Degassing Rate of CO

Estimates of present-day global volcanic degassing rates are under constant revision (e.g., see Gerlach, 1991; Brantley and Koepenick, 1995;

Sano and Williams, 1996; Marty and Tolstikhin, 1998; Kerrick, 2001). A compilation of recent estimated rates of most degassing processes is shown in table 4.1. A constraint on estimates is that none can exceed total global degassing. The latter can be determined from the steady-state assumption that CO2 release by global degassing must be balanced by global uptake by Ca and Mg silicate weathering (Berner, 1990; Berner and Caldeira, 1997). (This assumes essential balance of the organic C subcycle.) Global Ca and Mg silicate weathering, based on river fluxes of these elements to the sea, has been estimated to be about 6 ± 3 x 1018 mol/my (Berner, 1990). Gaillardet et al. (1999) estimate a minimum value for Ca and Mg silicate weathering of 3.6 x 1018 mol/my. The individual estimates shown in table 4.1, being less than 6 x 1018 mol/my, are thereby acceptable, and one can derive an estimated total degassing from these data that essentially agrees with the range derived from the rate of CO2 uptake by weathering.

The global rate of carbon degassing and how it has changed for all of geologic time has been modeled by Tajika and Matsui (1992), Sleep and Zahnle (2001), and Franck et al. (2002). A general result of these models is that over time the carbon content of the mantle has changed, but the rate of change during the past 500 million years has been small. This suggests that net loss of carbon to the mantle or gain from the mantle over the Phanerozoic has been sufficiently small that the mass of crustal carbon can be assumed to have remained essentially constant (Tajika and Matsui, 1992). This greatly simplifies Phanerozoic carbon cycle modeling. However, the rise of planktonic organisms in the Jurassic and the subsequent transfer of carbonate deposition from shelves and platforms to the deep sea probably has resulted in the addition of extra carbon to the mantle by seafloor subduction over the past 150 million years. This is taken into consideration in GEOCARB modeling (Berner, 1994; Berner and Kothavala, 2001; see chapter 3) and models of Wallmann (2001). Before this time, however, it is as-

Table 4.1. Present-day global degassing fluxes for CO2.



(1018 mol/my)


Release from mantle at mid-ocean rises (MOR)



Release in arc volcanoes from subducted CaCO3


1, 2

Release in arc volcanoes from mantle


1, 2

Release from mantle in intraplate volcanoes (plumes)


1, 3

Release from all subaerial volcanoes



Release during metamorphism of continental carbonates



Total global degassing


References: 1, Marty and Tolstikhin (1998); 2, Sano and Williams (1996); 3, Wallmann (2001); 4, Kerrick (2001); 5, Kerrick et al. (1995).

References: 1, Marty and Tolstikhin (1998); 2, Sano and Williams (1996); 3, Wallmann (2001); 4, Kerrick (2001); 5, Kerrick et al. (1995).

sumed that the mass of crustal carbon remained constant during the remainder of the Phanerozoic.

In many carbon cycle models (e.g., Berner et al., 1983; Fancois and Godderis, 1998; Tajika, 1998; Berner and Kothavala, 2001; Wallmann, 2001; Francois et al., 2002), changes in rates of Phanerozoic volcanic and metamorphic global degassing have been assumed to be directly proportional to rates of seafloor area creation, in other words spreading rate. For example, Tajika (1998) separates degassing into four types: (1) mid-ocean ridge volcanism, (2) hot-spot volcanism, (3) metamor-phism of carbonate at subduction zones, and (4) metamorphism of organic carbon at subduction zones. All but hot-spot volcanism are assumed to be driven by changes in seafloor spreading rate. A similar approach is used by Wallmann (2001). How spreading rate affects global CO2 degassing has been expressed in GEOCARB modeling (e.g., Berner, 1991) by the dimensionless parameter fSR(t):

fSR(t) = seafloor spreading rate (t)/seafloor spreading rate (0) (4.1)

where (0) represents the present. Spreading rate dependency makes sense for the degassing that occurs at spreading centers and at subduction zones, but it also has been extended to be a measure of total tec-tonically induced degassing, whether on land or at sea (Berner et al., 1983; Berner, 1991). In this case the parameter fSR(t) is assumed to be equal to fG(t) to represent all degassing. This may be incorrect because spreading rates are not a good measure of CO2 degassing resulting from magmatism and metamorphism accompanying continent-continent and arc-continent collisions (Kerrick and Caldeira, 1998).

Another source of CO2 that is probably not a function of seafloor spreading rate is that resulting from mid-plate volcanism. During the Cretaceous period large mid-plate volcanic plateaus arising from mantle "superplumes" were built on the Pacific seafloor (Larson, 1991). This could have led to extra CO2 emission to the atmosphere and has been called upon by many workers to explain Cretaceous warming and high CO2 levels (Larson, 1991; Kaiho and Saito, 1994; Tajika, 1999; Wallmann, 2001). Kaiho and Saito correlate mid-plate volcanism, along with ridge and back-arc volca-nism, with the temperature of Cretaceous seawater as recorded by oxygen isotopes. However, Heller et al. (1996) have challenged the quantitative significance of this "extra" volcanism on geophysical grounds.

Kerrick (2001) has pointed out additional problems with using spreading rate as a measure of global degassing: (1) In the circum-Pacific arc, there is no correlation between the number of volcanoes and subduction rate; (2) as oceanic basalts age, they undergo an increase in carbonate content by submarine weathering so that older subducting crust would release more CO2 than younger crust; (3) the rise of calcareous plankton in the Mesozoic caused an increase in the carbonate content of sub ducting sediments; and (4) the extent of subduction-zone metamorphic decarbonation depends on pressure and temperature gradients as they affect the stability of carbonate minerals. In a global and long-term context one can answer these additional criticisms. First, there is an overall qualitative correlation between subduction and metamorphic degassing as shown by a global map of the distribution of CO2-rich fumeroles and soda springs by Barnes et al. (1978). Volcanic density is not the sole measure of total degassing at any one place. Second, the average age and average carbonate content of subducting basalt should vary inversely with global average spreading rate. Third, the role of calcareous plankton in transferring carbonate deposition to the deep-sea floor is included in most carbon cycle models (Berner, 1991, 1994; Berner and Kothavala, 2001; Wallmann, 2001; see also chapter 3). Finally, much degassing at subduction zones is probably volcanic, not metamorphic, in origin, so the suggestion that metamorphic decarbon-ation varies inversely with subduction rate (Kerrick et al., 2003) applies only to a portion of arc degassing.

Following the earlier work of Budyko and Ronov (1979), an alternative to the use of seafloor spreading rate as a proxy for CO2 degassing was offered by Kerrick et al. (2003). They assume that the CO2 flux from volcanic arcs can be correlated with the rate of andesitic extrusion and that the rate of andesitic extrusion through time could serve as a measure of arc degassing. Examining modern volcanic data, they conclude that the present degassing rate per unit volume of andesite is 3.5 x 108 mol/my/km3. From this value and a volumetric estimate of Late Cretaceous andesites, they conclude that CO2 emission rates for the Late Cretaceous could have been more than twice that at present.

The idea of Kerrick et al. (2003) can be tested using published data on the abundance of not just andesites but total terrestrial and submarine volcanics extruded over Phanerozoic time (Ronov, 1993). Total volcanics should give a better idea of total global degassing. To correct for loss by erosion over time, the data can be fitted to an exponential decay curve, just as was done for sandstones and shales in chapter 2 (equation 2.2):


t = mean age of a given volume of rocks extruded within the time span At AV = volume of volcanic rocks within age span At (AV/At)o = rate of volcanic rock extrusion at present k = weathering and erosion decay coefficient for volcanics (assumed constant).

Values of AV/At are determined from the data of Ronov (1993) for 27 age spans ranging from the Miocene to the lower Cambrian. A plot fitted to the Ronov data is shown in figure 4.1. To avoid overly biasing the fitted curve by including excessive total rock abundance accompanying the Plio-Pleistocene glaciation, "present" is assumed to represent the mid-Miocene (15 Ma). Also, use of a constant value for k assumes that the probability of erosive loss does not change with time. Deviations of each time span data point above and below the exponential curve can then be interpreted as original increases or decreases in global vol-canism relative to that at present (Wold and Hay, 1990). In other words:

where the subscript v refers to volcanics, ron refers to the data of Ronov, and exp to the value of AV/At calculated from equation (4.2) for the same time. A plot of fV(t) derived in this manner is shown in figure 4.2. Values of fSR(t) used in the GEOCARB II and III models, based on the data of Gaffin (1987) and that of Engebretson et al. (1992) (for the past 150 Ma), are also included in figure 4.2. The rapid variations in fV(t) at some times, I believe, are due more to the inability to accurately quantify ancient rock abundance than to real changes in extrusion and degassing rates. At any rate, the two approaches to paleo-degassing rate crudely agree in the sense that they both give f values within the range 0.5-2.0.

In the absence of a better quantitative model for global degassing over geologic time, change in seafloor spreading rate is still probably the best

Extrusion Degassing

Time my

Figure 4.1. Plot of volcanic rock abundance versus time fitted by an eponential decay (equation 4.2) representing expected loss by erosion. (Data from Ronov, 1993.)

Time my

Figure 4.1. Plot of volcanic rock abundance versus time fitted by an eponential decay (equation 4.2) representing expected loss by erosion. (Data from Ronov, 1993.)



Figure 4.2. Plot of volcanic degassing parameters versus time. The parameter fSR(t) (Berner, 1994; Berner and Kothavala, 2001) is seafloor spreading rate at a past time divided by the spreading rate at present. The parameter fV(t) is the rate of volcanic eruption divided by that for the Miocene (as a representation of the "present"). Values of fV(t) are calculated from volcanic rock abundance data (Ronov, 1993) using equation (4.3).

first approximation of change in global degassing. If so, it is imperative to know how this rate has changed over geologic time. Estimates for spreading rate over the past 150 Ma have varied from essentially constant (Parsons, 1982; Heller et al., 1996; Rowley, 2002) to increases of a factor of 1.3-2 (Hays and Pitman, 1973; Pitman, 1978; Komincz, 1984; Engebretson et al., 1992; Gaina et al., 2003). Most carbon cycle models have used the Engebretson et al. formulation (which is based on subduction rates). This formulation has been challenged by Rowley (2002). However, the study by Gaina et al. (2003), which accounts for the areas of seafloor already subducted (not considered by Rowley), shows a definite overall reduction (about 50%) in spreading rate since the Jurassic. Considering the extreme amount of work, including use of marine gravity data, magnetic anomalies, bathymetry, seismic data, and geologic subduction data, that has gone into the Gaina et al. formulation, I believe it is the best one available at the present time.

The oldest seafloor is only 180 Ma. This means that there is no direct measure of spreading rate before this time. A proxy that has been used to extend calculations of global spreading rate is the history of sea level. Studies by Hays and Pitman (1973) and Kominz (1984) have demonstrated a good correlation between sea level and spreading rate for the past approximately 100 million years. This correlation comes about because an increase in the total volume of the mid-ocean ridges, which displaces seawater upward, accompanies the faster addition of new oceanic crust

(i.e., faster spreading rate). If this correlation applies to earlier times, then spreading rate can be deduced from estimates of paleo-sea level. This has been done by Gaffin (1987), and his spreading rate estimates, combined with the first-order sea-level data of Vail et al. (1977), are used in the GEOCARB models (Berner, 1991, 1994; Berner and Kothavala, 2001) to calculate spreading rate before 150 Ma (figure 4.2).

The explanation of changes in sea level in terms of seafloor spreading rate has been challenged by Heller et al. (1996). They maintain that there has been little change in spreading rate, at least over the past 150 million years, and that the high sea levels of the Jurassic and Cretaceous are a result of the rifting and breakup of the supercontinent Pangaea. The filling of proto-oceanic rifts with sediment is equivalent to the enlargement of the area of the continents at the expense of the oceans (Heller and Angevine, 1985), and decreasing ocean area, for a given volume of water, must result in a rise of sea level. An additional factor is that creation of new seafloor, as a result of continental breakup, results in a decreasing mean age of oceanic crust (Worsley et al., 1984; Heller and Angevine, 1985). A younger seafloor is a hotter seafloor; a hotter seafloor is a higher seafloor (Parsons and Sclater, 1977), and a higher seafloor means higher sea level.

Was this article helpful?

0 0
Organic Gardening

Organic Gardening

Gardening is also a great way to provide healthy food for you and your loved ones. When you buy produce from the store, it just isnt the same as presenting a salad to your family that came exclusively from your garden worked by your own two hands.

Get My Free Ebook


  • bellisima
    How to measure co2 levels extrusion?
    5 years ago

Post a comment