Simulating trends in soil organic carbon of an Acrisol under no-tillage ...
Geoderma 120 (2004) 283 – 295 www.elsevier.com/locate/geoderma
Simulating trends in soil organic carbon of an Acrisol under no-tillage and disc-plow systems using the Century model Luiz Fernando Carvalho Leite a, Eduardo de Sa´ Mendoncßa b,*, Pedro Luiz Oliveirade de Almeida Machado c, Elpı´dio Ina´cio Fernandes Filho b, Ju´lio Ce´sar Lima Neves b a
Empresa Estadual de Pesquisa Agropecua´ria da Paraı´ba-EMEPA, Rua Eurı´pedes Tavares, 210, 58013-290 Joa˜o Pessoa, Paraı´ba, Brazil b Departamento de Solos, Universidade Federal de Vicßosa, 36571-000 Vicßosa, Minas Gerais, Brazil c Embrapa Solos-EMBRAPA, Rua Jardim Botaˆnico 1024, 22460-000 Jardim Botaˆnico, Rio de Janeiro, Brazil Received 22 October 2002; received in revised form 22 July 2003; accepted 17 September 2003 Available online 10 December 2003
Abstract Soil organic matter (SOM) and its different pools have key importance in nutrient availability, soil structure, in the flux of trace gases between land surface and the atmosphere, and thus improving soil health. This is particularly critical for tropical soils. The rates of accumulation and decomposition of carbon in SOM are influenced by several factors that are best embodied by simulation models. However, little is known about the performance of SOM simulation model in an acid tropical soil under different tillage systems including no-tillage (NT). Our objective was to simulate soil organic matter dynamics on an Acrisol under no-tillage and different plowed systems using Century model. Tillage systems consisted of no-tillage, disc plow, heavy disc harrow followed by disc plow, and heavy disc harrow. Soil C stocks simulated by Century model showed tendency to recovery only under no-tillage. Also, simulated amounts of C stocks of slow and active pools were more sensitive to management impacts than total organic C. The values estimated by Century of soil C stocks and organic carbon in the slow and passive pools fitted satisfactorily with the measured data. Thus fitted, except for the active pool, Century showed acceptable performance in the prediction of SOM dynamics in an acid tropical soil. D 2003 Elsevier B.V. All rights reserved. Keywords: Acidic tropical soils; Soil carbon fractions; Long-term experiments; Century model
1. Introduction Soil organic matter (SOM) is an important component of acid tropical soils and its significance can be seen on the positive effects on Ferralsol cation exchange capacity, size and stability of aggregates, * Corresponding author. E-mail address: [email protected] (E. de Sa´ Mendoncßa). 0016-7061/$ - see front matter D 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.geoderma.2003.09.010
improved moisture and soil structure for plant growth (Goedert et al., 1997). Also, SOM has large influence on the flux of the trace greenhouse gases between land surface and the atmosphere (Batjes, 1996). Soil cultivation often leads to diminution of SOM content (Castro Filho et al., 1991), but conservation tillage such as no-tillage can improve soil conditions to those found in forest soils (Machado and Silva, 2001). Because of cost reductions and soil erosion control,
284
L.F. Carvalho Leite et al. / Geoderma 120 (2004) 283–295
no tillage is widely used in Brazil over an area of 14 million ha (Pereira, 2002). No tillage, combined with crop rotation involving cover crops, favors the accumulation of plant residues on the soil surface (Machado and Silva, 2001). In a large study conducted in the USA, Kern and Johnson (1993) reported that the widespread conversion of major field crop production from conventional tillage (mouldboard plow) to conservation tillage would change the soil system from a source to a sink of atmospheric carbon. Similar observations were also reported later for European soils (Smith et al., 1997a). These findings show the potential for agriculture to contribute to global carbon mitigation, particularly through no-tillage. The ability to predict the effects of environment (e.g. climate and atmospheric composition) and landuse change on SOM dynamics is of utmost importance in formulating environmental and agricultural policies (Smith et al., 1997b). Modeling is a powerful means to simulate a range of intricate processes and predict soil organic matter changes for long time periods (Paustian et al., 1992). Century is a model of terrestrial C, N, P, and S dynamics that use a four pool SOM submodel. Century has been successfully used in temperate ecosystems (Parton and Rasmussen, 1994; Kelly et al., 1997; Del Grosso et al., 2001), but in spite of some studies about modeling tropical agroecosystems (Parton et al., 1989; Motavalli et al., 1994), no information is available on the use of mathematical models on the dynamics of soil organic carbon in Acrisol under notillage. Our objective was to simulate soil organic matter dynamics on an Acrisol under no-tillage and different plowed systems commonly used in Brazil using Century model.
2. Material and methods 2.1. Experimental site Simulations with the version 4 of the Century model were carried out for an experimental site established in 1985 at the Experimental Station of the Federal University of Vicßosa, in Coimbra, State of Minas Gerais, Brazil (20j45S and 42j51W; 700 m asl). The mean annual temperature is 19 jC and
average rainfall is 1350 mm, and roughly two-thirds of this rain falls in the warmer season of the year from October to April. The area was covered by Atlantic Forest until 1930 and was cultivated for 54 years with subsistence crops, such as maize (Zea mays L.) and common bean species. The soil in the experimental area is a loamy Acrisol (Argissolo Vermelho-Amarelo, Brazilian Classification System; Typic Kandiudult, US Taxonomy) and some chemical and physical characteristics are shown in Table 1. The experiment started at 1985 and consisted of four soil management systems, arranged in a complete randomized block design, with four replications. Plots were 4 11 m under maize/fallow/soybean succession. The tillage treatments were: 1. No tillage (NT)—no disturbance to the soil other than sowing operation; 2. Disk plowing (DP)—plowing at 20 to 25 cm depth with a three-fixed disk plow, in a single pass; 3. Heavy disk harrow + disk plowing (HHDP)—one single pass at 0– 15 cm using a heavy disk harrow with 20 disks followed by disk plowing at 20– 25 cm depth, with a three-fixed disk plow; 4. Heavy disk harrow (HH)—one harrowing at 10 to 15 cm depth with a heavy disk harrower of 20 disks weighing approximately 2 t; In addition, as a reference, samples were taken from an area under secondary Atlantic Forest (AF), adjacent to the experiment, (100 m away) in the same soil type. At the middleslope, four areas (4 4 m) Table 1 Chemical and physical characteristics of an Acrisol (0 – 20 cm layer) under different tillage systems from the State of Minas Gerais, Brazil Treatmenta pH TOC TN Clay Bulk (H2O) (dag kg 1) (dag kg 1) (dag kg 1) density (mg m 3) NT DP HHPD HH AF
4.97 5.05 5.05 5.00 5.48
1.46 1.28 1.28 1.26 2.83
0.12 0.10 0.10 0.11 0.22
41 38 39 37 46
1.32 1.22 1.21 1.24 1.13
TOC: total organic carbon, TN: total nitrogen. a NT: no tillage; DP: disk plow; HHDP: heavy disk harrow + disk plow; HH: heavy harrow; AF: secondary Atlantic Forest.
L.F. Carvalho Leite et al. / Geoderma 120 (2004) 283–295
were placed along a 100-m transect in which soil samples were collected. 2.2. Chemical analysis Soil samples were collected, in April 2000, just after the harvest period. In each plot, eight topsoil samples (0 – 20 cm) were collected and combined into one composite sample. In the area under Atlantic Forest, 15 samples were collected from each sampling area (n = 4) and bulked into one sample. The samples were ground to pass a 2-mm sieve. An aliquot of 100 g was separated and kept refrigerated at 4 – 8 jC before microbial biomass analysis. For all remaining analyses, the soil samples were air-dried. Total organic carbon (TOC) was obtained by wet digestion with a mixture of potassium dichromate and sulfuric acid, under heating (Yeomans and Bremner, 1988). Total N was measured in the soil samples with a sulfuric digestion followed by determination in the Kjedahl distillation (Bremner, 1996). Measurement of microbial biomass was conducted by the irradiation – extraction method—using microwave (Islam and Weil, 1998), 0.5 mol l 1 K2SO4 as extractant and the biomass C was determined by wet combustion (Yeomans and Bremner, 1988). The factor (KC) used to convert the flow of C for the microbial biomass C (CMB) was 0.33 (Sparling and West, 1988). The CMB was used as an estimate of the active C pool (Paul, 1984; Motavalli et al., 1994). The free-light organic carbon fraction (CLF) was determined by flotation in NaI solution (d = 1.8 g cm 3) as proposed by Sohi et al. (2001). The isolated material was dried at 105 jC for 72 h. The CLF, quantified by dry combustion (Perkin Elmer 2400 CHNS/O elemental analyzer), was further used as an estimate of the slow C pool. Passive C pool was calculated using the following equation: Passive C ¼ TOC ðCMB þ CLF Þ Bulk soil density was determined on nondeformed soil samples collected from a single field replicate. The values of bulk soil density were used to calculate SOM pools based on an equivalent soil mass (Angers et al., 1997; Peterson et al., 1998).
285
2.3. Model parameterization The Century SOM model was originally developed and tested on data sets mainly from grassland and wheat – fallow agriculture in the US Great Plains (Parton et al., 1987, 1988). All parameter values determined from these previous studies were initially left unchanged to provide more criterious evaluations of the simulations. As given by Paustian et al. (1992), these general or non-site specific parameters include the maximum specific decomposition rates for each compartment, the constants that splits the flows of decomposition products and the parameters that control the effects of soil texture, lignin/N ratios, temperature, and moisture on decomposition rates. Site-specific parameters and initial conditions, such as soil texture (sand, silt and clay content), bulk density, soil depth and total soil C and N content, were given values obtained from the field experiment at Coimbra. Monthly precipitation and mean maximum and minimum monthly temperatures from 1967 to 2000 were obtained from the weather station at the Vicßosa Federal University. The parameter determining potential crop productivity was based on the maximum production level observed during the course of the field experiment in each treatment and temperature curve, C/N ratios and lignin contents of biomass pools were obtained through default crop parameterizations distributed with Century model. The main input data for the model are in Table 2. Some model adjustments were made to improve the tillage effects on SOM decomposition. First, the plowing option was adjusted to increase its effect on decomposition (Six et al., 1998). All clteff values (cultivation’s effect on decomposition) for the DP, HHDP and HH were changed from 1.6 to 5 (Table 2). The second change was added to increase the length of time which plowing effects decomposition. Since Century uses a monthly time-step each action only affects SOM dynamics for that specific month although some studies have shown that plowing affects decomposition for several months (Metherell et al., 1995). Thus, an option called ‘‘Additional plowing effect’’ was used in the months following plowing in order to keep the decomposition rates at higher levels (Manies et al., 2000). To initialize the percentage of total SOM in each of the three pools (active, slow and passive) used indirect
286
L.F. Carvalho Leite et al. / Geoderma 120 (2004) 283–295
Table 2 Model input for simulation of tillage systems using version 4 of the Century model Value Soil variables Texture (% sand, % silt, % clay) Bulk density (mg m 3) Initial SOM (g C m 2; C/N) Active surface Active soil Slow soil Passive soil Monthly weather variables Mean total precipitation (cm month 1) Mean maximum temperature (jC) Mean minimum temperature (jC) Cultivation variables Multiplier for increased decomposition Active pool
Slow pool Passive pool
38, 16, 46 1, 13
50, 12,5 159, 9 1776, 22 4460, 12
14 14.8 26.4
1.0 (NT); 1,8 (DP); 2.0 (HHDP); 1.6 (HH) 1.0 (NT); 4.0 (DP); 5.0 (HHDP); 3.0 (HH) 1.0 (NT); 1.8 (DP); 2.0 (HHDP); 1.6 (HH)
Carbon pool values were obtained from direct method. NT: no tillage; DP: disc plow; HHDP: heavy disk harrow + disk plow; HH: heavy harrow.
methods involving simulation of steady state organic matter levels and direct methods using analytical techniques. In the indirect method, Century model parameterized all data including carbon pools for a long term (6000 years) by equilibrium simulation. Simulated values for carbon pools were used as input variables for simulation of land use change. To each treatment, the model simulated SOM dynamics for 54 years representing forest conversion into cropland (e.g. maize and bean production) and later, for 66 years, representing different tillage systems. In the direct method, the initial soil carbon pool sizes under Atlantic Forest were estimated by laboratory analysis and similarly used in the indirect method. All Century estimates were based on a 20-cm depth. Both simulated and measured values for TOC, active, slow and
passive pools in 2000 were subjected to linear regression and Pearson’s correlation. An average of these data sets was taken from each treatment and subjected to a Student’s t-test to determine the significance of the coefficients at the 0.05 and 0.01 probability levels. Simulations for nitrogen pools were also done for NT and HHDP systems.
3. Results and discussion 3.1. Estimates of carbon pools by equilibrium values Century model simulated the equilibrium values of total organic carbon (TOC) and carbon pools (active, slow and passive) for 6000 years. Compared to the initial values, storage of both TOC and passive C pool increased while active and slow carbon pools remained generally constant (Fig. 1). The increase in the passive carbon pool, reflected by TOC, is probably due to carbon pools that are highly recalcitrant, physically protected against microbial attack and less prone to oxidation. After reaching equilibrium, stocks of TOC (64 mg ha 1) and active carbon pool (1.60 mg ha 1) were similar to those measured in the soil under Atlantic Forest (Table 3). The stocks of TOC are 3.9% higher than the value (61.5 mg ha 1) found by Silveira et al. (2000) in another soil under Atlantic Forest in the Piracicaba river basin, Brazil. The estimated value of the slow carbon pool (30.1 mg ha 1) was 42% higher than the measured value in the forest soil. On the other hand, the estimated value of
Fig. 1. Modeled stocks of soil organic carbon (TOC) and organic carbon pools of an Acrisol (0 – 20 cm) under Atlantic Forest.
L.F. Carvalho Leite et al. / Geoderma 120 (2004) 283–295 Table 3 Measured values of total organic carbon (TOC) and carbon pools (active, slow and passive) of an Acrisol under Atlantic Forest (Brazil) Pools
Stocks (mg ha 1)
Total organic carbon Active Slow Passive
63.95 1.59 17.4 44.9
( F 4.25) ( F 0.02) ( F 1.23) ( F 4.01)
Values in parenthesis represent the standard error of the mean.
the passive carbon pool (32.3 mg ha 1) was lower than the measured data (Table 3). Hence, model fit to measured slow and passive carbon pool was not good, probably due to sizes of the passive pool higher in
287
tropical soils than the assumption of the Century model. Starting at initial values of carbon pools obtained from the equilibrium simulation, Century estimates for the stocks of TOC and carbon pools after 120 years decreased after changing forest into agriculture (Fig. 2). Parfitt et al. (1997) studying the effects of clay minerals and land use on organic matter pools obtained similar results, i.e. a decrease of the TOC and C pools after forest clearance and subsequently land use under pasture and maize in an Inceptisol. In 1984, before setting up the experiment, TOC stock was 28 mg ha 1, which were 56% lower than the initial value predicted by the model (64 mg ha 1). In 2000, 15 years after the introduction of the tillage systems,
Fig. 2. Time variation of the stocks of TOC (A) and active (B), slow (C) and passive (D) carbon pools simulated by Century based on the initial values obtained from the equilibrium simulation in the no-till (NT), disc plow (DP), heavy harrow followed by disc plow (HHDP) and heavy harrow (HH).
288
L.F. Carvalho Leite et al. / Geoderma 120 (2004) 283–295
stocks of TOC in the soil under NT (29 mg ha 1) showed an increase compared to the stocks of TOC in the NT soil at the beginning of the experiment. This was in contrast with DP (24.5 mg ha 1), HSPD (23.4 mg ha 1), and HH (25.2 mg ha 1) (Fig. 2). In 2050, stocks of TOC in the soils under NT and plowed systems were estimated as 26 and 15 mg ha 1, respectively, thus showing less soil organic matter degradation under no tillage. Also in both active and slow pools, carbon stocks diminished with the change from forest to cropland. After the introduction of tillage systems, it was observed less significant changes in the carbon stock of active and slow pools. However, the amount of passive carbon stocks in the soils under NT was significantly higher. In 2000, the amount of organic carbon in the soil under NT was 5 mg ha 1 higher than the soils under DP, HHDP and HH systems. This difference increased with time and in 2050 the stocks of TOC in the soil under NT (23 mg ha 1) were estimated as two times higher than the amounts in the soils under DP (12.8 mg ha 1), HHDP (10.6 mg ha 1), and HH (13.2 mg ha 1). These results show the effect of soil disturbance by plowing, which favors higher SOM mineralization rate and thus leads to an increase in humification and in the passive carbon pool. 3.2. Estimates of carbon pools by the measured values In the beginning of the field experiment, the stock of TOC had declined to 40 mg ha 1 (Fig. 3), a decrease of 37%, compared to the 63.95 mg ha 1, initial value measured in the soil under Atlantic Forest. The measured stocks of TOC were 20% lower than the Century model equilibrium estimates of TOC stocks. This difference is probably due to the higher proportion of passive carbon pool in measured TOC. As Century was originally developed for temperate soils it is unlikely to be adjusted to tropical environments because organic –mineral associations in the tropics are different from that observed in the temperate grasslands. Tillage systems did not change the trend to decreasing stocks of carbon. Fifteen years after setting up the field experiment the stocks of TOC in the soils under NT, DP, HHDP and HH were 38, 32, 31, and 34
mg ha 1, respectively (Fig. 3). This tendency continued mainly due to conventional plowed systems and in 2050 the projected stocks of TOC will be 34 mg ha 1 in the soil under NT, and approximately 20 mg ha 1 in the soils under DP, HHDP and HH. Despite these results, soils under NT system were the only ones to show a slight recovery in long term. Our results are corroborated by Smith et al. (2001) that observed at several soils groups and cultures rotation that no tillage resulted, in long term, in the rate of TOC gain rising as high as 0.15 mg ha 1 C year 1 in Black Chernozem and Gleysolic soil groups whereas the conventional tillage showed a loss of TOC. Apparently, crop succession involving fallow, even with no tillage, will demand too long time to reach new equilibrium. This situation, however, may be changed if a crop rotation involving a cover crop to improve mulching (e.g. millet) is included or a ley farming system is adopted. Ley farming involving grass such as Brachiaria may greatly increase the stock of soil organic carbon. In 1984, carbon stocks of the active, slow and passive pools were 0.2, 2 and 36 mg ha 1. These values, compared to those initial values in the soil under forest, represented a decrease of 87%, 89%, and 19%, respectively (Fig. 3). This diminution trend was still observed even after the adoption of NT system, which shows that although the soil has been under NT for 15 years, no soil disturbance without cover crop management as pointed out by Machado and Silva (2001) hardly help to improve an increase of TOC in acidic tropical soils. As reported by Parton et al. (1987), Metherell et al. (1993) and Del Grosso et al. (2001) our results also indicate the high sensitivity of the active and slow carbon pools to changes in soil management parallel to a higher stability of the passive carbon pool. In 2000, compared to the amounts in the beginning of the experiment, the carbon stocks of the active pool in the plowed soils increased approximately 0.3 mg ha 1. However, Century simulation for 50 years showed a decrease in the carbon stocks even in the soil under NT (Fig. 3), indicating that, as reported by Machado and Silva (2001), no tillage without cover crop management (e.g. millet, sun hemp, black oat) hardly improve TOC content of an acid tropical soil. In 2000, the modeled carbon stocks of the slow pool in the soils under NT, DP, HSDP and HH were 2.3, 2.8, 2.9 and 2.7 mg
L.F. Carvalho Leite et al. / Geoderma 120 (2004) 283–295
289
Fig. 3. Time variation of organic carbon stocks (TOC) (A) and active (B), slow (C), and passive (D) pools simulated by Century based on the initial values obtained from direct method in the no-till (NT), disc plow (DP), heavy harrow followed by disc plow (HHDP) and heavy harrow (HH).
ha 1, respectively (Fig. 3). The passive carbon pool showed higher differentiation in carbon content among tillage systems than the active and slow pools. In 2000, the highest values of soil carbon stocks were found in the soils under NT (36 mg ha 1) and HH (31 mg ha 1) and the lowest amounts were found in the soils under DP (29.5 mg ha 1) and HHDP (28 mg ha 1). This shows that, after setting up the field experiment, only the carbon stocks of the soils under NT did not decrease. The highest proportion of soil TOC was found in the passive pool (90%). Similar results were also reported by Freixo et al. (2002) investigating tillage and crop rotation interactions on organic carbon fractions of a Ferralsol from southern Brazil. Increasing proportion of passive carbon pool with simulta-
neous decrease of active and slow carbon pools may indicate soil organic matter degradation because the latter pools are highly associated to microbial activity and decomposition and are the most relevant pools to nutrient cycling. 3.3. Nitrogen pools by the measured values The replacement of the Atlantic Forest by agriculture also led to a decrease in the contents of total nitrogen (TN). In 1984, TN stocks of the soil before setting up the experiment were 3.34 mg ha 1. This corresponds to a 32% decrease relative to the soil under forest (Fig. 4). Similar to what was observed for TOC, the stocks of TN decreased in the soils under NT and HHDP systems. However, this decrease was
290
L.F. Carvalho Leite et al. / Geoderma 120 (2004) 283–295
Fig. 4. Dynamics of stocks of total nitrogen (TN) (A) and active (B), slow (C) and passive (D) nitrogen pools simulated by Century model at 0 – 20 cm depth. NT: no-tillage; HHDP: heavy harrow followed by disc plow.
less pronounced in the soil under NT than under HHDP. Although N losses in the soil under no-tillage were lower than those observed under plowed systems, there is an apparent need to include legume plants in the crop rotation as cover crops. This would increase N inputs through biological N fixation in addition to returning plant residues to the soil to supplement soil fertilization. The use of cover crops to improve mulching in the tropics is strongly recommended in a crop rotation system to increase soil carbon stocks (Machado and Silva, 2001). From 1930 to 1984, N stocks in the active, slow and passive pools followed similar trend of what was observed for C stocks. The close relationship between C and N can be observed in the nitrogen mineralization. Nitrogen pools in the N submodel can only be mineralized if CO2 is lost in the corresponding pool of the C submodel, whose decomposition rate can be regulated by N, P and S availability. Soil tillage caused an increase of TN of the active pool of both NT and HHDP systems. However, until the end of the simulation period, TN stocks decreased to the same levels found in the beginning of the experiment. In the N submodel, the active pool
represents the microbial biomass N (Parton et al., 1987) and it can be assumed that in a short run the limited amount of available carbon substrates will not support biomass in soil under NT and HHDP systems. The slow pool N stocks were also decreased after deforestation. In 1984, N stock estimated by Century model was 0.1 mg ha 1, which is approximately 85% lower than the amount found in the soil under Atlantic Forest. In 2000, similarly to what was observed in the active pool, the N stocks in the slow pool were higher in the soil under HHDP (0.22 mg ha 1) than in the soil under NT (0.1 mg ha 1). From 2000 to 2050, N stocks in the soil under NT were projected to tend to increase up to 0.16 mg ha 1 while in the soil under HHDP, N stocks remained stable at 0.18 mg ha 1 (Fig. 4). Hence, the tendency in the short term is that the nitrogen stocks in the soil under NT will be similar or even higher than those found in the soil under HH. In the passive pool, the decrease in the N stocks after changing forest to agriculture was less pronounced than that observed in the active and slow pool. This is probably due to the high recalcitrance of the passive pool (Romanya´ et al., 2000). For the simulated period, contrary to the values observed in
L.F. Carvalho Leite et al. / Geoderma 120 (2004) 283–295
the active and slow pools, N stocks in the passive pool were higher in the soil under NT than under HHDP. After 15 years, soil N stocks were 3.1 mg ha 1 for NT system and 2.4 mg ha 1 for HHDP system. Estimates for 2050 showed a significant decrease in nitrogen stocks of soils under HHDP. In this year, N stocks in the soil under NT were 2.7 mg ha 1, which is approximately 47% higher than the N stocks in the soil under HHDP (Fig. 4). 3.4. Comparison between measured and simulated C pools using Century model In 2000, the Century model simulated TOC in different tillage systems showed similar patterns to those observed in the measured TOC data, especially in the soil under NT (Fig. 5). In the soils under DP, HHDP and HH the stocks of TOC estimated by Century were higher than those shown by measurements, but differences were 4%, 0.4% and 7%, respectively. Similar trends were observed dividing TOC into 70% passive carbon, 27% slow carbon and 3% active carbon. Falloon and Smith (2002), in their modeling of the arable site at Martonvasar (Hungary) divided TOC into similar proportions, but resulted in an overestimation of measured TOC. A reasonable fit to the measured data was found using 34% passive carbon, 63% slow carbon and 2.6% active carbon. In our study, the higher proportion of the passive carbon pool than the other pools enabled an optimum fit between measured and Century model values. Compared to labile carbon fractions, in tropical soils the largest TOC stocks are humified soil organic matter probably due climatic conditions favoring microbial decomposition all over the year and also chemical and physical stability (Bayer et al., 2002; Leite et al., 2003). It is well known that caolinitic soils, such as the Acrisol of our study, are originally well structured with adequate aggregate size distribution to promote drainage (El-Swaify, 1980). Furthermore, differences between tropical/subtropical and temperate soils suggest the need for parameterization of the Century model using regional data and measured carbon pools thus contributing to more adequate modeling in the tropics. However, reasonable Century model fit to the measured TOC values were reported for both temperate (Mikhailova et al., 2000; Alvarez, 2001) and tropical (Parton et al., 1989; Motavalli et al., 1994)
291
soils. This shows the potential of Century model to simulate soil organic carbon changes in soils under different tillage systems and its sensitivity to distinguish organic carbon changes due to different tillage systems. Compared to the measured data obtained by microbial biomass carbon, in all tillage systems, the C stocks of the active pool simulated by the Century model were underestimated (Fig. 5). The differences between simulated and measured data of C stocks were 70%, 52%, 52% and 51% for NT, DP, HHDP and HH, respectively. Similarly, Motavalli et al. (1994) studying forest soils from Colombia, Peru and Brazil with varying mineralogy reported that the values of dissolved and biomass carbon stocks were larger than Century simulated values. In oxidic soils, Motavalli et al. (1994) reported that the simulated values were 45%, 51% and 39% higher than the measured values in Valencß a, Ouro Preto e Una (Brazil), respectively. In both studies, the results are probably related to factors that control C to the active pool. The Century model uses soil humidity, soil temperature, soil texture and management as regulators of the active pool. However, besides environmental aspects, microbial growth is affected by substrate availability (organic matter) and soil chemical properties (e.g. soil pH, N content) that are not taken into account by the model. Also, the mechanisms that describe C exudation by roots and its microbial metabolism are not clearly defined by the model and the lack of these mechanisms in the model may contribute to its inaccuracy. Additionally, the decomposition rate of the active pool is likely to be overestimated or the kEC value needs to be calibrated for acid tropical soils. Joergensen (1996) investigated the effects of soil properties and different form of land use on the calibration of the kEC value and found that a kEC value of 0.38 can be recommended for C analysis by dichromate consumption and a kEC value of 0.45 for that by UV-per sulfate or oven oxidation. Compared to the measured data, apart from the soil under HHDP, the C values of the slow pool simulated by the Century model were underestimated (Fig. 5). However, differences between measured and simulated values were small: 13%, 14%, and 16% for NT, DP and HH systems, respectively (Fig. 5). This suggests that also in tropical soils, the carbon of the free-light fraction may represent the slow pool as proposed by
292
L.F. Carvalho Leite et al. / Geoderma 120 (2004) 283–295
Fig. 5. Measured and simulated total organic carbon (TOC) (A), active (B), slow (C) and passive (D) carbon pool, and total nitrogen (TN) (E) in different soil management systems. NT: no-tillage; DP: disk plow; HHDP: heavy disk harrow followed by disc plow and light harrowings; HH: heavy disk harrowing (n = 4 for measured values).
Cambardella and Elliott (1992). On the other hand, Motavalli et al. (1994) showed that measured stocks of the free-light organic carbon fraction were lower than Century simulated values. The discrepancies between measured and simulated values varied in soils showing 69% to 83% oxidic mineralogy. Motavalli et al. (1994) believed that the slow pool contains substances other than free light organic carbon or the
extraction procedure is not efficacious to isolate all free light organic carbon soils. Underestimation of the Century active and slow pool may be explained by the lack of important chemical processes in acid tropical soils not considered by the model. In tropical and subtropical acid soils, the organic matter – Al complex is relevant in the control of the Al toxicity and therefore in the soil organic
L.F. Carvalho Leite et al. / Geoderma 120 (2004) 283–295
matter mineralization (Haynes and Mokolobate, 2001; Meda et al., 2001). High acidity and high soil aluminum content are also responsible for the stabilization of organic matter in acid tropical soils (Mendoncßa and Rowell, 1994; Mendoncßa, 1995). Thus, parameterization of simulation models should improve knowledge about the effect of soil mineralogy, soil pH and soil exchangeable aluminum on SOM formation and decomposition in acid tropical soils. Thus more detailed investigations are needed to identify the underlying differences between the theoretical requirements of the Century pools and those analytically quantified. Century simulated C stocks of the passive pool and those estimated by difference showed similar patterns, especially in the soils under NT and HHDP. In the soils under DP and HH,
293
differences between measured and simulated C stocks were 7% and 10%, respectively. Apart from the soil under NT, N stock contents estimated by Century were higher (7 – 10%) than measured values (Fig. 5). These trends were also reported by Fernandes (2002) in an Acrisol from southern Brazil in the no and conventional tillage as in the native grass. Regression analysis showed that the Century simulated TOC stocks correlated well (R 2 = 0.91; p < 0.05) with the measured values and a good 1:1 correspondence between simulated and measured values (Fig. 6). This indicates that Century model is able to simulate the TOC dynamic from tropical soil under different management systems. On the other hand, correlations among simulated and measured C pools indicated that only passive C pool correlated signifi-
Fig. 6. Relationship between measured and simulated total organic carbon (A), C active pool (B), C slow pool (C), C passive pool (D) and total nitrogen (E). *, ** indicate significance at the 0.05 and 0.01 probability level, respectively; ns = not significant.
294
L.F. Carvalho Leite et al. / Geoderma 120 (2004) 283–295
cantly (R2 = 0.89; p < 0.05). Thus, it is essential to conduct more studies on acid tropical soils to optimize the relationship between the theoretical concepts of the Century model pools and the measured fractions. The regression coefficient of measured TN versus simulated TN has an R2 value of 0.97, but here is not a good 1:1 correspondence between measured and simulated TN values.
4. Conclusions Both active and slow carbon pools were more sensitive to soil management systems than total organic carbon and passive carbon pool. This indicates that active and slow pools can be used as early warning indicators of soil organic matter degradation. The Century model simulated changes in the total organic carbon content and obtained an excellent fit to measured data (only 5% contrast) and this shows the high potential of the model to simulate soil organic matter dynamics in the tropics. Similarly to the total organic carbon, the Century model simulated values of passive carbon pool showed similar patterns to those observed in the measured data. However, the Century model underestimated stocks of slow and especially active carbon pool and thus there is a need to include some important chemical process in the model in acid tropical soils.
References Alvarez, R., 2001. Estimation of carbon losses by cultivation from soils of the Argentine Pampa using Century model. Soil Use Manag. 17, 62 – 66. Angers, D.A., Bolinder, M.A., Carter, M.R., Gregorich, E.G., Drury, C.F., Liang, B.C., Voroney, R.P., Simard, R.R., Donald, R.G., Beyaert, R.P., Martel, J., 1997. Impact of tillage practices on organic carbon and nitrogen storage in cool, humid soils of eastern Canada. Soil Tillage Res. 41, 191 – 201. Batjes, N.H., 1996. Total carbon and nitrogen in soils of the world. Eur. J. Soil Sci. 47, 151 – 163. Bayer, C., Dick, D.P., Ribeiro, G.M., Scheuermann, K.K., 2002. Carbon stocks in organic matter fractions as affected by land use and soil management, with emphasis on no-tillage effect. Cieˆnc. Rural 32, 401 – 406. Bremner, J.M., 1996. Nitrogen total. In: Sparks, D.L. (Ed.), Methods of Soil Analysis: Part 3. America Society of Agronomy, Madison, pp. 1085 – 1121. SSSA Book Series:5.
Cambardella, C.A., Elliott, E.T., 1992. Particulate soil organic matter changes across a grassland cultivation sequence. Soil Sci. Soc. Am. J. 56, 777 – 783. Castro Filho, C., Henklain, J.C., Vieira, M.J., Casa˜o Jr., R. 1991. Tillage methods and soil and water conservation in southern Brazil. Soil Tillage Res. 20, 271 – 283. Del Grosso, S.J., Parton, W.J., Mosier, A.R., Hartman, J., Brenner, D.S, Ojima, D.S., Schimel, D.S., 2001. Simulated interaction of soil carbon dynamics and nitrogen trace gas fluxes using the DAYCENT model. In: Shaffer, M.J., MA, L., Hansen, S. (Eds.), Modeling Carbon and Nitrogen Dynamics for Soil Management. Lewis Publishers, Boca Raton, FL, pp. 303 – 332. EL-Swaify, S.A., 1980. Physical and mechanical properties of oxisols. In: Theng, B.K.G. (Ed.), Soil with Variable ChargeNew Zealand Soil Science Society, 303 – 324. Falloon, P., Smith, P., 2002. Simulating SOC changes in long-term experiments with RothC and Century: model evaluation for a regional scale application. Soil Use Manage. 18, 101 – 111. Fernandes, F.F., 2002. Uso do modelo Century para o estudo da dinaˆmica de carbono orgaˆnico de solos do Rio Grande do Sul. Universidade Federal do Rio Grande do Sul, Porto Alegre. Freixo, A.A., de Machado, P.L.O.A., Santos, H.P., Silva, C.A., Fadigas, F.S., 2002. Soil organic carbon and fractions of a Rhodic Ferralsol under the influence of tillage and crop rotation system in southern Brazil. Soil Sci. Soc. Am. J. 65, 1121 – 1128. Goedert, W.J., Lobato, E., Lourencß o, S., 1997. Nutrient use efficiency in Brazilian acid soils: nutrient management and plant efficiency. In: Moniz, A.C., Furlani, A.M.C., Schaffert, R.E., Fageria, N.K., Rosolem, C.A., Cantarella, H. (Eds.), Plant – Soil Interactions At Low pH: Sustainable Agriculture and Forestry Production. Brazilian Soil Science Society, Vicßosa, Brazil, pp. 97 – 104. Haynes, R.J., Mokolobate, M.S., 2001. Amelioration of Al toxicity and P deficiency in acid soils by addition of organic residues: a critical review of the phenomenon and the mechanisms involved. Nutr. Cycl. Agroecosyst. 59, 47 – 63. Islam, K.R., Weil, R.R., 1998. Microwave irradiation of soil for routine measurement of microbial biomass carbon. Biol. Fertil. Soils 27, 408 – 416. Joergensen, R.G., 1996. The fumigation – extraction method to estimate soil microbial biomass: calibration of the KEC value. Soil Biol. Biochem. 28, 25 – 31. Kelly, R.H., Parton, W.J., Crocker, G.J., Grace, P.R., Klir, J., Korschens, M., Poulton, P.R., Richter, D.D., 1997. Simulating trends in soil organic carbon in long-term experiments using the Century model. Geoderma 81, 75 – 90. Kern, J.S., Johnson, M.G., 1993. Conservation tillage impacts on national soil and atmospheric carbon levels. Soil Sci. Soc. Am. J. 57, 200 – 210. Leite, L.F.C., Mendoncß a, E.S., Machado, P.L.O.A., Matos, E.S., 2003. Total C and N storage and organic C pools of a red-yellow podzolic under conventional and no tillage at the Atlantic Forest Zone, Southeastern Brazil. Aust. J. Soil Res. 41, 717 – 730. Machado, P.L.O.A., Silva, C.A., 2001. Soil management under notillage systems in the tropics with special reference to Brazil. Nutr. Cycl. Agroecosyst. 61, 119 – 130.
L.F. Carvalho Leite et al. / Geoderma 120 (2004) 283–295 Manies, K.L., Harden, J.W., Kramer, L., Parton, W., 2000. Parameterizing Century to Model Cultivated and Noncultivated Sites in the Loess Region of Western Iowa. USGS, Fort Collins, p. 28. Meda, A.R., Cassiolato, M.E., Pavan, M.A., 2001. Alleviating soil acidity through plant organic compounds. Braz. Arch. Biol. Technol. 44, 185 – 189. Mendoncßa, E.S., 1995. Oxidacßa˜o da mate´ria orgaˆnica e sua relacßa˜o com diferentes formas de alumı´nio de Latossolos. R. Bras. Ci. Solo 19, 25 – 30. Mendoncß a, E.S., Rowell, D.L., 1994. Dinaˆmica do alumı´nio e de diferentes fracßo˜es orgaˆnicas de um Latossolo argiloso sob cerrado e soja. R. Bras. Ci. Solo 18, 295 – 303. Metherell, A.K., Harding, C.V., Cole, C.V., Parton, W.J., 1993. Century soil organic matter model environment. Technical Documentation Agroecosystems Version 4.0. Great Plains System Research Report 4 USDA-ARS. Colorado State University, Fort Collins, CO. Metherell, A.K., Cambardella, C.A., Parton, W.J., Peterson, G.A., Harding, L.A., Cole, C.V., 1995. Simulation of soil organic matter dynamics in dryland wheat – fallow cropping systems. In: Lal, R., Kimble, J., Levine, E., Stewart, B.A. (Eds.), Soil Management and Greenhouse Effect. CRC Press, Boca Raton, FL, pp. 259 – 270. Mikhailova, E.A., Bryant, R.B., Degloria, S.D., Post, C.J., Vassenev, I.I., 2000. Modeling soil organic matter dynamics after conversion of native grassland to long term continuous fallow using the Century model. Ecol. Model. 132, 247 – 257. Motavalli, P.P., Palm, P.C.A., Parton, W.J., Elliot, E.T., Freys, S.D., 1994. Comparison of laboratory and modeling simulation methods for estimating soil carbon pools in tropical forest soils. Soil Biol. Biochem. 26, 935 – 944. Parfitt, R.L., Theng, B.K.G., Whitton, J.S., Shepherd, T.G., 1997. Effects of clay minerals and land use on organic matter pools. Georderma 75, 1 – 12. Parton, W.J., Rasmussen, P.E., 1994. Long-term effects of crop management in wheat – fallow: II. Century model simulations. Soil Sci. Soc. Am. J. 58, 530 – 536. Parton, W.J., Schimel, D.S., Cole, C.V., Ojima, D.S., 1987. Analysis of factors controlling soil organic matter levels in great plains grasslands. Soil Sci. Soc. Am. J. 51, 1173 – 1179. Parton, W.J., Stewart, J.W.B., Cole, C.V., 1988. Dynamics of C, N, P e S in grassland soils: a model. Biogeochemistry 5, 109 – 131. Parton, W.J., Sanford, R.L., Sanchez, P.A., Stewart, J.W.B., 1989. Modeling soil organic matter dynamics in tropical soil. In: Coleman, D.C., Oades, J., Uehara, M. (Eds.), Dynamics of Soil Organic Matter in Tropical Ecosystems, 153 – 170.
295
Paul, E.A., 1984. Dynamics of organic matter in soils. Plant Soil 76, 275 – 285. Paustian, R., Parton, W.J., Persson, J., 1992. Modeling soil organic matter in organic-amended and nitrogen-fertilized long-term plots. Soil Sci. Soc. Am. J. 56, 476 – 488. Pereira, M.H., 2002. The second green revolution. In: Saturnino, H.M., Landers, J.N. (Eds.), The Environment and Zero Tillage. Associacßa˜o de Plantio Direto no Cerrado, Brası´lia, DF, Brazil, pp. 27 – 30. Peterson, G.A., Halvorson, A.D., Havlin, J.L., Jones, O.R., Lyon, D.J., Tanaka, D.L., 1998. Reduced tillage and increasing cropping intensity in the great plains conserves soil C. Soil Tillage Res. 47, 207 – 218. Romanya´, J., Cortina, J., Falloon, P., Coleman, K., Smith, P., 2000. Modeling changes in soil organic matter after planting fastgrowing Pinus radiata on Mediterranean agricultural soils. Eur. J. Soil Sci. 51, 627 – 641. Silveira, A.M., Victoria, R.L., Ballester, M.V., Camargo, P.B., Martinelli, L.A., Piccolo, M.C., 2000. Simulacß a˜o dos efeitos das mudancß as do uso da terra na dinaˆmica de carbono no solo na Bacia do Rio Piracicaba. Pesqui. Agropecu. Bras. 25, 389 – 399. Six, J., Elliot, E.T., Paustian, K., Doran, J.W., 1998. Aggregation and soil organic matter accumulation in cultivated and native grassland soils. Soil Sci. Soc. Am. J. 62, 1367 – 1377. Smith, P., Powlson, D.S., Glendining, M.J., Smith, J.U., 1997a. Opportunities and limitations for C sequestration in European agricultural soils through changes in management. In: Lal, R., Kimble, J.M., Follett, R.F., Stewart, B.A. (Eds.), Management of Carbon Sequestration in Soil. Advances in Soil Science CRC Press, Boca Raton, pp. 143 – 152. Smith, P., Powlson, D.S., Smith, J.U., Elliot, E.T., 1997b. Evaluation and comparison of soil organic matter models using longterm datasets. Geoderma 81, 1 – 255. Smith, W.N., Desjardins, R.L., Grant, B., 2001. Estimated changes in soil carbon associated with agricultural practices in Canada. Can. J. Soil Sci. 81, 221 – 227. Sohi, S., Mahieu, N., Arah, J.R.M., Polwson, D.S.P., Madari, B., Gaunt, J.L., 2001. A procedure for isolating soil organic matter fractions suitable for modeling. Soil Sci. Soc. Am. J. 65, 1121 – 1128. Sparling, G.P., West, A.W., 1988. A direct extraction method to estimate soil microbial C: calibration in situ using microbial respiration and 14C labelled cells. Soil Biol. Biochem. 20, 337 – 343. Yeomans, J.C., Bremner, J.M., 1988. A rapid and precise method for routine determination of organic carbon in soil. Commun. Soil Sci. Plant Anal. 19, 1467 – 1476.
Simulating trends in soil organic carbon of an Acrisol under no-tillage and disc-plow systems using the Century model Luiz Fernando Carvalho Leite a, Eduardo de Sa´ Mendoncßa b,*, Pedro Luiz Oliveirade de Almeida Machado c, Elpı´dio Ina´cio Fernandes Filho b, Ju´lio Ce´sar Lima Neves b a
Empresa Estadual de Pesquisa Agropecua´ria da Paraı´ba-EMEPA, Rua Eurı´pedes Tavares, 210, 58013-290 Joa˜o Pessoa, Paraı´ba, Brazil b Departamento de Solos, Universidade Federal de Vicßosa, 36571-000 Vicßosa, Minas Gerais, Brazil c Embrapa Solos-EMBRAPA, Rua Jardim Botaˆnico 1024, 22460-000 Jardim Botaˆnico, Rio de Janeiro, Brazil Received 22 October 2002; received in revised form 22 July 2003; accepted 17 September 2003 Available online 10 December 2003
Abstract Soil organic matter (SOM) and its different pools have key importance in nutrient availability, soil structure, in the flux of trace gases between land surface and the atmosphere, and thus improving soil health. This is particularly critical for tropical soils. The rates of accumulation and decomposition of carbon in SOM are influenced by several factors that are best embodied by simulation models. However, little is known about the performance of SOM simulation model in an acid tropical soil under different tillage systems including no-tillage (NT). Our objective was to simulate soil organic matter dynamics on an Acrisol under no-tillage and different plowed systems using Century model. Tillage systems consisted of no-tillage, disc plow, heavy disc harrow followed by disc plow, and heavy disc harrow. Soil C stocks simulated by Century model showed tendency to recovery only under no-tillage. Also, simulated amounts of C stocks of slow and active pools were more sensitive to management impacts than total organic C. The values estimated by Century of soil C stocks and organic carbon in the slow and passive pools fitted satisfactorily with the measured data. Thus fitted, except for the active pool, Century showed acceptable performance in the prediction of SOM dynamics in an acid tropical soil. D 2003 Elsevier B.V. All rights reserved. Keywords: Acidic tropical soils; Soil carbon fractions; Long-term experiments; Century model
1. Introduction Soil organic matter (SOM) is an important component of acid tropical soils and its significance can be seen on the positive effects on Ferralsol cation exchange capacity, size and stability of aggregates, * Corresponding author. E-mail address: [email protected] (E. de Sa´ Mendoncßa). 0016-7061/$ - see front matter D 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.geoderma.2003.09.010
improved moisture and soil structure for plant growth (Goedert et al., 1997). Also, SOM has large influence on the flux of the trace greenhouse gases between land surface and the atmosphere (Batjes, 1996). Soil cultivation often leads to diminution of SOM content (Castro Filho et al., 1991), but conservation tillage such as no-tillage can improve soil conditions to those found in forest soils (Machado and Silva, 2001). Because of cost reductions and soil erosion control,
284
L.F. Carvalho Leite et al. / Geoderma 120 (2004) 283–295
no tillage is widely used in Brazil over an area of 14 million ha (Pereira, 2002). No tillage, combined with crop rotation involving cover crops, favors the accumulation of plant residues on the soil surface (Machado and Silva, 2001). In a large study conducted in the USA, Kern and Johnson (1993) reported that the widespread conversion of major field crop production from conventional tillage (mouldboard plow) to conservation tillage would change the soil system from a source to a sink of atmospheric carbon. Similar observations were also reported later for European soils (Smith et al., 1997a). These findings show the potential for agriculture to contribute to global carbon mitigation, particularly through no-tillage. The ability to predict the effects of environment (e.g. climate and atmospheric composition) and landuse change on SOM dynamics is of utmost importance in formulating environmental and agricultural policies (Smith et al., 1997b). Modeling is a powerful means to simulate a range of intricate processes and predict soil organic matter changes for long time periods (Paustian et al., 1992). Century is a model of terrestrial C, N, P, and S dynamics that use a four pool SOM submodel. Century has been successfully used in temperate ecosystems (Parton and Rasmussen, 1994; Kelly et al., 1997; Del Grosso et al., 2001), but in spite of some studies about modeling tropical agroecosystems (Parton et al., 1989; Motavalli et al., 1994), no information is available on the use of mathematical models on the dynamics of soil organic carbon in Acrisol under notillage. Our objective was to simulate soil organic matter dynamics on an Acrisol under no-tillage and different plowed systems commonly used in Brazil using Century model.
2. Material and methods 2.1. Experimental site Simulations with the version 4 of the Century model were carried out for an experimental site established in 1985 at the Experimental Station of the Federal University of Vicßosa, in Coimbra, State of Minas Gerais, Brazil (20j45S and 42j51W; 700 m asl). The mean annual temperature is 19 jC and
average rainfall is 1350 mm, and roughly two-thirds of this rain falls in the warmer season of the year from October to April. The area was covered by Atlantic Forest until 1930 and was cultivated for 54 years with subsistence crops, such as maize (Zea mays L.) and common bean species. The soil in the experimental area is a loamy Acrisol (Argissolo Vermelho-Amarelo, Brazilian Classification System; Typic Kandiudult, US Taxonomy) and some chemical and physical characteristics are shown in Table 1. The experiment started at 1985 and consisted of four soil management systems, arranged in a complete randomized block design, with four replications. Plots were 4 11 m under maize/fallow/soybean succession. The tillage treatments were: 1. No tillage (NT)—no disturbance to the soil other than sowing operation; 2. Disk plowing (DP)—plowing at 20 to 25 cm depth with a three-fixed disk plow, in a single pass; 3. Heavy disk harrow + disk plowing (HHDP)—one single pass at 0– 15 cm using a heavy disk harrow with 20 disks followed by disk plowing at 20– 25 cm depth, with a three-fixed disk plow; 4. Heavy disk harrow (HH)—one harrowing at 10 to 15 cm depth with a heavy disk harrower of 20 disks weighing approximately 2 t; In addition, as a reference, samples were taken from an area under secondary Atlantic Forest (AF), adjacent to the experiment, (100 m away) in the same soil type. At the middleslope, four areas (4 4 m) Table 1 Chemical and physical characteristics of an Acrisol (0 – 20 cm layer) under different tillage systems from the State of Minas Gerais, Brazil Treatmenta pH TOC TN Clay Bulk (H2O) (dag kg 1) (dag kg 1) (dag kg 1) density (mg m 3) NT DP HHPD HH AF
4.97 5.05 5.05 5.00 5.48
1.46 1.28 1.28 1.26 2.83
0.12 0.10 0.10 0.11 0.22
41 38 39 37 46
1.32 1.22 1.21 1.24 1.13
TOC: total organic carbon, TN: total nitrogen. a NT: no tillage; DP: disk plow; HHDP: heavy disk harrow + disk plow; HH: heavy harrow; AF: secondary Atlantic Forest.
L.F. Carvalho Leite et al. / Geoderma 120 (2004) 283–295
were placed along a 100-m transect in which soil samples were collected. 2.2. Chemical analysis Soil samples were collected, in April 2000, just after the harvest period. In each plot, eight topsoil samples (0 – 20 cm) were collected and combined into one composite sample. In the area under Atlantic Forest, 15 samples were collected from each sampling area (n = 4) and bulked into one sample. The samples were ground to pass a 2-mm sieve. An aliquot of 100 g was separated and kept refrigerated at 4 – 8 jC before microbial biomass analysis. For all remaining analyses, the soil samples were air-dried. Total organic carbon (TOC) was obtained by wet digestion with a mixture of potassium dichromate and sulfuric acid, under heating (Yeomans and Bremner, 1988). Total N was measured in the soil samples with a sulfuric digestion followed by determination in the Kjedahl distillation (Bremner, 1996). Measurement of microbial biomass was conducted by the irradiation – extraction method—using microwave (Islam and Weil, 1998), 0.5 mol l 1 K2SO4 as extractant and the biomass C was determined by wet combustion (Yeomans and Bremner, 1988). The factor (KC) used to convert the flow of C for the microbial biomass C (CMB) was 0.33 (Sparling and West, 1988). The CMB was used as an estimate of the active C pool (Paul, 1984; Motavalli et al., 1994). The free-light organic carbon fraction (CLF) was determined by flotation in NaI solution (d = 1.8 g cm 3) as proposed by Sohi et al. (2001). The isolated material was dried at 105 jC for 72 h. The CLF, quantified by dry combustion (Perkin Elmer 2400 CHNS/O elemental analyzer), was further used as an estimate of the slow C pool. Passive C pool was calculated using the following equation: Passive C ¼ TOC ðCMB þ CLF Þ Bulk soil density was determined on nondeformed soil samples collected from a single field replicate. The values of bulk soil density were used to calculate SOM pools based on an equivalent soil mass (Angers et al., 1997; Peterson et al., 1998).
285
2.3. Model parameterization The Century SOM model was originally developed and tested on data sets mainly from grassland and wheat – fallow agriculture in the US Great Plains (Parton et al., 1987, 1988). All parameter values determined from these previous studies were initially left unchanged to provide more criterious evaluations of the simulations. As given by Paustian et al. (1992), these general or non-site specific parameters include the maximum specific decomposition rates for each compartment, the constants that splits the flows of decomposition products and the parameters that control the effects of soil texture, lignin/N ratios, temperature, and moisture on decomposition rates. Site-specific parameters and initial conditions, such as soil texture (sand, silt and clay content), bulk density, soil depth and total soil C and N content, were given values obtained from the field experiment at Coimbra. Monthly precipitation and mean maximum and minimum monthly temperatures from 1967 to 2000 were obtained from the weather station at the Vicßosa Federal University. The parameter determining potential crop productivity was based on the maximum production level observed during the course of the field experiment in each treatment and temperature curve, C/N ratios and lignin contents of biomass pools were obtained through default crop parameterizations distributed with Century model. The main input data for the model are in Table 2. Some model adjustments were made to improve the tillage effects on SOM decomposition. First, the plowing option was adjusted to increase its effect on decomposition (Six et al., 1998). All clteff values (cultivation’s effect on decomposition) for the DP, HHDP and HH were changed from 1.6 to 5 (Table 2). The second change was added to increase the length of time which plowing effects decomposition. Since Century uses a monthly time-step each action only affects SOM dynamics for that specific month although some studies have shown that plowing affects decomposition for several months (Metherell et al., 1995). Thus, an option called ‘‘Additional plowing effect’’ was used in the months following plowing in order to keep the decomposition rates at higher levels (Manies et al., 2000). To initialize the percentage of total SOM in each of the three pools (active, slow and passive) used indirect
286
L.F. Carvalho Leite et al. / Geoderma 120 (2004) 283–295
Table 2 Model input for simulation of tillage systems using version 4 of the Century model Value Soil variables Texture (% sand, % silt, % clay) Bulk density (mg m 3) Initial SOM (g C m 2; C/N) Active surface Active soil Slow soil Passive soil Monthly weather variables Mean total precipitation (cm month 1) Mean maximum temperature (jC) Mean minimum temperature (jC) Cultivation variables Multiplier for increased decomposition Active pool
Slow pool Passive pool
38, 16, 46 1, 13
50, 12,5 159, 9 1776, 22 4460, 12
14 14.8 26.4
1.0 (NT); 1,8 (DP); 2.0 (HHDP); 1.6 (HH) 1.0 (NT); 4.0 (DP); 5.0 (HHDP); 3.0 (HH) 1.0 (NT); 1.8 (DP); 2.0 (HHDP); 1.6 (HH)
Carbon pool values were obtained from direct method. NT: no tillage; DP: disc plow; HHDP: heavy disk harrow + disk plow; HH: heavy harrow.
methods involving simulation of steady state organic matter levels and direct methods using analytical techniques. In the indirect method, Century model parameterized all data including carbon pools for a long term (6000 years) by equilibrium simulation. Simulated values for carbon pools were used as input variables for simulation of land use change. To each treatment, the model simulated SOM dynamics for 54 years representing forest conversion into cropland (e.g. maize and bean production) and later, for 66 years, representing different tillage systems. In the direct method, the initial soil carbon pool sizes under Atlantic Forest were estimated by laboratory analysis and similarly used in the indirect method. All Century estimates were based on a 20-cm depth. Both simulated and measured values for TOC, active, slow and
passive pools in 2000 were subjected to linear regression and Pearson’s correlation. An average of these data sets was taken from each treatment and subjected to a Student’s t-test to determine the significance of the coefficients at the 0.05 and 0.01 probability levels. Simulations for nitrogen pools were also done for NT and HHDP systems.
3. Results and discussion 3.1. Estimates of carbon pools by equilibrium values Century model simulated the equilibrium values of total organic carbon (TOC) and carbon pools (active, slow and passive) for 6000 years. Compared to the initial values, storage of both TOC and passive C pool increased while active and slow carbon pools remained generally constant (Fig. 1). The increase in the passive carbon pool, reflected by TOC, is probably due to carbon pools that are highly recalcitrant, physically protected against microbial attack and less prone to oxidation. After reaching equilibrium, stocks of TOC (64 mg ha 1) and active carbon pool (1.60 mg ha 1) were similar to those measured in the soil under Atlantic Forest (Table 3). The stocks of TOC are 3.9% higher than the value (61.5 mg ha 1) found by Silveira et al. (2000) in another soil under Atlantic Forest in the Piracicaba river basin, Brazil. The estimated value of the slow carbon pool (30.1 mg ha 1) was 42% higher than the measured value in the forest soil. On the other hand, the estimated value of
Fig. 1. Modeled stocks of soil organic carbon (TOC) and organic carbon pools of an Acrisol (0 – 20 cm) under Atlantic Forest.
L.F. Carvalho Leite et al. / Geoderma 120 (2004) 283–295 Table 3 Measured values of total organic carbon (TOC) and carbon pools (active, slow and passive) of an Acrisol under Atlantic Forest (Brazil) Pools
Stocks (mg ha 1)
Total organic carbon Active Slow Passive
63.95 1.59 17.4 44.9
( F 4.25) ( F 0.02) ( F 1.23) ( F 4.01)
Values in parenthesis represent the standard error of the mean.
the passive carbon pool (32.3 mg ha 1) was lower than the measured data (Table 3). Hence, model fit to measured slow and passive carbon pool was not good, probably due to sizes of the passive pool higher in
287
tropical soils than the assumption of the Century model. Starting at initial values of carbon pools obtained from the equilibrium simulation, Century estimates for the stocks of TOC and carbon pools after 120 years decreased after changing forest into agriculture (Fig. 2). Parfitt et al. (1997) studying the effects of clay minerals and land use on organic matter pools obtained similar results, i.e. a decrease of the TOC and C pools after forest clearance and subsequently land use under pasture and maize in an Inceptisol. In 1984, before setting up the experiment, TOC stock was 28 mg ha 1, which were 56% lower than the initial value predicted by the model (64 mg ha 1). In 2000, 15 years after the introduction of the tillage systems,
Fig. 2. Time variation of the stocks of TOC (A) and active (B), slow (C) and passive (D) carbon pools simulated by Century based on the initial values obtained from the equilibrium simulation in the no-till (NT), disc plow (DP), heavy harrow followed by disc plow (HHDP) and heavy harrow (HH).
288
L.F. Carvalho Leite et al. / Geoderma 120 (2004) 283–295
stocks of TOC in the soil under NT (29 mg ha 1) showed an increase compared to the stocks of TOC in the NT soil at the beginning of the experiment. This was in contrast with DP (24.5 mg ha 1), HSPD (23.4 mg ha 1), and HH (25.2 mg ha 1) (Fig. 2). In 2050, stocks of TOC in the soils under NT and plowed systems were estimated as 26 and 15 mg ha 1, respectively, thus showing less soil organic matter degradation under no tillage. Also in both active and slow pools, carbon stocks diminished with the change from forest to cropland. After the introduction of tillage systems, it was observed less significant changes in the carbon stock of active and slow pools. However, the amount of passive carbon stocks in the soils under NT was significantly higher. In 2000, the amount of organic carbon in the soil under NT was 5 mg ha 1 higher than the soils under DP, HHDP and HH systems. This difference increased with time and in 2050 the stocks of TOC in the soil under NT (23 mg ha 1) were estimated as two times higher than the amounts in the soils under DP (12.8 mg ha 1), HHDP (10.6 mg ha 1), and HH (13.2 mg ha 1). These results show the effect of soil disturbance by plowing, which favors higher SOM mineralization rate and thus leads to an increase in humification and in the passive carbon pool. 3.2. Estimates of carbon pools by the measured values In the beginning of the field experiment, the stock of TOC had declined to 40 mg ha 1 (Fig. 3), a decrease of 37%, compared to the 63.95 mg ha 1, initial value measured in the soil under Atlantic Forest. The measured stocks of TOC were 20% lower than the Century model equilibrium estimates of TOC stocks. This difference is probably due to the higher proportion of passive carbon pool in measured TOC. As Century was originally developed for temperate soils it is unlikely to be adjusted to tropical environments because organic –mineral associations in the tropics are different from that observed in the temperate grasslands. Tillage systems did not change the trend to decreasing stocks of carbon. Fifteen years after setting up the field experiment the stocks of TOC in the soils under NT, DP, HHDP and HH were 38, 32, 31, and 34
mg ha 1, respectively (Fig. 3). This tendency continued mainly due to conventional plowed systems and in 2050 the projected stocks of TOC will be 34 mg ha 1 in the soil under NT, and approximately 20 mg ha 1 in the soils under DP, HHDP and HH. Despite these results, soils under NT system were the only ones to show a slight recovery in long term. Our results are corroborated by Smith et al. (2001) that observed at several soils groups and cultures rotation that no tillage resulted, in long term, in the rate of TOC gain rising as high as 0.15 mg ha 1 C year 1 in Black Chernozem and Gleysolic soil groups whereas the conventional tillage showed a loss of TOC. Apparently, crop succession involving fallow, even with no tillage, will demand too long time to reach new equilibrium. This situation, however, may be changed if a crop rotation involving a cover crop to improve mulching (e.g. millet) is included or a ley farming system is adopted. Ley farming involving grass such as Brachiaria may greatly increase the stock of soil organic carbon. In 1984, carbon stocks of the active, slow and passive pools were 0.2, 2 and 36 mg ha 1. These values, compared to those initial values in the soil under forest, represented a decrease of 87%, 89%, and 19%, respectively (Fig. 3). This diminution trend was still observed even after the adoption of NT system, which shows that although the soil has been under NT for 15 years, no soil disturbance without cover crop management as pointed out by Machado and Silva (2001) hardly help to improve an increase of TOC in acidic tropical soils. As reported by Parton et al. (1987), Metherell et al. (1993) and Del Grosso et al. (2001) our results also indicate the high sensitivity of the active and slow carbon pools to changes in soil management parallel to a higher stability of the passive carbon pool. In 2000, compared to the amounts in the beginning of the experiment, the carbon stocks of the active pool in the plowed soils increased approximately 0.3 mg ha 1. However, Century simulation for 50 years showed a decrease in the carbon stocks even in the soil under NT (Fig. 3), indicating that, as reported by Machado and Silva (2001), no tillage without cover crop management (e.g. millet, sun hemp, black oat) hardly improve TOC content of an acid tropical soil. In 2000, the modeled carbon stocks of the slow pool in the soils under NT, DP, HSDP and HH were 2.3, 2.8, 2.9 and 2.7 mg
L.F. Carvalho Leite et al. / Geoderma 120 (2004) 283–295
289
Fig. 3. Time variation of organic carbon stocks (TOC) (A) and active (B), slow (C), and passive (D) pools simulated by Century based on the initial values obtained from direct method in the no-till (NT), disc plow (DP), heavy harrow followed by disc plow (HHDP) and heavy harrow (HH).
ha 1, respectively (Fig. 3). The passive carbon pool showed higher differentiation in carbon content among tillage systems than the active and slow pools. In 2000, the highest values of soil carbon stocks were found in the soils under NT (36 mg ha 1) and HH (31 mg ha 1) and the lowest amounts were found in the soils under DP (29.5 mg ha 1) and HHDP (28 mg ha 1). This shows that, after setting up the field experiment, only the carbon stocks of the soils under NT did not decrease. The highest proportion of soil TOC was found in the passive pool (90%). Similar results were also reported by Freixo et al. (2002) investigating tillage and crop rotation interactions on organic carbon fractions of a Ferralsol from southern Brazil. Increasing proportion of passive carbon pool with simulta-
neous decrease of active and slow carbon pools may indicate soil organic matter degradation because the latter pools are highly associated to microbial activity and decomposition and are the most relevant pools to nutrient cycling. 3.3. Nitrogen pools by the measured values The replacement of the Atlantic Forest by agriculture also led to a decrease in the contents of total nitrogen (TN). In 1984, TN stocks of the soil before setting up the experiment were 3.34 mg ha 1. This corresponds to a 32% decrease relative to the soil under forest (Fig. 4). Similar to what was observed for TOC, the stocks of TN decreased in the soils under NT and HHDP systems. However, this decrease was
290
L.F. Carvalho Leite et al. / Geoderma 120 (2004) 283–295
Fig. 4. Dynamics of stocks of total nitrogen (TN) (A) and active (B), slow (C) and passive (D) nitrogen pools simulated by Century model at 0 – 20 cm depth. NT: no-tillage; HHDP: heavy harrow followed by disc plow.
less pronounced in the soil under NT than under HHDP. Although N losses in the soil under no-tillage were lower than those observed under plowed systems, there is an apparent need to include legume plants in the crop rotation as cover crops. This would increase N inputs through biological N fixation in addition to returning plant residues to the soil to supplement soil fertilization. The use of cover crops to improve mulching in the tropics is strongly recommended in a crop rotation system to increase soil carbon stocks (Machado and Silva, 2001). From 1930 to 1984, N stocks in the active, slow and passive pools followed similar trend of what was observed for C stocks. The close relationship between C and N can be observed in the nitrogen mineralization. Nitrogen pools in the N submodel can only be mineralized if CO2 is lost in the corresponding pool of the C submodel, whose decomposition rate can be regulated by N, P and S availability. Soil tillage caused an increase of TN of the active pool of both NT and HHDP systems. However, until the end of the simulation period, TN stocks decreased to the same levels found in the beginning of the experiment. In the N submodel, the active pool
represents the microbial biomass N (Parton et al., 1987) and it can be assumed that in a short run the limited amount of available carbon substrates will not support biomass in soil under NT and HHDP systems. The slow pool N stocks were also decreased after deforestation. In 1984, N stock estimated by Century model was 0.1 mg ha 1, which is approximately 85% lower than the amount found in the soil under Atlantic Forest. In 2000, similarly to what was observed in the active pool, the N stocks in the slow pool were higher in the soil under HHDP (0.22 mg ha 1) than in the soil under NT (0.1 mg ha 1). From 2000 to 2050, N stocks in the soil under NT were projected to tend to increase up to 0.16 mg ha 1 while in the soil under HHDP, N stocks remained stable at 0.18 mg ha 1 (Fig. 4). Hence, the tendency in the short term is that the nitrogen stocks in the soil under NT will be similar or even higher than those found in the soil under HH. In the passive pool, the decrease in the N stocks after changing forest to agriculture was less pronounced than that observed in the active and slow pool. This is probably due to the high recalcitrance of the passive pool (Romanya´ et al., 2000). For the simulated period, contrary to the values observed in
L.F. Carvalho Leite et al. / Geoderma 120 (2004) 283–295
the active and slow pools, N stocks in the passive pool were higher in the soil under NT than under HHDP. After 15 years, soil N stocks were 3.1 mg ha 1 for NT system and 2.4 mg ha 1 for HHDP system. Estimates for 2050 showed a significant decrease in nitrogen stocks of soils under HHDP. In this year, N stocks in the soil under NT were 2.7 mg ha 1, which is approximately 47% higher than the N stocks in the soil under HHDP (Fig. 4). 3.4. Comparison between measured and simulated C pools using Century model In 2000, the Century model simulated TOC in different tillage systems showed similar patterns to those observed in the measured TOC data, especially in the soil under NT (Fig. 5). In the soils under DP, HHDP and HH the stocks of TOC estimated by Century were higher than those shown by measurements, but differences were 4%, 0.4% and 7%, respectively. Similar trends were observed dividing TOC into 70% passive carbon, 27% slow carbon and 3% active carbon. Falloon and Smith (2002), in their modeling of the arable site at Martonvasar (Hungary) divided TOC into similar proportions, but resulted in an overestimation of measured TOC. A reasonable fit to the measured data was found using 34% passive carbon, 63% slow carbon and 2.6% active carbon. In our study, the higher proportion of the passive carbon pool than the other pools enabled an optimum fit between measured and Century model values. Compared to labile carbon fractions, in tropical soils the largest TOC stocks are humified soil organic matter probably due climatic conditions favoring microbial decomposition all over the year and also chemical and physical stability (Bayer et al., 2002; Leite et al., 2003). It is well known that caolinitic soils, such as the Acrisol of our study, are originally well structured with adequate aggregate size distribution to promote drainage (El-Swaify, 1980). Furthermore, differences between tropical/subtropical and temperate soils suggest the need for parameterization of the Century model using regional data and measured carbon pools thus contributing to more adequate modeling in the tropics. However, reasonable Century model fit to the measured TOC values were reported for both temperate (Mikhailova et al., 2000; Alvarez, 2001) and tropical (Parton et al., 1989; Motavalli et al., 1994)
291
soils. This shows the potential of Century model to simulate soil organic carbon changes in soils under different tillage systems and its sensitivity to distinguish organic carbon changes due to different tillage systems. Compared to the measured data obtained by microbial biomass carbon, in all tillage systems, the C stocks of the active pool simulated by the Century model were underestimated (Fig. 5). The differences between simulated and measured data of C stocks were 70%, 52%, 52% and 51% for NT, DP, HHDP and HH, respectively. Similarly, Motavalli et al. (1994) studying forest soils from Colombia, Peru and Brazil with varying mineralogy reported that the values of dissolved and biomass carbon stocks were larger than Century simulated values. In oxidic soils, Motavalli et al. (1994) reported that the simulated values were 45%, 51% and 39% higher than the measured values in Valencß a, Ouro Preto e Una (Brazil), respectively. In both studies, the results are probably related to factors that control C to the active pool. The Century model uses soil humidity, soil temperature, soil texture and management as regulators of the active pool. However, besides environmental aspects, microbial growth is affected by substrate availability (organic matter) and soil chemical properties (e.g. soil pH, N content) that are not taken into account by the model. Also, the mechanisms that describe C exudation by roots and its microbial metabolism are not clearly defined by the model and the lack of these mechanisms in the model may contribute to its inaccuracy. Additionally, the decomposition rate of the active pool is likely to be overestimated or the kEC value needs to be calibrated for acid tropical soils. Joergensen (1996) investigated the effects of soil properties and different form of land use on the calibration of the kEC value and found that a kEC value of 0.38 can be recommended for C analysis by dichromate consumption and a kEC value of 0.45 for that by UV-per sulfate or oven oxidation. Compared to the measured data, apart from the soil under HHDP, the C values of the slow pool simulated by the Century model were underestimated (Fig. 5). However, differences between measured and simulated values were small: 13%, 14%, and 16% for NT, DP and HH systems, respectively (Fig. 5). This suggests that also in tropical soils, the carbon of the free-light fraction may represent the slow pool as proposed by
292
L.F. Carvalho Leite et al. / Geoderma 120 (2004) 283–295
Fig. 5. Measured and simulated total organic carbon (TOC) (A), active (B), slow (C) and passive (D) carbon pool, and total nitrogen (TN) (E) in different soil management systems. NT: no-tillage; DP: disk plow; HHDP: heavy disk harrow followed by disc plow and light harrowings; HH: heavy disk harrowing (n = 4 for measured values).
Cambardella and Elliott (1992). On the other hand, Motavalli et al. (1994) showed that measured stocks of the free-light organic carbon fraction were lower than Century simulated values. The discrepancies between measured and simulated values varied in soils showing 69% to 83% oxidic mineralogy. Motavalli et al. (1994) believed that the slow pool contains substances other than free light organic carbon or the
extraction procedure is not efficacious to isolate all free light organic carbon soils. Underestimation of the Century active and slow pool may be explained by the lack of important chemical processes in acid tropical soils not considered by the model. In tropical and subtropical acid soils, the organic matter – Al complex is relevant in the control of the Al toxicity and therefore in the soil organic
L.F. Carvalho Leite et al. / Geoderma 120 (2004) 283–295
matter mineralization (Haynes and Mokolobate, 2001; Meda et al., 2001). High acidity and high soil aluminum content are also responsible for the stabilization of organic matter in acid tropical soils (Mendoncßa and Rowell, 1994; Mendoncßa, 1995). Thus, parameterization of simulation models should improve knowledge about the effect of soil mineralogy, soil pH and soil exchangeable aluminum on SOM formation and decomposition in acid tropical soils. Thus more detailed investigations are needed to identify the underlying differences between the theoretical requirements of the Century pools and those analytically quantified. Century simulated C stocks of the passive pool and those estimated by difference showed similar patterns, especially in the soils under NT and HHDP. In the soils under DP and HH,
293
differences between measured and simulated C stocks were 7% and 10%, respectively. Apart from the soil under NT, N stock contents estimated by Century were higher (7 – 10%) than measured values (Fig. 5). These trends were also reported by Fernandes (2002) in an Acrisol from southern Brazil in the no and conventional tillage as in the native grass. Regression analysis showed that the Century simulated TOC stocks correlated well (R 2 = 0.91; p < 0.05) with the measured values and a good 1:1 correspondence between simulated and measured values (Fig. 6). This indicates that Century model is able to simulate the TOC dynamic from tropical soil under different management systems. On the other hand, correlations among simulated and measured C pools indicated that only passive C pool correlated signifi-
Fig. 6. Relationship between measured and simulated total organic carbon (A), C active pool (B), C slow pool (C), C passive pool (D) and total nitrogen (E). *, ** indicate significance at the 0.05 and 0.01 probability level, respectively; ns = not significant.
294
L.F. Carvalho Leite et al. / Geoderma 120 (2004) 283–295
cantly (R2 = 0.89; p < 0.05). Thus, it is essential to conduct more studies on acid tropical soils to optimize the relationship between the theoretical concepts of the Century model pools and the measured fractions. The regression coefficient of measured TN versus simulated TN has an R2 value of 0.97, but here is not a good 1:1 correspondence between measured and simulated TN values.
4. Conclusions Both active and slow carbon pools were more sensitive to soil management systems than total organic carbon and passive carbon pool. This indicates that active and slow pools can be used as early warning indicators of soil organic matter degradation. The Century model simulated changes in the total organic carbon content and obtained an excellent fit to measured data (only 5% contrast) and this shows the high potential of the model to simulate soil organic matter dynamics in the tropics. Similarly to the total organic carbon, the Century model simulated values of passive carbon pool showed similar patterns to those observed in the measured data. However, the Century model underestimated stocks of slow and especially active carbon pool and thus there is a need to include some important chemical process in the model in acid tropical soils.
References Alvarez, R., 2001. Estimation of carbon losses by cultivation from soils of the Argentine Pampa using Century model. Soil Use Manag. 17, 62 – 66. Angers, D.A., Bolinder, M.A., Carter, M.R., Gregorich, E.G., Drury, C.F., Liang, B.C., Voroney, R.P., Simard, R.R., Donald, R.G., Beyaert, R.P., Martel, J., 1997. Impact of tillage practices on organic carbon and nitrogen storage in cool, humid soils of eastern Canada. Soil Tillage Res. 41, 191 – 201. Batjes, N.H., 1996. Total carbon and nitrogen in soils of the world. Eur. J. Soil Sci. 47, 151 – 163. Bayer, C., Dick, D.P., Ribeiro, G.M., Scheuermann, K.K., 2002. Carbon stocks in organic matter fractions as affected by land use and soil management, with emphasis on no-tillage effect. Cieˆnc. Rural 32, 401 – 406. Bremner, J.M., 1996. Nitrogen total. In: Sparks, D.L. (Ed.), Methods of Soil Analysis: Part 3. America Society of Agronomy, Madison, pp. 1085 – 1121. SSSA Book Series:5.
Cambardella, C.A., Elliott, E.T., 1992. Particulate soil organic matter changes across a grassland cultivation sequence. Soil Sci. Soc. Am. J. 56, 777 – 783. Castro Filho, C., Henklain, J.C., Vieira, M.J., Casa˜o Jr., R. 1991. Tillage methods and soil and water conservation in southern Brazil. Soil Tillage Res. 20, 271 – 283. Del Grosso, S.J., Parton, W.J., Mosier, A.R., Hartman, J., Brenner, D.S, Ojima, D.S., Schimel, D.S., 2001. Simulated interaction of soil carbon dynamics and nitrogen trace gas fluxes using the DAYCENT model. In: Shaffer, M.J., MA, L., Hansen, S. (Eds.), Modeling Carbon and Nitrogen Dynamics for Soil Management. Lewis Publishers, Boca Raton, FL, pp. 303 – 332. EL-Swaify, S.A., 1980. Physical and mechanical properties of oxisols. In: Theng, B.K.G. (Ed.), Soil with Variable ChargeNew Zealand Soil Science Society, 303 – 324. Falloon, P., Smith, P., 2002. Simulating SOC changes in long-term experiments with RothC and Century: model evaluation for a regional scale application. Soil Use Manage. 18, 101 – 111. Fernandes, F.F., 2002. Uso do modelo Century para o estudo da dinaˆmica de carbono orgaˆnico de solos do Rio Grande do Sul. Universidade Federal do Rio Grande do Sul, Porto Alegre. Freixo, A.A., de Machado, P.L.O.A., Santos, H.P., Silva, C.A., Fadigas, F.S., 2002. Soil organic carbon and fractions of a Rhodic Ferralsol under the influence of tillage and crop rotation system in southern Brazil. Soil Sci. Soc. Am. J. 65, 1121 – 1128. Goedert, W.J., Lobato, E., Lourencß o, S., 1997. Nutrient use efficiency in Brazilian acid soils: nutrient management and plant efficiency. In: Moniz, A.C., Furlani, A.M.C., Schaffert, R.E., Fageria, N.K., Rosolem, C.A., Cantarella, H. (Eds.), Plant – Soil Interactions At Low pH: Sustainable Agriculture and Forestry Production. Brazilian Soil Science Society, Vicßosa, Brazil, pp. 97 – 104. Haynes, R.J., Mokolobate, M.S., 2001. Amelioration of Al toxicity and P deficiency in acid soils by addition of organic residues: a critical review of the phenomenon and the mechanisms involved. Nutr. Cycl. Agroecosyst. 59, 47 – 63. Islam, K.R., Weil, R.R., 1998. Microwave irradiation of soil for routine measurement of microbial biomass carbon. Biol. Fertil. Soils 27, 408 – 416. Joergensen, R.G., 1996. The fumigation – extraction method to estimate soil microbial biomass: calibration of the KEC value. Soil Biol. Biochem. 28, 25 – 31. Kelly, R.H., Parton, W.J., Crocker, G.J., Grace, P.R., Klir, J., Korschens, M., Poulton, P.R., Richter, D.D., 1997. Simulating trends in soil organic carbon in long-term experiments using the Century model. Geoderma 81, 75 – 90. Kern, J.S., Johnson, M.G., 1993. Conservation tillage impacts on national soil and atmospheric carbon levels. Soil Sci. Soc. Am. J. 57, 200 – 210. Leite, L.F.C., Mendoncß a, E.S., Machado, P.L.O.A., Matos, E.S., 2003. Total C and N storage and organic C pools of a red-yellow podzolic under conventional and no tillage at the Atlantic Forest Zone, Southeastern Brazil. Aust. J. Soil Res. 41, 717 – 730. Machado, P.L.O.A., Silva, C.A., 2001. Soil management under notillage systems in the tropics with special reference to Brazil. Nutr. Cycl. Agroecosyst. 61, 119 – 130.
L.F. Carvalho Leite et al. / Geoderma 120 (2004) 283–295 Manies, K.L., Harden, J.W., Kramer, L., Parton, W., 2000. Parameterizing Century to Model Cultivated and Noncultivated Sites in the Loess Region of Western Iowa. USGS, Fort Collins, p. 28. Meda, A.R., Cassiolato, M.E., Pavan, M.A., 2001. Alleviating soil acidity through plant organic compounds. Braz. Arch. Biol. Technol. 44, 185 – 189. Mendoncßa, E.S., 1995. Oxidacßa˜o da mate´ria orgaˆnica e sua relacßa˜o com diferentes formas de alumı´nio de Latossolos. R. Bras. Ci. Solo 19, 25 – 30. Mendoncß a, E.S., Rowell, D.L., 1994. Dinaˆmica do alumı´nio e de diferentes fracßo˜es orgaˆnicas de um Latossolo argiloso sob cerrado e soja. R. Bras. Ci. Solo 18, 295 – 303. Metherell, A.K., Harding, C.V., Cole, C.V., Parton, W.J., 1993. Century soil organic matter model environment. Technical Documentation Agroecosystems Version 4.0. Great Plains System Research Report 4 USDA-ARS. Colorado State University, Fort Collins, CO. Metherell, A.K., Cambardella, C.A., Parton, W.J., Peterson, G.A., Harding, L.A., Cole, C.V., 1995. Simulation of soil organic matter dynamics in dryland wheat – fallow cropping systems. In: Lal, R., Kimble, J., Levine, E., Stewart, B.A. (Eds.), Soil Management and Greenhouse Effect. CRC Press, Boca Raton, FL, pp. 259 – 270. Mikhailova, E.A., Bryant, R.B., Degloria, S.D., Post, C.J., Vassenev, I.I., 2000. Modeling soil organic matter dynamics after conversion of native grassland to long term continuous fallow using the Century model. Ecol. Model. 132, 247 – 257. Motavalli, P.P., Palm, P.C.A., Parton, W.J., Elliot, E.T., Freys, S.D., 1994. Comparison of laboratory and modeling simulation methods for estimating soil carbon pools in tropical forest soils. Soil Biol. Biochem. 26, 935 – 944. Parfitt, R.L., Theng, B.K.G., Whitton, J.S., Shepherd, T.G., 1997. Effects of clay minerals and land use on organic matter pools. Georderma 75, 1 – 12. Parton, W.J., Rasmussen, P.E., 1994. Long-term effects of crop management in wheat – fallow: II. Century model simulations. Soil Sci. Soc. Am. J. 58, 530 – 536. Parton, W.J., Schimel, D.S., Cole, C.V., Ojima, D.S., 1987. Analysis of factors controlling soil organic matter levels in great plains grasslands. Soil Sci. Soc. Am. J. 51, 1173 – 1179. Parton, W.J., Stewart, J.W.B., Cole, C.V., 1988. Dynamics of C, N, P e S in grassland soils: a model. Biogeochemistry 5, 109 – 131. Parton, W.J., Sanford, R.L., Sanchez, P.A., Stewart, J.W.B., 1989. Modeling soil organic matter dynamics in tropical soil. In: Coleman, D.C., Oades, J., Uehara, M. (Eds.), Dynamics of Soil Organic Matter in Tropical Ecosystems, 153 – 170.
295
Paul, E.A., 1984. Dynamics of organic matter in soils. Plant Soil 76, 275 – 285. Paustian, R., Parton, W.J., Persson, J., 1992. Modeling soil organic matter in organic-amended and nitrogen-fertilized long-term plots. Soil Sci. Soc. Am. J. 56, 476 – 488. Pereira, M.H., 2002. The second green revolution. In: Saturnino, H.M., Landers, J.N. (Eds.), The Environment and Zero Tillage. Associacßa˜o de Plantio Direto no Cerrado, Brası´lia, DF, Brazil, pp. 27 – 30. Peterson, G.A., Halvorson, A.D., Havlin, J.L., Jones, O.R., Lyon, D.J., Tanaka, D.L., 1998. Reduced tillage and increasing cropping intensity in the great plains conserves soil C. Soil Tillage Res. 47, 207 – 218. Romanya´, J., Cortina, J., Falloon, P., Coleman, K., Smith, P., 2000. Modeling changes in soil organic matter after planting fastgrowing Pinus radiata on Mediterranean agricultural soils. Eur. J. Soil Sci. 51, 627 – 641. Silveira, A.M., Victoria, R.L., Ballester, M.V., Camargo, P.B., Martinelli, L.A., Piccolo, M.C., 2000. Simulacß a˜o dos efeitos das mudancß as do uso da terra na dinaˆmica de carbono no solo na Bacia do Rio Piracicaba. Pesqui. Agropecu. Bras. 25, 389 – 399. Six, J., Elliot, E.T., Paustian, K., Doran, J.W., 1998. Aggregation and soil organic matter accumulation in cultivated and native grassland soils. Soil Sci. Soc. Am. J. 62, 1367 – 1377. Smith, P., Powlson, D.S., Glendining, M.J., Smith, J.U., 1997a. Opportunities and limitations for C sequestration in European agricultural soils through changes in management. In: Lal, R., Kimble, J.M., Follett, R.F., Stewart, B.A. (Eds.), Management of Carbon Sequestration in Soil. Advances in Soil Science CRC Press, Boca Raton, pp. 143 – 152. Smith, P., Powlson, D.S., Smith, J.U., Elliot, E.T., 1997b. Evaluation and comparison of soil organic matter models using longterm datasets. Geoderma 81, 1 – 255. Smith, W.N., Desjardins, R.L., Grant, B., 2001. Estimated changes in soil carbon associated with agricultural practices in Canada. Can. J. Soil Sci. 81, 221 – 227. Sohi, S., Mahieu, N., Arah, J.R.M., Polwson, D.S.P., Madari, B., Gaunt, J.L., 2001. A procedure for isolating soil organic matter fractions suitable for modeling. Soil Sci. Soc. Am. J. 65, 1121 – 1128. Sparling, G.P., West, A.W., 1988. A direct extraction method to estimate soil microbial C: calibration in situ using microbial respiration and 14C labelled cells. Soil Biol. Biochem. 20, 337 – 343. Yeomans, J.C., Bremner, J.M., 1988. A rapid and precise method for routine determination of organic carbon in soil. Commun. Soil Sci. Plant Anal. 19, 1467 – 1476.
