SEARCH

SEARCH BY CITATION

Keywords:

  • agricultural soil;
  • mitigation;
  • Rothamsted carbon model;
  • soil carbon

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

We estimated the carbon (C) sequestration potential of organic matter application in Japanese arable soils at a country scale by applying the Rothamsted carbon (RothC) model at a 1-km resolution. After establishing the baseline soil organic carbon (SOC) content for 1990, a 25-year simulation was run for four management scenarios: A (minimum organic matter application), B (farmyard manure application), C (double cropping for paddy fields) and D (both B and C). The total SOC decreased during the simulation in all four scenarios because the C input in all four scenarios was lower than that required to maintain the baseline 1990 SOC level. Scenario A resulted in the greatest depletion, reflecting the effects of increased organic matter application in the other scenarios. The 25-year difference in SOC accumulation between scenario A and scenarios B, C and D was 32.3, 11.1 and 43.4 Mt C, respectively. The annual SOC accumulation per unit area was similar to a previous estimate, and the 25-year averages were 0.30, 0.10 and 0.41 t C ha−1 year−1 for scenarios B, C and D, respectively. The system we developed in the present study, that is, linking the RothC model and soil spatial data, can be useful for estimating the potential C sequestration resulting from an increase in organic matter input to Japanese arable soils, although more feasible scenarios need to be developed to enable more realistic estimation.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Implementing efficient agronomic practices to change the amount of soil organic carbon (SOC) stored in agricultural soil can serve to mitigate climate change (Paustian et al. 1997; Smith et al. 2008). The fourth assessment report of the Intergovernmental Panel on Climate Change (IPCC AR4) included details of the agricultural sector’s mitigation potential (Smith et al. 2007a) through practices such as carbon sequestration by soil management. The Kyoto Protocol allows carbon emissions to be offset by demonstrable removal of carbon from the atmosphere; this removal includes improved management of agricultural soils, as well as afforestation and reforestation (Intergovernmental Panel on Climate Change 2000).

It is essential for countries to estimate mitigation potential at a national scale. Several countries have published details of the mitigation potential of cropland management (e.g. Canada; Boehm et al. 2004). Such estimates, however, have not yet been published for Japan.

The most well-known agricultural practice that increases the sequestration of carbon in soils is no-tillage or reduced-tillage farming (Paustian et al. 1997). No-tillage techniques, however, would be difficult in many parts of Japan, as well as in other regions with humid climates, primarily because of problems with weeds. In these regions, other practices, including the application of compost and the use of “green manure” or multi-cropping, must be used to increase carbon input to soils and thus increase SOC storage.

Recently in Japan, the Ministry of Agriculture, Forestry and Fisheries (MAFF) calculated the potential carbon sequestration by compost application (Ministry of Agriculture, Forestry and Fisheries 2008). The MAFF roughly estimated that the application of compost at 10 and 15 t ha−1 year−1 (fresh weight) to all of Japan’s paddy soils and arable upland soils, respectively, would accumulate more SOC (by approximately 2 Mt C year−1) than if no compost were applied. This estimate was not, however, presented in a scientific paper and MAFF (2008) stated that further investigation was needed to develop a more reliable estimate.

There are several methods of estimating changes in SOC at a national scale. The IPCC Guidelines (Intergovernmental Panel on Climate Change 2006) provided a three-tiered approach. Tiers 1 and 2 are regression based. The tier 3 approach, with a dynamic model and spatially explicit data, is better if the data and model are available. The Rothamsted Carbon Model ([RothC] Coleman and Jenkinson 1996) is a leading model that is widely used, as are the CENTURY (Parton et al. 1987) and DNDC (Li et al. 1992) models. The RothC model has also been used for nation-scale estimation of carbon sequestration (Falloon et al. 2006; Smith et al. 2005, 2007b; van Wesemael et al. 2005).

The RothC model has been tested against data from long-term experiments in Japan. The model was modified for Andosols (Shirato et al. 2004) and for paddy soils (Shirato and Yokozawa 2005) so that changes in SOC with time can be well simulated at a plot scale. In addition, the spatial distribution of SOC in arable land in 1990 was recently calculated for Japan. By combining the modified RothC models and the spatial SOC distribution data, we attempted to estimate the country-scale changes in SOC that would result from the use of different soil management practices.

Specifically, our objective was to use the RothC model at a 1-km resolution to conduct a country-scale estimation of potential carbon sequestration through the application of organic matter to Japanese arable soils.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Model description

The current version of RothC, RothC-26.3, is derived from earlier versions developed by Jenkinson and Rayner (1977). It separates incoming plant residues into decomposable plant materials (DPM) and resistant plant materials (RPM), both of which undergo decomposition to produce microbial biomass (BIO) and humified organic matter (HUM) and to release CO2 (Coleman and Jenkinson 1996). The clay content of the soil determines the proportions of decomposed carbon allocated to CO2 and to BIO + HUM. The BIO and HUM fractions both undergo further decomposition to produce more CO2, BIO and HUM. The model also includes a pool of inert organic matter (IOM). Each compartment, except for IOM, undergoes decomposition by first-order kinetics at its own characteristic rate, which is determined by using modifiers for soil moisture, temperature and plant cover. The input parameters include monthly average air temperature, monthly precipitation, monthly open-pan evaporation, soil clay content, monthly C input from plant residues and/or farmyard manure (FYM), and monthly information on soil cover (whether the soil is bare or covered with vegetation).

Study area

We conducted a simulation by prescribing datasets required for running the RothC model, such as weather, soil and management data, on a 1-km grid basis. Grids in which 40% or more of the area was occupied by arable land were subjected to calculation. The percentage of arable land was derived from the 100-m-grid soil map of the Fundamental Soil Survey for Soil Fertility Conservation from 1953 to 1978 (Ministry of Agriculture, Forestry and Fisheries 1980). Concretely, if there were 40 or more 100-m-grids of arable soil within the 1-km grid, the 1-km grid was subjected to calculation. If the dominant soil in a grid was organic soil (locally known as muck soil or peat soil, corresponding to Histosols in the World Reference Base system; International Society of Soil Science et al. 1998), the grid was excluded from the calculation because the RothC model is not currently applicable to organic soils (Coleman and Jenkinson 1996).

Weather data

Monthly mean air temperature and monthly precipitation data were obtained from the gridded Automated Meteorological Data Acquisition System (AMeDAS) provided by the National Institute for Agro-Environmental Sciences. The original AMeDAS data were observed by the Japan Meteorological Agency. The station data were interpolated to grids of 45′ in longitude and 30′ in latitude (approximately 1 km × 1 km) on the terrain of Japan following the procedure developed by Seino (1993). Because open-pan evaporation data are not widely available in Japan, we instead estimated potential evapotranspiration from the air temperature (Thornthwaite 1948).

Soil data

Data from the dominant soil series in each 1-km grid were used for the simulation for simplicity, although there was often more than one soil series in a 1-km grid. The dominant soil series in each grid was derived from the 100-m-grid soil maps from MAFF’s Fundamental Soil Survey for Soil Fertility Conservation (1953–1978).

The mean value of each soil group (Table 1) was used to determine the clay content for all soil series belonging to the soil group because data were not available for all soil series. The percentage of pyrophosphate-extractable Al (Alp %), which is required for the modified model for Andosols (Shirato et al. 2004), was calculated from the total SOC concentration (%) as follows (Shoji et al. 1993):

Table 1. Soil organic carbon and clay content in 1990 for each soil group used for the simulations
Name of the soil groupCorresponding soil name in the WRB systemNo. soil seriesSOC in 0–30 cm (t ha−1)Clay (%)
AverageMinimumMaximum
  1. SOC, soil organic carbon; WRB, World Reference Base for Soil Resources.

LithosolsLeptsols273.346.086.516.0
Sand-dune RegosolsArenosols124.424.424.45.1
AndosolsAndosols61111.958.4212.714.0
Wet AndosolsAndosols48133.375.1283.116.1
Gley AndosolsAndosols, Gleysols14115.456.3162.713.3
Brown Forest soilsCambisols2369.550.6108.323.1
Gray Upland soilsGleysols, Planosols1573.545.8205.026.8
Gley Upland soilsGleysols, Planosols1156.549.080.740.0
Red soilsAcrisols, Alisols758.111.669.232.6
Yellow soilsAcrisols, Alisols2359.239.4166.521.0
Dark Red soilsAcrisols, Alisols645.236.550.234.1
Brown Lowland soilsFluvisols, Cambisols, Gleysols1954.824.087.116.6
Gray Lowland soilsFluvisols, Cambisols, Gleysols3862.444.292.218.5
Gley soilsFluvisols, Cambisols, Gleysols3766.137.697.625.3
  • image(1)

The 1990 value of SOC (t ha−1) in the top 30 cm of soil was calculated for each soil series and was used as the baseline in the present study. Therefore, a soil depth of 30 cm was used for all models. The carbon concentration and bulk density data for each soil series were obtained from the Basic Soil–Environment Monitoring Project (Stationary Monitoring) conducted by MAFF, in which SOC data were collected every 5 years from 1979 (e.g. first period, 1979–1982; second period, 1984–1987; third period, 1989–1992; fourth period, 1994–1997). The SOC data collected during the third period (1989–1992) were used to determine the 1990 value of SOC. Table 1 shows the average, minimum and maximum SOC (0–30 cm) of all soil series in each soil group.

Simulation procedure

In modeling each grid, we set the initial SOC content to the 1990 value (Table 1) and then simulated the changes in SOC with time for four management scenarios, which will be described in detail later. For each 1-km grid, the SOC of a dominant soil series in a grid in 1990 was set as the baseline, and the RothC was run to reach equilibrium with that baseline SOC under constant environmental conditions. As described by Jenkinson et al. (1999), assuming that the SOC content has reached equilibrium, the RothC model can be run inversely to calculate how much C needs to enter the soil annually to maintain a specified level of SOC. In this calculation process, the allocation of SOC into each of the five compartments (DPM, RPM, BIO, HUM and IOM) is determined. Therefore, we ran the model inversely to calculate the C input required to maintain the SOC content at 1990 levels, and we then set the SOC allocation in each of the five compartments to this equilibrium value. Soils were assumed to be covered with vegetation (summer crops) from May to October. For simplicity, the C input required to maintain the 1990 SOC level was added in October, the harvest month, as a single pulse because Coleman and Jenkinson (1996) reported that it makes little difference in the calculation of SOC content how the annual input is distributed, or even if it is all added in a single pulse. The DPM : RPM ratio was set at 1.44, a typical value for most agricultural crops and grasses (Coleman and Jenkinson 1996). The values of IOM were set by Eq. 2:

  • image(2)

Falloon et al. (1998), except for the IOM of Andosols, which was set to 0 because Andosols do not contain organic carbon when formed from fresh volcanic ash, although IOM is assumed to be present from the beginning of soil formation (Shirato et al. 2004).

The SOC content (t ha−1 in the top 30 cm), which was output by the model, was multiplied by the area (ha) of arable land within the grid to produce the total amount of SOC (t, 0–30 cm) in each grid. The area of arable land used for this calculation was the sum of the 1997 Digital National Land Information (land-use data) on the areas of paddy fields and other arable land.

Selection of three versions of RothC

The RothC model has been tested against data from long-term experiments in Japan, and has successfully simulated changes in SOC over time for non-volcanic upland soils (Shirato and Taniyama 2003). However, the original model was not successful in simulating carbon turnover in Andosols and paddy soils. Shirato et al. (2004) modified the model for Andosols by changing the HUM decomposition rate constant with concentration of pyrophosphate-extractable Al, taking the strong stability of humus in Andosols into account. Similarly, for paddy soils, Shirato and Yokozawa (2005) modified the model by tuning the decomposition rate constant of all pools separately for periods with and without submergence, on the basis of the slower decomposition rates of organic matter in paddy soils than in upland soils.

For grids in which the area of paddy fields was larger than that of upland crop fields, we used the RothC version modified for paddy soils (Shirato and Yokozawa 2005). If the dominant soil series was the Andosols group (Andosols, Wet Andosols and Gleyed Andosols), we used the model modified for Andosols (Shirato et al. 2004). For all other grids, the original version of the RothC model (Coleman and Jenkinson 1996) was used. The areas of paddy fields and upland crop fields were derived from 1997 Digital National Land Information (land-use data).

Soil management scenarios

Once the baseline SOC content had been established for each 1-km grid, four management scenarios were modeled for a 25-year period (Table 2).

Table 2. Carbon input as crop residue and farmyard manure in the four management scenarios
ScenariosCarbon input (t C ha−1 year−1)
PaddyUpland
Crop residueFYMCrop residueFYM
  1. Only roots and stubble enter the soil as residue. The amount was calculated from the yield as shown in Table 3; Fresh farmyard manure (FYM) was applied at a rate of 10 t C ha−1 year−1 and 15 t C ha−1 year−1 to paddy fields and upland fields, respectively. The concentration of C in the fresh FYM was assumed to be 10%.

A. Minimum organic matter application0.4600.410
B. FYM0.461.00.411.5
C. Double cropping for paddy fields0.46 + 0.7000.410
D. FYM + double cropping for paddy fields0.46 + 0.701.00.411.5
Scenario A (minimum organic matter application)

A minimum amount of crop residue (roots and stubble), 0.46 t C ha−1 year−1 for paddy fields and 0.41 t C ha−1 year−1 for upland fields, enters the soils. This scenario was compared with each of the other scenarios to assess the effects of the other scenarios on carbon sequestration.

Scenario B (farmyard manure application)

In this scenario, in addition to the minimum organic matter application of scenario A, FYM is applied at a rate of 1.0 t C ha−1 year−1 for paddy fields and 1.5 t C ha−1 year−1 for upland areas. This scenario was chosen because the application of FYM is regarded as a promising option to increase SOC in the Japanese agricultural system. The amount of FYM was set at the same level used by MAFF (2008), in which 10 and 15 t (fresh weight) ha−1 year−1 of FYM was applied to paddy and upland soils, respectively, and the carbon concentration of the fresh FYM was assumed to be 10%.

Scenario C (double cropping for paddy fields)

In this scenario, in addition to the treatment in scenario A, wheat residue (0.70 t C ha−1 year−1) from winter wheat cropping in all paddy fields was assumed to be input into the system. This scenario assesses the effects of increasing the carbon input from crop residues by mulch cropping, which is also considered to be an important option for increasing SOC.

Scenario D

In addition to the minimum treatment in scenario A, FYM is applied as described in scenario B and there is double cropping in the paddy fields, as described in scenario C.

The increase or decrease in SOC storage caused by each treatment was estimated by calculating the differences between scenarios B, C and D and scenario A.

The annual C inputs from crop residue and the application of FYM in each scenario are shown in Table 2. The parameters used to calculate the C inputs from crop residues are summarized in Table 3. The C input from crop residues (roots and stubble) in paddy soils was calculated from the average yield and the proportion of residue to yield of paddy rice, derived from the work of Ogawa et al. (1988). The C input from crop residue in upland soils was calculated by using similar data for wheat and soybeans (Ogawa et al. 1988), and the average of the two crops was used (Table 3) for simplicity.

Table 3. Calculation of the carbon input derived from crop residue (roots and stubble)
 Yield (t ha−1)(Roots + stubble)/yieldRoots + stubble (t C ha−1)
  1. Recent average yield. Moisture concentrations were assumed to be 15% for rice and soybean and 12.5% for wheat. Derived from the proportions of dry matter produced from each part of the crop as reported by Ogawa et al. (1988). Rice: grain, 37.2%; chaff, 8.2%; leaves and stalks, 44.6%: stubble, 6.7%; roots, 3.3%. Wheat: grain, 32.3%; chaff, 9.7%; leaves and stalks, 40.6%; stubble 9.4%; roots 8.1%. Soybean: grain, 33.1%; pod, 17.9%; leaves and stalks, 42.7%; stubble, 3.3%; roots, 3.0%. The C concentration of the dry matter was assumed to be 40%.

Paddy rice5.00.2300.46
Wheat3.70.4730.70
Soybean1.70.1620.11
Average of wheat and soybean  0.41

Again, for simplicity, the C input from plant residues was added to the soils only in the month of harvest: October for summer crops and April for winter crops. A DPM : RPM ratio of 1.44 was used for all types of plant residues, and the recommended values for FYM (DPM = 49%, RPM = 49% and HUM = 2%) were also used (Coleman and Jenkinson 1996). The months during which soil was covered by vegetation were set as May–October for summer crops and November–April for winter crops.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Soil organic carbon storage in 1990

The total area of arable land subjected to the simulation was 4.27 million ha: 2.39 million ha of paddy fields and 1.88 million ha of upland fields (Table 4). Gray Lowland soils, Gley soils and Wet Andosols occupied 76% of the area of paddy soils, and Andosols, Brown Forest soils and Brown Lowland soils occupied 69% of the upland soils area. These proportions are consistent with the findings of MAFF’s Fundamental Soil Survey for Soil Fertility Conservation (1953–1978), implying that the procedure used to represent a dominant soil series in a grid successfully approximated the actual area proportions for the soil types.

Table 4. Total baseline (1990) area and soil organic carbon of each soil group used for the simulation
Soil groupsArea (×1000 ha)Total SOC (Mt C) in 1990
PaddyUplandTotalPaddyUplandTotal
  1. SOC, soil organic carbon.

Lithosols0.64.24.80.040.320.37
Sand-dune Regosols2.716.319.00.070.400.46
Andosols120.7844.1964.813.9693.90107.86
Wet Andosols212.0107.4319.326.6116.1942.79
Gley Andosols24.28.933.12.691.123.81
Brown Forest soils41.8263.6305.42.8318.3821.21
Gray Upland soils60.990.1151.03.947.2511.19
Gley Upland soils40.08.648.62.260.482.75
Red soils4.029.533.50.231.711.94
Yellow soils127.296.1223.47.665.5813.24
Dark Red soils2.136.938.90.101.631.73
Brown Lowland soils156.0197.4353.39.439.9619.39
Gray Lowland soils873.7120.5994.254.667.5862.24
Gley soils725.357.8783.148.113.9152.03
Total2,391.11,881.44,272.4172.60168.40341.00

The total SOC of mineral soils in the top 30 cm of soil in 1990 (the sum of all of the grids) was 341 Mt; paddy soils and upland soils accounted for 173 Mt and 168 Mt, respectively (Table 4). This value is close to the MAFF (2008) estimate of 380 Mt, which included organic soils. Organic soils were excluded from our study, but if organic soils had been included then the total SOC would have been approximately 370 Mt.

The average SOC of all mineral soils was 79.8 t ha−1, and the average SOC for paddy soils and upland soils was 72.2 and 89.5 t ha−1, respectively. Soils in the Andosols group (Andosols, Wet Andosols and Gleyed Andosols) had high levels of C (>100 t ha−1).

Carbon input required to maintain the 1990 SOC level

On average, 2.8, 5.1 and 5.7 t C ha−1 year−1 of crop residues were required to maintain the 1990 level of SOC in paddy, upland (Andosols) and upland (non-Andosols) soils, respectively (Table 5).

Table 5. Amounts of carbon derived from crop residues and the carbon inputs required to maintain the 1990 soil organic carbon level
 Roots + stubbleOther residuesAll residuesRequired C input
t C ha−1t C ha−1 year−1
  1. Calculated from the yield data as shown in Table 3. Calculated by RothC. §5.06 for upland (Andosols) and 5.66 for upland (non-Andosols) soils.

Paddy rice0.462.402.862.76
Wheat0.702.002.705.06–5.66§
Soybean0.111.001.115.06–5.66§

The C inputs to soils in scenario A were 0.46 and 0.41 t ha−1 year−1 for paddy and upland soils, respectively (Tables 2,3). These values were much lower than the required inputs shown in Table 5, and it was obvious that SOC would decline (approximately 75 Mt C declined during the 25 years of our simulation) under this scenario.

In paddy soils, the total crop residue from a normal yield was calculated to be 2.86 t C ha−1, which included 0.46 t C ha−1 in roots and stubble and 2.40 t C ha−1 in leaves and straw (Table 5). This amount did exceed the required input of 2.8 t C ha−1, implying that SOC in paddy soils could be maintained or increased if all crop residues (not just the roots and stubble) were incorporated into the soils.

Conversely, the total amount of residue produced by wheat was estimated to be 2.7 t C ha−1 and that of soybeans was 1.1 t C ha−1; these amounts were clearly much lower than the required C inputs of 5.1 and 5.7 t ha−1, respectively (Table 5). These large amounts of required C input to maintain present SOC might result from the large C input (net primary production [NPP]) from original natural vegetation, such as forest (Kira 1975) or grassland (Caldwell 1975). For example, average annual above-ground net primary production rates (dry matter) of various types of Japanese forests of approximately 8–20 t ha−1 year−1 (Kira 1975) and 45% C concentration result in 3.6–9.0 t C ha−1 year−1 of C input. It is therefore difficult to maintain the SOC level of upland soils even if all residues are incorporated into the soils. In fact, the topsoil C of upland crop fields has been gradually decreasing recently, whereas that of paddy fields has remained almost constant (Nakai 2006).

Soil organic carbon accumulation effects

The SOC decreased during the 25-year simulation period in all four scenarios because the C inputs in all four scenarios (Table 2) were lower than the C inputs required to maintain the SOC level (Table 5). The rates of decrease were, however, different among the scenarios. As expected, scenario A resulted in the greatest depletion in SOC. The other scenarios had lower levels of depletion, reflecting the effects of organic matter application. The amount of SOC accumulated in scenario B (FYM application; 32.3 Mt) was greater than that in scenario C (double cropping in paddy fields; 11.1 Mt) as shown in Table 6. The accumulation in scenario D (using both techniques) was equal to the sum of the accumulations in B and C (43.4 Mt).

Table 6. Soil organic carbon accumulation under the three scenarios
ScenarioLand useTotal accumulation (25 years)Annual accumulation
1st year25th year25-year average1st 10-year average
Mt CMt C year−1
  1. FYM, farmyard manure.

B. FYMPaddy18.31.490.410.731.04
Upland14.01.460.320.560.83
All32.32.950.731.291.87
C. Double cropping for paddy fieldsPaddy11.10.930.250.450.63
Upland00000
All11.10.930.250.450.63
D. FYM + double cropping for paddy fieldsPaddy29.42.420.661.181.67
Upland14.01.460.320.560.83
All43.43.880.981.742.50

These accumulations increased rapidly in the early years of the simulation and then slowly increased toward equilibrium (Fig. 1a), although they had not reached equilibrium by the 25th year. The annual accumulations in scenarios B, C and D were 2.95, 0.93 and 3.88 Mt C year−1, respectively, for the first year, and decreased to 0.73, 0.25 and 0.98 Mt C year−1 for the 25th year, respectively (Table 6). A similar declining trend can be seen in the average accumulations for the first 10 years and the entire 25 years (Table 6).

image

Figure 1.  Soil organic carbon (SOC) accumulation effects in scenarios B, C and D, expressed as the difference between the SOC in each of the scenarios and that of scenario A (minimum organic matter application). (a) Total SOC accumulation effects, (b) SOC accumulation effects per unit area and (c) annual SOC accumulation effects per unit area. FYM, farmyard manure.

Download figure to PowerPoint

The effects of SOC accumulation per hectare are presented in Table 7. For the 25-year period, the per-hectare SOC accumulations in scenarios B, C and D were 7.55, 2.60 and 10.16 t ha−1, respectively. Again, there was a rapid increase at the beginning, followed by a slower rate of increase (Fig. 1b). Similarly, the annual SOC accumulations per hectare were large at the beginning and declined thereafter (Table 7; Fig. 1c).

Table 7. Soil organic carbon accumulation (per ha) under the three scenarios
ScenarioLand useAccumulation ha–1 (25 years)Annual accumulation ha–1
1st year25th year25-year average1st 10-year average
t C ha−1t C ha−1 year−1
  1. FYM, farmyard manure.

B. FYMPaddy7.660.620.170.310.43
Upland7.430.780.170.300.44
All7.550.690.170.300.44
C. Double cropping for paddy fieldsPaddy4.660.390.100.190.26
Upland00000
All2.600.220.060.100.15
D. FYM + double cropping for paddy fieldsPaddy12.311.010.280.490.70
Upland7.430.780.170.300.44
All10.160.910.230.410.58

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Total potential of carbon sequestration in Japanese arable land

The estimated annual SOC accumulation in response to the application of compost (scenario B) ranged from 0.73 to 2.95 Mt year−1 during the 25-year simulation, and the averages for the entire 25-year period and for the first 10 years were 1.29 and 1.87 Mt year−1, respectively (Table 6). These values are similar to MAFF’s (2008) estimation of a potential 2 Mt C year−1. As mentioned previously, the application rate of compost was the same in both estimates. The MAFF estimate was calculated on the basis of the difference between plots with and without compost and from continuous field observation datasets longer than 8 years, whereas our estimation was based on a 25-year model simulation. The two approaches are quite different, but both approaches produced similar estimates. The annual average for the first 10 years (1.87 Mt C year−1) was closer to the MAFF estimate (2.0 Mt C year−1) than was the 25-year average (1.29 Mt C year−1).

Per-area potential of carbon sequestration

The IPCC AR4 (Smith et al. 2007a) provided values for the per-area potential of mitigation technologies in agriculture. Its mitigation potentials for CO2 represent the net change in SOC derived from approximately 200 studies, primarily taken from the works of Ogle et al. (2005) and Smith et al. (2008). Mean estimates of annual per-area mitigation potentials for CO2 by agronomy, nutrient management, tillage and residue management range from 0.51 to 0.88 t CO2 ha−1 year−1 in cool-moist and warm-moist climate regions, which correspond to the Japanese climate. These values are equivalent to 0.14–0.24 t C ha−1 year−1. Our estimates (the 25-year average) of 0.30 t C ha−1 year−1 for compost application and 0.19 t C ha−1 year−1 for double cropping for paddy fields (Table 7) are similar to the IPCC AR4 estimates.

Trade-offs between soil carbon sequestration and CH4 or N2O emissions

In paddy soils, the increased C input to the soils with compost application causes increased CH4 emissions. MAFF (2008) referred to this trade-off between SOC accumulation and CH4 emissions and estimated that increased compost application (10 t ha−1 year−1; fresh weight) for all paddy fields may increase CH4 emissions by approximately 0.2 Mt C year−1 in terms of global warming potential (GWP), or approximately 10% of the SOC accumulation (2.0 Mt year−1). This suggests that the SOC accumulation effect of compost application vastly exceeds the negative effects of CH4 emissions. The rate of compost application was the same in our study, and we similarly expect that the positive effects of SOC accumulation will far outweigh the negative effects of increased CH4 emissions.

Emissions of N2O could also increase as the compost application rate increases because compost contains nitrogen. In our scenario, however, we assumed that an increase in the application rate of compost would coincide with a reduction in the use of chemical fertilizers; thus, total N2O emissions will not change.

Double cropping of paddy fields will cause an increase in N2O emissions from the use of fertilizers for wheat. A rough estimate of this increase is approximately 0.2 Mt C year−1, assuming an N fertilization rate of 100 kg N ha−1, an emission factor of 0.62 (Akiyama et al. 2006), an area of 2.4 million ha (Table 4) and a GWP of 298 (100-year time horizon; Forster et al. 2007). Thus, this negative effect is lower than the SOC accumulation (range, 0.25–0.93 Mt year−1; 25-year average, 0.45 Mt year−1).

Overall, the effect of SOC accumulation exceeded the negative effects of increasing CH4 and N2O emissions under all three scenarios.

Feasibility of the soil management scenarios

In scenario A, we assumed that only roots and stubble entered the soil and that other residues (e.g. leaves and stalks) were removed from the fields. Obviously, this amount of C input was smaller than what actually occurs. Hence this scenario was not business as usual because some of the above-ground residues are generally incorporated into the soils. This scenario was created as a basis of comparison with the other scenarios to assess the effects of the other scenarios on carbon sequestration; it was not set to simulate SOC under a business as usual scenario.

In scenario B (compost application), the rate of compost application (1.0 t C ha−1 year−1 for paddy fields and 1.5 t C ha−1 year−1 for upland crop fields) was set to be equal to the amount recommended by MAFF (2008). Although this application rate is feasible, the assumption that compost would be applied at the same rate to all arable soils is not realistic. The application rate has actually been decreasing and has declined from 4.51 t ha−1 in 1970 to 0.88 t ha−1 in 2005 for paddy rice and from 3.90 t ha−1 to 0.89 t ha−1 for wheat, although the application rate for vegetables has stayed relatively high (19.0 t ha−1 from 1994 to 1999; Ministry of Agriculture, Forestry and Fisheries 2008). A more realistic rate of compost application would be smaller than the rate used in our scenario. In addition, the assumption of compost application to all arable land is probably not realistic, and a smaller area of application would result in a smaller potential SOC accumulation. Our estimates of total C sequestration potential with this scenario might be an overestimate in terms of both the area of implementation and per-unit-area potential.

In scenario C (double cropping for paddy fields) we set the C input from winter wheat by assuming that only roots and stubble entered the soil and that the other residues were removed from the fields. The real amount of C input from winter wheat cropping would be much greater than the amount that we assumed because some of the above-ground residues are incorporated into the soil; hence, the C input per unit area would be greater than our scenario indicates. Conversely, we also assumed winter wheat cropping in all paddy fields, which is clearly not realistic. Some paddy fields are already double cropped, and winter cropping is difficult in some regions because of the climate. In this scenario, we may be underestimating the per-area potential and overestimating the area of implementation. In addition, mulch cropping occurs not only in paddy fields, but also in upland fields. From this perspective, our simulation should be considered as an example of this type of analysis.

More feasible scenarios of compost and crop residue application both for area of implementation and for rate per unit area need to be developed to generate better estimates of potential C sequestration. We conclude that, regardless of the problems listed above, the system we developed in the present study, that is, linking the RothC model and soil spatial data, can be useful to estimate the potential C sequestration resulting from an increase in organic matter input to Japanese arable soils.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

We thank Mr Kevin Coleman (Rothamsted Research, UK) for help and advice with the RothC model and Dr Hiroko Akiyama (National Institute for Agro-Environmental Sciences, Japan) for advice on calculating N2O emissions. This work was financially supported by the Ministry of Agriculture, Forestry and Fisheries, Japan (Evaluation, Adaptation and Mitigation of Global Warming in Agriculture, Forestry and Fisheries).

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  • Akiyama H, Yan X, Yagi K 2006: Estimations of emission factors for fertilizer-induced direct N2O emissions from agricultural soils in Japan: summary of available data. Soil Sci. Plant Nutr., 52, 774787.
  • Boehm M, Junkins B, Desjardins R, Kulshreshtha S, Lindwall W 2004: Sink potential of Canadian agricultural soils. Clim. Change, 65, 297314.
  • Caldwell M 1975: Primary production of grazing lands. In Photosynthesis and Productivity in Different Environments, Ed. JPCooper., pp. 4173, Cambridge University press, Cambridge.
  • Coleman K, Jenkinson DS 1996: RothC-26.3 – A model for the turnover of carbon in soil. In Evaluation of Soil Organic Matter Models, Ed. DSPowlson, PSmith and JUSmith., pp. 237246, Springer-Verlag, Berlin.
  • Falloon P, Smith P, Bradley RI et al. 2006: RothCUK – a dynamic modelling system for estimating changes in soil C from mineral soils at 1-km resolution in the UK. Soil Use Manage., 22, 274288.
  • Falloon P, Smith P, Coleman K, Marshall S 1998: Estimating the size of the inert organic matter pool from total soil organic carbon content for use in the Rothamsted carbon model. Soil Biol. Biochem., 30, 12071211.
  • Forster P, Ramaswamy V, Artaxo P et al. 2007: Changes in atmospheric constituents and in radiative forcing. Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge University Press, Cambridge, UK.
  • Intergovernmental Panel on Climate Change 2000: Land Use, Land-Use Change, and Forestry. Cambridge University Press, Cambridge, UK.
  • Intergovernmental Panel on Climate Change 2006: 2006 IPCC Guidelines for National Greenhouse Gas Inventories, Prepared by the National Greenhouse Gas Inventories Programme. Institute for Global Environmental Strategies (IGES), Japan.
  • ISSS, ISRIC, FAO 1998: World Reference Base for Soil Resources, World Soil Resources Reports 84. FAO, Rome.
  • Jenkinson DS, Meredith J, Kinyamario JI et al. 1999: Estimating net primary production from measurements made on soil organic matter. Ecology, 80, 27622773.
  • Jenkinson DS, Rayner JH 1977: The turnover of soil organic matter in some of the Rothamsted classical experiments. Soil Sci., 123, 298305.
  • Kira T 1975: Primary production in forests. In Photosynthesis and Productivity in Different Environments, Ed. JPCooper., pp. 540, Cambridge University press, Cambridge.
  • Li C, Frolking S, Frolking TA 1992: A model of nitrous oxide evolution from soil driven by rainfall events: 1. Model structure and sensitivity. J. Geophys. Res., 97, 97599776.
  • Ministry of Agriculture, Forestry and Fisheries 1980: Actual conditions of Japanese arable soils. Jpn. J. Soil Sci. Plant Nutr., 51, 520527 (in Japanese).
  • Ministry of Agriculture, Forestry and Fisheries 2008: Final report from the expert meeting for Sustainable agriculture. Available at: http://www.maff.go.jp/j/study/kankyo_hozen/pdf/h2004_report.pdf; accessed 3/11/2009 (in Japanese).
  • Nakai M 2006: The present condition and problems of Japanese arable soil survey. Nougyou (Agriculture), 1487, 3142 (in Japanese).
  • Ogawa K, Takeuchi Y, Katayama M 1988: Biomass production and the amounts of absorbed inorganic elements by crops in arable lands in Hokkaido, and its evaluation. Res. Bull. Hokkaido Natl. Agric. Exp. Stn., 149, 5791 (in Japanese with English summary).
  • Ogle SM, Breidt FJ, Paustian K 2005: Agricultural management impacts on soil organic carbon storage under moist and dry climatic conditions of temperate and tropical regions. Biogeochemistry, 72, 87121.
  • Parton WJ, Schimel DS, Cole CV, Ojima DS 1987: Analysis of factors controlling soil organic matter levels in Great Plains grasslands. Soil Sci. Soc. Am. J., 51, 11731179.
  • Paustian K, Andren O, Janzen HH et al. 1997: Agricultural soils as a sink to mitigate CO2 emissions. Soil Use Manage., 13, 230244.
  • Seino H 1993: An estimation of distribution of meteorological elements using GIS and AMeDAS data. J. Agric. Meteorol., 48, 379383.
  • Shirato Y, Hakamata T, Taniyama I 2004: Modified Rothamsted carbon model for Andosols and its validation: changing humus decomposition rate constant with pyrophosphate-extractable Al. Soil Sci. Plant Nutr., 50, 149158.
  • Shirato Y, Taniyama I 2003: Testing the suitability of the Rothamsted carbon model for long-term experiments on Japanese non-volcanic upland soils. Soil Sci. Plant Nutr., 49, 921925.
  • Shirato Y, Yokozawa M 2005: Applying the Rothamsted Carbon Model for long-term experiments on Japanese paddy soils and modifying it by simple tuning of the decomposition rate. Soil Sci. Plant Nutr., 51, 405415.
  • Shoji S, Nanzyo M, Dahlgren RA 1993: Volcanic Ash Soils. Elsevier, Amsterdam.
  • Smith P, Martino D, Cai Z et al. 2007a: Agriculture. Climate Change 2007: Mitigation. Contribution of Working Group III to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, UK.
  • Smith P, Martino D, Cai Z et al. 2008: Greenhouse gas mitigation in agriculture. Philos. Trans. R. Soc. B, 363, 789813.
  • Smith J, Smith P, Wattenbach M et al. 2005: Projected changes in mineral soil carbon of European croplands and grasslands, 1990-2080. Glob. Chang. Biol., 11, 21412152.
  • Smith J, Smith P, Wattenbach M et al. 2007b: Projected changes in the organic carbon stocks of cropland mineral soils of European Russia and the Ukraine, 1990-2070. Glob. Chang. Biol., 13, 342356.
  • Thornthwaite CW 1948: An approach toward a rational classification of climate. Geogr. Rev., 38, 5594.
  • Van Wesemael B, Lettens S, Roelandt C, Van Orshoven J 2005: Modelling the evolution of regional carbon stocks in Belgian cropland soils. Can. J. Soil. Sci., 85, 511521.