Carbon Stocks in Different Soil Types under Diverse Rainfed Production Systems in Tropical India

Soil carbon (C) pool plays a crucial role in the soil's quality, availability of plant nutrients, environmental functions, and global C cycle. Drylands generally have poor fertility and little organic matter and hence are candidates for C sequestration. Carbon storage in the soil profile not only improves fertility but also abates global warming. Several soils, production, and management factors influence C sequestration, and it is important to identify production and management factors that enhance C sequestrations in dryland soils. The objective of the present study was to examine C stocks at 21 sites under ongoing rainfed production systems and management regimes over the last 25 years on dominant soil types, covering a range of climatic conditions in India. Organic C stocks in the soil profiles across the country showed wide variations and followed the order Vertisols > Inceptisols > Alfisols > Aridisols. Inorganic C and total C stocks were larger in Vertisols than in other soil types. Soil organic C stocks decreased with depth in the profile, whereas inorganic C stocks increased with depth. Among the production systems, soybean‐, maize‐, and groundnut‐based systems showed greater organic C stocks than other production systems. However, the greatest contribution of organic C to total C stock was under upland rice system. Organic C stocks in the surface layer of the soils increased with rainfall (r = 0.59*), whereas inorganic C stocks in soils were found in the regions with less than 550 mm annual rainfall. Cation exchange capacity had better correlation with organic C stocks than clay content in soils. Results suggest that Indian dryland soils are low in organic C but have potential to sequester. Further potential of tropical soils to sequester more C in soil could be harnessed by identifying appropriate production systems and management practices for sustainable development and improved livelihoods in the tropics.


INTRODUCTION
Agricultural soils are among the earth's largest terrestrial reservoirs of carbon (C) and hold potential for expanded C sequestration. Thus, they provide a potential way to reduce atmospheric concentration of carbon dioxide (CO 2 ) (Lal 2004). At the same time, this process provides other important benefits in terms of increased soil fertility and environmental quality. Because of low C in the dryland soils, there is good potential for C sequestration (Wani et al. 2003). Low soil organic matter (SOM) in tropical soils, particularly those under the influence of arid, semi-arid, and subhumid climates, is a major factor contributing to their poor productivity (Syers et al. 1996;Katyal, Rao, and Reddy 2001). Therefore, proper management of SOM is important for sustaining soil productivity and ensuring food security and protection of marginal lands (Scherr 1999). Because fertilizer input in dryland agriculture is low, mineralization of organic matter acts as a major source of plant nutrients. Maintaining or improving organic C levels in tropical soils is more difficult because of rapid oxidation of organic matter under prevailing high temperatures (Lal 1997;Lal, Follett, and Kimble 2003). However, maintaining or improving soil organic matter is a prerequisite to ensuring soil quality, productivity, and sustainability.
The C balance of terrestrial ecosystems can be changed markedly by the impact of human activities, including deforestation, biomass burning, and land-use change, which result in the release trace gases that enhance the greenhouse effect (Biolin 1981;Trabalka and Reichle 1986;IPCC 1990;Batjes 1996;Bhattacharya et al. 2000;Lorenz and Lal 2005). Routinely soil surveys conducted for estimating soil organic C (SOC) pool consider a depth of about 1 m. However, the subsoil C sequestrating may be achieved directly by selecting plants/cultivars with deeper and thicker root systems that are high in chemical recalcitrant compounds such as suberin and lignin (Wani et al. 2003;Lorenz and Lal 2005). Information on global regional the SOC pool is limited (Eswaran, Vanden Berg, and Reich 1993;Batjes 1996). Moreover, conclusions on the effects of land-use changes on soil C stocks are often hampered by the narrow global database available (Lorenz and Lal 2005). The size and dynamics of C pools in soils of the developing world are still poorly understood (Batjes 1996). In Indian soils, earlier studies on SOM content (Jenny and Raychaudhuri 1960) and its stocks (Gupta and Rao 1994) lacked wide sampling bases (Bhattacharya et al. 2001). In recent times, though the role of SOC (Bhattacharya et al. 2000) and inorganic C (Sahrawat 2003) have been highlighted in sequestering C in drylands, relatively little data are available on it. The objective of this study was to determine C stocks in a range of Indian soils under diverse climatic and crop production systems.

Study Locations, Climate, and Soil Characteristics
Soil samples were collected from 21 locations, representing a wide range of climatic conditions in tropical India, which were under long-term cultivation of dryland production systems ( Figure 1). Climate varied from arid, semi-arid, to subhumid, with mean annual rainfall ranging from 412 mm to 1378 mm (Table 1) and maximum, minimum, and mean air temperature ranging from 27.8 uC to 42.4 uC, 16.4 uC to 26.7 uC, and 23.2 uC to 31.9 uC, respectively. Soils were alluvial, red, yellow, and black, and arid. Length of growing period varied widely from 60-90 days in the arid regions to 180-210 days in the subhumid regions. Among the soil types, Inceptisols/Entisols, Vertisols and Vertic subgroups, and Aridisols were neutral to alkaline in reaction, and Alfisols/Oxisols were acidic. Salinity was not a problem in most of the soils. Vertisols and associated soils and some Inceptisols were calcareous, whereas Alfisols/ Oxisols were noncalcareous. Except at Ranchi (Alfisol) and Indore (Vertisol), the remaining 19 locations were low in organic C (Table 2).
Depthwise sampling of soils (0.15-m intervals up to 1.05 m deep) was undertaken at 21 locations as depicted in Figure 1. At each location, sampling was done based on several dugout pits, and finally a composite sample was made for each horizon. The Walkley and Black (1934) method was used to estimate SOC, and calcium carbonate (CaCO 3 ) content in soils was determined by standard acid-base titration method (Jackson 1973). Bulk density of each horizon was determined by weight by volume. The size of C stock in each profile was calculated following the method described by Batjes (1996). It involved calculation of organic C by multiplying OC content (g C g 21 soil), bulk density (Mg m 23 ) of each layer, and thickness of this layer (m) for each horizon (0-0.15 m, 0.30 m, 0.45 m, 0.60 m, 0.75 m, 0.90 m, and 1.05 m). For the determination of carbonate (CO 3 ) carbon stock (inorganic), the calculation was made using 12% of C values in CaCO 3 using a similar procedure. Summation of C in all the horizons was taken as C stock for the individual profile, and summation of soil inorganic and organic C stocks was taken as total C stock and expressed on a per hectare basis. This information is relevant in terms of comparing the soil C stocks among soil 2340 C. Srinivasarao et al. types, production systems, and climate, and accordingly suitable management practices could be identified for better C sequestration in dryland soils.

Stocks in Relation to Soil Type
Organic, inorganic, and total C stocks varied between and within soil types (Table 3). Vertisols and associated soils contained greater C stocks, followed by Inceptisols , Alfisols , Aridisols ( Figure 2). In general, SOC content was greater than inorganic C content in Alfisols and Aridisols,   (Dalal and Mayer 1986). Wani et al. (2003) reported increased C sequestration in Vertisols with pigeonpea-based systems with improved Carbon Stocks in Tropical Indian Soils management options (32 kg OC ha 21 y 21 ) as compared to sorghumbased systems with farmer's management. The C concentrations reported from the Indian tropics are less than those reported by Dalal and Mayer (1986), Dalal (1989), Murphy et al. (2002), and Young et al. (2005). Significantly lower levels of organic C in these soils are attributed to high rates of oxidation of SOM resulting from high temperature in tropics and frequent cultivation (Dalal and Chan 2001;Wani et al. 2003). Young et al. (2005) reported that Vertisols with high clay content showed greater carbon stocks than other soils. Sahrawat (2003) stated that calcium carbonate is a common mineral in soils of the dry regions of the world, stretching from subhumid to arid zones, as soils of this region are calcareous in nature. According to an estimate by the National Bureau of Soil Survey and Land Use Planning, Nagpur, India, calcareous soils occupy about 230 million ha and constitute 69% of the total geographical area of India. It was further stated that SIC pool consists of primary inorganic carbonates or lithogenic inorganic carbonates and secondary inorganic carbonates. The reaction of atmospheric CO 2 with water (H 2 O) and calcium (Ca 2+ ) in the upper horizons of the soil, leaching into the subsoil, and subsequent reprecipitation results in formation of secondary carbonates and the sequestration of atmospheric CO 2 . This was the reason why deeper layers showed more inorganic C than surface soils in most profiles (Sahrawat 2003).

Stocks in Relation to Production System
Carbon stocks varied with production system and showed significant interaction with soil type. Organic, inorganic, and total C stocks under each production are presented in Figure 3. Soybean-based production system (62.31 Mg C ha 21 ) showed the most organic C stocks, followed by maize-based (47.57 Mg ha 21 ) and groundnut-based (41.71 Mg ha 21 ) systems. Pearl millet-and finger millet-based systems showed lower organic C stocks. On the other hand, cotton system (275.3 Mg ha 21 ) and post-rainy (rabi) sorghum production system (243.7 Mg ha 21 ), primarily on Vertisols and associated soils, showed the most SIC, whereas the SIC was the least in soils under lowland rice systems (18.15 Mg ha 21 ). Highest TCSs were found under the cotton-based based production system, followed by rabi sorghum-based and pearl millet-based system. However, percentage contribution of organic C to TCS was most under the rice-based system, whereas the greatest inorganic C contribution to total C was observed under the cotton-based system (Figure 4). On a regional scale, above-and belowground biomass production is probably the major determinant of the relative distribution of SOC with depth (Jobbagy and Jackson 2000). Aboveground organic matter has probably only limited effects on SOM levels compared to belowground organic matter, as has been demonstrated by long-term residue management studies (Clapp et al. 2000;Reicosky et al. 2002). The dominant role of root C in soil was also indicated by greater relative contribution of root vs. shoot tissue to the SOC pool (Rasse, Rumpel, and Dignac 2004). Root-to-shoot ratio of corn varied from 0.21 to 0.25; that of soybean was 0.23 and that of barley was 0.50 (Allmaroas, Linden, and Clapp 2004;Figure 3. Carbon stocks in soils under diverse rainfed production systems.
For each soil type, the effect of particular production system was also examined. For Vertisols and associated soils, cotton and sorghum systems showed larger SIC stocks ( Figure 5), whereas soybean and groundnut systems showed more SOC. Legume-based systems on Figure 4. Contribution of organic and inorganic pools to total carbon stocks in soils under diverse rainfed production systems.

2348
C. Srinivasarao et al. Vertisols showed more SOC than cereal-based systems in the tropics (Wani et al. 1995(Wani et al. , 2003. In Inceptisols, maize-based systems showed more inorganic as well as organic C content. In Alfisols, rice-based system (Ranchi and Phulbani) showed relatively more organic C content, whereas groundnut-based (Anantapur) system showed more inorganic C ( Figure 6). This could be due to larger carbonate deposits found in the deeper layers of the profile, frequent addition of gypsum to groundnut crop, and differences in rainfall, parent base material, and other management practices adapted at these locations. Under Aridisols, the pearl millet-based system at SK Nagar showed more total carbon than in Hisar.

Carbon Stocks in Relation to Rainfall and Temperature
In general, SOC stocks increased as the mean annual rainfall increased (Figure 7). Significant correlation (p , 0.05) was obtained between SOC stock and mean annual rainfall (r 5 0.59 * ; Figure 8). On the other hand, SIC stocks decreased with the increase in mean annual rainfall from 156.40 Mg ha 21 (,550 mm) to 25.97 Mg ha 21 (.1100 mm). As the SIC stocks were more dominant than SOC, TCS decreased with increase in mean annual rainfall from 183.79 Mg ha 21 in the arid environment (,550 mm) to 70.24 Mg ha 21 in subhumid regions (.1100 mm). As all the locations are under dryland conditions and belong to similar temperature regimes, air temperature showed nonsignificant negative correlation with organic C stocks ( Figure 9). However, cation exchange capacity (CEC) showed significant positive correlation (r 5 0.81**), whereas clay content in soil showed nonsignificant positive correlation with organic C stocks ( Figure 10). This indirectly indicates the type of clay mineral with larger surface area is largely responsible for greater C sequestration.
It has been postulated that aridity in the climate is responsible for the formation of pedogenic calcium carbonate, and this is a reverse process to the enhancement in soil organic C. Thus, increase in C sequestration via SOC enhancement in the soil would induce dissolution of native calcium carbonate, and the leaching of SIC would also result in C sequestration (Sahrawat 2003). In the present scenario of differing Carbon Stocks in Tropical Indian Soils climatic parameters such as temperature and annual rainfall in some areas of the country, it will continue to remain a potential threat for C sequestration in tropical soils of the Indian subcontinent. Therefore, the arid climate will continue to remain as a bane for Indian agriculture because this will cause soil degradation in terms of depletion of organic C and formation of pedogenic CaCO 3 with the concomitant development of sodicity and/or salinity (Eswaran, Vanden Berg, and Reich 1993;Bhattacharya et al. 2000).

CONCLUSIONS
Vertisols and associated soils had relatively greater SOC stocks than other soil types, whereas soils of regions with less rainfall showed larger inorganic C content than soils of regions with more rainfall. Amount of rainfall was significantly related with amounts of organic C stocks in the soils, and legume-based production systems showed more organic C sequestration. As soils of India are very low in organic C, its depletion  Carbon Stocks in Tropical Indian Soils occurs at a rapid rate because of continuous cultivation and exposure of the subsoil organic matter. However, long-term manure experiments under rainfed conditions showed marginal improvements in organic C levels with regular additions of organic manures. Most of the dryland farmers are not in a position to add manure or crop residue regularly without their own cattle. Therefore, alternative measures like minimum tillage, green manuring, cover cropping, green leaf manuring like gliricidia, compositing the farm waste, vermicomposting the farm and household waste, and including legumes in the system may have potential for improving C stocks in Indian soils. This could have long-term consequences in sustaining natural resources.