Previous PageTable Of ContentsNext Page

Soil carbon sequestration and mulch-based cropping in the Cerrado region of Brazil

Marc Corbeels1,2, Eric Scopel1,2, Alexandre Cardoso2, Jean-Marie Douzet1, Marcos Siqueira Neto3 and Martial Bernoux4

1 Agro-ecosystem Program, CIRAD, www.cirad.fr Email corbeels@cirad.fr
2
Embrapa-Cerrados, www.cpac.embrapa.br Email alexandre.cardoso@embrapa.br
4
CENA-USP, www.cena.usp.br Email msiquer@cena.usp.br
4
SeqC Research Unit, IRD, www.ird.fr Email: .martial.bernoux@ird.fr

Abstract

Direct seeding mulch-based cropping (DMC) systems have largely been adopted over the last 10 to 15 years in the Cerrado region of Brazil as a means to combat soil degradation. Continuous cropping and the absence of soil tillage under DMC favour the sequestration of C in soils. The objective of this study was to assess the soil C sequestration potential of the DMC systems in the Cerrados based on data from a chronosequence and simulation of soil C dynamics using the G’DAY plant-soil model that includes the CENTURY decomposition submodel. Soil C sequestration rate derived from linear regression of all soil C contents measured on a DMC chronosequence between age 0 and 12 in the upper 0-20 cm topsoil layer was 0.83 Mg/ha/yr. The corresponding change in soil total N was +79 kg/ha/yr. Model results indicate that C is lost at rates of about 1.25 Mg/ha/yr from the 0-40 cm soil layer after conversion of native vegetation to soybean or maize monocropping under conventional tillage for 30 years. In contrast, DMC systems with pearl millet as cover crop were adequate to maintain high soil C levels as under the native vegetation. This was largely attributed to the high net primary productivity (NPP) and relatively small removal of NPP as harvest. Gains in modeled C under DMC were sustained by increased N input and/or decreased N losses.

Media summary

Soil C sequestration using direct seeding mulch-based cropping (DMC) in the Cerrado region of Brazil is approximately estimated at 3.3 Tg/yr.

Key words

Soil organic matter, soil total nitrogen, cover crop, no-tillage, simulation modelling, agroecosystems

Introduction

The tropical wooded savannah of central Brazil (Cerrados) occupies approximately 23 % of the national territory. Since the 1970s the region has been the focus of intense agricultural expansion, and nowadays about 30 % of the native vegetation has been replaced by agricultural cropland or pasture. To combat soil erosion and degradation, no-till or direct seeding mulch-based cropping (DMC) systems have largely been adopted over the last 10 to 15 years, and today over 4 million ha are cultivated using DMC practices (Evers and Agostini 2001). In recent times, much attention has been given to alternative tillage and cropping systems as a means to mitigate agricultural emissions of CO2 (Paustian et al. 1997). DMC systems represent a potential for soil organic carbon (SOC) sequestration by increasing C inputs to the soil, reducing C losses due to soil erosion and by decreasing decomposition rates of SOC as a result of reduced mechanical soil disturbance. The objective of this study was to assess the soil C sequestration potential of the DMC systems in the Cerrado region of Brazil. The approach was firstly to use data from a long-term DMC chronosequence to quantitatively estimate soil C sequestration rates; and then to apply a simulation model of C and N transformations in soil-plant systems to further analyse and explore the soil C sequestration potentials for a restricted, but representative set of agricultural cropping systems.

Methods

Field site and sampling

The study area was located in the municipals of Rio Verde (17° 47‘S, 51° 55‘ W) and Montividiu (17° 24‘S, 51° 14‘ W) of the Goiás state, on a plateau in the centre of the Cerrado region. Mean annual rainfall is about 1,600 mm, with a dry season from May till September. Mean minimum and maximum temperatures during the growing season are 17 and 27°C, respectively. The study area covered about 5,000 ha of cropland that had been converted from native savannah woodland about 25 years ago. Early cropping consisted of soybean monoculture with ‘disk’ tillage, while the first DMC systems were introduced in the region about 13 years ago. DMC entails no-tillage with the implementation of a cover crop (millet, sorghum or maize) following the main crop (soybean or maize). The chronosequence included 45 fields of 0 to 12 years under continuous DMC. All croplands of the chronosequence had the same soil type (Geri-Gibbsic Ferralsol) and were situated on a similar topography. In 2002, soil samples were collected from the 0 to 5, 5 to 10 and 10 to 20 cm surface soil layers at 3 locations in each field. Samples were air dried and sieved before analysis for organic C and total N by dry combustion in a CHN Perkin-Elmer elemental analyser. Bulk density was measured in the field with volumetric steel rings. Soil particle-size analyses were conducted using the pipette method.

The G’DAY model

We used the Generic Decomposition and Yield (G’DAY, described in detail by Comins and McMurtrie 1993) model to simulate changes in soil C and N with agricultural management. G’DAY is a linked plant-soil model that incorporates the well established CENTURY organic matter decomposition submodel (Parton et al. 1993). The plant submodel in G’DAY represents the C and N content of foliage, wood and roots. The soil submodel represents C and N in four litter pools (structural and metabolic, both above and below-ground) and three soil organic matter (SOM) pools (active, slow and passive). Processes represented include plant C assimilation, plant N uptake, allocation, tissue senescence and N retranslocation, litter and SOM decomposition, soil N mineralisation and immobilisation, N input by atmospheric deposition, biological fixation and chemical fertilisation, and N loss by leaching or gaseous emission.

Model parameterisation

In this study the model was run using mean annual temperature and solar radiation as the climatic driving variables. Net primary productivity (NPP) was simulated based on radiation use efficiency (RUE). For each crop the value for RUE was adjusted to reflect the local crop varieties and growing conditions, and to match site-specific crop yields. We assumed optimal crop yields during the simulation period, based on yield data recorded on farmer’s fields. The fraction of the total plant production allocated below-ground was based on crop-specific average values for root production obtained from the literature. Initial values for the various model pools of organic soil C and N as they might have existed before taking the land into agricultural production were derived from ‘equilibrium’ simulations under native savannah vegetation (type cerrado sensu strictu). General and non-site specific parameters were those from earlier G’DAY model testing (Commins and McMurtrie 1993; Halliday et al. 2003). Based on mean monthly precipitation and potential evapotranspiration an average annual moisture reduction factor of 0.7 was introduced for all decomposition rate constants. Decomposition rates of soil C pools (active, slow and passive) were further decreased with 25% to account for soil depth effects (0-40 cm). Tillage effects were simulated by transferring a fraction of above-ground crop residue material (80 %) into the soil, and by increasing the decomposition rates of the soil C pools. In this study, they were set 1.2 times higher relative to no-tillage in the month after the tillage operation. Model parameters associated with crop residue quality were obtained from measurements on the sites.

Results and discussion

Soil C and N in surface soil

Soil organic C concentrations in surface soils (0-20 cm) were related to clay+silt content and years under DMC (76 % of variation accounted for). A similar model explained 52 % of the variation in soil total N concentrations (0-20 cm) across fields. There was a high correlation between silt+clay content and bulk density (r = 0.76, P < 0.001), and a significant (P < 0.05) change in bulk density with years under DMC. The average increase (P < 0.001) in SOC contents (0-20 cm, corrected for change in bulk density) along the chronosequence was 0.83 Mg/ha/yr (Fig. 1a). Total soil N contents (0-20 cm) increased (P < 0.05) on average by 79 kg/ha/yr under DMC (Fig. 1b). There was no significant (P > 0.1) change in soil C:N (0-20 cm) along the chronosequence. Our figure for soil C sequestration is in line with the 0.86 Mg/ha/yr (0-20 cm depth), estimated by de Sá et al. (2001) for DMC systems on oxisols in sub-tropical southern Brazil. The value is higher than the ‘average’ 0.57 Mg/ha/yr estimated by West and Post (2002) in a global data-analysis of effects of DMC management on soil C sequestration.

Figure 1. Soil (a) organic C and (b) total N contents in surface soils (0-20cm) in a DMC chronosequence in the Cerrado region of Brazil.

Model simulations

Model simulations were based on the assumption that if the model reasonably accurately estimates crop total biomass and grain production, then appropriate C input to soil, which is the major factor in determining soil C, can be assessed, and subsequently SOC dynamics can be adequately simulated.

As an example of model output, Figure 2 shows modeled changes in SOC and soil total N for a set of continuous cropping systems on a typical clayey soil (15 % silt, 70 % clay) of the region during 30 years following conversion of native Cerrado vegetation to cropland.

Figure 2. Modelled changes in soil organic C (a) and soil total N (b) contents in the 0-40cm soil layer for S/F-CT (soybean-bare fallow under conventional tillage); M/F-CT (maize-bare fallow under conventional tillage); S/F-NT (soybean-bare fallow under no-tillage); S/M-DMC (soybean-maize under DMC) and S/P-DMC (soybean-pearl millet under DMC). Time 0 represents steady state conditions under native savannah vegetation (type cerrado sensu strictu).

As expected, soil C sequestration was highest under DMC (S/P-DMC and S/M-DMC, Fig. 2a) and modelled C and N stocks under soybean with pearl millet as cover crop (S/P-DMC) were comparable to those under native vegetation. In contrast, soil C and N contents under systems with bare fallow (S/F-CT, M/F-CT and S/F-NT) were simulated to decline and resulted in 40 to 25 % soil C loss after 30 years. The modelled gains in soil C under DMC compared to systems with bare fallow were attributed to a higher NPP caused by greater cropping frequency, and less removal of NPP as harvest. Despite the higher biomass production and much larger above-ground residue input under maize than under soybean, modelled decline in soil C after conversion from native vegetation was comparable. This is in agreement with experimental findings (Sisti et al. 2004) and was attributed in the model to the lower root:shoot ratio for maize compared to soybean. Root production and turnover is considered to be an important determinant of soil C storage (Balesdent and Balabane 1996). Simulated effects of tillage on soil C sequestration were quite substantial: no-till soybean (S/F-NT, Fig. 2a) cropping reduced the C losses with about 40 % compared to soybean cropping with conventional tillage (S/F-CT) after 30 years. Gains in soil C were related to gains in soil N (Fig 2b): i.e. gains in NPP were sustained by increased N inputs and reduced N losses under DMC. This confirms the findings by Sisti et al. (2004) that when the net N balance of the cropping systems is close to zero or negative no long-term accumulation of soil C is to be expected.

Conclusions

This study shows the potential of DMC systems for promoting the conservation of SOC in the tropical Cerrado region of Brazil. This is attributed to high crop residue input and the lack of soil disturbance. Gains in SOC are sustained by gains in soil total N, as a results of increase N input or reduced N losses. The C sequestration rate is estimated to be 0.83 Mg/ha/yr. In the Cerrados over 4 million ha (in 2001) of cropland are cultivated using DMC system. Therefore, soil C sequestration with DMC in this region can be roughly estimated at 3.3 Tg/yr, which is equivalent to CO2 12.2 Tg/yr.

References

Balesdent J and Balabane M (1996). Major contribution of roots to soil carbon storage inferred from maize cultivated soils. Soil Biology and Biochemistry 28, 1261-1263.

Comins HN and McMurtrie RE (1993). Long-term response of nutrient-limited forests to CO2 enrichment: equilibrium behavior of plant-soil models. Ecological Applications 3, 666-681.

Evers G and Agostini A (2001). No-tillage farming for sustainable land management: lessons from the 2000 Brazil Study Tour. TCI Occasional paper Series N° 12, Soil Fertility Initiative 26pp.

de Sá JCM, Cerri CC, Dick WA, Lal R, Filho SPV, Piccolo MC and Feigl BE (2001). Organic matter dynamics and carbon sequestration rates for a tillage chronosequence in a Brazilian Oxisol. Soil Science Society of America Journal 65, 1486-1499.

Halliday JC, Tate KR, McMurtrie RE and Scott NA (2003). Mechanisms for changes in soil carbon storage with pasture to Pinus radiata land-use change. Global Change Biology 9, 1294-1308.

Parton WJ, Scurlock JMO, Ojima DS, Gilmanov TG, Scholes RJ, Schimel DS, Kirchner T, Menaut J-C, Seastedt T, Garcia Moya E, Kamnalrut A and Kinyamario JI (1993). Observations and modeling of biomass and soil organic matter dynamics for the grassland biome worldwide. Global Biogeochemical Cycles 7, 785-809.

Paustian K, Collins HP and Paul EA (1997). In 'Soil organic matter in temperate agroecosystems'. (Eds. EA Paul, K Paustian, ET Elliott and CV Cole) pp. 15-49. (CRC Press, Boca Raton, FL).

Sisti CPJ, dos Santos HP, Kohhann R, Alves BJR, Urquiaga S, and Boddey RM (2003). Change in carbon and nitrogen stocks in soil under 13 years of conventional or zero tillage in southern Brazil. Soil and Tillage Research 76, 39-58.

West TO and Post WM (2002). Soil organic carbon sequestration rates by tillage and crop rotation: a global data analysis. Soil Science Society of America Journal 66, 1930-1946.

Previous PageTop Of PageNext Page