EFFECT OF ENVIRONMENTAL FACTORS ON THE CHEMICAL
WEATHERING OF PLAGIOCLASE IN HAWAIIAN BASALT
Steven J. Gordon
Department of Economics and Geography
United States Air Force Academy
HQ USAFA/DFEG
2354 Fairchild Drive, Suite 6K102
United States Air Force Academy, Colorado 80840-6299
Abstract: Much of the literature that analyzes the effect of various factors on basaltic
weathering rates does not assess the synergistic effects of several factors at once. Those
studies that do address such synergies generally report their results in formats that are
difficult to integrate into geographic analysis. The research outlined in this paper employs
multivariate statistics to assess the effect of several environmental variables on the weathering rate of basalts on the Island of Hawaii. The presence of lichens on the surface of the
basalt increases weathering rate by an order of magnitude. Weathering that occurs in the
absence of lichen cover is regulated by high elevation/low temperature conditions, moisture availability, and the age of the flow. Weathering that occurs in the presence of lichen
cover is regulated primarily by moisture availability only. The statistical results reported
here are consistent with the results of a geochemical-style analysis of the same dataset
reported elsewhere, suggesting that a multivariate approach is appropriate for assessing
the simultaneous effects of multiple weathering factors on the weathering rate. A multivariate statistical approach allows for the construction of quantitative weathering rate
models that build on previously published qualitative models for describing geographic
variation in weathering rates. [Key words: chemical weathering, basalts, Hawaii.]
INTRODUCTION
The weathering of basaltic minerals and glasses relates to many areas of physical
geography, including landform change, dating of surface features, nutrient cycling
and ecosystem development, chemical composition of surface- and sea-waters, and
soil development. The process liberates nutrients into geochemical cycles and the
biosphere, playing a major role in ecological succession, especially in the early colonization of new basalt flows by plants (Likens et al., 1977; Chadwick et al., 1999).
Further, weathering contributes to soil development (Jenny, 1941; Pedro and Sieffermann, 1979) as it is the process by which bedrock is converted into regolith and
ultimately soil mineral matter. It is a prime factor in landscape change over 105 to
107 years (Twidale, 1982; McLennan, 1993). Weathering rinds have been used as
a dating tool for geomorphic surfaces (Colman and Pierce, 1981; Adams et al.,
1992; Lynch and Stevenson, 1992; Ambrose, 1994; Freidman et al., 1994). Weathering also affects the composition of continental waters (White et al., 1980; White,
1984) and oceanic waters (Bischoff and Dickson, 1975; Mottl and Holland, 1978;
Seyfried and Mottl, 1982). Further, studies of basaltic glass weathering provide
69
Physical Geography, 2005, 26, 1, pp. 69–84.
Copyright © 2005 by V. H. Winston & Son, Inc. All rights reserved.
70
STEVEN J. GORDON
information about the safe storage of high-level nuclear waste (HLNW; Ewing and
Haaker, 1979; Lutze et al., 1985; Werme et al., 1990; Gordon and Brady, 2002).
An understanding of the spatial variability in weathering at a given scale is
therefore useful. Pope et al. (1995) provided a conceptual model that addresses the
spatial variability in weathering as a function of numerous factors. Researchers from
a variety of disciplines address the quantitative effect of various factors on weathering rates of rocks and minerals. For instance, temperature increases the rate of
chemical weathering of feldspars, as reported from field studies (Velbel, 1993; Dorn
and Brady, 1995; White et al., 1999) and also laboratory experiments (White et al.,
1999; Welch and Ullman, 2000).
Many field-based studies substitute space for temperature variations by analyzing weathering rates at different elevations and assuming that a lapse rate exists
(e.g.,Velbel, 1993). While changes in elevation can exact changes in the temperature at which weathering reactions occur, changes in other environmental variables
may depress or entirely mask an expected rate/temperature relationship (Gordon,
1996), creating a complex relationship between weathering rate and elevation.
Elevation-dependent weathering factors other than temperature include precipitation and/or other available moisture sources (Hay and Jones, 1972; Dorn and Brady,
1995; Brady et al., 1999), thermal stress (e.g., freeze/thaw cycles) (Gordon, 1996;
Hall, 1999), duration and thickness of snow cover (Gordon, 1996), and insolation
(Paradise and Yin, 1993; Paradise, 1995).
Further, variables that change with elevation may have opposing effects on
weathering rates. Elevation has a negative effect upon temperature, and temperature has a positive effect upon weathering rate (mentioned earlier). Elevation has a
positive effect upon insolation, as higher elevations receive greater amounts of
incoming shortwave radiation (Barry, 1992). Insolation has a positive relationship
with weathering rate of limestone (Paradise, 1994), granite (Paradise and Yin, 1993),
and sandstone (Paradise, 1995). Precipitation and available moisture will vary with
elevation, therefore either enhancing or decreasing rates on a case-by-case basis.
Precipitation and available moisture are observed to increase physical and chemical weathering rates in both arid and humid climates (Roth, 1965; Hay and Jones,
1972; Brady et al., 1999).
Lichen activity affects chemical weathering in a positive manner. Biotic (plant)
activity decreases reaction solution pH through secretion of organic acids (Jackson
and Keller, 1970; Wasklewicz, 1994). Several studies identify the importance of
lichens in determining of the rate and nature of chemical weathering. These include
a reversal in the order of mineral weathering in Hawaii (Wasklewicz, 1994), and the
acceleration of chemical weathering (Jackson and Keller, 1970; Wilson and Jones,
1983; Berner, 1992; Brady and Carroll, 1994; Paradise, 1997; Banfield et al., 1999;
Brady et al., 1999; Chen et al., 2000). The effect of lichens on weathering rates (via
low pH) may vary with other weathering factors. For example, it is observed that the
response of mineral dissolution rates to pH is more pronounced at elevated temperatures (Casey and Sposito, 1992; Hellmann, 1994). Therefore, numerous factors
may have variable effects on weathering rates in the presence or absence of lichen
cover.
f5d p3.5
f5d c8.2
f5d c8.2
f5c o7
f5e b0.7
f5c o7
f5c o7
f5c o7
f5c o7
f5c o7
f5c o7
f5c o7
Hualalai Alt. Tran
Hualalai Mixed Age Isohyet
Hualalai Mixed Age Isohyet
Hualalai Mixed Age Isohyet
Hualalai Mixed Age Isohyet
SheepStation Flow
SheepStation Flow
SheepStation Flow
SheepStation Flow
SheepStation Flow
SheepStation Flow
SheepStation Flow
Source: Brady et al. (1999).
f5c o7
f5d p3.5
Hualalai Alt. Tran
f5c o7
f5d p3.5
Hualalai Alt. Tran
SheepStation Flow
f5d p3.5
Hualalai Alt. Tran
SheepStation Flow
f5d p3.5
Hualalai Alt. Tran
f5c o7
f5d p3.5
Hualalai Alt. Tran
f5c o7
f5d p3.5
Hualalai Alt. Tran
SheepStation Flow
f5d p3.5
Hualalai Alt. Tran
SheepStation Flow
Lava flow
Study
2200–2300
2200–2300
2200–2300
2200–2300
2200–2300
2200–2300
2200–2300
2200–2300
2200–2300
2200–2300
2200–2300
2670±80
2200–2300
2140±100
2030±80
2885±150
2885±150
2885±150
2885±150
2885±150
2885±150
2885±150
2885±150
C-14 age
1520
1370
1200
910
820
700
600
450
300
180
20
2377
20
100
500
1539
1250
975
580
450
300
143
61
Elevation
(m)
14.4
15.3
16.2
18.0
18.6
19.4
20.1
21.1
22.1
22.8
23.9
9.8
23.9
23.4
20.7
14.3
15.9
17.6
20.2
21.1
22.1
23.1
23.6
Temperature
(°C)
Table 1. Abiotic Data Matrix
1000
1200
1350
1500
1500
1500
1500
1200
1000
700
500
500
500
500
500
1200
1100
1000
750
700
600
500
350
Precipitation
(mm/yr)
200
350
500
430
370
300
-50
-300
-700
-1000
-1100
-1000
-1600
-1500
-300
-630
-850
-1250
-1375
-1550
-1725
-1950
0
Available
moisture
(mm)
0.94
1.40
1.96
3.02
3.26
3.61
3.87
3.62
3.24
2.71
2.43
0.45
2.43
2.22
1.63
4.77
4.44
3.92
3.01
2.95
2.89
2.82
2.90
Average %
porosity
CHEMICAL WEATHERING
71
Lava flow
f5d p3.5
f5d p3.5
f5d p3.5
f5d p3.5
f5d p3.5
f5d p3.5
f5d p3.5
f5d p3.5
f5d c8.2
f5d c8.2
f5c o7
f5e b0.7
f5c o7
f5c o7
f5c o7
f5c o7
f5c o7
f5c o7
f5c o7
f5c o7
f5c o7
f5c o7
f5c o7
Study
Hualalai Alt. Tran
Hualalai Alt. Tran
Hualalai Alt. Tran
Hualalai Alt. Tran
Hualalai Alt. Tran
Hualalai Alt. Tran
Hualalai Alt. Tran
Hualalai Alt. Tran
Hualalai Mixed Age Isohyet
Hualalai Mixed Age Isohyet
Hualalai Mixed Age Isohyet
Hualalai Mixed Age Isohyet
SheepStation Flow
SheepStation Flow
SheepStation Flow
SheepStation Flow
SheepStation Flow
SheepStation Flow
SheepStation Flow
SheepStation Flow
SheepStation Flow
SheepStation Flow
SheepStation Flow
2200–2300
2200–2300
2200–2300
2200–2300
2200–2300
2200–2300
2200–2300
2200–2300
2200–2300
2200–2300
2200–2300
2670±80
2200–2300
2140±100
2030±80
2885±150
2885±150
2885±150
2885±150
2885±150
2885±150
2885±150
2885±150
C-14 age
1520
1370
1200
910
820
700
600
450
300
180
20
2377
20
100
500
1539
1250
975
580
450
300
143
61
Elevation
(m)
14.4
15.3
16.2
18.0
18.6
19.4
20.1
21.1
22.1
22.8
23.9
9.8
23.9
23.4
20.7
14.3
15.9
17.6
20.2
21.1
22.1
23.1
23.6
Temperature
(°C)
Table 2. Lichen Data Matrix
1000
1200
1350
1500
1500
1500
1500
1200
1000
700
500
500
500
500
500
1200
1100
1000
750
700
600
500
350
Precipitation
(mm/yr)
0
200
350
500
430
370
300
-50
-300
-700
-1000
-1100
-1000
-1600
-1500
-300
-630
-850
-1250
-1375
-1550
-1725
-1950
Available
moisture
(mm)
16.69
21.53
24.81
27.74
28.28
29.11
29.74
26.66
21.37
13.54
7.66
0.83
5.66
4.89
2.74
28.10
25.16
18.02
11.72
8.39
8.10
7.05
7.22
Average %
porosity
72
STEVEN J. GORDON
73
CHEMICAL WEATHERING
Table 3. Matrix Descriptive Statistics
Variable
Maximum
z1
z2
16.305
9.864
0.83
29.74
-1.624
-0.009
Porosity abiotic (%)
2.804
1.055
0.45
4.77
0.170
-0.380
Age (yr. B.P.)
2448.696
342.760
2030
2885
-1.820
10.480
Elevation (m)
711.522
606.941
20
2377
0.486
0.959
19.461
3.759
9.80
23.9
-0.049
-0.779
919.565
394.778
350
1500
-1.405
0.188
-640.435
793.048
-1950
500
-1.376
0.007
Mean annual temperature (°C)
Precipitation (mm/yr)
Available moisture (mm/yr)
Mean
SD
Porosity lichen (%)
Minimum
The examples of elevation and lichen cover illustrate that it is possible to miss
complex relationships among weathering factors and weathering rates in studies
where only one or two factors are varied and the remainder are held constant.
Brady et al. (1999) illustrated how multiple weathering factors (elevation, temperature, available moisture, lichens) are observed to synergistically regulate dissolution
of plagioclase. The authors report their results in the form of activation energies,
which is common in the geochemical community. Activation energies, while useful
to illustrate the effects of the variables on dissolution, does not lend itself to
application to a synthetic “Pope et al.” type weathering model, nor to geospatial
modeling and mapping. Paradise and Yin (1993) illustrated how multivariate statistics may be used to address weathering rates. Some researchers resist the use of
multivariate statistics, however, in favor of the perceived “more reliable” geochemical lexicon. The purpose of the research reported in this paper is to apply
multivariate statistical analyses to an existing weathering dataset (that of Brady et
al., 1999) to assess the applicability of such statistical analyses for use in the
geospatial modeling of weathering.
DATA AND METHODS
The dataset (reported in part in Brady et al., 1999) consists of two matrices, each
consisting of twenty-three rows and five columns. Each row represents one specific
location of sampling and each of four columns represents a site-specific variable
(mean annual temperature, elevation, annual precipitation, and available
moisture), and an additional column represents the mean porosity of plagioclase
samples collected at each station (Tables 1 and 2). The term “available moisture”
refers to the value of precipitation minus potential evapotranspiration for each site.
The “abiotic” matrix (Table 1) reflects plagioclase samples collected from surfaces
free of visible rock coatings, such as rock varnish, silica glaze, and biofilms of algae,
mosses, or lichens. The “lichen” matrix (Table 2) is constructed from plagioclase
samples collected from lichen-covered (Stereocaulon vulcani) surfaces. Table 3 outlines some descriptive statistics of the dataset.
Plagioclase samples were collected from flows located on the Island of Hawaii
(Fig. 1) with a rock hammer and placed in plastic bags. The approach to sample collection was to avoid places where water might collect. Thus, only the tops of rock
74
STEVEN J. GORDON
Fig. 1. Location of the study. Shading represents elevation.
outcrops were sampled. For example, a knobby protuberance on an a’a flow would
be sampled, or the top of a dome on a pahoehoe flow.
It is not possible to be absolutely certain that lichens did not impact an “abiotic”
sample. Certainly, lichens theoretically could have been present and died without
leaving evidence of their presence. It is possible, however, to use existing evidence
to limit the possibility of truly “biotic” samples being mistaken for “abiotic”
samples. The first step is using a hand lens in the field. In the lab, visual inspections
of SEM (scanning electron microscopy) versus BSE (backscattered electron) images
are useful in identifying the presence of lichens. Lichens will appear conspicuously
on the surface of the sample in SEM, while in BSE mode they will appear very dark
in the same space on the sample (resulting from the low atomic number of carbon).
The oxalate crystals secreted by lichens are visible in backscatter, and allow for the
determination of lichen presence (Figure 2A). It is possible that the act of polishing
cross sections for imaging in SEM and BSE may remove the lichens in the Z-axis.
Despite this possibility, however, it is possible to observe the presence or effect of
lichens (Fig. 2B), even if the lichen is up to a centimeter away from the surface of
the cross section (Fig. 2C). Finally, the general “shape” or “nature” of the porosity
CHEMICAL WEATHERING
75
Fig. 2. Evidence of lichen influence in weathering of basalt samples, flow f7dh7.9 (610 m). Backscattered electron (BSE) images, where higher atomic number appears brighter in the image. Scale bar
in all images is 10 microns. (A) Oxalate crystals, secreted by lichens, appear very bright within the
darker organic (lichen) matrix in the upper left one-third of the image. The surface of the sample (and
the flow) runs roughly diagonally across the image from lower left to upper right. (B) Evidence of lichens
is apparent even if the lichens have been removed by sample preparation from the cross section. Bright
“speckles” are seen near the sample (and flow) surface in the upper right one-third of the image. The
speckles appear to “float” above the surface because surface lichens hold them in place, but the lichen
is not visible in this cross section. (C) This is true even when the lichen is up to a centimeter away from
the surface of cross section. Bright speckles of oxalates are seen along the surface, as is a large void left
from dissolution—further evidence of lichen activity even when lichens are not visible.
created by dissolution underneath lichens looks different than porosity created in
non-lichen conditions. The large areas of porosity in Figure 2 are examples of dissolution porosity found underneath lichens.
Plagioclase samples were processed to determine individual percent porosity
using procedures outlined in detail by Brady et al. (1999). Briefly, images of plagioclase grains were collected in the backscattered electron (BSE) mode of an electron
microprobe. These images reveal dark holes or pores in the mineral grain that are
the result of chemical dissolution. The images are processed with image processing
76
STEVEN J. GORDON
Table 4. Factor Analysis Loadingsa
Variable
Factor 1
Factor 2
Factor 3
Age
0.11535
-0.23588
0.96417
Elevation
0.97904
0.19301
0.06287
Temperature
0.19686
0.97089
-0.09321
Precipitation
0.26999
0.88512
-0.35814
-0.97076
-0.22918
-0.06760
Available moisture
a
Bold type shows emphasis of loadings on each factor.
software to determine the number of pixels that classify as pore space and the total
number of pixels in the entire grain. Division of pore space by total pixels yields a
percent porosity, which can also be divided by time since flow emplacement to
yield a weathering rate of percent area lost per unit of time.
It is worthwhile to note that the microporosity used as an indicator of chemical
dissolution is fundamentally different than vesicles commonly found in basalt.
Vesicles in basalt, usually the result of gas bubbles trapped in the cooling lava,
appear very large in comparison to microporosity created by chemical dissolution.
Vesicles may be further differentiated from dissolution microporosity by their ellipsoid shapes with smooth perimeters, while dissolution microporosity takes the form
of “etch pits” (Rowe and Brantley, 1993), “networks and holes” (Pope, 1995), and
“cracks” (Gordon, 1996). These features do not have smooth perimeters. Calculation of porosity from dissolution allows for easy avoidance of vesicles based upon
the characteristics above, and therefore the porosity values used in this study do not
include void space from vesicles.
Principal components analysis was used in order to compress the dataset and
reduce multicollinearity problems among independent variables in each of the two
matrices. Extraction at eigenvalues greater than one resulted in two factors for each
matrix, however, the statistical software defaults were overridden and three factors
were extracted. The software default of two values is set because most analyses will
only yield two factors. The default was overridden to allow the inclusion of the age
of the lava flow to be included in the analysis. Varimax rotation was employed to
achieve high loadings of a few variables on each factor.
Factor scores for each of the three resultant factors were saved and used in
multiple regression analysis. Two multiple regression analyses were performed,
with one regressing the dependent variable “abiotic percent porosity” against the
three factors (independent variables) described in more detail below. The second
analysis regressed “lichen percent porosity” against the same three factors. Each
stepwise multiple regression analysis used pairwise deletion of missing cases, an Fprobability of .05 to enter variables, and an F-probability of .10 to remove variables.
RESULTS
Principal components analysis was performed to compress the independent variable matrix (Tables 4 and 5). Recognizable patterns appeared for the three factors,
77
CHEMICAL WEATHERING
Table 5. Final Factor Analysis Statistics
Communality
Factor
Eigenvalue
Percentage of
variance
Cumulative
percentage
Age
0.99857
1
2.83007
56.6
56.6
Elevation
0.99972
2
1.59675
31.9
88.5
Temperature
0.99946
3
0.54559
10.9
99.4
Precipitation
0.99006
n.a.
n.a.
n.a.
n.a.
Available moisture
0.98459
n.a.
n.a.
n.a.
n.a.
Variable
Table 6. Abiotic Multiple Regression Results
Metric
Value
Adjusted r2
0.67529
Significance level of r2
0.00
Standard error of estimate
0.60
Sum of squares
17.61000
Intercept
2.80
Durbin-Watson value
0.30867
Table 7. Abiotic Regression Coefficients
Factor
b
Beta
-0.407258
-0.386145
Moist conditions
0.55939
0.530387
0.0003
Longer exposure
0.56713
0.537729
0.0003
High elevation, low temperature
Significance
0.00049
with the first factor representing low temperature, high elevation conditions
(describes 56.6% of the variation), the second factor representing moist conditions
(describes 31.9%), and the third factor representing increasing age of the lava flow
(describes 10.9%). These three factors describe 99.4% of the variance in each
matrix.
The factor scores were used in multiple regression analyses. The first analysis
involved the regression of abiotic porosity against the independent variable factors,
described in Tables 6 and 7. The low temperature, moist conditions, and age of lava
flow factors describe approximately 68% of the variation in porosity in plagioclase
(adjusted r2 = 0.68), which is significant at the 0.0049–0.0003 level. The standard
error is low (0.60), as is the sum of squares (17.61). A low Durbin-Watson statistic
(0.31) suggests a degree of autocorrelation among the error terms in the regression.
The regression coefficients for the abiotic analysis are shown in Table 7. The
moist and increasing age factors have approximately the same weight in the determination of plagioclase porosity, with beta values of approximately 0.53 each.
These coefficients are highly significant (each is significant to the 0.00 levels). The
low temperature, high elevation factor has a less pronounced, negative effect on
78
STEVEN J. GORDON
Table 8. Lichen Multiple Regression Results
Metric
Adjusted r
Value
2
0.93934
Significance level of r2
0.00
Standard error of estimate
2.42
Sum of squares
2028.30
Intercept
16.30
Durbin-Watson value
0.86966
Table 9. Lichen Regression Coefficients
Factor
b
Beta
Significance
High elevation, low temperature
0.87516
0.08873
0.1074
Moist conditions
9.56155
0.96937
0.0000
Longer exposure
0.08073
0.00819
0.8778
plagioclase porosity (beta = -0.39). This coefficient is also significant, to the 0.01
level.
The second multiple regression analysis regressed lichen porosity against the
independent variable factors, shown in Tables 8 and 9. The three factors account for
approximately 94% of the variation in porosity of plagioclase in lichen-encrusted
environments (adjusted r2 = 0.94), which is significant to the 0.0000–0.8778 level.
The standard error of the estimate is low (2.43). A Durbin-Watson statistic of 0.87
suggests less of a concern with error autocorrelation than in the abiotic analysis.
The regression coefficients for the lichen analysis are reported in Table 9. The
moist conditions factor is the most important control on plagioclase porosity (beta =
0.97, significant to the 0.00 level). The other two factors are less important in the
determination of plagioclase porosity, and the regression coefficients are much less
significant as well.
DISCUSSION
The results of this analysis agree with those reported in Brady et al. (1999). The
lichens enhance the rate of dissolution of plagioclase by an order of magnitude
(Table 3). Therefore, the use of multivariate statistical analysis did not alter the fundamental result of the previous study that employed activation energies. This falls
into the extreme lower end of the range of published lichen enhancement values
(from an order of magnitude [Jackson and Keller, 1970; McCarroll and Viles, 1995])
to 2–3 orders of magnitude (Schwartzman and Volk, 1989). The generally small
enhancement supports the ideas of Berner (1992) and Cochran and Berner (1993)
that vascular plants, not lichens, drive global carbon dioxide cycling as a result of
enhanced silicate weathering. The results also underscore the conclusions of
Wasklewicz et al. (1993) and Waskelwicz (1994) that the presence of lichens on
CHEMICAL WEATHERING
79
Hawaiian lava flows is instrumental in determining the nature and order of chemical weathering of component minerals.
Lichens are important weathering accelerators in other contexts. The above
results reinforce the importance of biotic acceleration of dissolution rates of building stones (e.g., Saiz-Jimenez, 1999) and nuclear waste glasses (e.g., Staudigel et
al., 1995). While vascular plants may play a stronger role in weathering enhancement in a carbon cycle context, it is more likely that a lichen, not a vascular plant,
will adopt a statue, fresco, upright cemetery stone, or glass host in a federal waste
repository as a growth substrate. Lichens therefore represent the highest order
organisms that would probably subject building stones and nuclear waste glasses to
biotic acceleration of dissolution.
Colder, higher elevation locations resulted in lower rates of weathering while
low elevation, warmer locations resulted in more rapid rates. Brady et al. (1999)
found that higher temperatures favor weathering. This agrees with published results
from the field (e.g., Velbel, 1993) and the laboratory (e.g., Welch and Ullman,
2000). The advantage of the statistical approach is that we know that the high
elevation/low temperature factor in the regression had a negative effect on rate,
most notably in the abiotic data matrix. Because the factor contained variation in
both temperature and elevation, we can conclude that temperature, as it varies with
elevation, outweighs other factors not specifically named that vary with elevation
(e.g., insolation, temperature range) in the determination of dissolution rate. In
other words, we know that in this case, the expected increase in insolation with
elevation (and therefore expected increase in weathering rate with elevation) is
overshadowed by the expected decrease in temperature (and therefore expected
decrease in dissolution rate).
Higher elevations typically experience increases in physical weathering, which
may also lead to increased chemical weathering by exposing fresh surfaces to
weathering solutions. This study uses porosity values that are the result of an
attempt to avoid porosity created from physical weathering, while acknowledging
that this is difficult to do. The results show, however, that the increased physical
weathering that is expected at higher elevations does not increase chemical weathering rates enough in this setting to overcome the effect of warmer temperatures at
lower elevations. The above two types of results are extremely useful in mapping
weathering potential of real-world locations.
The abiotic and biotic analyses show the importance of moisture in increasing
weathering rates. This is also in agreement with the results of Brady et al. (1999).
The importance of water in weathering is underscored in arid environments (Roth,
1965; Pope et al., 1995) and the same importance may be applied here. Special
attention may be paid to the observation that in biotic conditions, the “moist conditions” factor is clearly more important to weathering rate than age or elevation/
temperature. Brady et al. (1999) suggested possible reasons for this, including: (1)
water availability maximizes secretion of organic acids or hyphal activity, (2)
lichens increase water retention in pore spaces, and (3) the complexing ability of
organic acids to remove dissolved mineral components from the weathering solution. Additionally, because lichens can survive extended periods of desiccation and
reabsorb water and swell when wet (Saiz-Jimenez, 1999), areas of greater moisture
80
STEVEN J. GORDON
may facilitate intensified physical weathering from shrink/swell. This physical
action may then create fresh surfaces for accelerated dissolution. The significance
of physical weathering in accelerating chemical weathering is discussed by Hoch et
al. (1999).
The results of the statistical analysis reveal that the age of the lava flow is a factor
in the determination of lichen-free dissolution of plagioclase, but is not a significant
factor in the lichen-covered rate. This expands upon the findings of Jackson and
Keller (1970) who reported that a range of <100 years among different flows
resulted in no perceptible dependence of the rate of basalt weathering underneath
lichens. Logic suggests that weathering will be more pronounced in materials that
have been exposed to the subaerial environment longer. This is not the case in the
lichen-mediated dissolution matrix. It appears that moisture availability is such a
robust factor for these settings that the range of ages of the flows (all within ~1000
yrs of each other) is negligible, and that it is only where moisture is not as crucial
(in the lichen-free settings) that age variation exerts an influence. This addresses the
importance of lichens in the weathering environment.
This result may, however, be the result of sampling flows within a very narrow
range of ages. It is probable that a similar analysis performed on a dataset with a
greater range of ages would not replicate this result, but rather show that lava flow
age is a significant factor in determining weathering rate. This underscores the utility of the statistical approach, in that the method allows data from a narrow range
of ages to be effectively and “safely” co-mingled without invalidating the overall
result. It also allows the age of samples, in a context where the samples come from
a large rage of ages, to contribute vigorously to the determination of weathering rate
in relation to simultaneous variation in other weathering factors (moisture, temperature, etc.).
The suite of multivariate statistical analyses employed in this project (principal
components analysis, multiple regression) appears to be a valid and useful method
of modeling the spatial variability of chemical weathering of plagioclase, and an
alternative to methods that require control (constant) variables. Studies of weathering of earth materials generally follow a form necessary for their application. The
“geochemical” approach to weathering analysis, driven by the need to cast environmental variables in a form usable by climate models, are generally not directly
comparable or applicable to studies that address landscape processes in the context
of qualitative models, such as Pope et al. (1995). This paper outlines a method to
address this lack of applicability.
Application of multivariate statistics provides a quantitative expression of the
impact of several weathering factors on weathering rates as they vary simultaneously. This equation is suitable for use in raster-based analysis (Geographic
Information Systems) of a given area in a way that is difficult for geochemical
studies. It may be possible to apply results of multivariate modeling to the development of a multi-layered array for a given area, perhaps of Hawaii Volcanoes
National Park. Each layer of the array would contain information on an individual
weathering factor (insolation, age, elevation, etc.) whereby the value of each pixel
in the raster layer is the average value of that factor (e.g., 2345 m elevation, 3030
years B.P., etc.). Such an array could then be easily transformed into a map of
CHEMICAL WEATHERING
81
weathering rate, based upon mathematical operations. The logical extension for
further research is then to field-test the model for Hawaii or any other given area.
CONCLUSION
Conceptual models which seek to address the simultaneous effects of many environmental factors on weathering rates, such as outlined by the Pope et al. “Boundary-Layer Model” (1995), suffer from an inability to quantify these influences.
Indeed, the closing sentences of that work contain an invitation to weathering
researchers to refine the model. The multivariate approach outlined in this work
offers the possibility of converting the Pope et al. model into a quantitative equation
useful for empirical study and for geospatial modeling and mapping. Further, the
results of this project show that the technique derives results that do not differ substantially from those obtained using the same dataset via geochemical-based analyses. The results of this study are as follows.
(1) Lichens enhance the rate of chemical weathering of plagioclase in
Hawaiian basalts by an order of magnitude.
(2) Cooler, higher elevation environments experience slower chemical
weathering, whereas warmer, lower elevation conditions result in more
rapid chemical weathering.
(3) Moisture availability has a strong influence in determining chemical
weathering rate, especially for weathering in the presence of lichens.
(4) Moisture availability may overshadow the effects of other variables,
such as age, elevation, and temperature.
The use of a multivariate approach, such as the one outlined in this paper, allows
the above results to be easily exchanged with other researchers in a quantitatively
meaningful way, and lends itself to geospatial modeling and prediction of chemical
weathering rates via a raster-based Geographic Information System.
Acknowledgments: The author wishes to thank Pat Brady, Ron Dorn, Bob Balling, and anonymous
reviewers for their critical readings of, and ultimate improvements to, this manuscript. Figure 1 was
generously provided by Thomas Paradise. The BSE images in Figure 2 were provided by Ron Dorn.
REFERENCES
Adams, K. D., Locke, W. W., and Rossi, R. (1992) Obsidian-hydration dating of
fluvially reworked sediments in the West Yellowstone region, Montana.
Quaternary Research, Vol. 38, 180–195.
Ambrose, W. R. (1994) Obsidian hydration dating of a Pleistocene age site from the
Manus Islands, Papua-New-Guinea. Quaternary Science Reviews, Vol. 13, 137–
142.
82
STEVEN J. GORDON
Banfield, J. F., Barker, W. W., Welch, S. A., and Taunton, A. (1999) Biological
impact on mineral dissolution: Application of the lichen model to understanding
mineral weathering in the rhizosphere. Proceedings of the National Academy of
Sciences of the United States of America, Vol. 96, 3404–3411.
Barry, R. G. (1992) Mountain Weather and Climate. New York, NY: Routledge.
Berner, R. A. (1992) Weathering, plants, and the long-term carbon cycle. Geochimica
et Cosmochimica Acta, Vol. 56, 3225–3231.
Bischoff, J. L. and Dickson, F. W. (1975) Seawater-basalt interaction at 200 C and
500 bars: Implications for origin of sea-floor heavy-metal deposits and regulation
of seawater chemistry. Earth and Planetary Science Letters, Vol. 25, 385–397.
Brady, P. V. and Carroll, S. A. (1994) Direct effects of CO2 and temperature on silicate weathering: Possible implications for climate control. Geochimica et Cosmochimica Acta, Vol. 58, 1853–1856.
Brady, P. V., Dorn, R. I., Brazel, A. J., Clark, J., Moore, R. B., and Glidewell, T. (1999)
Direct measurement of the combined effects of lichen, rainfall, and temperature
on silicate weathering. Geochimica et Cosmochimica Acta, Vol. 63, 3293–3300.
Casey, W. H. and Sposito, G. (1992) On the temperature dependence of mineral
dissolution rates. Geochimica et Cosmochimica Acta, Vol. 56, 3825–3830.
Chadwick, O. A., Derry, L. A., Vitousek, P. M., Huebert, B. J., and Hedin, L. O.
(1999) Changing sources of nutrients during four million years of ecosystem
development. Nature, Vol. 397, 491–497.
Chen, J., Blume, H.-P., and Beyer, L. (2000) Weathering of rocks induced by lichen
colonization–A review. Catena, Vol. 39, 121–146.
Cochran, M. F. and Berner, R. A. (1993) Enhancement of silicate weathering rates
by vascular land plants: Quantifying the effect. Chemical Geology, Vol. 107,
213–215.
Colman, S. M. and Pierce, K. L. (1981) Weathering Rings on Andesitic and Basaltic
Stones as a Quaternary Age Indicator, Western United States. Washington, DC:
United States Geological Survey.
Dorn, R. I. and Brady, P. V. (1995) Rock-based measurement of temperature-dependent plagioclase weathering. Geochimica et Cosmochimica Acta, Vol. 59,
2847–2852.
Ewing, R. C. and Haaker, R. F. (1979) Natural Glasses: Analogues for Radioactive
Waste Forms. Richland, WA: Battelle Pacific Northwest Laboratories.
Freidman, I., Trembour, F. W., Smith, F. L., and Smith, G. I. (1994) Is obsidian hydration dating affected by relative-humidity? Quaternary Research, Vol. 41, 185–
190.
Gordon, S. J. (1996) Assessment of the Effects of Temperature on Chemical Weathering of Plagioclase, Iztaccihuatl, Mexico. Unpublished master’s thesis, Department
of Geography, Arizona State University, Tempe. Arizona, 62.
Gordon, S. J. and Brady, P. V. (2002) In situ determination of long-term basaltic glass
dissolution in the unsaturated zone. Chemical Geology, Vol. 190, 113–122.
Hall, K. (1999) The role of thermal stress fatigue in the breakdown of rock in cold
regions. Geomorphology, Vol. 31, 47–63.
Hay, R. L. and Jones, B. F. (1972) Weathering of basaltic tephra on the Island of
Hawaii. Geological Society of America Bulletin, Vol. 83, 317–332.
CHEMICAL WEATHERING
83
Hellmann, R. (1994) The albite-water system: Part I. The kinetics of dissolution as a
function of pH at 100°, 200°, and 300°C. Geochimica et Cosmochimica Acta,
Vol. 58, 595–611.
Hoch, A. R., Reddy, M. M., and Drever, J. I. (1999) Importance of mechanical disaggregation in chemical weathering in a cold alpine environment, San Juan
Mountains, Colorado. Geological Society of America Bulletin, Vol. 111, 304–
314.
Jackson, T. A. and Keller, W. D. (1970) A comparative study of the role of lichens
and “inorganic” processes in the chemical weathering of recent Hawaiian lava
flows. American Journal of Science, Vol. 269, 446–466.
Jenny, H. (1941) Factors of Soil Formation. New York, NY: McGraw-Hill.
Likens, G. E., Bormann, F. H., Pierce, R. S., Easton, J. S., and Johnson, N. M. (1977)
Biogeochemistry of a Forested Ecosystem. New York, NY: Springer-Verlag.
Lutze, W., Malow, G., Ewing, R. C., Jercinovic, M. J., and Keil, K. (1985) Alteration
of basalt glasses—Implications for modeling the long-term stability of nuclear
waste glasses. Nature, Vol. 314, 252–255.
Lynch, T. F. and Stevenson, C. M. (1992) Obsidian hydration dating and temperature
controls in the Punta-Negra region of Northern Chile. Quaternary Research, Vol.
37, 117–124.
McCarroll, D. and Viles, H. A. (1995) Rock weathering by the lichen Lecidea auriculata in an arctic alpine environment. Earth Surface Processes and Landforms,
Vol. 20, 199–206.
McLennan, S. M. (1993) Weathering and global denudation. Journal of Geology,
Vol. 101, 295–303.
Mottl, M. J. and Holland, H. D. (1978) Chemical exchange during hydrothermal
alteration of basalt by seawater—I. Experimental results for major and minor components of seawater. Geochimica et Cosmochimica Acta, Vol. 42, 1103–1115.
Paradise, T. and Yin, Z.-Y. (1993) Weathering pit characteristics and topography on
Stone Mountain, Georgia. Physical Geography, Vol. 14, 68–80.
Paradise, T. R. (1994) Limestone weathering rate analysis. In P. M. Bikai, ed., The
Great Temple of Amman. The Architecture. Amman, Jordan: American Center of
Oriental Research, 110–114.
Paradise, T. R. (1995) Sandstone weathering thresholds in Petra, Jordan. Physical
Geography, Vol. 16, 205–222.
Paradise, T. R. (1997) Disparate sandstone weathering beneath lichens, Red
Mountain, Arizona. Geografiska Annaler, Vol. 79A, 177–184.
Pedro, G. and Sieffermann, G. (1979) Weathering of rocks and formation of soils. In
F. R. Siegel, ed., Review of Research on Modern Problems in Geochemistry. Paris,
France: UNESCO.
Pope, G., Dorn, R. I., and Dixon, J. (1995) A new conceptual model for understanding geographical variations in weathering. Annals of the Association of American
Geographers, Vol. 85, 38–64.
Pope, G. (1995) Internal weathering in quartz grains. Physical Geography, Vol. 16,
315–338.
Roth, E. S. (1965) Temperature and water content as factors in desert weathering.
Journal of Geology, Vol. 73, 454–468.
84
STEVEN J. GORDON
Rowe, G. L. and Brantley, S. L. (1993) Estimation of the dissolution rates of andesitic
glass, plagioclase and pyroxene in a flank aquifer of Poás Volcano, Costa Rica.
Chemical Geology, Vol. 105, 71–87.
Saiz-Jimenez, C. (1999) Biogeochemistry of weathering processes in monuments.
Geomicrobiology Journal, Vol. 16, 27–37.
Schwartzman, D. W. and Volk, T. (1989) Biotic enhancement of weathering and the
habitability of Earth. Nature, Vol. 340, 457–460.
Seyfried, W. E. and Mottl, M. J. (1982) Hydrothermal alteration of basalt by seawater under seawater dominated conditions. Geochimica et Cosmochimica
Acta, Vol. 46, 985–1002.
Staudigel, H., Chastain, R. A., Yayanos, A., and Bourcier, W. (1995) Biologically
mediated dissolution of glass. Chemical Geology, Vol. 126, 147–154.
Twidale, C. R. (1982) Granite Landforms. Amsterdam, The Netherlands: Elsevier.
Velbel, M. A. (1993) Temperature dependence on silicate weathering in nature:
How strong a negative feedback on long-term accumulation of atmospheric
CO2 and global greenhouse warming? Geology, Vol. 21, 1059–1062.
Wasklewicz, T., Dorn, R., Clark, S., Pope, G., Hetrick, J., Dixon, J., Moore, R., and
Clark, J. (1993) Olivine does not necessarily weather first. Singapore Journal of
Tropical Geography, Vol. 14, 72–80.
Wasklewicz, T. (1994) Importance of environment on the order of mineral weathering in olivine basalts, Hawaii. Earth Surface Processes and Landforms, Vol. 19,
715–735.
Welch, S. A. and Ullman, W. J. (2000) The temperature dependence of bytownite
feldspar dissolution in neutral aqueous solutions of inorganic and organic
ligands at low temperature (5–35°C). Chemical Geology, Vol. 167, 337–354.
Werme, L., Bjorner, I. K., Bart, G., Zwicky, H. U., Grambow, B., Lutze, W., Ewing,
R. C., and Magrabi, C. (1990) Chemical corrosion of highly radioactive borosilicate nuclear waste glass under simulated repository conditions. Journal of
Materials Research, Vol. 5, 1130–1146.
White, A. F. (1984) Weathering characteristics of natural glasses and influences on
associated water chemistry. Journal of Non-Crystalline Solids, Vol. 67, 225–244.
White, A. F., Blum, A. E., Bullen, T. D., Vivit, D. V., Schulz, M., and Fitzpatrick, J.
(1999) The effect of temperature on experimental and natural chemical
weathering rates of granitoid rocks. Geochimica et Cosmochimica Acta, Vol. 63,
3277–3291.
White, A. F., Claasen, H. C., and Benson, L. V. (1980) The Effect of Dissolution of
Volcanic Glass on the Water Chemistry in a Tuffaceous Auifer, Rainier Mesa,
Nevada. Washington, DC: United States Geological Survey, USGS Water-Supply
Paper 1535-Q.
Wilson, M. J. and Jones, D. (1983) Lichen weathering of minerals: Implications for
pedogenesis. In R. C. L. Wilson, ed., Residual Deposits: Surface Related Weathering Processes and Materials. Oxford, UK: Blackwell.