| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Nicholas School of the Environment, Duke University, Durham, NC 27708, currently at Ashoka Trust for Research in Ecology and the Environment (ATREE), P.O. Box 2402, HA Farm Post, Hebbal, Bangalore 560 024, India
b Nicholas School of the Environment, Duke University, Durham, NC 27708
* Corresponding author (jagdish{at}atree.org)
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSite DescriptionMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONREFERENCES |
|---|
Abbreviations: BS, base saturation • ECEC, effective cation-exchange capactiy • SOM, soil organic matter
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSite DescriptionMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONREFERENCES |
|---|
25000 km2 yr-1 in recent decades. The conversion is especially prominent in Central and South America (William and Turner, 1994). Forest conversion to nonforest uses has significant effects on biogeochemical and physical properties of soils (Buol, 1994). Resultant soil erosion, loss of water-stable aggregates, loss of nutrients, and change in soil C storage, all have implications for the sustainability of land use (Pimental et al., 1995; Richter and Markewitz, 2001) and future climate change (Schlesinger, 1997). Many areas currently under conversion are dominated by soils with variable charge. Many of these are in an advanced state of weathering as soils proceed through developmental stages (Jackson and Sherman, 1953). This advanced stage of weathering is characterized by Ultisols and Oxisols and the commonly associated soil minerals kaolinite, gibbsite, and hydrous Fe-oxides. Such soils are found on over 8 million km2 in Central and South America alone (Richter and Babbar, 1991). Variable charge affects many of the biogeochemical and hydrologic properties of advanced weathering stage soils (El-Swaify and Dangler, 1977; Sollins et al., 1988). This includes nutrient retention, ion exchange, aggregation, erodibility, drainage, aeration, and infiltration.
The total surface charge density (
T) and total charge (
T) of variably charged soil systems can be represented by:
 | [1] |
Intra-aggregate bonds are formed and maintained by a variety of mechanisms in which functional OH groups on metal oxides and SOM are key participants (Sposito, 1989). Stability of the soil aggregates can be affected by changes in SOM and pH effects on net charge, positive or negative (El-Swaify and Dangler, 1977; Suarez et al., 1984; Chorover and Sposito, 1995). Soil organic matter also plays a role in nonelectrostatic intra-aggregate forces (Bartoli et al., 1992b). Larger aggregates are protected by roots and fungal hyphae (Oades, 1984). Thus, SOM dynamics and changes in surface charge because of pH are the most important determinants of aggregate stability in soils undergoing forest conversion in the tropics.
Advanced weathering-stage soils are well represented by many soils of the Terraba basin (the Zona Sur) in southern Costa Rica. Under forest cover such soils are characterized by stable aggregation, rapid drainage, and resistance to erosion (Sollins et al., 1988; Richter and Babbar, 1991). Forest conversion and subsequent agricultural land uses typically eliminate the protective litter layer, reduce the organic matter content of the mineral soil, and alter the soil electro-chemical environment. Any disturbance such as the conversion of permanent, deeply rooted forest cover to pasture have complex biogeochemical effects on soils and the ecosystem. However, changes in the biogeochemistry of variable charge soils are poorly documented. Current knowledge of the chemical processes of cation and anion retention, soil surface chemistry and soil aggregate stability of advanced weathering-stage soils at the landscape scale is rather rudimentary. Relatively little quantitative data are available on aggregate stability of highly weathered soils such as Oxisols and Ultisols and their response to forest and to nonforest uses (Castro and Logan, 1991; Bartoli et al., 1992a).
This paper compares chemical and physical properties of forests and pastures on acidic, advanced weathering-stage soils dominated by variable charge properties in the Rio General Valley of Costa Rica. The specific objectives were to (i) quantify the differences in soil chemical and physical properties, specifically SOM and pH, between forests and pastures; and (ii) assess the secondary effects of differences in organic matter and pH on chemical properties such as cation exchange and physical properties such as bulk density and water-stable aggregates.
|
Site Description |
|---|
|
TOPABSTRACTINTRODUCTIONSite DescriptionMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONREFERENCES |
|---|
|
Although the valley has been inhabited by Native Americans for millenia, since about 1950, forest cover in the Terraba basin has been reduced by >80% in the 3500 km2 outside the high-elevation forest reserves. Fire-maintained savannah woodlands occur in areas with a longer dry season (Kesel and Spicer, 1985). Vegetation prior to about 1950 was tropical moist forest much of which is present only as relicts in the main valley, alluvial fans, and river terraces (Skutch, 1971). The region's dominant current land use is cattle (Bos taurus) grazing. Pastures now occupy about 2000 km2 of the basin.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONSite DescriptionMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONREFERENCES |
|---|
The constraints of space-for-time-substitution were considered in the field selection of blocks and plots. The five interfluves were treated as experimental blocks and 20 by 20 m plots installed within forests and pastures of each block. Farmers may well have selected certain types of areas to clear, but we attempted to select soils on the same or similar landforms and all sites within blocks had no visual indications that they had been different from the forest sites prior to clearing.
Forests had closed canopies and no obvious evidence of recent human disturbance. Most forest conversion to pasture took place between 1950 and 1980 based on land-use history, maps, satellite imagery, and local information (Krishnaswamy, 1999). When forests were cleared, wood was used for construction timber, furniture, and fuel, but much or most was probably burned and decomposed on site. Pastures have dense grass cover, scattered trees, and are often subject to intense cattle grazing. Chemical inputs however have been minimal.
Additional soils were also sampled in a tree plantation, two additional pastures, and three intensive agricultural sites (one in sugar-cane [Saccharum L.] and two in pineapple [Ananas P. Mill.] plantations) in the same region to enlarge the data set used to evaluate general soil properties and relationships between parameters. These sites were all from upland interfluves, and were within a few kilometers of the other sites. The slopes of all the sites were all <10%.
Soil Sampling
Bulk density was estimated with a 60-mm slide-hammer sampler (Blake, 1983) from three replicates in each soil plot in forests and pastures. In forests and pastures, bulk-density was estimated in the 0- to 10- and 10- to 30-cm layers. Undisturbed subsamples of soil were carefully collected by breaking aggregates along planes of weakness when field moist to give clods less than about 40-mm diam. These were air-dried for subsequent aggregate analyses, and carefully packed for shipment to the soils lab at Duke University.
Samples for other chemical parameters were collected with a 110-mm diam. soil auger. Samples came from the corners of the 20-m square plot at four depths: 0 to 10, 10 to 30, 30 to 70, and 70 to 120 cm. In one plot out of the 16, several depths >30 cm were not sampled because of the presence of stones. Soil samples were air-dried and passed through a 2-mm screen prior to chemical and textural analysis. Soils were air-dried on site, carefully packed for shipment, imported to the USA on a USDA permit that allowed no entry-treatment of the samples.
Soil Analysis
Soil samples were analyzed for pH (2:1 in water and 0.01 M CaCl2), 1 M KCl-exchangeable acidity and Al using NaF titration (Thomas, 1982). Cations (Na, K, Ca, and Mg) were extracted with 1 M ammonium acetate buffered at pH 7 and analyzed by atomic absorption spectrophotometry. Effective cation-exchange capacity was estimated from the sum of extracted base cations and KCl-acidity, and base saturation (BS) from the quotient of the sum of exchangeable base cations and ECEC. A subset of samples were analyzed for cation-exchange capacity (CEC) at pH 8.2 using a BaCl2 extraction and triethanolamine buffer (Mehlich, 1938; Thomas, 1982). Total C and N were determined using Dumas dry combustion (Perkin Elmer CHN/S analyzer, Perkin Elmer, Norwalk, CT) on pulverized samples. Texture was determined using the pipette-gravimetric method.
Throughout the chemical analyses at least 10% of the samples were randomly replicated for quality monitoring. The differences between replicates was usually <5% (>90% of samples) and was never >10%.
Three clayey (>50%) samples from 70 to 120 cm with lowest ECEC (<0.5 cmolc kg-1) were chosen for mineralogical analyses. The clay fraction (<2 mm) was obtained by Calgon dispersion and was treated with Mg acetate before oriented X-ray diffraction (Whittig, 1983). The diffractograms (intensity vs. 2
) were plotted to detect peaks.
Air-dried samples for aggregate analyses were carefully sieved for 30 s using a 5-mm sieve to collect aggregates retained on a 1-mm sieve positioned below the 5-mm sieve. Aggregate stability was measured using a modified single-sieve method (Kemper and Rosenau, 1986; Beare and Bruce, 1993) with 1-h shaking at 39 strokes per min in water. Sieve sizes were 0.25, 1, and 2 mm. Sand, roots, and stones remaining on the sieve after shaking were determined after NaOH (0.05 M) treatment to disperse aggregates. Two replicates were performed for each sieve size. Tests with larger aggregates (>1 and >2 mm) were performed on aggregate samples pooled across all five blocks, whereas the 1- to 5-mm aggregates were tested using the 0.25-mm sieves seperately for each block. In addition, about 50 g of air-dry aggregates from the pooled sample were used for estimating the distribution of aggregate sizes by using a nest of sieves (5, 4, 2.8, 2, and 1 mm) and gentle shaking for 10 s.
Contents and changes in contents of soil constituents for pastures were estimated using the equal mass method to correct for bulk density changes following compaction (Veldkamp, 1994).
Box and whisker plots displaying lower and upper quartiles and median, were utilized to compare land uses. This method has the advantage of estimating central tendencies better when the number of replications is small (
5), and in addition gives an idea of the spread about the central tendency. Wherever appropriate, 95% confidence intervals about the median were computed. These approximate intervals are a function of the quartiles and are little affected by outliers (McGill et al., 1978). This was supplemented with ANOVA (Kleinbaum et al., 1988). Contrasts between land uses were analyzed by the LSD multiple comparison technique. The results were compared with other techniques (Duncan, Tukey and Waller) to ensure that results were not sensitive to choice of method. It is to be noted that whereas the ANOVA accounts for block differences if present while detecting land-use treatment effects, the box and whisker plots indicates relative effects of land-use across the landscape. Unless reported otherwise, all estimates given are medians with 95% confidence intervals if they are positive.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONSite DescriptionMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONREFERENCES |
|---|
|
The contribution of clay minerals to ECEC is remarkably small in these soils, 1.48 cmolc kg-1clay at 70 to 120 cm, another indication of the advanced state of weathering. The mineralogical and chemical properties described above confirms to the so-called mineralogical monotony (i.e., the domination of Fe oxides and kaolinite) of many Ultisols and Oxisols (Schwertmann and Herbillon, 1992).
The soils are well aggregated and friable. Bulk density of soils under forest (Fig. 2) were 0.72 ± 0.148 (0–10 cm) and 0.98 ± 0.113 (10–30 cm). Water-stable aggregates >0.25 mm were 94.5% in the upper 30 cm.
|
|
The bulk density of pasture soils was higher than those of forest soils (Fig. 2). Conversion to pasture has apparently increased bulk densities from 0.72 ± 0.148 to 1.01 ± 0.042 g cm-3 in the upper 10 cm and from 0.98 ± 0.113 to 1.01 ± 0.007 g cm-3 for the 10- to 30-cm depth. Means were significantly different (P < 0.05) for the 0- to 10-cm depth. The greater variability of bulk density in pasture soils than under forest (Fig. 2) may be associated with variable ages of pastures and cumulative compaction because of cattle. Despite the change, the overall bulk density in pastures (1.0 g cm-3) is still relatively low, and assuming a particle density of 2.5 g cm-3, the estimated pore space would be 60%. Thus, in the event of management practices destroying the aggregation drastically such as in cultivation, there would appear to be great potential for pore space to collapse.
Organic Carbon and pH
The concentration of soil C was numerically greater under forest than pasture throughout the upper 120 cm (Fig. 3). The median concentration difference ranged from 10.6% at 0 to 10 cm to 30.3% at 70 to 120 cm. However, means were not significantly different except at 30 to 70 cm (P < 0.1). Median C contents under forests and pastures in the upper 30 cm were 79.5 ± 11.7 and 63.5 ± 17.9 Mg ha-1, respectively. Overall, there seems to be a small to moderate loss of soil C and this is attributed to the inputs from grass roots, manure, and fires in the pastures.
Soil pH was lower under forest than under pasture throughout the upper 120 cm (Fig. 4) . Means were significantly different at 10 to 30 and 30 to 70 cm (P < 0.1), and at 70 to 120 cm. (P < 0.05). The conversion from forest to pasture has apparently elevated soil pHH2O by 0.44 ± 0.102 pH units in the upper 120 cm. Effective BS was correlated with pH, based on the larger data pooled across all depths (R2 = 0.60, P < 0.0001, n = 60). Median base saturation on a content basis (30 cm) was 73.9% in pastures and 22.8% in forests. However, means were not statistically different although effective BS tended to be higher in pastures at all depths.
|
Soil Cation-Exchange Capacity
Differences in ECEC between soils under forests and pastures are evident throughout the upper 120 cm (Fig. 3 and 4). Mean ECEC of pastures were significantly higher (P < 0.05) than those in forests at 30- to 70-cm and 70- to 120-cm depths. The difference in ECEC in soils of pastures from those of forests is an estimated 1.41 ± 0.102 cmolc kg-1 in the upper 120 cm, an increase of 34%.
Another index of net charge apart from ECEC is 
pH (pHCaCl2 - pHH2O) (Mendonca and Rowell, 1996). In general, ECEC was negatively correlated with this index of effective charge, at depth >30 cm (R2 = 0.23, n = 28, P < 0.02). The trends in pH with land-use are similar to that for ECEC (Fig. 4) which hints at a real change in soil charge.
The increase in cation-storage capacity apparently because of forest conversion to pasture is estimated to be 39.0 ± 19.9 kmolc ha-1 in the upper 30 cm (Fig. 3). Although the ANOVA results indicated no significant differences between pasture and forest, this was attributed to the effects of one outlier block (Fig. 3). The median cation-storage capacity without considering this block are 114.3 ± 19.32 kmolc ha-1 in pastures and 81.2 ± 19.6 kmolc ha-1 in forests, an increment of 33 kmolc ha-1. Mean ECEC in pastures (110 kmolc ha-1) was also significantly different (P < 0.05) from forests (77 kmolc ha-1) when this block was omitted.
Organic Matter, pH, and Effective Cation-Exchange Capacity
A regression model of ECEC as response with the only the interaction of pH and C as a covariate was significant (ECEC = 1.30 + 0.18 (pHH2O x C), R2 = 0.40, P < 0.0001). The simulated effect of pH on ECEC was analyzed graphically. The pKa of carboxylic groups in organic matter is 4 (Thomas and Hargrove, 1984). The data were split into two groups based on CaCl2 pH: one pH 4 and below, the other above pH 4.0. Above pH 4, increases in soil C are associated with increases in ECEC, whereas below, there is a weak or no response (Fig. 5)
. At elevated pH deprotonation of organic functional groups is enhanced. Above a soil pHCaCl2 of about 4, the contribution of SOM to pH-dependent variable charge is well expressed (Fig. 5).
|
|
Effects of Land Use on Effective Charge
Changes in ECEC because of land-use in variable charge soils have been estimated in varied ecosystems (Alegre et al., 1988; Ghuman and Lal, 1991; Richter et al., 1994; Koutika et al., 1997; Menzies and Gillman, 1997). The presence of low-activity soil minerals with low constant charge makes the ECEC of the General Valley soils strongly linked to SOM and its response to change. The results presented indicate that although there is a small to moderate reduction in organic matter, soils under pastures have higher negative charge than forests. In addition, the pH in pasture soils is higher than that in forests. The regression models presented earlier clearly demonstrate the role of enhanced pH on variable charge derived mainly from SOM. The apparent increase in pH after conversion suggests that pH-dependant charge is responsible for the increase in ECEC.
To test whether this mechanism (effects of pH on variable charge) could explain the higher ECEC in pasture soils as compared with forests, ECEC was plotted against the percentage of C separately for forests and pastures (Fig. 7) . The ECEC–soil C relationship appears to be shifted upwards for pastures. Intercepts were significantly different at P < 0.05. This change is primarily attributed to the pH effects on SOM acid functional groups in addition to enhancing the charge of mineral components. In pastures, the impact of moderate reduction in SOM on ECEC (lesser number of potential exchange sites) has been more than compensated by an elevation in pH, which enhances deprotonation of the acidic functional hydroxyls and the inorganic mineral hydroxyls. This mechanism has been proposed for highly weathered soils elsewhere (Mendonca and Rowell, 1996).
|
pH are somewhat similar to the trends in ECEC discussed above (Fig. 4). This suggests that forest conversion to pasture has changed the nature and magnitude of the exchange complex with an increase in negative charge and a decrease in positive charge.
A part of the differences in ECEC, especially at >70-cm depth, could be attributed to differences in inherited properties between pastures and forests, since certain types of areas may have been cleared preferentially. However, there is no evidence that this factor played a major role in the current study. Ionic strength differences because of land-use change could also affect effective charge, however the methods used in this study cannot explicitly account for this. We however attribute the major part of the observed differences in measures of effective charge (ECEC and pH) to effects of land-use change and subsequent management on variable charge properties especially in the upper 30 cm.
Exchangeable Cations
Additional ECEC sites at the elevated pH under pastures have retained an increment of 19.3 ± 17.68 kmolc ha-1 exchangeable Ca (see also Fig. 3). Pasture soils also tended to have higher levels of exchangeable K (10.82 kmolc ha-1) than forests (3.33 kmolc ha-1). If one block was left out, soils under pasture had significantly higher overall mean levels compared with pastures (P < 0.1). Pasture tended to have less exchangeable Al than forests in the upper 30 cm (Fig. 3). In general, nutrient cations have replaced some of the exchangeable Al and acidity on cation-exchange sites. Calcium saturation of exchange sites on a content basis increased from 23 to 47% (P < 0.06) in the upper 30 cm.
The generation of additional ECEC in pastures has retained Ca from the burnt forest biomass and subsequent grass-fire dynamics. A compilation of 20 studies of humid tropical forests world-wide (Ewel et al., 1981; Alegre et al., 1988; Nykvist, 1998) indicates that Ca2+ in tropical forest biomass and burnt ash is 62.6 ± 46.85sd kmolc ha-1. Thus, the difference in exchangeable ECEC (39 kmolc ha-1) in forest and pastures would potentially be able to retain a substantial part of the Ca from forest biomass alone.
Although nutrient cations associated with ECEC may have been derived from the ash of forest biomass, and subsequent cattle manure inputs, there are other potential sources as well. Nutrient retention may well be efficient in the grass pastures, and nutrients taken up by roots at depths >1 m in the soil have often been underestimated (Markewitz and Davidson, 1997). These inputs of nutrient cations would be retained by extensive grass rooting and enhanced ECEC of near surface soils.
Aggregate Size and Stability
The size distribution of the air-dry aggregates (1–5 mm) is generally similar for pastures and forests throughout the 0- to 30-cm soil depth (Table 2). Nearly 90% by weight of the 1- to 5-mm aggregates have effective diameters of 1 to 3 mm under pastures and forests. Any observed land-use differences in water stability of aggregates cannot be attributed to initial differences in aggregate size or soil texture.
|
|
pH, pHCaCl2 – pHH2O) was investigated in explaining the observed variability in water-stable aggregates >0.25 mm. Of these two, soil C emerged as a stronger covariate to aggregate stability (Fig. 8). A nonlinear fit explained 34% of the variability in percentage of water-stable aggregates >0.25 mm (P < 0.05), whereas only 19% was attributed to pH (P < 0.06). In the aggregate-size range 1 to 5 mm, land-use differences between forests and pastures are detectable for all three sieve sizes: 0.25, 1, and 2 mm (Fig. 9) . This is significant because air-drying may suppress land-management effects on aggregate stability of highly weathered soils (Kemper and Rosenau, 1986; Beare and Bruce, 1993). Larger soil aggregates (e.g., >2 mm) appeared to be much less water stable under pastures than under forest (Fig. 9).
|
|
CONCLUSION |
|---|
|
TOPABSTRACTINTRODUCTIONSite DescriptionMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONREFERENCES |
|---|
This study, like many other ecological studies, has been based on the space-for-time substitution or chronosequence approach (Pickett, 1989). Results from such studies need to be cautiously interpreted. Although they produce useful knowledge and insight, they are no substitute for direct observations of ecosystems undergoing change (Richter and Markewitz, 2001). Ecological scientists need an efficient network of soil-ecosystem experiments that will quantify soil and ecosystem change over time scales of decades. A global network of efficiently run field experiments is urgently needed across a range of soil textures, clay mineralogies, plant species, climates, and management regimes.
|
ACKNOWLEDGMENTS |
|---|
Financial and logistic support was obtained from the Duke-UNC Center for Latin American Studies, Duke's Center for International Studies, SIDG of the Nicholas School of the Environment, and the TRIALS III project of U.S. Aid for International Development and OET (Organization Para Estudios Tropicales). Paul Heine supervised analytical work at the soils laboratory at Duke University, and Jane Raikes assisted in mineralogical analyses. Michael Hofmockel helped with figures and formatting, and Dan Markewitz commented on an earlier draft. The Wildlife Institute of India provided support and facilitated the revision of this paper.
Received for publication August 5, 1999.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONSite DescriptionMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. E. Russell, J. W. Raich, O. J. Valverde-Barrantes, and R. F. Fisher Tree Species Effects on Soil Properties in Experimental Plantations in Tropical Moist Forest Soil Sci. Soc. Am. J., June 29, 2007; 71(4): 1389 - 1397. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Vadose Zone Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
under forest in the Rio General Valley, Costa Rica.
