of course, on rates of volcanic activity, which undoubtedly varied in Earth history. Nonetheless, this sodium input is far less than the uncertainties of other large fluxes, so it matters little whether the flux is included in the total calculation.
10. Groundwater seepage
According to Austin and Humphreys, large amounts of groundwater are seeping into the oceans, carrying some 96 million tons/year of sodium with them. This is the second largest input of sodium from Table 1—how is it calculated? Citing Garrels and Mackenzie (1971), they take the difference between global runoff and global rainfall minus evaporation to be the amount of groundwater seeping from continent to oceans every year. They then multiply this mass of water by what they assume to be the average sodium concentration of groundwater.
This almost makes sense, intuitively. Imagine you poor 100 liters of water into a large wooden planter, of which 10 liters evaporate into the open air. Now, 90 liters of water remain somewhere in the planter. Imagine now that 80 liters leaked out of the planter onto the lawn through cracks between the wood (much like rivers discharging into the oceans). What about the remaining 10 liters of water? We must assume that this mass of water infiltrated through the planter and seeped into the ground on which the planter is situated, right?
Not entirely. We can be certain that some of this water will be stored in the planter itself. Likewise, some 3.3×1020 kg of water on Earth is now stored on the continents in underground reservoirs, because not all precipitation ends up in the oceans. Therefore, Garrels and Mackenzie (1971) take the difference (used by Austin and Humphreys) as a maximum estimate of groundwater flow to the oceans. Given the large errors in calculating global precipitation, evapotranspiration, and runoff, they further write:
“Conceivably this excess could be delivered by subsurface flow. If so, and if these ground waters have about the same total salinity as streams, approximate 4×1014 g/year of dissolved solids could be entering the ocean basins from subterranean flow. Both required assumptions are shaky; from the preceding discussion of stream discharge it is clear that a 10 percent difference between total precipitation minus evaporation and stream discharge could be accounted for by errors in either estimate. Also, we do not have good numbers for the dissolved solid content of those ground waters reaching the sea.” (emphasis mine; from this quote, we learn that groundwater may or may not be seeping into the oceans in large quantities)
So Garrels and Mackenzie (1971), writing in an era before satellite constraints on the global hydrological cycle, proceed with caution in estimating the maximum plausible influx of sodium to the oceans from groundwater (which they estimate to be 20 million tons/year, a meager 20% of the value used by Austin and Humphreys). Regardless, Austin and Humphreys use a high-end estimate of groundwater seepage with confidence and further imagine that groundwater seeping into the ocean is, on average, 5 times saltier than river water. They provide no direct evidence of this figure, to which they attach almost no uncertainty (unlike Garrels and Mackenzie, whom they cite). On the contrary, they suggest only that it might be even higher!
Since groundwater seepage to the oceans occurs mainly from shallow, coastal aquifers, it is rather reasonable to assume that groundwater seeping into the oceans is about as fresh as rivers draining into the oceans, and not five times saltier. Very saline groundwater is found only in deep, continental aquifers, or coastal aquifers where recent salt deposits exist (e.g. around the Gulf of Mexico). The strategy of Austin and Humphreys, therefore, is one of selective sampling of ballpark estimates from rather old scientific literature, after which errors/uncertainties are ignored or minimized unrealistically.
Since Garrels and Mackenzie reviewed estimates of global precipitation, evapotranspiration, and runoff in 1971, ongoing research and technological development has provided the scientific community with far more accurate and comprehensive data. A more recent assessment of the global water cycle is presented by Trenberth et al. (2007), from which I took the figure below.
|Figure 1 from Trenberth et al. (2007); summary of the modern water cycle.|
According to their review of data published within the last decade, the difference between surface runoff (40 thousand cubic km) and precipitation minus evapotranspiration (113 – 73 thousand cubic km) is precisely zero. In other words, groundwater seepage is not a significant flux of water to the oceans, and should occur only locally or in response to minor climate fluctuations.
Before concluding, we should be thorough scientists and ask: what is the source of sodium dissolved in this groundwater seeping into the oceans? We cannot answer precisely, but we can be certain that much of the sodium in groundwater (like in river water) derives from either sea spray or dissolved halite deposits underground. Since the sodium in sea spray or halite deposits derives directly from the oceans, we should remove that amount from any long-term model of the sodium cycle (again, we are simply re-depositing money withdrawn from the same account).
Taking all of these factors into account, we may conclude that the total influx of sodium from groundwater seepage cannot be higher than 20 million tons/year, as estimated by Garrels and Mackenzie (1971, Table 4.11). More likely, however, the total long-term input is effectively 0 tons/year.
11. Seafloor hydrothermal vents
The final sodium input used by Austin and Humphreys (1990) constitutes their most egregious error in accounting. They claim that ~15 million tons/year of sodium are added to the oceans from water cycled through hydrothermal vents on the seafloor. In fact, a wide base of scientific literature from the past 3 decades, including papers cited by Austin and Humphreys, proves just the opposite: hydrothermal vent systems remove sodium from the oceans, and they do so in massive quantities. This major error was first documented by Glenn Morton in an open letter entitled Salt in the sea. Dr. Snelling even acknowledges the error (though subtly) in his summary article:
“Long-agers also argue that huge amounts of sodium are removed during the formation of basalts at mid-ocean ridges, but this ignores the fact that the sodium returns to the ocean as seafloor basalts move away from the ridges.” (notice, he makes no attempt to refute the claim that sodium is removed during basalt formation, but only to misdirect the accusation)
Unfortunately, Dr. Snelling offers no evidence for his claim that sodium taken up at mid-ocean ridges eventually returns to the ocean (mainly because he is wrong—it does not). The process by which sodium is removed from oceans through hydrothermal vent systems is called albitization. In short, feldspar minerals in oceanic crust (being created constantly at mid-ocean ridges) are converted from calcium-rich feldspar to sodium-rich feldspar in the presence of hot seawater. This chemical alteration releases calcium into the oceans in exchange for sodium, balancing the global cycle. Bach and Früh-Green (2010) write:
“Alkali elements [e.g. sodium] are leached from the rocks by seawater-derived fluids in high-temperature, axial, hydrothermal processes, while in low-temperature ridge-flank systems, they are transferred from the circulating seawater to the oceanic crust. The net effect is that oceanic crust is a prominent sink for alkali elements…” (emphasis mine)
As oceanic crust moves away from mid-ocean ridges, the crust’s temperature drops and hydrothermal vents become less active. The majority of newly formed albite is crafted deep within the oceanic crust, however, and is not exposed to seawater once hydrothermal waters cease to circulate. Bach and Früh-Green (2010) add:
“Hydrous minerals (smectites, zeolites) and carbonates form in these ridge-flank systems and slowly seal the crust, which also becomes increasingly insulated from the ocean by the accumulation of sediments.” (emphasis mine)
Snelling’s misdirection is thus wildly inaccurate; this major sodium sink does not return to the oceans. Therefore, Austin and Snelling (1990) have listed a sodium input that should be counted as a sodium output. So what is the magnitude of sodium lost to oceanic ridge systems?
The uptake of dissolved sodium by mid-ocean ridge processes was noted by Holland (2005), who follows Berner and Berner (1997) and estimates that it accounts for ~25.3 million tons/year of sodium drawn out of the oceans. I devised my own calculation using chemical data from 152 hydrothermal vents (documented by 5 separate papers, listed below), and multiplied the average sodium loss through hydrothermal vents by the estimated volume of water circulated through those vents. Using this method and taking all uncertainties into account, I estimated that the total sodium loss via albitization is 11–47 million tons/year. This figure encompasses the estimate by Berner and Berner (1997), so I am fairly confident in the results. (Note: contact me if you would like to see my original data/calculations, which are too large to paste here).
The major error of Austin and Humphreys (1990) is one of basic geochemistry. They concluded that hydrothermal vents add sodium to the oceans because water emitted by those vents contains a higher concentration than seawater, but this approach ignores the fact that water itself is lost in the process of hydrothermal alteration. In other words, when newly formed oceanic basalt is exposed to hot seawater, not only does it take up sodium into its mineral structure, but it also absorbs water. Therefore, we cannot use the concentration of sodium (i.e. total grams of sodium per liter of water) as a guide to estimate sodium loss/gain, because we know that water itself is lost in the process. Instead, we must use the ratio of Na/Cl in hydrothermal vent water relative to that of average seawater (chlorine is not lost or gained, so it will stay constant). As Reeves et al., 2011 put it:
“Endmember Na/Cl ratios… are all lower than the seawater ratio, consistent with the removal of Na during albitization…”
Despite this basic error in geochemistry, YEC ministry sites continue to reference the work by Austin and Humphreys unreservedly, propagating the false notion that sodium is constantly added to the oceans through hydrothermal vents. I hope you can sympathize with the challenge that we critics of YEC face: it is far easier to spread misinformation than to correct it.
Thus far, I have only addressed the inputs of sodium estimated by Austin and Humphreys (1990), but we can see already that these authors employ a rather deceptive strategy to win over their young-Earth audience. Most of these fluxes are calculated by ignoring basic geochemistry or selectively citing high end estimates, even when the cited authors advise against it. In the next article, I will briefly examine their estimates of sodium outputs to see if the integrity of their research improves. Concluding there, I will provide a revised table that more accurately reflects the sodium cycle and proves that world’s oceans are just as salty as we might expect on a 4.5-billion-year-old Earth.
(to be continued…)
References for hydrothermal vent calculations: