long-term soil chemistry changes in aggrading forest ecosystems
TRANSCRIPT
DIVISION S-2-SOIL CHEMISTRY
Long-Term Soil Chemistry Changes in Aggrading Forest EcosystemsJennifer Donaldson Knoepp* and Wayne T. Swank
ABSTRACTAssessing potential long-term forest productivity requires identifi-
cation of the processes regulating chemical changes in forest soils. Weresampled the litter layer and upper two mineral soil horizons, A andAB/BA, in two aggrading southern Appalachian watersheds 20 yrafter an earlier sampling. Soils from a mixed-hardwood watershedexhibited a small but significant decrease in soil pH. Extractable basecation content declined substantially in both mineral horizons. Forexample, Ca2+ levels in the A horizon fell from 236 kg ha'1 in 1970to 80 kg ha"1 in 1990. Proportionally, the decline was greatest forMg2+, which dropped from 111 to 20 kg ha-1. A white pine (Pinusstrobus L.) plantation was planted in 1956, after clear-felling hard-woods and recutting sprouts for 15 yr. Soil pH and base cationconcentrations declined in the A horizon from 1970 to 1990. Soil pHdeclined from 5.9 to 5.0 and Ca2+ levels from 534 to 288 kg ha'1.Cation content did not change significantly in the AB/BA soil horizon.Nutrient budgets were constructed using these soil and litter data plusexisting data on weathering rates, forest productivity, and hydrologicfluxes and associated chemistry. Decreases in soil base cations andsoil pH are attributed to leaching and to the sequestration of nutrientsin biomass.
T ONG-TERM SITE PRODUCTIVITY may depend on bothI J long- and short-term changes in soil chemical com-
position. In soils that have been depleted of nutrients byyears of exploitative agricultural practices, reforestation
increases nutrient and organic matter content (Fisher,1990). Planting forest species on normal production agri-culture soil can decrease soil pH and base cation concen-trations (Binkley et al., 1989). Only a few studies in theUSA have examined long-term changes in soils thatremained forested (Johnson et al., 1991). Decreases insoil pH and base cation concentrations appear common.These changes have been attributed to cation leachingand biomass accumulation.
Cation leaching can result from natural processes oracidic deposition. Leaching as a result of natural pro-cesses was reported in two N2-fixing red alder (Alnusrubra Bong.) stands in western Washington by VanMiegroet and Cole (1984, 1985) and Binkley and Sollins(1990). Large accumulations of N during a 55-yr periodunder alder resulted in high rates of N mineralizationand nitrification. High levels of soil solution NOf in-creased cation leaching and decreased soil pH in surfacehorizons. This resulted in the displacement of cations tolower soil horizons on one site but cation removal fromanother site.
In a long-term study, Johnson et al. (1988) attributedsoil cation losses to biomass accumulation and leaching.Calcium was translocated from subsoil horizons to above-ground biomass. Decreases in soil Mg2+ content resulted
USDA Forest Service, Coweeta Hydrologic Lab, 999 Coweeta Lab Road,Otto, NC 28763. Received 10 Feb. 1993. *Corresponding author.
Published in Soil Sci. Soc. Am. J. 58:325-331 (1994).
Abbreviations: HDWD, mixed-hardwood watershed; WP, white pinewatershed; AAS, atomic absorption spectrophotometer; CEC, cation-exchange capacity; BS, base saturation; cmol H+, exchangeable acidity;ANOVA, Analysis of Variance; SE, Standard error.
326 SOIL SCI. SOC. AM. J., VOL. 58, MARCH-APRIL 1994
from leaching losses possibly caused by SOl~ deposition(Johnson and Todd, 1990).
We studied long-term changes in forest soil nutrientcontent in two aggrading forest ecosystems. Our studyobjectives were to (i) examine the 20-yr changes in surfacesoil and forest floor chemistry in an undisturbed mixed-oak forest and a white pine plantation and (ii) quantifybiogeochemical processes underlying any chemicalchanges.
MATERIALS AND METHODS
The study was conducted on Watersheds 17 and 18 at theCoweeta Hydrologic Laboratory in western North Carolina.The HDWD is a 13-ha, north-facing reference watershed withmixed hardwood vegetation. In the 1920s, HDWD was selec-tively logged and subsequently has not been disturbed byhumans. Basal area of the forest is about 26 m2 ha"1, and theoverstory is dominated by three Quercus spp.: Acer rubrumL., Liriodendron tulipifera L., and Carya glabra (Miller)Sweet. The understory is dominated by Rhododendron maxi-mum L. and Kalmia latifolia L., both evergreen species. OnWP, also a north-facing watershed adjacent to HDWD, the13-ha hardwood forest was clearcut in 1940 and regrowth wascut annually until 1955. White pine was planted in 1956, andpines were released from hardwood competition as necessaryby both cutting and chemical means until 1970. Basal area ofWP in 1990 averaged about 50 m2 ha~'.
Slopes on both watersheds average =50%. Two soil serieswere sampled on both watersheds. The Saunook series, afine-loamy, mixed, mesic Humic Hapludult, occupies stream-side positions. Ridge positions are occupied by Cowee-Evardcomplex soils, fine-loamy, mixed-oxidic, mesic, Typic Haplu-dults.
Sampling was conducted at four positions on two sampletransects established in each watershed by Yount (1975a,b) in1970. Each transect led from a stream channel to a point 60 mupslope on both facing slopes. In HDWD, transects were atmid-watershed level, with one transect bisecting the perennialstream and one bisecting an intermittent stream. In WP, tran-sects were across the lower and upper thirds of the singlestream channel. Two plots (2 by 15 m) were established oneach slope of a transect for a total of eight plots per watershed.
Methods of sample collection and preparation used in 1990were the same as those of Yount (1975a,b). In 1990, sampleswere collected in March and November. Data from these 2 moof 1970 only were used for comparison to avoid biases causedby seasonal variation in soil and forest floor cation concentra-tion and forest floor quantity.
Following Yount's procedures, triplicate 0.25-m2 quadratswere randomly selected for sample collection on each plot.The Oi and Oe litter layers were collected as one sample,after which composite soil samples were collected with a1.8-cm-diam. soil probe sampler. Yount collected Oi, Oe, andOa layers individually for WP. The Oa on HDWD was notcollected but was "partly incorporated into the Al horizon"(Yount 1975b). We quantitatively collected the Oa layer onHDWD at the November sampling date. The Oa extractablecations were determined using soil extraction methods. TheA horizon cation content represents soil cation content plusOa extractable cations. Extractable Oa cations were 20, 12,and 7% of total extractable A horizon Ca2+, Mg2+, and K+,respectively. The top two soil surface horizons were sampled,and the depth of the A horizon, which was present on all plots,was recorded. The second horizon varied from an AB to aBA and is therefore referred to as AB/BA. Average A horizon
depths were 5.4 and 5.8 cm for HDWD and WP, respectively,in 1970. The HDWD A horizon depth increased significantly(P < 0.05) to 7.3 cm in 1990 and, although WP A horizondepth also increased (6.4 cm), the change was not significant(P < 0.05).
Soil samples were air dried and sieved to <2 mm. Five-gramsubsamples were extracted with 20 mL 0.05 M HC1 plus0.0125 M H2SO4. Sample analyses in 1970 were performedby the Soil Testing and Plant Analyses Laboratory, CooperativeExtension Service, Athens, GA. Methods used in 1990 werethe same, except AAS was used for cation determinationsinstead of colorimetric methods. Cation concentrations weredetermined by AAS, following establishment of a standardcurve. All diluted samples were analyzed in duplicate. SoilpH was determined in a 1:1 soil/H2O slurry. The cmol H+
was determined by pH change in a 1:2 soil/Adams and Evansbuffer slurry. The cmol H+ was calculated from the changein buffer pH (8.0), 0.8 cmol H+ kg'1 soil for every 1.0 unitpH decrease. The CEC was calculated as the sum of double acidextracted cations and cmol H+. Percentage BS was calculated asthe sum of base cations divided by total CEC X 100.
Litter samples were oven dried at 50 °C, weighed, andground to pass a 1-mm sieve. Samples (0.5 g) were dry ashedat 480 °C for 4 h, dissolved in 20% HNO3, and brought upto volume with 2% HNOs solution. Cation concentrations weredetermined by AAS.
Means of die three quadrats per plot (« = 8 plots X 2 dates= 16) in 1970 and 1990 were used for statistical analysis.Watershed means in 1970 and 1990 were compared using theANOVA procedure (SAS Institute, 1985).
Nutrient budgets were calculated from the 1970 and 1990soil and litter data. Extractable soil cation contents (kilogramsper hectare) were calculated using the measured depth of theA horizon and a total depth for the A plus AB/BA horizonsfor 20 cm (Table 1). Bulk density values (<2-mm fraction)are 0.87 (SE = 0.05; n = 13) and 1.00 Mg nr3 (McGinty,1976) for the A and AB/BA horizons, respectively. Total Aplus AB/BA horizon depth and A horizon bulk densities weredetermined in soil pits in other studies in this portion of theCoweeta basin.
Deposition inputs were calculated from the bulk precipitationchemistry sampling network at Coweeta (Swank and Hender-son, 1976). Chemistry and precipitation amounts have beenmeasured weekly since 1972 and thus provide a high degreeof precision for estimating inputs of most dissolved solutes(Swank and Waide, 1988). Geochemical weathering rates werecalculated by the hydrologic balance method (Berry, 1977,unpublished data; Velbel, 1988). This method combines geo-chemical mass balance data with empirically determined min-
Table 1. Characteristic soil profiles showing horizon depths andNHUOAc extractable base cations for Saunook and Cowee-Evard soil series. Values represent means of four soil pits locatedin sites in the Coweeta basin similar to the location of sampleplots within the mixed-hardwood and white pine watersheds.
Horizon Depth
-kgha~Saunook: fine-loamy, mixed, mesic Humic Hapludult
A 0-12AB 12-25BC 25-45 +
Cowee-Evard: fine-loamy,A 0-9E 9-19EB/AB 19-33Bt/Bw 33-54BC 54-74 +
79 10854 121
291 296mixed, mesic Typic Hapludult
49 1176 48
24 15164 23029 460
4236
239
326
22151270
KNOEPP & SWANK: FOREST ECOSYSTEM SOIL CHEMISTRY CHANGES 327
Table 2. Analysis of variance showing effects of year, month, year x month, and soil type on total Ca2+, K+ , and Mg2+ to a 20-cm depthfor mixed-hardwood (HDWD) and white pine watersheds (WP). Degrees of freedom for each effect equals one.
MS
HDWDP<F MS
WPP<F
Ca2+
YearMonthYear x monthSoil type
3596693269
10247198386
42.820.391.22
23.62
0.000.540.280.00
31417060163
968453460
5.571.070.028.05
0.020.310.900.01
YearMonthYear x monthSoU type
35482137368955
20.050.080.210.54
0.000.780.650.47
165079
6131275
0.730.040.270.56
0.400.850.610.46
YearMonthYear x monthSoil type
137574106608941
43090
44.453.442.89
13.92
0.000.070.100.00
1606557505666384
271
14.286.675.900.02
0.010.010.020.88
era! weathering reactions specific to the primary minerals atCoweeta. Estimates of mineral weathering rates at Coweetacompare favorably with mineral weathering rates determinedin laboratory experiments under similar hydrogeochemical con-ditions (Velbel, 1988).
Outputs from the watersheds include dissolved exports calcu-lated from continuous weir discharge measurements combinedwith chemical analyses of weekly grab samples of stream watersince 1972 (Swank and Waide, 1988). Analysis of Coweeta
streams has shown that grab sample estimates of export formost constituents are within 1 to 5 % of values calculated fromflow-proportional stream samples (Swank and Waide, 1988);thus, export estimates are considered to be highly reliable.Other components of output are also based on long-term dataand include sediment export (Swank and Waide, 1988) andbiomass accumulation above ground (Monk and Day, 1988;Swank and Schreuder, 1973; Swank, 1988, unpublished data)and below ground (McGinty, 1976; Vose, 1992, unpublished
O
600
500
400
300
200
100
0
1£331970E321990
* x\;H US ^**
V///////////A
A
iCa Mg pH
6.0
5.5
5.0
4.5
4.0
12
10
8o* 6O
o 4
2
0
**
70
60
50
40
30
10
0CEC CMOL H BS
o
300
250
200
150
100
50
0 ICa K Mg
6.0
5.5
5.0
4.5
4.0
o
12
10
8
6 •O2o 4
0CEC CMOL H BS
70
60
50
40
30
20 ftCD
10
0
Fig. 1. Double acid extractable Ca2+, K+ , and Mg2+ contents, soil pH, cation-exchange capacity, exchangeable acidity, and base saturation for(a and c) A horizon and (b and d) AB/BA horizon soils for a mixed-hardwood watershed (HDWD). Values represent means of eight sampleplots at two sample dates. *, ** Significant difference between 1970 and 1990 means at P == 0.05 and 0.01, respectively.
328 SOIL SCI. SOC. AM. J., VOL. 58, MARCH-APRIL 1994
data). Aboveground biomass accumulation was calculated fromremeasurement of permanent vegetation plots in both water-sheds in combination with allometric equations for estimatingbiomass (Monk and Day, 1988; Swank and Schreuder, 1973;Swank, 1988, unpublished data). Belowground biomass incre-ment for WP was calculated from measurement of coarse rootbiomass in 1975 (McGinty, 1976) and 1991 (Vose, 1992,unpublished data) on another white pine watershed at Coweeta.Initial belowground biomass for HDWD was measured in 1974(McGinty, 1976); biomass accumulation was estimated usingthe ratio of aboveground/belowground biomass accumulationmeasured for WP. Fine root biomass was assumed to be insteady state. Estimates of woody litter accumulation measuredat Walker Branch Watershed (Johnson et al., 1990) were usedin the nutrient budget because these data were not availablefor Coweeta.
RESULTS AND DISCUSSIONSoil and Litter Chemical Composition
Characteristic soil profiles for the Saunook andCowee-Evard soil series are shown in Table 1. Dataincluded are horizon depth and Ca2+, K+, and Mg2+
contents. Profile descriptions represent the means of foursoil pits for each series located in the same area asHDWD and WP in the Coweeta basin.
Table 2 shows ANOVA for total kilograms per hectareCa2+, K+, and Mg2+ (20-cm depth), including mean
square, F statistics, and probability of values greaterthan F. Year is significant for Ca2+ and Mg2+ on bothHDWD and WP and for K+ on HDWD only. Soil typehas a significant effect on changes in Ca2+ content forboth HDWD and WP and in Mg2+ content only onHDWD, indicating some spatial variability. However,spatial variability in bulk density was slight. Bulk densitywas determined in 13 soil pits in HDWD and an adjacentwatershed. A horizon bulk density was 0.87 Mg m~3
with ±6% SE. The B horizon was less variable; bulkdensity mean was 1.27 Mg m~3 with ±4% SE.
In HDWD A horizon soils, extractable Ca2+, K+, andMg2+ and soil pH declined substantially and significantly(P < 0.05) between 1970 and 1990 (Fig. la). Soil Mg2+
declined 82% from 111 kg ha"1 in 1970 to 20 kg ha"1
in 1990. Soil K+ showed the least decline, droppingfrom 84 to 44 kg ha"1. Soil pH decreased from 4.9 to4.7. In the AB/BA horizon, all three cations and soilpH also decreased significantly (P < 0.05) (Fig. Ib).Extractable Ca2+ decreased most, from 78 to 22 kg ha"1.Soil K+ again showed the smallest proportional decreasein pool size, from 69 to 43 kg ha"1. Soil pH decreasedfrom 5.4 to 4.9.
The CEC and BS decreased in HDWD A horizon soilsfrom 9.5 to 6.4 cmol kg"1 and 49 to 14% in 1970 and1990, respectively (Fig. Ic). The cmol H+ increasedfrom 4.4 to 5.4 cmol kg"1. The AB/BA horizon also
o
600
500
400
300
200
100
0
ESS 1970
Ca Mg pH
6.0
5.5
5.0*
4.5
4.0
o
12
10
8
o2o 4
2
0CEC CMOL H
70
60
50
40
30
20
10
0BS
300
250
200
150
100
50
Ca
1K
1Mg
6.0
5.5
5.0
4.5
4.0
o•
12
10
8
o2o 4
CEC CMOL H BS
70
60
50
40
30
20
10
§5LJ
<
Fig. 2. Double acid extractable Ca2+, K+ , and Mg2+ contents, soil pH, cation-exchange capacity, exchangeable acidity, and base saturation for'(a and c) A horizon and (b and d) AB/BA horizon soils for a white pine watershed (WP). Values represent means of eight sample plots attwo sample dates. *, ** Significant difference between 1970 and 1990 means at P < 0.05 and 0.01, respectively.
KNOEPP & SWANK: FOREST ECOSYSTEM SOIL CHEMISTRY CHANGES 329
showed a decrease in BS along with an increase in cmolH+, but CEC did not change (Fig. Id).
In WP A horizon soils, extractable Ca2+, K+, andMg2+ and soil pH decreased between 1970 and 1990(Fig. 2a). Soil Mg2+ levels fell from 145 to 35 kg ha~',and K+ content declined the least, from 68 to 48 kgha"1 (P = 0.06). In the AB/BA horizon, only Mg2+
decreases were significant (Fig. 2b). Soil pH decreasedin both A and AB/BA horizons from 5.9 to 5.0 and 5.5to 5.0, respectively.
In the A horizon soils of WP (Fig. 2c), BS decreasedfrom 66 to 36%, CEC decreased, and cmol H+ increased.The AB/BA horizon soils had an increase in CEC andcmol H+ but no change in BS (Fig. 2d).
The total weight of Oi plus Oe layers increased onboth HDWD and WP (Fig. 3a and 3b). On HDWD,total litter weight increased 28% from 7200 to 9300 kgha"1, and on WP, it increased 45%, from 12200 to17 800 kg ha"1. On HDWD and WP, Ca2+, K+, andMg2+
concentrations decreased (P < 0.05) in the combined Oiand Oe layers (Fig. 3a and 3b).
Nutrient BudgetsComplete nutrient budgets were constructed for each
watershed using soil and litter chemistry, nutrient inputsand outputs, leaching, and biomass accumulation. Thisallowed us to identify the biogeochemical processes re-sponsible for cation losses (Table 3). On HDWD, cationoutputs to streamflow were equal to, or slightly greaterthan, biomass accumulation. Ratios of biomass accumu-lation/leaching were 3:4 for Ca2+, 1:1 for K+, and 2:3for Mg2+. In contrast, on WP, large proportions of basecations depleted from the surface soils were sequesteredin the vegetation rather than lost through leaching. Ratiosof biomass accumulation to leaching were 2:1 for Ca2+,3:1 forK+ , and 2:1 for Mg2+.
Net nutrient balance values were about the same onHDWD and WP but varied for individual cations. Thesmall discrepancies for Ca2+ budgets, +17 kg ha"1 onHDWD and +27 kg ha"1 on WP, indicate that mostlosses from surface soils are accounted for by vegetationincrement and leaching. The negative Mg2+ budget bal-ances indicate unmonitored losses on both watersheds.In contrast, balances for K+ in both watersheds arepositive. Apparently, an unmonitored source of K+ isavailable for vegetative uptake and leaching loss.
Eleven-year changes in B horizon cation concentra-tions were measured in soils from Walker Branch, amixed-deciduous forested watershed in Tennessee (John-son et al., 1988). Exchangeable Ca2+ and Mg2+ bothdecreased. Changes in Ca2+ corresponded with accumu-lation in forest biomass. However, Mg2+ losses did notagree with calculated amounts sequestered in biomass.Later studies indicated that Mg2+ losses were the resultof leaching by SO?T from atmospheric inputs (Johnsonand Todd, 1990).
Leaching was also an important loss mechanism forMg*+ from HDWD in our study. The biomass accumula-tion/leaching ratio was 2:3. However, biomass accumu-lation and leaching as streamflow do not account for all
the Mg2+ lost. It is possible that Mg2+ was leached outof the A and AB/BA horizons and retained in the deepersubsurface horizons. Displacement of cations from sur-face to subsurface soils was found by Van Miegroet andCole (1985). Rates of cation leaching from surface soilsof the alder stand they studied were high because ofhigh rates of NOf formation. In that case, base cationsmoved to the B horizons at 15 to 45 cm.
The subsurface soil horizons can act not only as apotential cation sink, as in the case of Mg2+, but alsoas a potential source. Nutrient budgets indicate that therewas a source of K+ that was not included in our budgets.Examination of soil chemical profiles characteristic ofstudy site soils showed substantial amounts of all basecations, and particularly K+, available below the A andAB/BA horizons (Table 1). Cations at depths between20 and 65 cm are readily available to plants. Approxi-mately 30% of the root biomass of hardwood tree speciesat Coweeta is at depths of 20 to 80 cm (McGinty, 1976).Subsoil fertility was also found to be important to thenutrition of trees and other deep-rooted plants by Com-erford et al. (1984) and Nowak et al. (1991).
Direct mineral weathering by roots may be anothersource of nutrients available to plants that is not includedin standard nutrient budgets. April and Keller (1990)investigated the effects of roots on soil minerals and the
15
1970B221990
20
15
Mg
10
107
Fig. 3. Total weight and cation concentration of Oi plus Oe litterlayers for (a) mixed-hardwood and (b) white pine watersheds.Values represent means of eight sample plots at two sample dates.* and ** represent significant difference between 1970 and 1990means at P = 0.05 and 0.01, respectively.
330 SOIL SCI. SOC. AM. J., VOL. 58, MARCH-APRIL 1994
Table 3. Nutrient budgets for mixed-hardwood (HDWD) and white pine watersheds (WP). Budget is balanced in respect to the surfacesoil (20 cm) cation concentration differences between 1970 and 1990. Inputs and outputs represent 20-yr totals based on annual averages.' +' and ' — ' balance values indicate unmonitored additions and removals of cations to the watershed, respectively.
1970 poolsSoiltLitter):
InputsDepositions§Bedrock weathering^
TotalOutputs
VegetationAbovegroundl, ftBelowgroundit
Stream§,§§SedimentDissolved
1990 poolsSoULitterWoody litterl 1
Losses or additions
Ca2+
HDWDK* Mg*+ Ca2+
WP
K+ Mg2*lr« k.-l
31499
8082
575
9059
3189
1029356
+ 17
15315
4076
284
10036
0.1134
861116
+ 99
17013
1756
256
2019
0.260
39134
-101
77492
8282
1030
21555
3117
5758111
-27
15322
4076
291
20044
0.191
1381110
+ 203
22512
1756
310
5615
0.231
831110
-104
t Yount, 1975a.t Yount, 1975b.§ Swank and Waide, 1988; Swank, 1988, unpublished data.I Berry, 1977, unpublished data.# Monk and Day, 1988.tt Swank and Schreuder, 1973; Swank, 1988, unpublished data.tt McGinty, 1976; Vose, 1992, unpublished data.§§ Swank, 1988.II Johnson et al., 1990.
potential for nutrient uptake at several sites, includingCoweeta. They found evidence for differential weather-ing of biotite, a source of K+, adjacent to root surfaces.
Watershed History and Species EffectsEffects of the older mixed-hardwood forest and
younger white pine plantation, on both soils and biomass,differed during the 20-yr period. All cation concentra-tions in the HDWD AB/BA horizon decreased signifi-cantly , but changes in soil cation concentrations in the WPAB/BA horizon were not significant. In 1970, beneath thepine, Ca2+ pools in the A plus AB/BA horizons weremore than two times greater than in HDWD (Table 3).The higher initial cation content on WP results from thetreatment before and during white pine establishment.The watershed was clear-cut in 1940 with no woodremoval from the site. The vegetation was recut annuallyfrom 1940 through 1955 as part of a water yield experi-ment. After planting, the pines were released throughmanual and chemical means. The continued addition ofvegetation material may have acted as a slow-releasefertilizer treatment for the surface soils. The greaterquantity of soil Ca2+ available in 1970 under the pineresulted in a smaller proportional change in soil Ca2+
during the 20-yr period, i.e., 25 and 75% decreases forWP and HDWD, respectively.
Biomass accumulation/leaching ratios of all three cat-ions were greater for WP than for HDWD. This dissimi-larity is due to both lower cation leaching rates and greater
rates of biomass accumulation in WP. Two processes areresponsible for reduced leaching rates in WP. First, 25 %greater evapotranspiration in WP (Swank et al., 1988)decreases total water movement through the watershed.Second, net primary production of woody material ismuch greater in the white pine than the hardwood forest,9.0 vs. 4.0 Mg ha"1 yr"1. Thus, the younger, morevigorous white pine stand accumulated about 80% moreCa2+, K+, and Mg2+ than the mixed hardwood forest.
Components of the K+ budgets for these ecosystemsalso differ. For example, there was little or no decreasein K+ pools in the upper 20 cm of soil during the 20-yrperiod for WP. It appears that a greater amount of K+
was accumulated from subsurface soils in WP than inHDWD (204 vs. 99 kg ha'1). This difference is approxi-mately equal to the increased sequestration of K+ in thewhite pine plantation (108 kg ha"1).
Alban (1982) and Binkley and Valentine (1991) foundthat many tree species change soil pH and cation concen-trations, especially in the surface horizons. Speciesdiffered in their effects on soil acidity. Binkley andValentine (1991) found more acidic soils under whitepine and Norway spruce [Picea abies (L.) karsten] thanunder green ash (Fraxinus pennsylvanica Marshall var.pennsylvanica). Alban (1982), on the other hand, foundthat the acidifying effect of aspen (Populus tremuloidesMichaux) was greater than red pine (Pinus resinosaAiton) or jack pine (Pinus banksiana Lambert) and aboutequal to white spruce [Picea glauca (Moench) Voss].In our study, soil pH of A horizon soils in WP decreased
KNOEPP & SWANK: FOREST ECOSYSTEM SOIL CHEMISTRY CHANGES 331
from 5.9 to 5.0 whereas, in HDWD, it fell from 4.9 to4.7. The decrease hi BS indicates that accumulation ofcations in biomass is largely responsible for the pHdecrease. The decline in CEC indicates that the exchangecomplex may have changed (Binkley, 1992). This canresult from changes in pH-dependent exchange sites onorganic compounds, in organic matter quality, or insecondary minerals. Soil pH also decreased in AB/BAhorizons in both watersheds. Decreases in HDWD soilpH resulted from cation losses. This is evident becauseof the decrease hi BS with no change hi CEC (Fig. Id).The WP shows a different pattern where AB/BA soilpH decreased but CEC increased and BS did not change.This suggests a change in the exchange complex (Binkley,1992).
Our results show the importance of monitoring biomassaccumulation and cation leaching as well as changes insoil chemical composition. Without this information, itwould be impossible to attribute changes in soil nutrientcontent to the appropriate regulating processes. This under-standing is essential to assessments of long-term produc-tivity and interpretations of anthropogenic influences onforest ecosystems. The compilation of nutrient budgetsshows that each base cation cycles differently and eachwatershed reacts distinctively.
ACKNOWLEDGMENTSThe authors thank Dr. Dan Binkley, Colorado State Univ.,
Fort Collins, and Dr. Dale Johnson, Desert Research Inst,Univ. of Nevada, Reno, for their critical reviews, which im-proved this manuscript.