245
Base cation mineral weathering and total release rates from soils in three calibrated forest
watersheds on the Canadian Boreal Shield
Rock Ouimet and Louis Duchesne
Direction de la recherche forestière, Ministère des Ressources naturelles, de la Faune et des Parcs, Gouvernement du Québec, 2700, rue Einstein, Sainte-Foy, Québec, Canada
G1P 3W8 (e-mail : [email protected]). Received 23 September 2004, accepted 5 February 2005.
Ouimet, R. and Duchesne, L. 2005. Base cation mineral weathering and total release rates from soils in three calibrated for-est watersheds on the Canadian Boreal Shield. Can. J. Soil Sci. 85: 245–260. Total release rates of base cations (Ca, Mg, K,and Na) from soils and from watersheds were evaluated using three methods. Three methods, one of which is new, were also usedto evaluate mineral weathering rates of soils for three calibrated forest watersheds in the forest ministry’s monitoring network(Réseau d’étude et de surveillance des ecosystems forestiers: Quebec Forest Ecosystem Research and Monitoring Network;RESEF) on the Canadian Boreal Shield. We also compiled an extensive literature review of forest soil base cation release rates,focussed on northeastern North American forest soils of granitic lithology. With the exception of the total release and mineralweathering of Ca from soils at the Lake Laflamme Watershed site, and the total release of K from the three watersheds, soils andwatershed release rates for the three watersheds were within the confidence interval of release rates compiled for forest ecosys-tems with similar granitic environment (compiled data for solum [mmol (+) m–2 yr–1 ± 95% CI], Ca: 33.8 ± 16.3, Mg: 16.8 ± 4.2,K: 13.0 ± 5.6, Na:11.1 ± 3.0, and sum of base cations (BC): 61.2 ± 11.0; compiled data from watersheds, Ca: 82.8 ± 24.6, Mg:50.8 ± 17.0, K: 7.8 ± 2.2, Na: 44.7 ± 12.8, and BC: 186.0 ± 49.9). Given the uncertainties associated with the calculations, therewas little overall difference between total release rates and weathering release rates from soils for two of the three watersheds.
Key words: Weathering rate, forest soil, granitic environment, watershed, PMB method, compositional trends
Ouimet, R. et Duchesne, L. 2005. Base cation mineral weathering and total release rates from soils in three calibrated for-est watersheds on the Canadian Boreal Shield. Can. J. Soil Sci. 85: 245–260. Nous avons employé trois méthodes pour évaluerle taux de largage total des cations basiques (Ca, Mg, K et Na) des sols et des bassins versants et trois méthodes (dont une nou-velle) pour évaluer le taux d’altération minérale des cations basiques des sols dans trois bassins versants forestiers du Réseauquébécois d’étude et de surveillance des écosystèmes forestiers (RESEF) situés sur le Bouclier canadien. Nous avons aussi com-pilé des données à partir d’une revue de littérature extensive sur le taux de dégagement des cations basiques des sols, centrée surles sols forestiers du nord-est de l’Amérique du Nord ayant une lithologie granitique. Mis à part le taux de largage total et d’altéra-tion minérale du Ca des sols dans le bassin versant du Lac Laflamme et le taux de largage total du K des trois bassins versants, lestaux de largage en cations basiques se trouvaient à l’intérieur des limites de confiance de ceux compilés pour des écosystèmesforestiers similaires ayant un environnement granitique (compilation pour le solum [mmol (+) m–2 a–1 ± 95 % IC], Ca: 33.8 ± 16.3,Mg: 16.8 ± 4.2, K: 13.0 ± 5.6, Na:11.1 ± 3.0 et la somme des cations basiques (CB): 61.2 ± 11.0; compilation pour les bassins ver-sants, Ca: 82.8 ± 24.6, Mg: 50.8 ± 17.0, K: 7.8 ± 2.2, Na: 44.7 ± 12.8, et CB: 186.0 ± 49.9). Étant donné les incertitudes associéesau calculs, il y avait peu de différence globalement entre le taux de largage total et celui d’altération minérale de CB des sols dedeux des trois bassins versants.
Mots clés: Taux d’altération, sols forestiers, environnement granitique, bassin versant, méthode PMB, évolution compositionnelle
There is a growing body of evidence that some forest soilshave been experiencing depletion of available base cations,particularly calcium, in northeastern US and easternCanada, mainly on the basis of nutrient mass balance stud-ies conducted at the watershed scale (Foster et al. 1992;Houle et al. 1997; Lawrence et al. 1997; Likens et al. 1998;Friedland and Miller 1999; Martin et al. 2000; Bélanger et
al. 2002b; Bailey et al. 2003; Watmough and Dillon2003a,b; Watmough et al. 2005). These observations sug-gest that external sources cannot compensate for lossesthrough vegetation immobilization, harvesting, and leach-ing. Therefore, available soil base cations would make upthe difference between ecosystem inputs and outputs, andtheir pool would become depleted over time. Along with
Abbreviations: BC, base cations; CLT, Cl tracer; CT, compositional trends; LCW, Lake Clair Watershed; LLW, LakeLaflamme Watershed; LTW, Lac de la Tirasse Watershed; IOB, input-output balance; PBM, pedological mass balance;QA/QC, quality assurance/quality control; TR, total release; TR(X), TR of element X; XRD, X-ray diffraction; XRF, X-rayfluorescence; WR, weathering release
246 CANADIAN JOURNAL OF SOIL SCIENCE
watershed-monitoring studies, evidence is accumulatingthat reveals increased forest soil acidification in northeast-ern North America (Miller et al. 1993; Johnson et al. 1994;Lawrence et al. 1997; Watmough and Dillon 2003b).Many forest ecosystems in these regions have been experi-encing mineral nutrient deficiencies and health problems(Ouimet and Camiré 1995; Sharpe et al. 1999; Horsley et al.2000), which can be qualified as a new historical phenome-non (Bondietti et al. 1990). This phenomenon is thought tobe due partly to depletion of soil-available base cations(Sharpe et al. 1999; Driscoll et al. 2001; Duchesne et al.2002).
Mineral weathering, along with atmospheric deposition,are the only long-term “net” sources of base cations to ter-restrial ecosystems. Other ecosystem sources (throughfallleaching, organic matter mineralization in the forest floorand in the mineral soil, root exudation, and soil desorption)are only base cations being recycled through the so-called“biogeochemical cycle”.
However, interpretation of watershed monitoring massbalance calculations have been based on various mineralweathering estimation methods that are based on inherentassumptions. The release rates of base cations by soil min-erals have always been difficult to measure and quantifybecause of the lack of long-term soil studies (Richter andMarkewitz 2001). The quantification of mineral weatheringof soils is of importance for determining long-term avail-ability of plant nutrients, buffering capacity of soils and sur-face water quality. Weathering within the solum is achemical process that is controlled by soil mineral composi-tion and its exposed surface area, soil temperature and mois-ture regime, CO2 concentration, organic acid concentration,and soil depth (Sverdrup and Warfvinge 1993). The lastthree factors are in turn influenced by the vegetation (Kellyet al. 1998). Mineral weathering refers here to chemical andmineralogical transformations induced by the interaction ofwater and soil rock minerals. Thus, soil mineral “weatheringrelease” (WR) of base cations will be distinguished fromsoil “total release” (TR) as this latter process includes notonly base cation weathering from the soil mineral pool, butalso changes in net storage of base cations in the vegetationbiomass, in the soil organic matter pool, and in the soilexchangeable pool (Likens and Bormann 1995).
In this paper, we report estimates of TR rates of basecations from soils and watersheds and WR rates from soilslocated on the Boreal Shield in Quebec, in eastern Canada.Based on field data, three evaluation methods (or variants)for determining TR from soils and watersheds and threemethod variants to assess soil WR rates were used. Theresults were compared to those obtained from studies con-ducted on northeastern North American forest soils andwatersheds of granitic lithology.
MATERIAL AND METHODS
Study SitesWatershed characteristics are shown in Table 1. The catch-ments are part of the Quebec Forest Ecosystem Researchand Monitoring Network (Réseau d’étude et de surveillance
des écosystèmes forestiers; RESEF). The sites are locatedon the Boreal Shield ecozone (Fig. 1), underlain by agranitic gneiss series of the Grenville Province. The domi-nant soils are mainly well-developed loamy sand to sandyloam Orthic Ferro-Humic or Humo-Ferric Podzols accord-ing to the Canadian System of Soil Classification (SoilClassification Working Group 1998; Orthic Podzols accord-ing to the FAO) developed from sandy basal tills. The tillscontain angular rock fragments originating from the region-al bedrock.
The Lake Clair Watershed (LCW) has been monitoredsince 1987. Its vegetation is typical of the northern hard-wood forest of northeastern North America. Lake LaflammeWatershed (LLW) is located slightly north of LCW, but at ahigher altitude. The vegetation type is characteristic of theboreal coniferous balsam fir-white birch forest. This water-shed has been the subject of intensive biogeochemical mon-itoring since 1998. The third site, the Lac de la TirasseWatershed (LTW), is located at a higher latitude and hasbeen monitored since 1997. It is located in the black spruce-moss forest of the northern boreal coniferous range. Thiswatershed contains a first-order stream only, no lake, but asmall open bog (2.5 ha) is located at the stream source.
SamplingBulk precipitation was measured and sampled weekly in thethree watersheds in a forest opening using 2-L plastic bottles(n = 3) fitted with 577-cm2 plastic funnels. During winter(November to April, inclusively), bulk collectors werereplaced with 708-cm2 pails for snow collection on a 2-wkschedule. In a main study plot located at the midslope of thewatersheds, throughfall precipitation was measured andsampled at the same time as bulk precipitation, using 12 col-lectors in LLW and LTW, and 36 collectors in the LCW.Throughfall collectors were replaced by six pails for snowcollecting according to precipitation sampling methods. Soilwater percolating below the solum was sampled weekly dur-ing the snow-free period using ceramic cup lysimeters (n =4/watershed), located at an average depth of 80 cm (range of50–95 cm depending on solum depth) within the main studyplots. Watershed stream flow was measured at a gauged sta-tion that continuously monitored its flow. The stream wasalso sampled weekly or biweekly throughout the monitoringyears. Soil volumetric water content was monitored contin-uously at each site using time-domain reflectometry (n = 4;Tecktronix model 1502B). Ten soil profiles were sampled atthe LCW, and four each at LLW and LTW. Selected soilsamples were located at mid-slope, sampled by horizon, andtheir depth was recorded. Bulk density was determined foreach horizon using a volumetric soil sampler with a doublecylinder stainless steel core. Finally, 6 to 34 trees perspecies in different diameter classes were harvested in eachwatershed to determine biomass and nutrient content.
Laboratory AnalysisAll water samples were passed through a 45-µm filter priorto analysis for the following chemical species: SO4
2-, NO3-,
and Cl- by ion chromatography (Dionex), and NH4+ by
colourimetry (Technicon). Samples were lightly acidified
OUIMET AND DUCHESNE — FOREST SOIL WEATHERING 247
for the determination of Ca2+, Mg2+, K+, and Na+ by induc-tively coupled plasma emission spectroscopy (ICP JarrelAsh). Quality assurance/quality control (QA/QC) of soilwater analysis included ionic balance and measured versuscalculated electrical conductivity values. The error associat-
ed with the water chemical analyses was ≤ 3%. Soil sampleswere air dried and sieved at 2 mm prior to laboratory analy-ses. Soil pH was measured using a soil-to-water ratio of1:2.5. Forest floor organic matter was measured by dry losson ignition (Gallardo et al. 1987), while mineral soil organ-
Table 1. Characteristics of the three watersheds, and input and calibration parameters, for assessing the mineral weathering and total release ratesof soils in the three watersheds
Characteristic LCW LLW LTW
Time period for data compilation 1997–2001 1999–2001 1997–2001Longitude 71°40′W 71°14′W 73°39′WLatitude 46°57′N 47°17′N 49°13′NElevation a.s.l. (m) 350 800 400Mean annual air temperature (°C) 3.4 0.4 1.2Basin area (ha) 226 62.3 56.0Lake area (ha) 36 7.7 0Forest type Northern hardwoods Balsam fir-moss Black spruce-mossForest age (yr) >120 60 60Forest cover (%) 97 100 95Mean till depth (m) 1–3 3–15 > 1Average soil depth (m) 0.75 0.86 0.50Maximum rooting depth (m) 0.60 0.60 0.30Mean annual soil temperature (°C) 5.0 3.6 4.0Time since deglaciation (yr) 11 000 11 000 8 400Till geology Grey quartzic gneiss, plagioclase, Charnockitic gneiss Charnockitic gneiss, plagioclase,
biotite, and hornblende and granitoid orthopyroxene biotite, and hornblende
Hydrologic fluxes for the compilation periodAnnual precipitation (mm yr–1) 1210 1133 823Annual precipitation throughfall (mm yr–1) 1059 952 674Annual drainage from soil (mm yr–1) 787 837 594Annual drainage from watershed (mm yr–1) 805 860 618
Wet atmospheric depositions (mmol (+) m–2 yr–1) and dry deposition multiplication factor (within parentheses) for the compilation periodCa 6.53 (1.23) 3.83 (1.28) 2.96 (1.49)Mg 1.80 (1.30) 1.19 (1.24) 0.98 (1.42)K 0.71 (1.35) 0.47 (1.23) 0.38 (1.42)Na 2.40 (1) 1.79 (1) 1.15 (1)NH4 22.65 (1) 14.6 (1) 7.72 (1)NO3 28.72 (1) 28.9 (1) 10.81 (1)Cl 3.03 (1.58) 2.61 (2.55) 1.20 (2.58)SO4 39.90 (1.19) 27.45 (1.48) 16.14 (1.53)
Soil water volume weighted-mean concentrations at 80 cm depth (µeq L–1) for the compilation periodCa 28.02 52.40 23.82Mg 7.41 13.10 10.69K 2.74 2.29 3.54Na 19.54 28.73 25.33NH4 0.58 0.25 0.58NO3 6.47 5.07 0.23Cl 8.62 6.78 5.28SO4 75.02 72.43 21.43
Stream water volume-weighted mean concentrations (µeq L–1) for the compilation periodCa 103.56 80.78 106.30Mg 36.19 30.73 62.36K 8.59 4.56 7.86Na 37.72 36.73 37.32NH4 0.88 0.75 0.68NO3 5.97 2.62 2.70Cl 6.69 5.22 4.56SO4 81.74 57.45 44.41
Forest current net annual accumulation (mmol (+) m–2 yr–1) for the compilation periodCa 31.84 60.03 56.99Mg 6.34 11.19 9.79K 9.16 10.46 9.31N 69.32 113.80 45.41
248 CANADIAN JOURNAL OF SOIL SCIENCE
ic carbon was determined by wet combustion (Yeomans andBremner 1988). Total contents of Ca, Mg, K, and Na weredetermined for all soil samples by plasma emission spec-troscopy following digestion with concentrated H2SO4.Total concentrations of Mg, Ca, K, Na, Ti, Zr, and Cl weremeasured for only the most representative soil profile ateach watershed by X-ray fluorescence (XRF) spectroscopy(Philips, model 2400) after 4-g soil subsamples were groundin an agate mortar and pelletized with 1 g cellulose in ahigh-pressure hydraulic cylinder. Also, for these samples,mineral species in separated clay, silt, and sand fractionswere identified by X-ray diffraction (XRD). Pretreatmentfor particule size and XRD analyses included destruction ofhumus, and dispersion with sodium pyrophosphate during30 min of ultrasonic treatment. For XRD analysis, the soilfraction preparation included K, Mg saturation, and heattreatment of the K-saturated probes to 550°C. Aliquots ofeach preparation were comminuted in tungsten carbidegrinding dishes. One random powder mount of each prepa-ration was analysed for the presence of minor and majorminerals on a Diano-8535 semi-automated X-ray diffrac-tometer. Scans were run using Ni-filtered, Cu Kα radiation
produced at 45 kV and 15 mA. The proportions of eachidentified mineral were calculated by integration of peakareas without internal standards; overall mineral proportionswere calculated as soil fraction-weighted proportions ofminerals. Thus, mineral assemblages have to be interpretedas semi-quantitative. For all chemical analyses, local-refer-ence soil samples were used as QA/QC. The error associat-ed with the soil chemical laboratory analyses was ≤ 5%.
Data on these representative soil profiles are presented inTable 2. Finally, samples for the major tree compartmentsfrom each watershed were analysed for total N, K, Ca, andMg contents.
Data HandlingBulk wet deposition was calculated by summing precipita-tion input on a yearly basis. Total deposition is the sum ofwet, dry-, fog- and cloud-water deposition. Dry- and fog-deposition to plant canopies can be estimated via throughfallfluxes corrected for canopy interactions. Throughfall Na+,SO4
2- and Cl- deposition were considered to represent totaldeposition because of their conservative behaviour (Houleet al. 1999). Dry deposition impact factors for Ca, Mg, and
Fig. 1. Location of the three watersheds in Québec, Canada.
OUIMET AND DUCHESNE — FOREST SOIL WEATHERING 249T
able
2. P
rope
rtie
s of
the
rep
rese
ntat
ive
soil
prof
iles
used
for
the
cal
cula
tion
, by
the
PR
OF
ILE
and
PM
B m
etho
ds, o
f m
iner
al w
eath
erin
g ra
tes
in s
oils
of
the
thre
e w
ater
shed
s
LC
W p
rofi
le B
-3L
LW
pro
file
A-3
LT
W p
rofi
le A
–2
FFA
hB
hfB
f1B
f2B
CC
FFA
eB
hfB
f1B
f2B
CC
FFA
eB
f1B
f2B
CC
Thi
ckne
ss (
cm)
54
1019
2022
206
1832
1020
147
518
20D
epth
(cm
)0
414
3353
750
624
5666
860
712
3050
Coa
rse
frag
men
tfr
actio
n >
5 cm
(%
)45
4545
4545
453
33
33
30
1010
1010
10pH
(H2O
)3.
43.
94.
54.
95.
05.
05.
14.
24.
54.
85.
05.
15.
15.
42.
94.
24.
85.
05.
15.
3O
rgan
ic m
atte
r (g
kg–1
)54
283
.018
2.0
84.3
65.8
11.9
6.2
875
37.5
26.1
8.8
6.1
3.5
2.8
837
7.1
26.1
4.2
1.3
0.9
Min
eral
fra
gmen
t fr
actio
n 2–
50 m
m (
%)
33
34
1214
46
2213
512
15
40
4B
ulk
dens
ity<
2 m
m (
kg m
–3)
837
484
761
854
1324
1406
1168
1291
1536
1583
1631
1790
1364
1006
1442
1530
1543
Tot
al c
onte
nt
Mg
(g k
g–1)
1.69
2.89
3.38
4.58
4.95
3.92
6.75
7.72
6.88
6.45
7.54
6.21
2.77
6.82
7.42
6.69
6.88
Ca
(g k
g–1)
10.4
10.9
12.2
14.3
14.7
14.9
20.9
22.8
22.9
19.7
24.9
23.4
16.5
22.4
24.2
24.2
25.5
Na
(g k
g–1)
19.0
17.0
18.4
20.6
22.3
26.0
24.2
21.5
22.5
21.7
24.3
24.2
27.2
27.7
30.9
31.5
30.9
K (
g kg
–1)
27.1
23.5
24.2
25.2
25.8
28.0
29.5
26.7
28.2
28.0
27.6
29.9
12.0
9.5
10.0
10.7
9.8
Ti (
g kg
–1)
7.1
6.4
6.4
5.3
5.3
3.3
11.9
15.8
14.8
15.0
11.5
10.1
2.0
3.1
2.6
2.2
2.2
Zr
(mg
kg–1
)94
370
370
065
751
937
092
696
094
578
010
0075
223
120
219
417
220
6C
l (m
g kg
–1)
5012
080
120
100
110
8011
010
080
110
9060
110
100
100
90T
i/Zr
7.56
9.12
9.16
8.12
10.1
68.
9112
.82
16.4
915
.67
19.2
911
.51
13.3
98.
5615
.14
13.6
012
.90
10.4
8
Min
eral
ogic
al c
ompo
siti
on (
%)
Qua
rtz
59.3
35.8
33.4
39.4
45.4
36.9
52.9
40.2
31.1
27.3
23.5
31.6
35.6
32.6
36.1
39.0
42.5
Pyro
xene
1.0
5.1
5.0
6.0
7.0
3.3
1.4
4.8
1.6
2.7
3.8
2.8
1.4
2.2
2.3
2.0
1.8
Plag
iocl
ase
12.8
28.3
27.4
24.1
20.7
28.4
21.1
25.0
24.3
27.4
30.4
33.9
37.9
40.6
37.5
39.0
36.0
Mus
covi
te0
05.
52.
70
00
00
00
00
00.
50
0.0
Mg-
verm
icul
ite0.
42.
01.
82.
02.
30.
81.
01.
11.
31.
92.
41.
11.
01.
01.
00.
81.
4K
-fel
dspa
rs22
.523
.66.
313
.320
.421
.220
.120
.636
.730
.023
.225
.120
.112
.911
.212
.27.
7K
aolin
ite0
0.4
0.3
0.2
0.2
0.2
00
0.3
0.1
00.
20
0.3
0.3
0.1
0.1
Epi
dote
00.
55.
92.
90
00
00
0.2
0.4
00
00
00.
0C
hlor
ite0.
30.
40.
30.
40.
40.
70.
20.
30.
20.
30.
30.
30.
20.
80.
50.
50.
4B
iotit
e1.
20.
70.
40.
60.
91.
60.
10.
20.
30.
40.
40.
20.
61.
20
1.1
0.8
Apa
tite
00.
30.
40.
30.
20.
30.
10.
20.
10.
61.
10.
10.
10.
10.
10
0.1
Hor
nble
nde
0.7
1.7
3.5
2.7
1.9
2.7
0.9
2.3
1.4
1.7
1.9
1.8
0.9
3.6
7.4
3.2
6.2
250 CANADIAN JOURNAL OF SOIL SCIENCE
K were then estimated using the Na-element ratio methodadapted by Duchesne et al. (2001). Briefly, dry depositionwas calculated by the following equation:
(1)
where P represents deposition (mmol (+) m–2 yr–1), the[X]/[Na] refers to the mean weighted ratio of monthly con-centration (µeq L–1) of corresponding elements, and sub-scripts “bulk”, “dry”, and “throughfall” refer tocorresponding pathways.
For the three watersheds, water fluxes through soils wereestimated using the FORHYM2 model (Arp and Yin 1992).The hydrological model was previously calibrated for LLW(Arp and Yin 1992) and LCW (Houle et al. 2002). Themodel had been validated for the period under study basedon soil moisture and stream flow measurements. Soil watervolume-weighted mean concentration was calculated bycombining average monthly soil water concentrations andmonthly predicted water fluxes below the rooting zone.Annual leaching from solum and from the watershed wasdetermined by summing monthly predicted fluxes andhourly recorded observations, respectively.
Local allometric equations relating stem wood and barkdry biomass to tree height and diameter were constructed foreach site (Tremblay et al., in prep). Using the periodic sur-veys (5-yr intervals over 10 yr) of stand tree height anddiameter distribution, we compiled the net nutrient accumu-lation rate as the difference in amount of element stored inliving trees at the first inventory and the one stored in livingand dead trees at the final inventory. In this study, we con-sidered that the net Na storage rate in biomass was negligi-ble over the computation period.
Methods for Determining Total and WeatheringRelease Rates We determined TR rates from soils and from watershedsand WR rates from soils using the methods described below.
Methods for Determining Total Release RatesInput-output BalanceThe input-output balance (IOB) method can be described bythe following equation:
(2)
where TR(X) is the TR rate of cation X (mmol (+) m–2 yr–1),DX is total cation loss through leaching from soil or streamoutflow, PX is atmospheric cation wet plus dry deposition,and UX is net cation accumulation by vegetation. TR esti-mation for each base cation was determined from volume-weighted annual mean concentrations and water fluxes aspresented in Table 1.
Cl as TracerThe use of chlorine (Cl) as a tracer can provide similar actu-al release estimates as with the input-output balance
approach, but without potential errors caused by unmea-sured hydrological fluxes. The TR of element X [TR(X)]estimated by the Cl tracer (CLT) method is given by [adapt-ed from Lockwood et al. (1995)]:
(3)
where PCl is the total (wet + dry) atmospheric Cl deposition(mmol (+) m–2 yr–1), brackets refer to volume-weighted aver-age annual concentrations, and the subscript l refers to leachingfrom soil or stream outflow. We used throughfall Cl concen-trations (Table 1) to estimate PCl because we found that thiselement behaved conservatively (i.e., with no apparent interac-tion with the ecosystem biochemical and biogeochemicalcycle) in these forest types (Houle et al. 1999). This methodassumes that soil WR rate of Cl = 0; the XRF analyses of totalCl content in the representative soil profiles were measured toverify the validity of this assumption.
MAGICThe MAGIC model describes the long-term impact ofatmospheric deposition and release of cations through soilchemical weathering and cation exchange (Cosby et al.1985). In the model, soil and watershed total release rateswere estimated by iteration. We selected the “two soil hori-zons” (for soil and subsoil) and the “no wetland” options asflow routing parameters. Lake characteristics were added.Deposition history was based on actual average depositiondata (Table 1) scaled to long-term deposition trends innortheastern North America (Bélanger et al. 2002a). Foreach watershed, the model parameters were adjusted inter-actively to match the average water chemistry at solum endand at stream outlet. Data are provided in Table 1. Hence,the values used for model adjustment caused the WR para-meters to be assimilated to TR according to our definitions.The MAGIC model v. 7.77 was used for the calculations.
Methods for Determining Weathering ReleaseRates from Soils The determination of WR rates from soils considered allmineral soil horizons down to the top of the first C horizon(i.e., the solum).
Pedological Mass BalanceThe pedological mass balance (PMB) method consists ofcomparing the depletion in the labile versus non-labile ele-ment ratio between upper weathered soil horizons of a soilprofile and its unweathered C horizon by compiling theirchanges in volume (or height), bulk density, and chemicalcomposition (Brimhall et al. 1991). The following equationwas used [Eq. 14 of Egli and Fitze (2000)] to calculatecation depletion relative to soil total Zr and to total Ti con-centrations, two of the most resistant elements to weatheringin cold climates:
(4)
P P PX
NaX Ca Mg K
X dry Na througfall Na bulkbulk
bulk, , ,– ;
, ,
= ( ) ×[ ][ ]
= ( )
TR X D P UX X X( ) = − +
TR XP X
ClP UCl l
lX X( ) =
[ ][ ]
− +
W X ZX p pi w
X w wi
n
= [ ] × ×+
× ×
=∑0 1
1
11
.,
,ρε
τ ∆
OUIMET AND DUCHESNE — FOREST SOIL WEATHERING 251
where WX is the depletion of element X since deglaciation(kg ha–1), i is the soil horizon number from 1 to n, [X]p is theconcentration (g kg–1) of element X, ρp is the bulk density(kg m–3), εi,w is the strain, i.e., the volumetric changes (unit-less), τX,w is the open-system mass transport function basedon X (unitless), and ∆Zw corresponds to the weathered hori-zon height (cm). The subscripts p and w refer to the regolithand the weathered horizons, respectively. WX was convert-ed to WR by dividing WX by its mass equivalent and by thenumber of years since retreat of the Laurentian Ice Sheet, asassessed by Dyke and Priest (1987) (Table 1).
The resistant elements Zr and Ti were used to calculatestrain (εi,w):
(5)
where [I] refers to concentration of the resistant element (Zror Ti).The open-system mass transport function for element X wasdefined by
(6)
Data from XRF analyses of the representative soil profileswere used to calculate the average WR rates. This calcula-tion method assumes the parent material from which the soilwas derived was homogeneous, and that all released cationswere those exported out of the solum. The PMB gives anaverage of WR over the Holocene.
Compositional Linear TrendWe also applied a new method to estimate WR rates of soilsby calculating their individual compositional trends (CT)(von Eynatten et al. 2003) using elemental compositionsobtained from all soil profiles. A compositional trend isdetermined by an initial composition a and a unitary com-position p, which define the direction of the weatheringtrend (L(a;p)). For each watershed, we used the elementalcomposition of the C horizons (geometric average) as initialcomposition a, assuming that the parent material composi-tion was homogeneous. The linear compositional trend wascalculated by the following formula (von Eynatten et al.2003):
(7)
where v1 is the first unitary eigenvector obtained from non-central principal component analysis of the vector clr(x1) –clr (a),…, clr(xn) – clr(a), where x represents element com-position (Ca, Mg, K, and Na) in the D-4 dimensional space,indices 1 to n represent the horizon number, and clr repre-sents the centered logratio transformation from the simplexspace SD to the real space RD, given by clr(x) = log(x/g(x))for each x in SD. The variable g(x) is the geometric mean ofthe components of the composition x. The inverse x–1 of a
composition x = (x1,…, xD) is given by the compositionC(1/x1,…, 1/xD) where C denotes the closure operator (i.e.,each component of the vector is divided by the sum of all itscomponents). The p vector was estimated, including totalanalyses obtained of H2SO4 digestion from all soil profilesfrom each watershed.
According to the theory, the change of composition isdirectly proportional to the ratio pj/pr, since soil weatheringis assumed to be mostly a linear process. Thus, the releaserate for a given element j through chemical weatheringwould be equal to the inverse of the ratio pj/pNa · Na releasedfrom soils. We used the average Na currently released fromsoils, estimated from the IOB and CLT method, to estimatebase cation weathering rates, assuming Na released = Naweathering (Bailey et al. 2003). The values obtained wouldthus represent chemical weathering averaged over theHolocene, but adjusted for current Na release conditions.
We also estimated the number of perturbations, denominat-ed by k, that the initial composition a has to be perturbed by pto yield a given soil horizon composition as calculated by:
(8)
where i represents the horizon number, and ⟨,⟩ stands for thecommon inner product in RD. The variable k represents anunbiased indicator of the degree of weathering for a soilsample (von Eynatten et al. 2003). The compositional dataused for computation were expressed on a mole of chargebasis.
PROFILEThe PROFILE soil chemistry model (Sverdrup andWarfvinge 1993) provides estimates of the current weather-ing rate from independent data on geochemistry and soilconditions. The model requires climate, soil properties, andvegetation properties as input variables. Many of these vari-ables had been collected for use in the other WR or TR rateestimation methods (Tables 1 and 2). The model was adjust-ed individually for the three representative soil profiles forwhich mineralogical data were available. Surface area wasestimated from bulk density, coarse fragment content, andsoil particle size distribution (Sverdrup and Warfvinge1993). Gibbsite equilibrium coefficients and dissolvedorganic carbon in soil solution were calibrated to fit indi-vidual soil horizon pH. The anorthite composition in the pla-gioclase minerals was calibrated to best fit soil solution Caand Na concentration ratios. The international model pro-vided by ÖKO-DATA (version 4.2, Strausberg, Germany)was used for the computations. The main model assumptionis that ecosystem hydrological and thermal processes werein steady-state.
Statistical and Sensitivity AnalysesTotal base cations (BC) were calculated by summing Ca,Mg, K, and Na. A two-factor analysis of variance withoutreplication was performed on individual and sum of basecation TR and WR, and at the two measuring scales (soiland watershed) separately with watershed and method as
ερ
ρi wp p
w w
I
I, =[ ][ ]
− 1
τρρ
εX ww w
p pi w
X
X, ,=[ ][ ]
+( )
−1 1
p clr v= −11
ki clr x a clr pi= ⊕( ) ( )−1 ,
252 CANADIAN JOURNAL OF SOIL SCIENCE
factors. The analyses are entirely valid on the condition it isassumed that there was no significant interaction effectbetween the two factors (Zar 1974). This assumption waschecked graphically prior to carrying out the analyses.
The use of only one soil profile, albeit representative, forWR calculations with the PMB and PROFILE methodsmade it difficult to compare our analyses with other multi-ple sampling (CT) or watershed-scaled (IOB, CLT,MAGIC) methods. The accuracy of WR rates obtained withthe PROFILE model has been widely studied (Jönsson et al.1995; Hodson et al. 1996, 1997; Zak et al. 1997). In the caseof the PMB method, a sensitivity analysis was performed onthe main variables used to calculate WRBC (Ti and BC con-tent, horizon thickness, bulk density, and time sincedeglaciation) of the representative soil profiles. The range ofthe majority of the main soil variables measured in eachwatershed was ≤ 20%; therefore, an error range of ±20%was assigned to all main variables. A Monte Carlo analysiswas run for each representative watershed soil. Variableswere varied individually in 1000 runs, and simultaneously in5000 runs from a uniform (rectangular) distribution of theerrors. To enable comparison of the results for the threewatershed soils, the deviations at the 10th and 90th per-centiles were standardized.
RESULTS
Soil Chemical Composition and Weathering TrendSoil characteristics at each watershed are presented in Table2. The solum depth-weighted total content of Ca and Mg atLCW was 38% lower than at the two other watersheds(LCW vs. average of LLW and LTW: for Ca, 14 vs. 23 gkg–1, for Mg, 4.2 vs, 6.8 g kg–1, respectively). In LCW soils,quartz and pyroxene were 23 and 140% more abundant,respectively, than in soils of LLW and LTW, while theirplagioclase abundance was similar to the one at LLW(25%). At LTW the average soil plagioclase abundance wasabout 38% higher than for the two other watersheds.However, solum thickness was smaller at LTW by 36%compared to LCW and LLW.
The compositional linear trends explained between 60and 92% of element composition variability in soils withineach watershed (Table 3). The p vectors indicate that Na andK are relatively more recalcitrant elements than are Ca andMg, having the highest values in the three watershed soils.As a result, all ratios pj/pNa will be lower than one, meaningthat the proportion of all elements considered shoulddecrease relative to Na in these soil profiles in the course ofthe weathering process (with the exception of K at LCW).
The estimated number of k perturbations for each soilhorizon in soils from the three watersheds showed adecrease with soil horizon depth (Fig. 2). The k estimatescorrespond to the amount of weathering each horizon isdeemed to have experienced, i.e. the soil horizons shouldfollow the order from the most to the least weathered in rela-tion to soil depth (Ae, Bhf, Bf1, Bf2, and finally BC forPodzols). Such a pattern was observed at the three water-sheds. Eluviated Ae horizons at LLW and LTW are the mostweathered horizons at these watersheds. However, at LCW,
the three Bf horizons display similar or higher k values thantheir superficial Ah horizon. The k value of the BC horizonsin the three watersheds does not differ from the C horizons.
Total Release and Weathering Release RatesThe ANOVA results are presented in Table 4, and factormeans in Table 5. There was no difference in TR rates ofany individual soil cations among the three watersheds (P ≥0.297), with the exception of Ca (P = 0.002). The averagesoil TR and WR rates of Ca were between 35 and 60% high-er at LLW than at the two other watersheds, amounting to102.7 and 48.1 mmol (+) m–2 yr–1, respectively. Calciumwas the predominant released cation in soils at the threewatersheds, accounting for 65% of soil TR rates of BC; Mgrepresented 13%, K 11%, and Na 11%.
There were significant differences among estimates ofsoil TR and WR for Ca, Na, and overall BC. On average,estimated soil WR rates of Ca were lower by at least 25%(average = 48.7%) compared to estimated soil TR rates.Thus, soil WR rates for BC also were lower than soil TRrates. Conversely, the average soil WR rate of Na was 77%higher than the corresponding TR rate at LCW and LLW.
At the watershed level, no difference existed amongwatersheds regarding TR rates of BC (P ≥ 0.107), except forMg (P = 0.026). Watershed TR rates of Mg for LCW wereat least 29.5% lower than for LLW and LTW. WatershedTR rates of BC were 1.9 times higher on average (range:1.51–2.38) than from sola only.
Comparison of MethodsAveraged over the three watersheds, differences existedbetween soil TR and WR rates of individual cations and ofoverall soil BC (P ≤ 0.026; Table 4), with the exception ofK (P = 0.25). At the soil scale, the PMB method, deemed toestimate historical WR rates, tended to yield higher WRrates of Na for the LCW and LLW relative to the PROFILEmethod. The PMB method with Zr as the resistant elementeven yielded negative values for WR rates of cations atLTW. The CT method, also based on the elemental compo-sition of the soils, tended to yield higher WR rates of Mg,but lower WR rates of Ca, relative to the other WR methodsused.
At the watershed level, the IOB, CLT, and MAGIC meth-ods yielded similar estimates of TR rates for the three catch-ments (P ≥ 0.091).
Table 3. Unitary composition p, proportion of total variability (r2) captured by the compositional linear trend of the four components (Ca,Mg, K and Na), and unitary composition p relative to Na
Site nz Ca Mg K Na r2 (%)
LCW 51 0.2339 0.0983 0.3403 0.3276 81.8LLW 15 0.1190 0.1558 0.3401 0.3851 60.2LTW 16 0.2251 0.0999 0.3037 0.3713 91.7
Unitary composition relative to NaLCW 0.7139 0.3000 1.0388LLW 0.3089 0.4045 0.8832LTW 0.6063 0.2691 0.8181zNumber of soil horizon samples used in the computations.
OUIMET AND DUCHESNE — FOREST SOIL WEATHERING 253
DISCUSSION
Composition and Weathering TrendsThe identification of relatively fast-release elements (Ca andMg) is consistent mineralogically with the decomposition ofsilicates (amphiboles, micas, and pyroxenes), while themore recalcitrant Na and K may be mainly found infeldspars, which are relatively more stable to weathering.The WR compositional trends of soils at LCW showed
strong differences compared to LLW and LTW. The soil atLCW is at a more advanced stage of weathering than are thesoils at the other two watersheds. In fact, the LLW and LTWsoils had a higher TR rate of BC per centimetre of solum(>1.6 mmol (+) m–2 yr–1 cm–1 for LLW and LTW; 0.9 mmol(+) cm–2 yr–1 cm–1 for LCW; Table 5).
The differences observed between estimated TR vs. WRrates suggest that some chemical or biological mechanisms ofselective retention or accumulation for some elements in theupper soil profiles play a major role in the release of basecations. For instance, Bailey et al. (2003) suggested that othersources of Ca besides minerals, mainly organic sources, arecontributing to TR of Ca in the Hubbard Brook watershed soils.
Comparison of Base Cation Release Rates forForest SoilsWe compiled forest soil TR or WR published values for arange of sites with similar forest cover, soil properties, andbedrock acidity in North America, calculated averages andconfidence limits, and compared them to the results fromour watershed soils (Table 6). It appears that release rates ofBC in soils at LLW are higher than the averages compiledfrom the literature by 115% (means ± 95% CI: 131.8 ± 24.1vs. 61.2 ± 11.0). Higher soil total Ca content combined withthe presence of a deep solum for this boreal coniferousecosystem (Table 2) may explain the higher soil BC releaserate estimated at LLW. However, at the watershed level, theTR of BC at LLW did not differ from the average for simi-lar watersheds. Meanwhile, the TR rates of K from the threewatersheds under study were about twice as higher as theaverage reported in the literature for watersheds with a sim-ilar granitic environment. Potassium levels, as well as othercations in streams, may be influenced by biotic and abioticfactors, particularly insect outbreaks (Furman et al. 1998)and climatic conditions (Courchesne et al. 2001). For theother cations, soil and watershed release rates at LCW andLTW were within the range of values for forest ecosystemswith similar lithology (Table 6). The soil WR rates estimat-ed in this study for LCW were similar to one published ear-lier by Houle et al. (1997) when the value is adjusted for thedistinct forest net nutrient uptake values used.
Fig. 2. Box plots of the index of relative amount of weathering (k)in the soil profiles of the three watersheds.
Table 4. Results of the two-way ANOVA, testing the significance of dif-ferences between TR and WR rates of soils and of watersheds as afunction of evaluation method and watershed. For soils, TR refers toinput-output (IOB), Cl-tracer (CLT), and MAGIC methods; WRrefers to the pedological mass balance (PMB), PROFILE, and compo-sitional trend (CT) methods. For Na, the CT method was excludedfrom the analysis
P > F
Source of variation Ca Mg K Na BC
SoilsWatershed site 0.002 0.818 0.297 0.548 0.471Method 0.012 0.056 0.250 0.026 0.024TR vs. WR <0.001 0.191 0.081 0.006 0.008
WatershedsSite 0.107 0.026 0.272 0.204 0.174Method 0.143 0.091 0.429 0.230 0.149
254 CANADIAN JOURNAL OF SOIL SCIENCE
The ANOVA (Table 6) produced using site data with agranitic or gneissic bedrock, indicated that, after parent mate-rial type, the evaluation method used to assess cation releasewas the factor most related to the release rate of Ca (P <0.001). The published Ca release rate values showed thegreatest variations among the base cations examined. As datain Table 6 showed that the percentage of release for a givenbase cation relative to total BC release was not related to totalBC release (P ≥ 0.118), individual cation release rates as afunction of BC were simply determined by calculating geo-metric means of their proportions. The results indicated thatpublished Ca, Mg, and K WR rates obtained from Table 6(primarily obtained using the PMB method) were proportion-al to WR rate of BC (Fig. 3). Sodium WR rate, however, wasnot related to WR rate of BC based on these values from theliterature. This observation contradicts the results of the com-positional trend analysis, which showed significant linearweathering trends for the three soils investigated. These con-tradictory findings suggest that (1) other ecosystem processescontributed to WR rates of BC from soils, or (2) non-stoi-chiometric WR of Na vs. other cations from soils. It is likelythat the first hypothesis holds, because it is supported by evi-dence that plants can selectively dissolve minerals throughseveral mechanisms (Bormann et al. 1998; Kelly et al. 1998;Hagerberg et al. 2003) and that considerably higher amountsof Ca, Mg, and K circulate in forest ecosystems through thebiochemical and biogeochemical cycles, and some may belost. Increased losses through soil leaching have been report-ed to occur since the 1960s, due partly to increased atmos-pheric acid deposition (Foster et al. 1992; Houle et al. 1997;Lawrence et al. 1997; Likens et al. 1998; Friedland and Miller
1999; Martin et al. 2000; Bélanger et al. 2002b; Bailey et al.2003; Watmough and Dillon 2003a,b). The second hypothe-sis does not appear to be supported by the significant linearweathering trends observed with the CT method, and withobservations on the Ca/Na ratio range, which can be con-strained in stream water (Bailey et al. 2003).
In the three studied watersheds, WR rates of Ca from soilswere found to be lower than the observed TR by a factor of two.However, large uncertainties are associated with these values(see next section). As already mentioned, the methods we usedfor determining TR rates of BC could not distinguish per sebetween ecosystem processes (e.g., depletion of base cationsfrom the soil exchangeable pool) and true WR from soil min-erals. Other base cation sources may have contributed substan-tially to release of soil base cations, including vegetationby-products [e.g., calcium oxalate crystals as suggested byBailey et al. (2003)], capillary rise cycles in the soil, and short-and long-term variation in the decomposition rates of organicmatter and saturation of the exchangeable soil complex. TheTR rates of BC values presented in this paper thus represent theoverall contribution of these processes.
The PMB and CT methods assumed a linear trend, eitherin time or in soil composition, to assess WR of base cationsfor the Holocene. The CT analysis showed a significant lin-ear trend in compositional changes (Table 3), and, moreglobally, such a linear trend appears to provide a good esti-mate of the chemical weathering in the earth’s upper crust(von Eynatten et al. 2003). However, although soil compo-sition can follow a chemical linear trend, changes in soilporosity, permeability, and production of secondary miner-al phases through soil development can alter the rate over
Table 5. Total release (TR) rates of base cations from soils and from watersheds, and mineral weathering (WR) rates from soils, according to dif-ferent evaluation methods
Soil TR rate (mmol (+) m–2 yr–1) Soil WR rate (mmol (+) m–2 yr–1) Watershed TR rate (mmol (+) m–2 yr–1)
Input- Mean Mean Input- Meanoutput release weathering output release
Site balance Cl tracer MAGIC (SD) Zr PMB Ti PMB PROFILE CT (SD) balance Cl tracer MAGIC (SD)
LCWCa 45.8 39.4 37.7 41.0 (4.3) 36.3 38.1 34.2 14.3 30.7 (11.0) 107.2 97.9 86.6 97.2 (10.3)Mg 9.8 8.1 7.8 8.6 (1.1) 12.3 13.2 25.1 34.0 21.2 (10.4) 33.1 29.9 25.6 29.5 (3.8)K 10.4 9.7 10.2 10.1 (0.4) 35.2 37.0 4.4 9.8 21.6 (16.9) 15.1 14.3 14.6 14.7 (0.4)Na 13.0 8.4 9.2 10.2 (2.5) 63.8 66.6 6.1 – 45.5 (34.2) 28.0 24.6 23.5 25.4 (2.3)BC 79.0 65.6 64.9 69.8 (7.9) 147.6 154.9 69.8 68.3 110.2 (47.6) 183.4 166.7 150.3 166.8 (16.6)
LLWCa 99.0 106.5 102.7 102.7 (3.8) 37.1 56.3 69.8 29.1 48.1 (18.4) 124.6 157.9 102.7 128.4 (17.8)Mg 20.7 22.6 11.6 18.3 (5.9) 6.0 14.4 33.4 22.2 19.0 (11.6) 36.1 48.8 29.6 38.2 (9.8)K 11.8 12.1 10.8 11.6 (0.7) 28.3 40.9 27.3 10.2 26.7 (12.6) 13.8 15.7 13.7 14.4 (1.1)Na 9.2 11.1 6.7 9.0 (2.2) 37.8 55.2 20.7 – 37.9 (17.2) 29.8 44.9 22.0 32.2 (11.6)BC 140.6 152.2 131.8 141.5 (10.2) 109.2 166.8 151.2 70.6 124.5 (43.4) 204.3 267.4 168.0 213.2 (50.2)
LTWCa 66.7 66.6 66.4 66.6 (0.2) –z 22.5 17.4 22.8 20.9 (3.0) 118.2 124.9 108.9 117.1 (8.9)Mg 14.8 14.7 14.1 14.5 (0.4) – 8.4 4.4 51.3 21.3 (26.0) 46.9 50.8 41.6 46.4 (5.2)K 10.9 10.9 11.0 10.9 (0.1) – 0.8 1.2 16.9 6.3 (9.2) 13.6 14.1 13.3 13.6 (0.5)Na 13.9 13.7 13.9 13.8 (0.1) – 14.5 2.9 – 8.7 (8.2) 21.9 24.2 18.1 21.5 (3.5)BC 106.3 105.8 105.4 105.8 (0.5) – 46.1 25.9 104.7 58.9 (40.9) 200.7 214.0 181.9 198.5 (17.9)
BC mean 108.6 107.9 100.7 105.7 128.4 122.6 82.3 81.2 101.4 194.6 217.2 166.7 192.8(SD) (30.9) (43.4) (33.7) (31.7) (27.1) (66.5) (63.6) (20.4) (48.6) (10.5) (50.3) (15.8) (34.7)zThe method yielded negative values.
OUIMET AND DUCHESNE — FOREST SOIL WEATHERING 255T
able
6. T
otal
rel
ease
or
min
eral
wea
ther
ing
rate
s of
soi
ls a
nd o
f w
ater
shed
s fo
r fo
rest
eco
syst
ems
on a
cid
subs
trat
e lo
cate
d m
ainl
y in
nor
thea
ster
n N
orth
Am
eric
a
Wea
ther
ing
estim
atio
n W
eath
erin
g ra
te (
mm
ol(+
)m–2
yr–1
)
Site
Soi
lsFo
rest
type
Pare
nt m
ater
ial
Geo
logy
Soil
text
ure
met
hodz
Ca
Mg
KN
aB
CR
efer
ence
Plai
nfie
ld, W
IR
ed p
ine
plan
tatio
nA
cid
glac
ial o
utw
ash
Gra
nit
Loa
my
sand
110
9.8
41.1
17.6
––
(Bok
heim
et a
l. 19
83)
Whi
tefa
ce M
tn.,
NY
Spru
ce-f
irG
laci
al ti
llA
nort
hosi
teSa
ndy
loam
142
.921
.414
.6–
–(F
ried
land
and
Mill
er 1
999)
126
0.0
50.0
50.0
––
(Fos
ter
et a
l. 19
92)
Cal
houn
Exp
. For
., SC
Lob
lolly
pin
eR
esid
ual
Mix
ed G
rani
teSa
ndy
loam
1≈
017
.530
.7–
–(R
icht
er e
t al.
1994
)pl
anta
tion
and
gnei
ss
Tur
key
Lak
es, O
NN
orth
ern
hard
woo
dG
laci
al ti
llM
ixed
gra
nite
Sa
ndy
loam
129
0.0
70.0
60.0
––
(Fos
ter
et a
l. 19
92)
and
basa
lt7
100.
0(M
oaye
ri e
t al.
2001
)
Hub
bard
Bro
ok, N
HN
orth
ern
hard
woo
dG
laci
al ti
llG
neis
s, g
rani
te,
Sand
y lo
am1,
2,7
17.4
12.5
1.4
23.5
54.8
(Gbo
ndo-
Tug
baw
a et
al.
2001
);
and
schi
sts
(Joh
nson
et a
l. 19
68)
213
.0–
––
90.0
(Lik
ens
et a
l. 19
98);
(Joh
nson
et a
l. 19
68)
65.
08.
27.
726
.147
.0(A
dam
s an
d B
oyle
197
9)
Rou
ndto
p M
tn.,
QC
Spru
ce-f
irG
laci
al ti
llL
oam
6–
––
–26
.3(M
cCou
rt a
nd H
ende
rsho
t 199
2)
Clo
quet
Exp
. For
.A
spen
Fluv
io-g
laci
al
Mix
edSa
ndy
loam
1,3,
6,7
26.0
16.5
4.5
––
(Kol
ka e
t al.,
199
6)(C
loqu
et),
MN
outw
ash
Clo
quet
Exp
. For
. Fl
uvio
-gla
cial
M
ixed
Sand
y lo
am21
.718
.65.
7–
–(O
meg
a), M
Nou
twas
hPi
ke B
ay E
xp. F
or.
Gla
cial
till
Mix
edSa
ndy
loam
27.8
10.7
4.7
––
(Ont
onag
on),
MN
Bru
le r
iver
Sta
te F
or.
Gla
cio-
lacu
stri
ne ti
llM
ixed
Cla
y19
.213
.74.
5–
–(S
olon
Spr
ing)
, WI
Bru
le r
iver
Sta
te F
or.
Fluv
io-g
laci
al
Mix
edL
oam
y sa
nd9.
78.
12.
2–
–(W
arba
), W
Iou
twas
h
Her
min
e E
xp.
Nor
ther
n ha
rdw
ood
Gla
cial
till
Ano
rtho
site
Sand
y lo
am3
6.6
4.6
3.6
4.2
19.0
(Cou
rche
sne
et a
l. 20
02)
Wat
ersh
ed, Q
CR
ouyn
-Nor
anda
, QC
Asp
enG
laci
o-la
cust
rine
Mix
edC
lay
8.5
7.5
1.5
10.0
27.5
Plas
tic L
ake,
ON
Nor
ther
n ha
rdw
ood
Gla
cial
till
Gra
nitic
gne
iss
Sand
y lo
am3
13.3
7.9
7.2
14.0
42.4
(Kir
kwoo
d an
d N
esbi
tt 19
91)
RE
SEF,
QC
(n
= 3
)N
orth
ern
hard
woo
dG
laci
al ti
llSh
ale
Loa
m32
.122
.562
.67.
112
4.3
(Oui
met
et a
l. 20
01);
R
. Oui
met
, unp
ubl.
data
RE
SEF,
QC
(n
= 5
)N
orth
ern
hard
woo
dG
laci
al ti
llG
rani
te a
nd g
neis
sSa
ndy
loam
32.6
16.5
2.2
3.5
54.7
RE
SEF,
QC
(n
= 1
)B
alsa
m f
irG
laci
al ti
llSa
ndst
one
Sand
y lo
am13
.511
.614
.91.
341
.3R
ESE
F, Q
C (
n =
1)
Bal
sam
fir
Gla
cial
till
Gne
iss
Loa
m21
.27.
42.
22.
233
.0
Ada
ms,
NY
Nor
ther
n ha
rdw
ood
Fluv
io-g
laci
al
Gra
nite
and
gne
iss
Sand
y lo
am3
24.1
11.7
14.9
25.0
75.7
(Jer
sak
et a
l. 19
95)
outw
ash
Bec
ket,
NY
Mix
ed f
ores
tG
laci
al ti
llG
rani
te a
nd g
neis
sSa
ndy
loam
7.1
3.6
4.3
10.9
25.9
Can
adia
n sh
ield
site
s,
Mos
tly n
orth
ern
Gla
cial
till
Gra
nite
and
gne
iss
Sand
y lo
am7
––
––
23.0
(Arp
et a
l. 19
96)
ON
(n
= 5
)ha
rdw
ood
256 CANADIAN JOURNAL OF SOIL SCIENCET
able
6. c
onti
nued
Inte
rmed
iate
aci
d-M
ostly
nor
ther
nG
laci
al ti
ll,
Mix
ed ig
neou
s
Loa
m–
––
–10
9.2
sens
itive
site
s,
hard
woo
deo
lian
sand
and
sedi
men
tary
ON
(n
= 6
)
Pond
Bra
nch
Wat
ersh
ed, M
DH
ardw
ood
Wea
ther
ed b
edro
ck
Schi
st a
nd q
uart
zL
oam
1,3
6.4
9.4
5.9
11.4
33.1
(Cle
aves
et a
l. 19
70)
Pant
her
Lak
e, N
YG
laci
al ti
llG
rani
tic g
neis
sSa
ndy
loam
310
.08.
415
.016
.549
.9(A
pril
et a
l. 19
86)
Woo
ds L
ake,
NY
Gla
cial
till
Gra
nitic
gne
iss
Sand
y lo
am3
17.8
12.2
12.9
18.9
61.8
(Bai
ley
et a
l. 19
96)
Con
e Po
nd, N
HM
ixed
con
ifer
ous
Gla
cial
till
Met
ased
imen
tary
Sa
ndy
loam
46.
0sc
hist
s an
d qu
artz
ites
110
.73
5.5
4.5
2.8
13.0
25.7
(Hym
an e
t al.
1998
)
Dor
set,
ON
(n
= 7
)N
orth
ern
Har
dwoo
dG
laci
al ti
llG
rani
tic g
neis
sSa
ndy
loam
334
.124
.68.
219
.386
.2(W
atm
ough
and
Dill
on 2
003a
)
Ave
rage
and
95%
con
fide
nce
limits
33.8
16
.813
.011
.161
.2±
16.3
±4.
2±
5.6
±3.
0±
11.0
Lak
e C
lair
Nor
ther
n H
ardw
ood
Gla
cial
till
Gra
nitic
gne
iss
Sand
y lo
am1,
2,3,
735
.1
15.8
16.7
27.9
92.9
Thi
s st
udy
Wat
ersh
ed, Q
C±
7.3
±7.
4±
10.0
±23
.2±
29.8
Mea
ns ±
95%
con
fide
nce
limits
Lak
e L
afla
mm
e
Bal
sam
fir
Gla
cial
till
Cha
rnoc
kitic
gne
iss
Sand
y lo
am1,
2,3,
771
.5
18.7
20.2
23.5
131.
81W
ater
shed
, QC
±23
.8±
6.6
±8.
9±
15.4
±24
.1L
ake
Tir
asse
Bla
ck s
pruc
eG
laci
al ti
llG
rani
tic g
neis
sL
oam
y sa
nd1,
2,3,
743
.7
18.0
8.6
11.8
82.4
Wat
ersh
ed, Q
C±
20.1
±13
.5±
5.1
±4.
4±
29.2
Wat
ersh
eds
Ave
rage
rel
ease
fro
m 7
4 w
ater
shed
s w
ith g
rani
tic li
thol
ogyy
82.8
50
.87.
844
.786
.0(O
liva
et a
l. 20
03)
±24
.6±
17.0
±2.
2±
12.8
1±
49.9
Mea
ns ±
95%
con
fide
nce
limits
Lak
e C
lair
H
ardw
ood
Gla
cial
till
Gra
nitic
gne
iss
Sand
y lo
am1,
2,7
97.2
29.5
14.7
25.4
166.
8T
his
stud
yW
ater
shed
, QC
±10
.1±
3.7
±0.
4±
2.3
±16
.3M
eans
±95
% c
onfi
denc
e lim
itsL
ake
Laf
lam
me
Bal
sam
fir
Gla
cial
till
Cha
rnoc
kitic
Sa
ndy
loam
1,2,
712
8.4
38.2
14.4
32
.2
213.
2 W
ater
shed
, QC
gnei
ss±
17.4
±9.
6±
1.1
±11
.4±
49.2
Lak
e T
iras
seB
lack
spr
uce
Gla
cial
till
Gra
nitic
gne
iss
Loa
my
sand
1,2,
711
7.1
46.4
13.6
21.5
198.
5W
ater
shed
, QC
±8.
7±
5.1
±0.
5±
3.4
±17
.5z R
elea
se e
stim
atio
n m
etho
d: 1
Inp
ut-o
utpu
t m
ass
bala
nce;
2 I
ndic
ator
ele
men
ts (
Na,
Si,
Cl)
; 3
Pedo
logi
cal
mas
s ba
lanc
e (P
MB
); 4
Str
ontiu
m i
soto
pes;
5 W
eath
erin
g ba
gs;
6 L
abor
ator
y ex
peri
men
ts;
7M
odel
s.
OUIMET AND DUCHESNE — FOREST SOIL WEATHERING 257
time. Studies of soil chronosequences have shown thatchemical weathering rates in soils is a non-linear processover time (Bain et al. 1993; Taylor and Blum 1995; Hodsonand Langan 1999a). The soil BC weathering rate is not con-stant and decreases with soil age. However, individual base
cation weathering rates measured by the PMB method donot seem to follow a similar decay rate or trend (Taylor andBlum 1995; Lichter 1998). As a result, soil weathering val-ues obtained by averaging over the Holocene period shouldoverestimate actual soil weathering rates, and the resultsshowed that this would be the case for K and Na in two ofthe three watershed soils.
The unreasonable values obtained at LTW for the PMBmethod with Zr as the resistant element also indicate thatmuch care is needed when employing this method. Althoughthe analysis was based on only one soil profile, close exam-ination of the data for each horizon indicated base cationdepletion in the first two mineral horizons, but accumulationin the next two (Bf2 and BC), thus yielding an overall netaccumulation of base cations (i.e., negative WR rate) in thesoil rather than a depletion. Also, calculating weatheringrates over a single time interval is problematic as discussedabove. In comparing four different methods for estimatingmineral WR of five soils (Kolka et al. 1996), the results indi-cated that the methods, including PMB, yielded individualbase cation weathering estimates that differed from soil tosoil. This finding is corroborated by other studies (Bain andLangan 1995; Hodson and Langan 1999b).
The IOB and CLT methods yielded very similar results,since only the drainage calculation procedure differedbetween the two methods. The depth-weighted average Clconcentrations in soils were sufficiently low that the releaseof Cl through weathering could be considered negligible ascompared to the deposition input. For instance, at the LCW,the WR rate of Ca in the soil was 30.7 ± 11 mmol (+) m–2
yr–1 (Table 5) and its depth-weighted solum content was13.24 g kg–1; this means that the soil weathered at a rate of
Fig. 3. Individual base cation WR rates (mostly obtained by thePMB method) as a function of total soil base cation (BC) WR rate.Data from Table 6.
Fig. 4. Distribution of the calculated WR rate of total base cations(BC) using the PMB method in the three watershed soils resultingfrom a simultaneous 20% variation in the variables (Ti content,individual base cation content, soil horizon thickness, bulk densi-ty, and time since deglaciation). The figure shows the average(filled circle in box), the median (line within box), the Q1 and Q3(lower and upper side of box), the 10th and 90th percentile (lowerand upper whisker), and the 5th and 95th percentile (filled circle atthe whisker ends).
258 CANADIAN JOURNAL OF SOIL SCIENCE
46.5 ± 16.65 g soil m–2 yr–1. Given a depth-weighted Cl con-tent of 100.3 mg kg–1 (Table 2), the soil weathering rate ofCl would be 0.131 ± 0.047 mmol (+) m–2 yr–1. This valuerepresents only 3% of the Cl flux brought through atmos-pheric deposition (4.79 mmol (+) m–2 yr–1).
The PROFILE and MAGIC models are widely used tosimulate ecosystem fluxes. Both models yielded soil releaserate values for all elements similar to the other correspond-ing WR and TR methods. In the MAGIC model, individualcations were added as a source of element to match soilsolution and stream chemistry at equilibrium. Therefore,MAGIC could not differentiate among other potentialsources of elements. It appears that both short-term(Bormann et al. 1998) and long-term (Likens et al. 1998)soil studies are needed to better understand the role and con-tribution of chemical weathering in the supply of elementsto ecosystems.
Uncertainties Associated with WR EstimationSoil mineralogy was determined from only one soil profileby watershed; therefore, the mineral proportions could havediffered substantially. Zak et al. (1997) showed that the cal-culated WR rate from PROFILE was mostly sensitive to soilmineralogy and other physical parameters (soil moisturecontent, exposed mineral surface area, and soil bulk densi-ty). According to Jönsson et al. (1995), the uncertainty asso-ciated with the PROFILE estimation of WR for threePodzols (similar to this study) was ±40%. Most methods forWR determination are accurate at approximately ±50%according to Hodson and Langan (1999b). Assuming theerror associated with the WR estimates obtained from PRO-FILE for the three watershed soils are of the same magni-tude, this means that the WR estimation of BC was not
different from the TR estimation of BC for LCW and LLW,with the exception of LTW, for which WR rates of BCremained about a third of its TR estimation.
To our knowledge, the uncertainty associated with thePMB method, which is widely used to estimate averages ofWR rates of soils over the Holocene, has not been reported.Therefore, we simulated the error associated with a 20%change in variables used in the PMB calculations. The sen-sitivity analysis indicated that the CVs of the WR rates asso-ciated with the assumed 20% uncertainty range for allvariables were 22, 33, and 73% for the LCW, LLW, andLTW soils, respectively (Fig. 4). The 10th to the 90th pre-diction intervals suggest that among the main variables stud-ied, the resistant element content in soils has the mostinfluence on the WR calculation of BC with this method(average error range: ±25% for LCW to ±87% for LTW),followed by the base cation contents (error range: ±9% forLCW to ±47% for LTW) (Fig. 5). A 20% variation in theother variables (solum and individual horizon thickness,bulk density, and time since deglaciation) yielded predictionintervals of less than 10% of the mean values. Given thesevariations, the PMB estimates of BC for LCW and LLWsoils are not different from their corresponding average TRestimates. While for LTW soils, WR of BC calculated withthe PMB method remains significantly lower than its aver-age TR estimate (5–100 vs. 105.8 mmol (+) m–2 yr–1,respectively). The error associated with WR rates using thePMB method may, however, be smaller than the simulationresults presented here, as resistant element content is corre-lated with K or Na content in the soils being studied (rTi-K =0.73***, rZr-K = 0.87***, rZr-Na = –0.75***) and since indi-vidual base cation content is also intercorrelated (rCa-Mg =0.92***, rCa-Na = 0.70***, rK-Na = –0.73***).
Fig. 5. Relative prediction intervals of calculated WR rates of total base cations (BC) using the PMB method in the three watershed soilsresulting from a 20% variation in individual variables and simultaneously in all variables. In the case of individual base cation content, soilCa, Mg, K and Na content were varied simultaneously
OUIMET AND DUCHESNE — FOREST SOIL WEATHERING 259
CONCLUSIONUsing various methods to assess base cation total releaseand chemical weathering rates, and their associated uncer-tainties, of three Podzolic soils in distinct forest ecosystemsindicated little overall difference between total release rateand weathering rate for assessing Ca, Na, and total BCmobilization rates from soils, with the exception of the LTWsoils. For soil Mg and K there was clearly no significant dif-ference among weathering or total release rates. The weath-ering trend within a soil profile could be explained in greatpart by a linear change in composition of Ca, Mg, K, and Naof the solum horizons relative to the C horizon. Mineralweathering rates of soils remains a process that is difficult toassess. Thus, in steady-state conditions, total net releaserates of base cations can be viewed at best as maximum val-ues for mineral weathering rates.
ACKNOWLEDGEMENTSThis study was supported by the Ministère des Ressourcesnaturelles, de la Faune et des Parcs du Québec (project no.0200-3053). We would like to thank J. Gagné, M. Saint-Germain, and the Saint-Félicien management unit personnelfor having performed the field work to monitor the watersheds.
Adams, P. and Boyle, J. R. 1979. Cation release from MichiganSpodosols leached with aspen leaf extracts. Soil Sci. Soc. Am. J.43: 593–596.April, R., Newton, R. and Coles, T. 1986. Chemical weatheringin two Adirondack watersheds: past and present day rates. Geol.Soc. Am. Bull. 97: 1232–1238.Arp, P. A. and Yin, X. 1992. Predicting water fluxes throughforests from monthly precipitation and mean monthly air tempera-ture records. Can. J. For. Res. 22: 864–877.Arp, P. A., Oja, T. and Marsh, M. 1996. Calculating critical Sand N loads and current exceedances for upland forests in southernOntario, Canada. Can. J. For. Res. 26: 696–709.Bailey, S. W., Buso, D. and Likens, G. 2003. Implications ofsodium mass balance for interpreting the calcium cycle of a forest-ed ecosystem. Ecology 84: 471–484.Bailey, S. W., Hornbeck, J. W., Driscoll, C. T. and Gaudette, H.E. 1996. Calcium inputs and transport in a base-poor forest ecosys-tem as interpreted by Sr isotopes. Water Resour. Res. 32: 707–719.Bain, D. C. and Langan, S. J. 1995. Weathering rates in catch-ments calculated by different methods and their relationship toacidic inputs. Water Air Soil Pollut. 85: 1051–1056.Bain, D. C., Mellor, A., Robertson-Rintoul, M. S. E. andBuckland, S. T. 1993. Variations in weathering processes andrates with time in a chronosequence of soils from Glen Feshie,Scotland. Geoderma 57: 275–293.Bélanger, N., Côté, B., Courchesne, F., Fyles, J. W., Warfvinge,P. and Hendershot, W. H. 2002a. Simulation of soil chemistryand nutrient availability in a forested ecosystem of southernQuebec. I. Reconstruction of time-series files of nutrient cyclingusing the MAKEDEP model. Environ. Model. Soft. 17: 427–445.Bélanger, N., Courchesne, F., Côté, B., Fyles, J. W.,Warfvingue, P. and Hendershot, W. H. 2002b. Simulation ofsoil chemistry and nutrient availability in a forested ecosystem ofsouthern Quebec. Part II. Application of the SAFE model. Environ.Mod. Soft. 17: 447–465.Bokheim, J. G., Lee, S. W. and Leide, J. E. 1983. Distributionand cycling of elements in a Pinus resinosa plantation ecosystem.Can. J. For. Res. 13: 609–619.
Bondietti, E. A., Momoshima, N., Shortle, W. C. and Smith, K.T. 1990. A historical perspective on divalent cation trends in redspruce stemwood and the hypothetical relationship to acidic depo-sition. Can. J. For. Res. 20: 1850–1858.Bormann, B., Wang, D., Bormann, F., Benoit, G., April, R. andSnyder, M. 1998. Rapid, plant-induced weathering in an aggrad-ing experimental ecosystem. Biogeochemistry 43: 129–155.Brimhall, G. H., Lewis, C. J., Ford, C., Bratt, J., Taylor, G. andWarin, O. 1991. Quantitative geochemical approach to pedogene-sis: importance of parent material reduction, volumetric expansion,and eolian influx in lateritization. Geoderma 51: 51–91.Cleaves, E. T., Godfrey, A. E. and Bricker, O. P. 1970.Geochemical balance of a small watershed and its geomorphicimplications. Geol. Soc. Am. Bull. 81: 3015–3032.Cosby, B. J., Hornberger, G. M., Galloway, J. N. and Wright,R. 1985. Modeling the effects of acid deposition: assessment of alumped-parameter model of soil water and stream water chemistry.Water Resour. Res. 21: 51–63.Courchesne, F., Roy, G. A., Biron, P. M., Côté, B., Fyles, J. W.and Hendershot, W. H. 2001. Fluctuations of climatic conditions,elemental cycling and forest growth at the watershed scale.Environ. Monitor. Assess. 67: 161–177.Courchesne, F., Hallé, J.-P. and Turmel, M.-C. 2002. Bilansélémentaires holocènes et altération des minéraux dans les solsforestiers du Québec méridional. Géogr. Phys. Quatern. 56: 5–17.Driscoll, C. T., Lawrence, G. B., Bulger, A. J., Butler, T. J.,Cronan, C. S., Eagar, C., Lambert, K. F., Likens, G. E.,Stoddard, J. L. and Weathers, K. C. 2001. Acidic deposition inthe northeastern United States: Sources and inputs, ecosystemseffects, and management strategies. BioScience 51: 180–198.Duchesne, L., Ouimet, R. and Houle, D. 2002. Basal area growthof Acer saccharum in relation to acid deposition, stand health andsoil nutrients. J. Environ. Qual. 31: 1676–1683.Duchesne, L., Ouimet, R., Camiré, C. and Houle, D. 2001.Seasonal nutrient transfers by foliar resorption, leaching, and litterfall in a northern hardwood forest at Lake Clair Watershed,Quebec, Canada. Can. J. For. Res. 31: 333–344.Dyke, A. S. and Priest, V. K. 1987. Late Wisconsin and Holocenehistory of the Laurentide ice sheet. Géogr. Phys. Quatern. 41:237–263.Egli, M. and Fitze, P. 2000. Formulation of pedologic mass balancebased on immobile elements: a revision. Soil Sci. 165: 437–443.Foster, N. W., Mitchell, M. J., Morrison, I. K. and Shepard, J.P. 1992. Cycling of acid and base cations in deciduous stands ofHuntington Forest, New York, and Turkey Lakes, Ontario. Can. J.For. Res. 22: 167–174.Friedland, A. and Miller, E. 1999. Major-element cycling in ahigh-elevation Adirondack forest: Patterns and changes,1986–1996. Ecol. Appl. 9: 958–967.Furman, T., Thompson, P. and Hatchl, B. 1998. Primary miner-al weathering in the central Appalachians: A mass balanceapproach. Geochim. Cosmochim. Acta 62: 2889–2904.Gallardo, J. F., Saavedra, J., Martin-Patino, T. and Millen, A.1987. Soil organic matter determination. Commun. Soil Sci. PlantAnal. 18: 69–707.Gbondo-Tugbawa, S. S., Driscoll, C. T., Aber, J. D. and Likens,G. E. 2001. The evaluation of an integrated biogeochemical model(PnET-BGC) at a northern hardwood forest ecosystem. WaterResour. Res. 35: 1057–1070.Hagerberg, D., Thelin, G. and Wallander, H. 2003. The productionof ectomycorrhizal mycelium in forests: Relation between forestnutrient status and local mineral sources. Plant Soil 252: 279–290.Hodson, M. and Langan, S. J. 1999a. The influence of soil age oncalculated mineral weathering rates. Appl. Geochem. 14: 387–394.
260 CANADIAN JOURNAL OF SOIL SCIENCE
Hodson, M. and Langan, S. J. 1999b. Considerations of uncer-tainty in setting critical loads of acidity of soils: the role of weath-ering rate determination. Environ. Pollut. 106: 73–81.Hodson, M. E., Langan, S. J. and Wilson, M. J. 1996. A sensi-tivity analysis of the PROFILE model in relation to the calculationof soil weathering rates. Appl. Geochem. 11: 835–844.Hodson, M. E., Langan, S. J. and Wilson, M. J. 1997. A criticalevaluation of the use of the PROFILE model in calculating miner-al weathering rates. Water Air Soil Pollut. 98: 79–104.Horsley, S. B., Long, R. P., Bailey, S. W., Hallett, R. A. and Hall,T. J. 2000. Factors associated with the decline disease of sugar mapleon the Allegheny Plateau. Can. J. For. Res. 30: 1365–1378.Houle, D., Duchesne, L., Ouimet, R., Paquin, R., Meng, F. andArp, P. A. 2002. Evaluation of the FORHYM2 model for predictionof hydrologic fluxes and soil temperature at the Lake Clair Watershed(Duchesnay, Quebec). For. Ecol. Manage. 159: 249–260.Houle, D., Paquin, R., Camiré, C., Ouimet, R. and Duchesne,L. 1997. Response of the Lake Clair Watershed (Duchesnay,Québec) to changes in precipitation chemistry (1988–1994). Can.J. For. Res. 17: 1813–1821.Houle, D., Ouimet, R., Paquin, R. and Laflamme, J.-G. 1999.Interactions of atmospheric deposition with a mixed hardwood andconiferous forest canopy at the Lake Clair Watershed (Duchesnay,Quebec). Can. J. For. Res. 29: 1944–1957.Hyman, M. E., Johnson, C. E., Bailey, S. W., Hornbeck, J. W.and April, R. H. 1998. Chemical weathering and cation loss in abase-poor watershed. Geol. Soc. Am. Bull. 110: 85–95.Jersak, J., Amundson, R. and Brimhall, Jr., G. 1995. A massbalance analysis of podzolization: examples from the northeasternUnited States. Geoderma 66: 15–42.Johnson, A. H., Andersen, S. B. and Siccama, T. . 1994. Acidrain and soils of the Adirondacks. 1. Changes in pH and availablecalcium, 1930–1984. Can. J. For. Res. 24: 39–45.Johnson, N. M., Likens, G. E., Bormann, F. H. and Pierce, R.S. 1968. Rate of chemical weathering of silicate minerals in NewHampshire. Geochim. Cosmochim Acta 32: 531–545.Jönsson, C., Warfvinge, P. and Sverdrup, H. 1995. Uncertaintyin predicting weathering rate and environmental stress factors withthe PROFILE model. Water Air Soil Pollut. 81: 1–23.Kelly, E. F., Chadwick, O. A. and Hilinski, T. E. 1998. The effectsof plants on mineral weathering. Biogeochemistry 42: 21–53.Kirkwood, D. E. and Nesbitt, H. W. 1991. Formation and evolu-tion of soils from an acidified forest watershed: Plastic Lake,Ontario, Canada. Geochim. Cosmochim. Acta 55: 1295–1308.Kolka, R. K., Grigal, D. F. and Nater, E. A. 1996. Forest mineralweathering rates: use of multiple approaches. Geoderma 73: 1–21.Lawrence, G. B., David, M. B., Bailey, S. W. and Shortle, W. C.1997. Assessment of calcium status in soils of red spruce forests inthe northeastern United States. Biogeochemistry 38: 19–39.Lichter, J. 1998. Rates of weathering and chemical depletion insoils across a chronosequence of Lake Michigan sand dunes.Geoderma 85: 255–282.Likens, G. E., Driscoll, C. T., Buso, D. C., Siccama, T. G.,Johnson, C. E., Lovett, G. M., Fahey, T. J., Reiners, W. A.,Ryan, D. F., Martin, C. W. and Bailey, S. W. 1998. The biogeo-chemistry of calcium at Hubbard Brook. Biogeochemistry 41:89–173.Likens, G. E. and Bormann, F. H. 1995. Biogeochemistry of aforested ecosystem. 2nd ed. Springer-Verlag, New York, NY. Lockwood, P. V., McGarity, J. W. and Charley, J. L. 1995.Measurement of chemical weathering rates using natural chlorideas a tracer. Geoderma 64: 215–232.Martin, C. W., Driscoll, C. T. and Fahey, T. J. 2000. Changes instreamwater chemistry after 20 years from forested watersheds inNew Hampshire, U.S.A. Can. J. For. Res. 30: 1206–1213.
McCourt, G. H. and Hendershot, W. H. 1992. A new method fordetermining mineral weathering rates in soils. Commun. Soil Sci.Plant Anal. 23: 939–952.Miller, E. K., Blum, J. D. and Friedland, A. J. 1993.Determination of soil exchangeable-cation loss and weatheringrates using Sr isotopes. Nature 362: 438–441.Moayeri, M., Meng, F.-R., Arp, P. A. and Foster, N. W. 2001.Evaluating critical soil acidification loads and exceedances for adeciduous forest at the Turkey Lakes watershed. Ecosystems 4:555–567.Oliva, P., Viers, J. and Dupré, B. 2003. Chemical weathering ingranitic environments. Chem. Geol. 202: 225–256.Ouimet, R. and Camiré, C. 1995. Foliar deficiencies of sugarmaple stands associated with soil cation imbalances in the QuebecAppalachians. Can. J. Soil Sci. 75: 169–175.Ouimet, R., Duchesne, L., Houle, D. and Arp, P.A. 2001.Critical loads of atmospheric S and N deposition and currentexceedances for Northern temperate and boreal forests in Quebec.Water Air Soil Pollut. Focus 1: 119–134.Richter, D. D. and Markewitz, D. 2001. Understanding soilchange. Cambridge University Press, Cambridge, UK.Richter, D. D., Markewitz, D., Wells, C. G., Allen, H. L., April,R., Heine, P. R. and Urrego, B. 1994. Soil chemical change dur-ing three decades in an old-field loblolly pine (Pinus taeda L.)ecosystem. Ecology 75: 1463–1473.Sharpe, W. E., Swistock, B. R. and Sunderland, T. L. 1999. Soilacidification and sugar maple decline in northern Pennsylvania.Pages 191–197 in W. E. Sharpe and J. R. Drohan, eds. The effectsof acidic deposition on Pennsylvania’s forests. Vol. 1.Environmental Resources Research Institute, University Park, PA. Soil Classification Working Group. 1998. The Canadian systemof soil classification. 3rd ed. Agriculture and Agri-Food Canada,Ottawa, ON.Sverdrup, H. and Warfvinge, P. 1993. Calculating field weath-ering rates using a mechanistic geochemical model PROFILE.Appl. Geochem. 8: 273–283.Taylor, A. and Blum, J. D. 1995. Relation between soil age andsilicate weathering rates determined from the chemical evolutionof a glacial chronosequence. Geology 23: 979–982.von Eynatten, H., Barcelo-Vidal, C. and Pawlowsky-Glahn, V.2003. Modelling compositional change: The example of chemicalweathering of granitoid rocks. Math. Geol. 35: 231 - 251.Watmough, S. A. and Dillon, P. J. 2003a. Base cation and nitro-gen budgets for seven forested catchments in central Ontario,1983–1999. For. Ecol. Manage. 177: 155–177.Watmough, S. A. and Dillon, P. J. 2003b. Calcium losses from aforested catchment in South-Central Ontario, Canada. Environ.Sci. Technol. 37: 3085–3089.Watmough, S. A., Aherne, J., Alewell, C., Arp, P. A., Bailey, S.W., Clair, T. A., Dillon, P. J., Duchesne, L., Eimers, C.,Fernandez, I., Foster, N., Larssen, T., Miller, E. K., Mitchell,M. and Page, S. 2005. Sulphate, nitrogen and base cation budgetsat 21 forested catchments in Canada, the United States and Europe.Environ. Sci. Technol. (in press).Yeomans, J. C. and Bremner, J. M. 1988. A rapid and precisemethod for routine determination of organic carbon in soil.Commun. Soil Sci. Plant Anal. 19: 1467–1476.Zak, S. K., Beven, K. J. and Reynolds, B. 1997. Uncertainty inthe estimation of critical loads: a practical methodology. Water AirSoil Pollut 98: 297–316.Zar, J. H. 1974. Biostatistical analysis. Prentice-Hall, Inc.,Englewood Cliffs, NJ.
Top Related