ORGANIC CARBON STORAGE IN SOILS OF TROPICAL AND SUBTROPICAL CHINA LI ZHONG∗ , JIANG XIAO, PAN XIANZHANG and ZHAO QIGUO Institute of Soil Science, Academia Sinica, P.O. Box 821, Nanjing 210008, P.R. China (∗ author for correspondence, e-mail: [email protected])

(Received 26 July 1999; accepted 17 May 2000)

Abstract. Tropical and subtropical China occupies a total area of about 215 Mha, but the contents and storage of soil organic carbon (C) on this land have not been well studied. This research was conducted in this concern by jointly using soil survey data and the regional soil map. It was estimated that the 0–20 cm C density of the soil subgroups in the region varied from 1.2 to 9.7 kg C m−2 . In the major zonal soils, soil subgroup ‘young ∼’ had generally a C density about 1.2–4.4 times lower than its other natural counterparts. Furthermore, subgroups of the cultivated soil gave a C density about 20–63% lower than their natural counterparts (except subgroup ‘young ∼’), showing remarkable C losses caused by cultivation. For soil groups in the east, brown soil had the highest C density (6.8 and 21.4 kg C m−2 , for the upper 20 and 100 cm soil, respectively), followed by yellow soil, yellow brown soil and limestone soil (4.6–5.5 and 12.3–14.5 kg C m−2 , idem). Torrid red soil, purplish soil and fluvo-aquic soil had the lowest C density (2.0–2.2 and 5.8–7.5 kg C m−2 , idem) with the others ranking at a medium level. In the west, podzolic soil ranked the highest (17.9 and 55.8 kg C m−2 , idem), followed by dark brown soil, brown soil and limestone soil (6.9–11.6 and 21–29 kg C m−2 , idem), and again torrid red soil, purplish soil and fluvo-aquic soil ranked the lowest (3.9–5.1 and 11.4–14.1 kg C m−2 , idem). Finally, a total of about 26.8±7.4 Pg organic C was estimated being stored in soils of the entire region. Keywords: C density, C storage, soil, tropical and subtropical China Abbreviations: BS – brown soil, LS – limestone soil, PS – purplish soil, RS – red soil, YS – yellow soil, CSS – coastal solonchak and sandy soil, DBS – dark brown soil, FQS – fluvo-aquic soil, LOT – lotosol, LRS – lateritic red soil, MMS – mountain meadow soil, PDS – paddy soil, PZS – podzolic soil, TRS – torrid red soil, YBS – yellow brown soil, OTS – other soils.

1. Introduction With the increasing concern of the global climate change, the global C cycle has been addressed more ever than before by interdisciplinary scientists. However, we still cannot balance the global C budget (Schimel et al., 1995; Schlesinger, 1993), and our knowledge in predicting future changes is challenged by a number of unknowns. Soil organic C makes a major pool of the global C cycle (Schlesinger, 1977; Post et al., 1982; Eswaran et al., 1993), and plays a critical role in regulating the atmospheric CO2 (Post et al., 1990; Schlesinger, 1991). But an accurate estimate of the global soil C storage and its geographic distribution is hampered by lack of reliable data (Eswaran et al., 1995; Arnold, 1995), and the capacity of the soil Water, Air, and Soil Pollution 129: 45–60, 2001. © 2001 Kluwer Academic Publishers. Printed in the Netherlands.

46

L. ZHONG ET AL.

to sequester C is still not sufficiently elucidated. An improved understanding of the C status of the regional soils can help narrow our knowledge gaps in this concern, and provide options to assess regional soil management in C sequestration. Tropical and subtropical China covers an area of about 215 Mha, and divers largely in soil types, vegetation, geography and land uses. But its soil organic C contents and variations have not been well documented. This study was intended to estimate the organic C content, storage and geographic distribution of soils in the region based on data from the Second National Soil Survey and the regional soil map.

2. Materials and Methods 2.1. DATA SOURCE AND COLLECTION The whole area confined by Map of Soils in the Red Soil and Yellow Soil Region of China (1:4 M) was defined as tropical and subtropical China in this study, which matches roughly the common concept by Chinese meteorologists and soil scientists. As evident difference exists in climate, topography, vegetation, landuse and social economy, the region is traditionally divided into two parts: the east and the west, each of which was addressed separately in this article (Figure 1). The soil map mentioned above was used to characterize the geographic distribution of soils in the entire region. The primary data we employed were based on the Second National Soil Survey collected from Soil Species of China Volume I, III and VI (National Soil Survey Office, 1993, 1994, 1996). Provincial records of soil species from each related area (some unpublished) were also used to enlarge the database of soil pedons and to enhance representativity. As a result, about 15450 soil profiles were totally collected, of which 6326 were natural pedons and the others were cultivated soils. Because soil species data of Taiwan were not available, mean values of the C density of the east soils were correspondingly adopted to ensure integrality. The soil classification system employed in the Second National Soil Survey was not completely consistent with the legend of the Map of Soils in the Red Soil and Yellow Soil Region of China (1:4 M). The mergence of soil units, especially at subgroup level, was conducted for some soils to make the two systems identical and compatible. Basically, soil subgroups were merged in light of the principle of approximation. 2.2. C OMPUTATION OF SOIL ORGANIC C DENSITY 2.2.1. To Profile Depth For a soil with n layers, its organic C density (SOCd , in kg C m−2 ) was calculated by using the classical layer-based method as:

ORGANIC CARBON STORAGE IN SOILS

47

Figure 1. The area of study.

SOCd =

n


Ti ϑi Ci (1 − δi )/10

(1)

i=1

where Ti is the thickness (cm) of layer i, ϑi is the bulk density (g cm−3 ) of layer i, Ci is the C content (%) in layer i, and δi is the volumetric percentage of fragments larger than 2 mm. For soils without measured bulk density (about 70% of the total), mean values of their corresponding subgroups or groups were alternatively used. For soil horizons without measured stoniness but defined qualitatively as one of slight stony, medium stony or heavy stony, the median of 0–15%, 15–50% or 50–100% was appointed, respectively, keeping in view the original definition. C content was converted from organic matter concentration using a conversion coefficient of 1.724 since organic matter was determined by wet combustion with Cr2 O2− 7 (Wen, 1984). The C density of a given soil subgroup (SOCs ) supported by n soil species was computed by using

48

L. ZHONG ET AL.

SOCs =

n


(2)

SOCi f (i)

i=1

where SOCi is the C density of soil species i, and f (i) is the area fraction of species i in the total subgroup. The C density of a given soil group was averaged in the same way but based on all the subgroups it included. The total area of a soil subgroup or a soil group obtained by adding up all the collected species under the class was commonly much less than the total we got by digitizing the soil map, but the former was assumed to be a representative sample of the latter. 2.2.2. To 20 and 100 cm Depth In the Second National Soil Survey, soil profiles were divided into genetic horizon A, B, C, D, . . . . The upper 0–20 cm layer was not particularly identified, and profiles did not go exactly to 100 cm depth. To estimate C density in the upper 20 cm soil, we used only horizon A when A
20 cm, or used both horizon A and B when A < 20 cm. In the former case, it was computed by using all the other characteristics of horizon A but its depth. In case of A < 20 cm, a compensative layer from B was taken with its properties assigned unchangedly from the original B and then the upper 20 cm soil C density was able to be approximately estimated. The C density of 0–100 cm soil (SOC100 ) was estimated for each soil group by SOC100 = SOCp + SOCp−100

(3)

where SOCp is the C density to profile depth, and SOCp−100 is the C density from profile depth to 100 cm. SOCp was calculated by using the layer-based method as discussed in the previous section, but to estimate SOCp−100 , a quantitative relationship between organic C content and depth was first established (see Section 3.2), and then with other characteristics of the lowest horizon extrapolated to this layer, SOCp−100 was able to be calculated by employing the integral calculous approach. Area-weighted means were presistently used throughout this study whenever C density and profile depth needed to be averaged.

3. Results an Discussion 3.1. M AIN SOILS AND THEIR DISTRIBUTION Defined as given above, tropical and subtropical China is located from latidude 4 to 32 degrees north. The mean annual temperature of the region varies from 14 to 28 ◦ C and the annual rainfall from 1200–2500 mm, with both variables appearing larger in the south than in the north. Moreover, as the hilly east is to the pacific sea while the plateau and mountainous west is shielded by the Himalayas, district differences in climate exist between the two subregions (CGRY, 1985).

ORGANIC CARBON STORAGE IN SOILS

49

Figure 2. Main soils in tropical and subtropical China.

Affected mainly by climate, there are as many as 18 soil groups in the region according to the Chinese Soil Taxonomy. By digitizing calculation, Map of Soils in the Red and Yellow Soil Region of China (1:4 M) gives a total land area of about 215 Mha, of which around 90% is covered by red soil (29%), yellow soil (16%), lateritic red soil (13%), limestone soil (8%), purplish soil (6%) and paddy soil (14%). Lotosol as a tropical zonal soil amounts only to about 2.5% (Figure 2). 3.2. V ERTICAL DISTRIBUTION OF SOIL ORGANIC C It is well known that soil organic C level declines commonly with depth, but as far as the quantitative relationship between the two is concerned, quite limited information we can find from the literature. Simple power and exponential functions in particular have been proven applicable in describing C accumulation in profiles for some soils (Bennema, 1974; Arrouays and Pelissier, 1994; Bernoux et al., 1998). But, more complex patterns of C with depth have been identified in others (Elzein and Balesdent, 1995; Stone et al., 1993). Most soils in tropical and subtropical China supported the C-declining observation, but also showed that there are soils such as coastal sandy soil and coastal solonchaks, in which C distributes unanimously throughout the entire profile (Figure 3). The latter is likely due to the low plant debris controlled by limited plant growth and the frequently variable hydrological conditions of these soils. For the former, our simulation analysis at both soil group and subgroup levels indicated

50

L. ZHONG ET AL.

Figure 3. Soil organic C distribution with depth. (a) Uncultivated lateritic red soil, western; (b) yellow red soil, eastern; (c) cultivated yellow soil, western; (d) coastal solonchak and sandy soil.

that the relationship between organic C content and soil depth can be described by equation: C = (c + D)/(a + bD)

(4)

C = 1/(a + bD)

(5)

or

where C is organic C content (%), and D is depth (cm). For soils in the east, Equation (4) gave correlation coefficients (R2 value) varying from 0.28–0.85, averaged 0.47; Equation (5) yielded R2 values varying from 0.24–0.85, averaged 0.45. In the west, Equation (4) was supported by R2 values ranging from 0.22–0.62, averaged 0.39, while Equation (5) had R2 values from 0.22–0.55, averaged 0.38. The results show that Equation (4) seems better than Equation (5), however, sometimes the values of a, b and c were too large, causing inconvenience to use.

51

ORGANIC CARBON STORAGE IN SOILS

TABLE IA Soil organic C density categorized by subgroup

Soil group

FAO classification

Subgroup

Latosol

Haplic Acrisols

Lateritic red soil

Haplic Acrisols/ Alisol

Lotosol Redish ∼ Yellowish ∼ Cultivated ∼ Lateritic red soil Yellowish ∼ Young ∼ Cultivated ∼ Red soil Yellow ∼ Cinnamon red ∼ Young ∼ Cultivated ∼ Yellow soil Surface gleyed ∼ Young ∼ Cultivated ∼ Yellow brown soil Young ∼ Purplish soil Cultivated ∼ Red ∼ Brown ∼ Yellow ∼ Black ∼ Cultivated ∼ Fluvo-aquic soil Cultivated ∼

Red soil

Haplic Alisol/ Haplic Acrisol

Yellow soil

Haplic Alisol

Yellow brown soil Purplish soil

Ferric/Haplic Luvisol Calcaric Regosol

Limestone soil

Regosol/Leptisol

Fluvo-aquic soil

Fluvisol

Eastern part No. of Profile depth profiles (cm) 672 – 197 28 204 39 140 684 326 42 162 618 533 91 80 10 47 16 369 21 32 53 36 12 104 250 255

99.0±6.8 – 100.0±7.3 90.5±8.1 101.3±6.8 – 97.3±10.7 101.5±7.7 99.4±2.8 96.6±5.6 97.8±9.1 83.9±17.1 98.9±8.4 96.1±9.5 86.5±11.6 56.7±14.8 99.6±7.9 92.0±9.0 54.0±5.7 86.2±11.0 92.9±6.6 100.3±12.3 53.7±16.0 89.6±11.7 62.6±15.1 97.3±4.9 100.5±2.3 96.8±3.4

C density, kg C m−2 Profile 0–20 cm 8.01±2.71 – 10.63±3.73 4.85±0.84 8.80±2.31 – 7.62±2.55 6.39±1.62 8.46±2.37 11.27±2.71 7.22±2.70 3.88±1.79 9.42±3.62 13.43±4.28 12.74±4.44 7.30±3.56 9.55±3.00 11.34±4.82 9.11±3.15 6.36±2.84 5.82±1.17 16.67±6.78 7.51±3.03 12.66±4.35 13.71±2.99 10.53±3.36 6.97±1.53 8.07±2.11

2.29±1.00 – 3.29±1.51 1.17±0.19 3.00±1.17 – 2.49±0.95 1.95±0.47 3.38±1.02 4.60±1.30 2.69±0.88 1.54±0.74 2.71±0.83 5.78±2.11 6.03±2.49 3.99±1.62 2.18±0.59 4.36±1.18 5.60±2.74 2.22±0.47 2.01±0.62 4.51±2.80 3.59±1.35 4.81±1.96 6.31±2.03 2.65±1.24 2.02±0.60 2.37±0.49

3.3. S OIL ORGANIC C DENSITY IN ACCORDANCE WITH SOIL SUBGROUPS Table I gives the C density of the soil subgroups, and soils which were not classified to the subgroup level due to their small area were not included here. One point we can draw from Table I is that profiles of subgroup ‘young ∼’ generally went shallower than the others within the same group, which led, therefore, to a smaller C density. The C density of the upper 20 cm soil seemed not having been affected so strongly, but it occurred inevitably much less than that of its natural counterparts. In the east, for example, of the three major zonal soils as red soil, yellow soil and lateritic red soil, the C density of their natural subgroups, both to 20 cm and profile

52

L. ZHONG ET AL.

TABLE I Continued

Soil group

FAO classification

Latosol

Haplic Acrisols

Subgroup

Lotosol Redish ∼ Yellowish ∼ Cultivated ∼ Lateritic red soil Haplic Acrisols/ Lateritic red soil Alisol Yellowish ∼ Young ∼ Cultivated ∼ Red soil Haplic Alisol/ Red soil Haplic Acrisol Yellow ∼ Cinnamon red ∼ Young ∼ Cultivated ∼ Yellow soil Haplic Alisol Yellow soil Surface gleyed ∼ Young ∼ Cultivated ∼ Yellow brown soil Ferric/Haplic Yellow brown soil Luvisol Young ∼ Purplish soil Calcaric Regosol Purplish soil Cultivated ∼ Limestone soil Regosol/Leptisol Red ∼ Brown ∼ Yellow ∼ Black ∼ Cultivated ∼ Fluvo-aquic soil Fluvisol Fluvo-aquic soil Cultivated ∼

Western part No. of Profile depth C density, kg C m−2 profiles (cm) Profile 0–20 cm 20 25 27 14 134 40 1 126 254 185 79 9 355 456 66 52 283 111 17 268 267 8 31 141 27 103 50 50

69.7±13.4 99.2±5.8 92.0±8.1 97.9±4.6 90.1±9.6 88.7±8.44 100.0 96.5±5.0 96.9±4.62 88.2±7.9 85.4±12.3 52.5±17.8 97.1±7.0 83.8±6.5 81.9±19.6 76.8±15.2 78.0±12.7 86.7±16.8 90.9±9.6 76.6±17.7 73.4±15.2 81.4±16.5 97.9±13.0 74.1±21.1 70.3±20.4 100.8±7.8 104.0±3.7 104.0±3.9

4.22±1.31 11.65±3.56 13.37±5.02 5.73±1.19 12.45±3.39 21.46±8.75 5.96 12.36±3.44 10.99±1.77 11.81±3.36 9.79±2.78 5.38±1.90 11.35±3.13 11.76±3.65 12.35±4.05 7.27±3.28 9.00±0.47 16.43±4.78 4.91±2.03 5.10±1.11 4.77±0.92 13.63±2.85 27.59±7.37 12.23±3.22 21.80±10.09 11.61±2.46 9.80±2.15 9.80±2.22

1.61±0.40 4.18±1.44 4.49±1.67 1.65±0.25 4.67±0.86 8.37±3.33 1.47 4.78±0.79 4.23±0.73 4.68±1.29 3.52±1.04 2.40±0.89 3.94±1.00 5.45±1.31 6.99±2.42 3.13±1.35 3.25±0.66 6.82±0.68 1.67±0.70 1.82±0.33 1.59±0.55 5.18±1.21 8.08±2.04 6.39±1.17 9.68±5.3 3.17±0.77 2.19±0.66 2.19±0.65

depth, was 1.2–2.3 times that of the ‘young ∼’ mates. In the west, it occurred 1.7–2.3 times higher to profile depth, and 1.7–4.4 times higher to 20 cm depth. Soils within the same great group are formed under similar climatic and geological conditions, so the low C density of this skeletally characterized subgroup can partly be attributable to anthropogenic factors (Li and Zhao, 1988; Zhao et al., 1997). Numerous studies indicated that deforestation, shifting cultivation and other land use changes can cause considerable soil C losses (Woomer et al., 1994; Nye and Greenland, 1960; Tiessen and Santos, 1989), and it is probably applicable to the present study. In addition, soil erosion induced by native vegetation destruction is

ORGANIC CARBON STORAGE IN SOILS

53

also a key contributor to the low C of this subgroup, but the fate of the ‘off-land C’ is not known in context of the global change. Except for the subgroup of ‘young ∼’, cultivated soil subgroups had relatively low C densities too, especially in the upper 20 cm soil. Of the three major zonal soils in the east, they occurred about 20–63% less in the upper 20 cm soil as compared with the averages of their natural counterparts; in the west, except the red soil, a decline of about 20–47% was observed. Land conversion for agriculture generally results in losses in organic C (Detwiler, 1986; Mann, 1986; Anderson, 1995), and results from various authors vary widely, dependent on climate, soil and managements (Sanchez et al., 1989; Lal, 1976; Woomer et al., 1994). On average, a 20–50% loss of organic C was generally reported. The difference in C density between the natural and cultivated subgroups could give information on to what extent cultivation had caused C losses in the region. The reason of C losses after cultivation is due mainly to an increased rate of decomposition and the reduced amount of return of plant residues. Figure 4 presents spatially the 0–20 cm soil C density in the entire region. It was based on C density at the subgroup level. It shows clearly that in the east, soil C density in the upper 20 cm soil was dominantly within 2–4 kg C m−2 , while in the west, almost half was at the level of 4–6 kg C m−2 . It is notable that the majority of the soil subgroups in the west had higher C densities than their east counterparts. It can be partly attributed to variations in climatic conditions between the two subregions, but more intensive human disturbance in the east than in the west is also responsible. As the Chinese soil classification system differentiates from the widely cited FAO/UNESCO soil unit and the USDA taxonomy, we can hardly compare our results with others, especially at the subgroup level. Another problem with our results is that most profiles of ours could not reach 100 cm depth, although they are close to, which makes it disputable even to compare one with another. Fortunately, most soils gave C densities close to that at the 100 cm depth (see the following section). 3.4. S OIL ORGANIC C DENSITY IN ACCORDANCE WITH SOIL GROUPS After knowing the quantitative relationship between C content and depth of each soil group, we can estimate its C density to 100 cm depth based on the methodology stated in the previous section. Figure 5 presents the estimated C density of the east soils to profile and to 100 cm depth. It indicates that for east soils, the largest diference between SOC100 and SOCp fell to brown soil and yellow brown soil (about 10%), with less variations occurring to the others. West soils differentiated themselves by about 3–19%. Figures 6 and 7 give C density in 0–20 and 0–100 cm soil, respectively. In the east, brown soil had the highest C density (6.78 and 21.4 kg C m−2 , respectively, in the upper 20 cm and 100 cm soil), followed by yellow soil, yellow brown soil and

Figure 4. Organic C density in the 0–20 cm soils of tropical and subtropical China.

54 L. ZHONG ET AL.

ORGANIC CARBON STORAGE IN SOILS

Figure 5. C density to profile and to 100 cm depth, east soil.

Figure 6. Organic C density of the 0–20 cm horizon.

55

56

L. ZHONG ET AL.

Figure 7. Soil C density to 100 cm depth.

limestone soil (4.6–5.5 and 12.3–14.5 kg C m−2 , idem). Torrid red soil, purplish soil and fluvo-aquic soil had the lowest C density (2.0–2.2 and 5.8–7.5 kg C m−2 , idem) with the others ranking at a medium level. In the west, the C density of podzolic soil ranked the highest (17.9 and 55.8 kg C m−2 , idem), followed by dark brown soil, brown soil and limestone soil (6.9–11.6 and 21–29 kg C m−2 , idem) and again torrid red soil, purplish soil and fluvo-aquic soil exhibited the lowest C density (3.9–5.1 and 11.4–14.1 kg C m−2 , idem). Figure 8 gives the spatial distribution of soil organic C density in the upper 100 cm layer. Different from Figure 4, it was based on the averaged C density at the soil group level. It indicates that the majority of the eastern soils had a C density of 8–10 kg C m−2 , while the western soils were diversity supported by values from 11–12, 12–14, 14–16, and 16–60 kg C m−2 . Our zonal soils as lotosol, lateritic red soil, red soil and yellow soil, gave C densities of 2.4–5.5 and 8.2–13.4 kg C m−2 in the east, respectively for the upper 20 and 100 cm soil, while in the West, they showed values of 3.9–5.1 and 11.4– 14.1 kg C m−2 , respectively. Sanchez et al. (1982) reported a value of 8.3 kg C m−2 for 61 randomly-selected tropical profiles, which is relatively in consistent with the present findings for eastern lotosol, lateritic red soil and red soil, but occurs quite less than that of the eastern yellow soil and all the western zonal soils. The C density of Oxisols given by Sanchez et al. (1982) is 11.3 kg C m−2 , which is in accordance with that of the Ferralsols or Oxisols in North America estimated by

Figure 8. Organic C density in the 0–100 cm soils of tropical and subtropical China.

ORGANIC CARBON STORAGE IN SOILS

57

58

L. ZHONG ET AL.

Figure 9. Organic C storage in the upper 100 cm soil of the region.

Bohn (1976) and matches the results of our western lotosol and red soil. But it occurs higher than our estimates for the eastern zonal soils (except yellow soil) and is less than the values of our western lateritic red soil and yellow soil. Bohn’s (1976) value of 8 kg C m−2 for Ultosol (or Acrisol) is quite close to Kimble et al. (1990a) global Ultisol C density (8.3 kg C m−2 ) but slightly less than Kimble’s estimate for global Oxisol (9.7 kg C m−2 ), and all these results are comparable with the C density of our eastern lotosol, lateritic red soil and red soil, but less than that of our western zonal soils and the eastern yellow soil. 3.5. S OIL ORGANIC C STORAGE Multiplied by the area of a soil group, its C density gives the amount of organic C stored in the soil, and then by summing the C storage of all the soils in a region, we can easily estimate the regional C stocks. Based on our study, the total C storage in the upper 100 cm soils of tropical and subtropical China is about 26.8±7.4 Pg, in which about 52% is stored in the four zonal soils as red soil, yellow soil, alteritic red soil and lotosol. As the four zonal soils occupy about 61% of the total land area, about 10% larger than their C storage percentage, their C density occurs relatively lower than that of the regional average. Lotosol, as the only tropical soil in the region, contributes only 2% to the total due to its smaller area. Paddy soil, with an area of 13.8%, stores 13% of the total C, indicating that its C density is just around the regional average (Figure 9). Various authors have estimated the amount of organic C in the soil of tropics. Kimble et al. (1990b) estimated a value of 496 Pg C in the upper 100 cm soil, which matches well with the estimation of 506 Pg C given by Eswaran et al. (1993). Batjes (1996) reported a relative lower result of 384–403 Pg C for 0–100 cm soil. Accordingly, the amount of organic C stored in the soil of tropical and subtropical China amounts about to 5.5–7.5% of the world’s tropical total.

ORGANIC CARBON STORAGE IN SOILS

59

4. Conclusions National soil survey has the great potential in providing a regional database for soil organic C, but it has disadvantages. For the Second National Soil Survey of China, soil profiles were not extended to the well-accepted depth of 100 cm, which impeded greatly the effort in estimating the C density in the upper 100 cm soil. However, simple models characterizing the vertical distribution of soil organic C in profiles can help solve the problem. In tropical and substropical China, it was found that soil organic C content (C in %) changes with depth (D in cm) following a pattern of: C = (c+D)/(a+bD), or C = 1/(a+bD). Extrapolation of what founded here to other regions should be done with caution. A soil map at an appropriate scale contributed greatly to the accuracy of the regional soil C density and storage estimation, and GIS techniques is proved to be a very useful means for characterizing the spatial distribution of C on a regional or national scale. By using the Map of Soils in the Red Soil and Yellow Soil Region of China (1:4 M) as the ‘blueprint’ and data from the Second National Soil Survey, we estimated that there are about 26.8±7.4 Pg organic C stored in soils of tropical and substropical China. The C density of soil subgroups in the region varied from 1.2 to 9.7 C m−2 in the upper 20 cm soil. In the east, the C density of soil groups ranged from 2.0–6.8 and 5.8–21.4 C m−2 for the upper 20 and 100 cm soil, respectively, while in the west, it occurred from 3.9–17.9 and 11.4–55.8 kg C m−2 correspondingly. On average, cultivation was estimated having caused a carbon loss of about 20–63% in the region. Acknowledgements The authors are grateful to Dr. Z. C. Cai for his helpful suggestions through this study, and Dr. C. F. Xi for making provincial soil data available and helpful explanation to the principles guiding the Second National Soil Survey. This research was founded partly by a National Natural Sciences Foundation Grant 49631010 and partly by the Laboratory of Material Cycling in Pedosphere, Academia Sinica (grant number 2–951505). This study was also supported by the President Foundation of the Chinese Academy of Sciences. References Anderson, D. W.: 1995, ‘Decomposition of Organic Matter and Carbon Emissions from Soils’, in R. Lal, J. Kimble, E. Levine, and B. A. Stewart (eds.), Soils and Global Change, CRC/Lewis publishers, Boca Raton, FL, U.S.A., pp. 165–175. Arnold, R. W.: 1995, ‘Role of Soil Survey in Obtaining a Global Carbon Budget’, in R. Lal, J., Kimble, E. Levine and B. A. Stewart (eds.), Soils and Global Change, CRC/Lewis Publishers, Boca Raton, FL, U.S.A., pp. 257–263. Arrouays, D. and Pelissier, P.: 1994, Soil Science 157, 185. Batjes, N. H.: 1996, European Journal of Soil Science 47, 151.

60

L. ZHONG ET AL.

Bennema, J.: 1974, Pedologie XXIV 2, 119. Bernoux, M., Arrouays, D., Cerri, C. C. and Bourennane, H.: 1998, Soil Science 163, 941. Bohn, H. L.: 1976, Soil Sci. Soc. Am. J. 140, 468. CGRY (Cooperative Group for Soil Improvement and Utilization Division in Red and Yellow Soil Regions): 1985, Soil Utilization and Amelioration Division in Red and Yellow Soil Regions, Agriculture Press, Beijing, 192 pp. (in Chinese). Detwiler, R. P.: 1986, Biogeochemistry 2, 67. Elzein, A. and Balesdent, J.: 1995, Soil Sci. Soc. Am. J. 59, 1328. Eswaran, H., Van den Berg. E. and Reich, P.: 1993, Soil Sci. Soc. Am. J. 57, 192. Eswaran, H. E., Van den Berg, E., Reich, P. and Kimble, J.: 1995, ‘Global Soil Carbon Resources’, in R. Lal, J. Kimble, E. Levine and B. A. Stewart (eds.), Soils and Global Change, CRC/Lewis Publishers, Boca Raton, FL, U.S.A., pp. 27–43. Kimble, J., Cook, T. and Eswaran, H.: 199a, ‘Organic Matter in Soils of the Tropics’, in Proc. Symp. Characterization and Role of Organic Matter in Different Soils, 14th International Congress of Soil Science, Tokyo, Japan, 12–18 Augustus 1990, pp. 250–258. Kimble, J. M., Eswaran, H. and Cook, T.: 1990b, ‘Organic Carbon on a Volume Basis in Tropical and Temperate Soils’, in Transactions of the 14th International Congress of Soil Science 5, 248–253. Lal, R.: 1976, Proc. Soil Sci. Soc. Amer. 40, 762. Li, Z. and Zhao, O. G.: 1998, Climatic Change 40, 119. Mann, L. K.: 1986, Soil Sci. 42, 279. Nye, P. H. and Greenland, D. J.: 1960, ‘The Soil under Shifting Cultivation’, in Commonwealth Bureau of Soils (ed.), Tech. Comm. No. 51, Hapenden, England. Post, W. M., Peng, T. H., Emanuel, W. R., King, A. W., Dale, V. H. and DeAngelis, D. L.: 1990, American Scientist 78, 310. Sanchez, P. A., Gichuru, M. P. and Katz, L. B.: 1982, ‘Organic Matter in Major Soils of the Tropics and Temperate Regions’, in Non-Symbiotic Nitrogen Fixation and Organic Matter in the Tropics, 12th International Congress of Soil Science, New Delhi, pp. 99–114. Sanchez, P. A., Palm, C. A., Szott, L. T., Cuevas, E. and Lal, R.: 1989, ‘Organic Input Management in Tropical Agroecosystems’, in Coleman, D. C., Oades, J. M., Uehara, G. (eds.), Dynamics of Soil Organic Matter in Tropical Ecosystems, Univ. Hawaii Press, Honolulu, pp. 125–152. Schimel, D., Enting, I. G., Heimann, M., Wigley, T. M. L., Raynaud, D., Alves, D. and Siegenthaler, U.: 1995, ‘CO2 and the Carbons Cycle’, in Houghton, J. T., Meira Filho, L. G., Bruce, J. et al. (eds.), Climate Change 1994, Cambridge University Press, Cambridge, pp. 35–71. Schlesinger, W. H.: 1977, Annual Revue of Ecology and Systematics 8, 51. Schlesinger, W. H.: 1991, Biogeochemistry: An Analysis of Global Change, Academic Press, San Diego. Schlesinger, W. H.: 1993, Vegetatio 104, 295. Stone, E. L., Harris, W. G., Brown, R. B. and Kuehl, R. J.: 1993, Soil Sci. Soc. Am. J. 57, 179. The National Soil Survey Office: 1993, Soil Species of China, Vol. I, China Agricultural Press, Beijing, 924 pp. (in Chinese). The National Soil Survey Office: 1994, Soil Species of China, Vol. III, China Agricultural Press, Beijing, 744 pp. (in Chinese). The National Soil Survey Office: 1996, Soil Species of China Vol. VI, China Agricultural Press, Beijing, 880 pp. (in Chinese). Tiessen, H. and Santos, M. C. D.: 1989, Plant and Soil 119, 337. Wen, Q. W.: 1984, Study Methods of Soil Organic Matter, (in Chinese), Agricultural press, Beijing, 318 pp. Woomer, P. L., Martin, A., Albrecht, A., Resck, D. V. S. and Scharpenseel, H. W.: 1994, ‘The importance and Management of Soil Organic Matter in the Tropics’, in Woomer, P. L. and Swift, M. J. (eds.), The Biological Management of Tropical Soil Fertility, John Wiley & Sons, Chichester, U.K., pp. 47–80. Zhao, Q. G., Li, Z. and Xia, Y. F.: 1997, Nutrient Cycling in Agroecosystems 49, 229.