DISTRIBUTION AND VERTICAL STRATIFICATION OF CARBON AND NITROGEN IN SOIL UNDER DIFFERENT MANAGEMENTS IN THE PAMPEAN REGION OF ARGENTINA

Álvarez, Carina Rosa; Costantini, Alejandro Oscar; Bono, Alfredo; Taboada, Miguel Ángel; Gutiérrez Boem, Flavio Hernán; Fernández, Patricia Lilia; Prystupa, Pablo DISTRIBUTION AND VERTICAL STRATIFICATION OF CARBON AND NITROGEN IN SOIL UNDER DIFFERENT MANAGEMENTS IN THE PAMPEAN REGION OF ARGENTINA Revista Brasileira de Ciência do Solo, vol. 35, núm. 6, 2011, pp. 1985-1994 Sociedade Brasileira de Ciência do Solo Viçosa, Brasil


INTRODUCTION
Temperate agricultural regions across the globe are under increasing pressure to maintain, if not raise, production levels in order to meet the rising food demand (Watson & Zakri, 2005).One of the most productive areas is the Rolling Pampas region of Argentina (about 4.6 Mha), where cereal and oil crop production has increased three-fold in the last two decades, at the expense of pasture for cattle production (Satorre, 2005).Intensive agricultural production can adversely affect soil quality, as shown by decreasing total and active organic C levels (Echeverria et al., 1993;Casanovas et al., 1995;Costantini et al., 1996;Diaz Zorita, 1999).The recognition that crop production under intensive tilling can promote soil degradation and raise fuel costs, decreasing net profit, has led to the implementation of no-tillage (NT) systems on about 95 Mha around the world (Lal et al., 2007), including 14 Mha in Argentina (Satorre, 2005).One of the expected benefits for soil under NT is a higher soil C sequestration (Lal et al., 2007).However, higher C retention in soil is not always apparent when NT is applied, due mainly to substantial differences in climatic conditions, soil type, initial C content and crop rotation patterns (Steinbach & Alvarez, 2006).Much of the cultivated area of the Pampas was converted to NT management after decades of conventional tillage, during which at least half of the original total soil organic C (TOC) was lost (Soriano et al., 1991;Senigagliesi & Ferrari, 1993).Despite the known advantages of NT farming, few data are available on the TOC recovery potential of NT farmland in the Pampas with a long-term history of conventional tillage.
The erratic response of soils to NT management could be ascribed to the characteristics of TOC, which does not respond as readily to alterations in tillage management as other more active C pools with shorter turnover times (Quiroga et al., 2001).This active C pool can be analyzed in different ways, for example by quantifying the in vitro soil respiration, measured as C-CO 2 production under controlled temperature and moisture conditions (Álvarez et al., 1995).Another method of quantifying the C pool is by measuring particulate organic matter in wet-sieved (53 µm opening) aggregates (Quiroga et al., 1996).The C pool determined with this method indicates the initial loss of organic matter when undisturbed, never cropped soils are first taken into agricultural production (Cambardella & Elliott, 1992).Soil organic matter stratified with depth is of the indicators of soil quality (Franzluebbers, 2002).The main goal of the current study was to evaluate the potential of no-tillage management to increase the total organic C stock and its fractions in soils of the Pampas region of Argentina.It was hypothesized that the active C pools are more influenced by changes in soil management than the total organic C pool and that they are clearly stratified with depth.

Study area and experimental design
The mean annual rainfall in the Rolling Pampa of Argentina is 940 mm, concentrated in spring and summer, and the mean annual temperature is 17 °C (Soriano et al., 1991).The soils, developed over aeolian sediments (loess) and grassland vegetation (Soriano et al., 1991), belong predominantly to the Mollisol Order (US Soil Taxonomy).In the West, the soils have a loamy A and a Bt horizon (i.e.Typic Argiudoll), while in the East they have silty loam and silty clay loam A horizons and a Bt horizon with higher clay content than the western horizons (i.e.Abruptic and Vertic Argiudolls;Salazar Lea Plaza & Moscatelli, 1989).The soil pH is slightly acidic (around, 6 ± 0.2) and base saturation higher than 70 %.Forty cropped fields (22 under no-tillage and 18 under conventional tillage) and seven undisturbed locations, of the selected soil types and management systems, were sampled across the region (Figure 1).The surface (0-30 cm) texture of the studied soils is shown in table 1.Three management types were compared: conventional tillage (CT, using chisel or disk plow), no-tillage (NT) and undisturbed soils (UD).The undisturbed soils corresponded to enclosed, undisturbed areas around country houses and long-term pastures (> 20 yr old) excluded from grazing or grazed at a very low stocking rate.Vegetation consisted predominantly of perennial grasses.
The selected production farms used similar technologies and procedures in terms of, for example, fertilizer rates and herbicide and pesticide application.On some fields, a 2-year crop sequence of maize/mainmain-season soybean planted, while on others a 3 year rotation of maize/wheat-soybean/ main-season soybean or year-round soybean was maintained.Mean fertilizer rates were 50 kg ha -1 N (as urea or urea-NH 4 NO 3 ) and 16 kg ha -1 of P (as diammonium or monoammonium phosphate), applied to maize and wheat.Periods of UD and CT management always lasted several decades, while continuous NT periods lasted 5-18 yr.At least four years of continuous NT management is required to improve soil physical conditions in the surface layers (Voorhees & Lindstrom, 1984;Rhoton, 2000).Under NT, weeds were controlled with herbicides (mainly glyphosate and atrazine) and under CT mechanically and chemically.R. Bras. Ci. Solo, 35:1985-1994, 2011

Analytical determinations
Soil samples were collected in previously outlined homogeneous areas (2-3 ha) of each field.Composite samples contained at least 30 sub-samples from the 0-5, 5-15 or 15-30 cm soil layers, collected with an auger.Samples were taken during the fallow period, between May and July.The measurements were performed in this period to ensure soil reconsolidation after tillage under CT.These composite samples were analyzed for TOC by the Walkley and Black method (Nelson & Sommers, 1982) and total nitrogen (TN) by the Kjeldhal method.Particle size was determined by the pipette method; samples were pre-treated with HCl and H 2 O 2 to destroy carbonates and organic matter, respectively.Additionally, soil aggregates in the composite samples were fractionated by wet sieving to determine particulate organic carbon (POC) and particulate organic nitrogen (PON) in the > 53 μm aggregate-size fraction (Quiroga et al., 1996;Cambardella & Elliot, 1992).The resistant organic carbon (ROC) and resistant organic nitrogen (RON) concentrations were calculated as the difference between TOC and POC or TN and PON, respectively.In vitro C mineralization was measured after 10 d at 30 ºC and 50 % of water-holding capacity in the layers 0-5 and 5-15 cm.C-CO 2 production was determined by alkali absorption (Álvarez et al., 1995).In this study, C respired in 10 d was considered representative of the potentially mineralizable C (Franzluebbers et al., 1995).Soil bulk density was determined by the core method (Burke et al., 1986) using 230 cm 3 volume cores (n = 4) of the layers 0-5, 5-15 and 15-30 cm.
Stratification ratios (SR) were calculated as the quotient between concentrations in 0-5 cm and 5-15 cm or 0-5 cm and 15-30 cm of TOC, TN and their labile pools.

Statistical analysis
Data were analyzed by ANOVA for each sampling layer.Homogeneity of variance was always tested.The associations between soil properties were evaluated using simple regressions (Neter & Wasserman, 1974).Principal component (PC) analysis was performed to find the principal gradients in variability among sites.Gradients are linear combinations of the properties TOC, POC, ROC, TON, PON, RON, and PMC, expressed as equivalent mass and their SR 0-5/5-15 cm.

RESULTS
The organic C and N pools were not affected by "soil management x soil texture" interactions (Table 2).The effect of both factors can therefore be analyzed separately.The total quantities of TOC, TN and labile C and N pools (per hectare, equivalent soil mass, Neill et al., 1997) in soil were always significantly higher in UD than in agricultural soils.NT and CT soils were not significantly different (Table 2).
The cumulative concentrations of the C and N fractions are shown by figure 2a,b.In the 0-5 cm and 5-15 cm layers, both soil TOC and soil TN pools were significantly (p ≤ 0.05) higher in UD soils than in both agricultural soils (NT and CT), although differences between UD and the cropped soils were less substantial in 5-15 cm soil than in the surface layer.TOC concentration did not vary among soil managements in the 15-30 cm layer, while TN was significantly lower in NT soils.The active C fraction (POC) differed significantly among the three management methods in the 0-5 cm layer (UD > NT > CT) and was significantly higher in UD than in agricultural soils in the 5-15 and 15-30 cm layers.The ROC was also significantly higher in UD in the 0-5 and 5-15 cm layers but there were no significant differences in the 15-30 cm layer.Similarly, both N fractions (PON and RON) were significantly higher in UD than in the agricultural soils in the 0-5 and 5-15 cm layers, but there were no significant differences in the 15-30 cm layer.
Soil PMC was significantly higher in UD soils than in agricultural soils in the 0-5 cm (UD > NT = CT) and in the 5-15 cm (UD > NT < CT) layers (Table 3).More than 60 % of the variation in PMC was explained by either TOC or POC (Figure 3a,b).Although there also seemed to be a positive relationship between PMC and soil ROC, there was considerably more variation in the data, with a coefficient of determination (R 2 ) of 0.233 (Figure 3c).
Stratification ratios from the 0-5 cm to the 5-15 cm layers ranged from 0.9 to 1.6 when calculated with total and resistant C and N pools, such as TOC, TN, ROC and RON (Figure 4a,b).When calculated using labile C and N pools such as POC, PMC and PON, SR were higher, with values exceeding 2 (Figure 4a,b).The TOC 0-5/5-15 cm SR differed significantly between soil management strategies (1.59 in UD, 1.36 in NT and 1.13 in CT soils).The same trend was observed for TN.For SR 0-5/15-30 cm, differences between the tillage systems disappeared and the ratios increased in response to substantial decreases of TOC and TN concentration with depth.Values were rarely > 2 when calculated with total and resistant C and N pools and reached 4.7-5.3when calculated with the labile C and N pools (Figure 4c,d).

DISCUSSION
The absence of soil disturbance is critical to maintain C and N levels high (Hermle et al., 2008).The studied UD soils partially reflect the theoretical conditions of the Pampas prior to the introduction of agriculture by the end of the 19 th century (Soriano et al., 1991).Long-term agriculture during most of the 20 th century caused a significant decrease in TOC and TN stocks, as evidenced by the data of the cropped soils (Table 2).Differences in C and N levels between UD and the cropped soils were highest in the surface soil layers and essentially absent in the lowest 15- R. Bras. Ci. Solo, 35:1985-1994, 2011 30 cm layers.This pattern is consistent with similar temperate soils (Dalal & Mayer, 1986).
Repeated tillage and surface traffic alter the nature of the soil, causing compression and compaction.Reducing tillage and the associated vehicular traffic would expectedly cause differences in soil bulk density.Therefore, soil TOC and TN stocks were calculated to a depth of 30 cm, according to recommendations of Sisti et al. (2004).TOC in UD soils was significantly higher than in CT and NT soils, with no differences between the latter two (Table 2).A decrease of around 23 % in TOC was observed as a consequence of agricultural activity.Hermle et al. (2008) found similar C stocks at depths down to 40 cm under different tillage systems; only pasture had higher C stocks than agricultural soils.D´Haene et al. ( 2009) reported no differences in C sequestration under different tillage methods for the 0-0.6 m layer in Belgium, which represents a colder climate than the Rolling Pampa of Argentina.
It was hypothesized that after some years of continuous NT, organic C in soils could recover, slowly rising to pre-tillage levels.In a review of long-term tillage trials in the Pampas, Steinbach & Alvarez (2006) reported that C stocks (equivalent soil mass) under no-tillage were 5 % higher than traditionally plowed soil.Since our study included only production fields, variables other than tillage may have affected results, such as the history of a given area prior to the introduction of no-tillage.In addition, in the CT systems studied here chisel and harrow disk tillage were used, the most commonly used tillage systems in the region.Minimized tillage with these techniques, as opposed to moldboard plowing, induces C sequestration (West & Post, 2002).However, seemingly contradictory results regarding the impact of tillage method on C sequestration can be attributed to the form of data processing, specifically, calculation on a concentration or a mass basis (Sisti, 2005).Baker et al. (2007) reported that in studies in which the deeper soil layers were evaluated generally no higher C sequestration was observed in NT than in CT soils.In fact, TOC may be higher in the deeper layers of CT than NT soils, although this case was not observed in our study.
The accumulation of the labile and resistant C and N fractions in the upper layer (0-5 cm) of UD soils may be due to the higher C inputs from perennial grasses, the decrease of C input by roots with depth (Jobbagy & Jackson, 2000), and the lack of disturbance, compared with cropped soils.In general, when soil is cropped, the organic C content is adversely affected, because the organic matter input is reduced, and the decomposition of soil organic matter is accelerated by tillage (Gupta & Germida, 1988;Richter et al., 1990).
The amount of labile organic matter is generally higher under conservation tillage systems since residue and organic matter decomposition rises in concert with the amount of tillage ?increase according to the degree of tillage disturbance (Richter et al., 1990).Our data showed higher soil POC, an active C pool, in the surface layer of NT compared to CT soils (Figure 2).Because crop residues are not incorporated into the soil in no-tillage systems and due to the absence of mechanical disruption, there is an increase in POC in the top 5 cm (Alvarez et al., 1998;Fabrizzi et al., 2003;Steinbach & Alvarez, 2006).Because of this increase and in view of the greater sensitivity of POC compared with TOC or ROC, soil POC could be considered an early indicator of the potential future TOC increase under NT, as well as in other C labile fractions such as microbial biomass C (Sparling et al., 1992) or potentially mineralizable C (Franzluebbers et al., 1995).In contrast to POC, soil PON, the active nitrogen pool, was not enriched in the surface soil layer under NT.
The TOC and TN concentrations in NT and CT soils were similar in the top 5 cm (Figure 2), in contrast to some other studies which have reported that TOC and TN concentrations were higher in NT than in CT soils (Franzluebbers, 2002;Hermle et al., 2008), but in agreement with another study developed in the Argentine Pampas, in a 10 year experiment sampled at three depths (Costantini et al., 2006).The proportion of POC in TOC was higher in the surface of NT and UD than of CT soils, a trend that correlated with the lack of soil mechanical disturbance (Alvarez et al., 1998;Fabrizzi et al., 2003;Steinbach & Alvarez, 2006).
There were some differences in soil ROC and RON pools among the management strategies (Figure 2).In UD soils, more ROC and RON was found in the 0-5 cm layer than at the agricultural sites.Hermle et al. (2008) also observed this trend for the ROC fraction.The differences among management strategies they reported were not as pronounced as ours, but their study considered the 0-10 cm layer.The soil management strategy did not affect soil ROC below 5 cm.Soil PMC was also significantly higher in UD than in agricultural soils.Within this group, soil PMC was similar in CT and NT soils in the 0-5 cm layer but was significantly higher in CT, relative to NT, in the 5-15 cm layer.Similar results were found by Costantini et al. (2006) for soils of the humid Pampean region of Argentina.Costantini, (2003) also reported higher TOC and PMC values in uncropped than in agricultural soils, regardless of the tillage system in use.Much of the variation in soil PMC was explained by TOC and POC concentration (Figure 3).Alvarez & Alvarez (2000) detected a close relationship between mineralized C and the light C fraction (density < 1.6 g cm -3 ) and TOC.Several studies have suggested that the physical fractionation of the C components is useful to assess short-term changes in soil conditions induced by modifications of the management strategies (Alvarez & Alvarez, 2000).Mc Lauchalan & Hobbie (2004) found a high correlation between the light C fraction (< 1.7 g cm -3 ), hydrolysable C, soil microbial biomass, 12 d respired C and the labile C pool.According to the correlation found between POC and PMC (Figure 3), soil POC can be considered a good indicator of the C-labile pool.
Stratification ratios varied between the three soil management methods (Figure 4).Stratification was always highest in UD soils, which can be considered the most representative of the original, pre-agriculture soil conditions.Stratification was generally lowest in CT soils, in which tillage homogenized the soil components, leading to a more even distribution of C  and N. Other studies reported similar trends (Franzluebbers, 2002).Stratification was expected to increase in NT soils where crop residues are not mechanically incorporated into the soil.Franzluebbers (2002) proposed using the stratification rate as a soil quality indictor.Sá & Lal (2009) found a close relationship between the stratification rate and soil organic C stock.In our study, stratification ratios in NT were only significantly higher than in CT soils when calculated from the 0-5 to the 5-15 cm layers.Some variables (i.e.ROC, RON and PON) showed no differences between CT and NT soils, while others (i.e.TOC, POC and PMC) seemed to recover under NT management, returning to the high stratification ratio characteristic of UD soils.The degree to which stratification increased in UD soils was higher for soil POC and PMC, indicating these C pools as the most suitable parameters for assessments of variations in topsoil organic matter quality.This increase in stratification is noteworthy, since major differences are generally identified in soils of warm, humid climates with low TOC content (Franzluebbers, 2002); these conditions were however not observed in the studied areas.The magnitude of the stratification patterns identified in our study was similar to results reported by Sá & Lal (2009).
The principal component (PC) analysis for organic C and nitrogen stocks (TOC; POC; ROC; TN; PON; RON and PMC) and their SR 0-5/5-15 cm (Figure 5) showed that 50 % of the variance was explained by the first two axes and 33 % by the first axis.Soil total organic pools (TOC and TN) and the labile pools (PCM, POC and PON) were positively weighted.This first PC axis clearly distinguished UD from agricultural sites (CT and NT).This result is in agreement with the ANOVA analysis.The second PC axis explained 22 % of variance.The main components of this axis were the SR of TOC, TN and their fractions.This second PC axis distinguished NT from CT sites.

CONCLUSIONS
Continuous NT management indicated no recovery of TOC in the Pampas of Argentina.However, there was a substantial stratification of active organic C pools at the NT sites, mainly when the stratification ratio was based on the two surface layers, indicating POC and PMC as useful properties for assessments of variations in topsoil organic matter quality.Much of the variation in soil PMC was explained by POC, so this latter fraction can be considered a good indicator of the C-labile pool.PC analysis showed that total organic pools and their labile fractions allowed a distinction of UD from agricultural sites, while SR were higher in NT than in CT.

Figure
Figure 2. Concentrations of (a) total organic carbon content (TOC) as the sum of particulate organic carbon (POC) and resistance organic carbon (ROC) and (b) total nitrogen (TN) as the sum of particulate organic nitrogen (PON) and resistance organic nitrogen (RON) in soil from three different layers and for different management strategies.NT: untilled soils (n=22); CT: conventionally tilled soils (n=18); UD: undisturbed soils (n=7).Different letters indicate statistical differences between management strategies within a soil layer (Tukey test p < 0.05).Each C and N component, as well as total C and N, were analyzed separately.

Table 1 . Clay, silt and sand content of the studied soils of the different texture groups (0-30 cm) Table 2. Quantities (stocks) of total organic carbon (TOC), total nitrogen (TN), particulate organic carbon (POC), resistant organic carbon (ROC), particulate organic nitrogen (PON) and resistant organic nitrogen (RON) per unit of area corrected by equivalent mass for the different managements and soil texture groups
UD (undisturbed soil), NT (no tillage), CT (conventional tillage).Different letters in the same column indicate significant differences for management systems and soil texture (Tukey test, p < 0.05).