Urvashi Tomar · Ratul Baishya

Abstract Understanding the dynamics of soil carbon is crucial for assessing the soil carbon storage and predicting the potential of mitigating carbon dioxide from the atmosphere to the biomass and soil. The present study evaluated variations of soil carbon stock in semi-arid forests in India under different moisture regimes. Soil organic carbon (SOC)and soil inorganic carbon (SIC) stocks were determined in different moisture regimes i.e. monsoon, post-monsoon,winter and pre-monsoon seasons at 0—10 and ＞ 10—20 cm depths. SOC stock showed signif icant variations under different moisture regimes. The highest SOC stock was during winter (22.81 Mg C ha -1 ) and lowest during the monsoon season (2.34 Mg C ha -1 ) among all the ridge forests under study. SOC and SIC stock under different moisture regimes showed signif icant negative correlation with soil moisture( p ＜ 0.05), as a sudden increase in soil moisture after rainfall results in an increase in carbon loss due to microbial decomposition of accumulated carbon during the dry period.There was an increase in annual SOC stock and a decrease(or no change in some cases), in SIC stock at both the depths during the study period. The SOC and SIC sequestration rates were estimated as any increase/decrease in the respective stock during each successive year. SOC sequestered ranged between 0.046 and 0.741 Mg C ha -1 y -1 . Similarly,SIC sequestration ranged between 0.013 and 0.023 Mg C ha -1 y -1 over all ridge forests up to 20 cm depth. The Delhi ridge forests, which accounts to 0.007% of the semi-arid regions of India, contribute 0.25—0.32% of the national potential (semi-arid region) for SOC sequestration up to 20 cm depth. The estimates of the rate of C sequestration in this study provide a realistic image of carbon dynamics under present climatic conditions of semi-arid forests, and could be used in developing a database and formulating new strategies for carbon dioxide mitigation by enhancing soil C sequestration rates.

Keywords Soil or ganic c arbon · Soil i norganic c arbon ·Carbon s equestration · Moisture r egime · Semi-arid f orests

Introduction

Forest ecosystems cover a large part of the world’s land surface and play an important role in the terrestrial carbon cycle (Conforti et al. 2016). These forests have the potential to sequester and store atmospheric carbon in vegetation and in the soil, and are sinks for atmospheric carbon (Frolking et al. 1996; Perruchoud et al. 1999; Halliday et al. 2003;Lal 2005; Jandl et al. 2007). In terrestrial ecosystems, soil carbon is greater than the amount of carbon stored in living vegetation (Post and Kwon 2000).

Soil carbon is an important parameter of soil quality, soil fertility, water holding capacity and productivity (Bhattacharya et al. 2008). Soil organic carbon (SOC) and soil inorganic carbon (SIC) are the two components which comprise the soil carbon pool. Highly active yet (chemically) inert humus, including plant, animal and microbial residues in different stages of decomposition, contribute to the SOC pool(Post and Kwon 2000), and elemental C and carbonate minerals are included in the SIC pool (Lal and Kimble 2000).SOC is the major indicator of soil fertility and soil health(Venkanna et al. 2014), and even a small variation in soil organic carbon can affect soil properties, global C cycling,and climate change (Lal 2005; Powlson et al. 2011; Conforti et al. 2016). The composition, structure, and function of terrestrial ecosystems are inf luenced by the SOC pool (Bhattacharya et al. 2016; Wu et al. 2009). Under the inf luence of arid, semi-arid and sub-humid climates, soils in tropical regions are low in SOC (Venkanna et al. 2014). Hence, it is important to quantify and understand the special distribution of soil carbon in order to support management policies for forest ecosystems (Conforti et al. 2016). Soils of forest ecosystems are an important medium for their contribution to biomass production by enhancing nutrient availability for plant growth, thus contributing to carbon sequestration(Mishra et al. 2019). Therefore, it is important to quantify both aboveground and belowground rates of carbon sequestration. Previous studies have estimated aboveground carbon sequestration (Ramachandran et al. 2007; Kaul et al. 2010;Jana et al. 2011; Pant and Tewari 2014; Jha 2015; Devi and Yadava 2015) in India. But very few studies have estimated soil carbon sequestration rate of Indian forests.

In arid and semi-arid ecosystems, dry—wet cycles are coupled with strong seasonal precipitation (Austin et al. 2004).Due to irregular rainfall patterns in semi-arid areas, forest soils are susceptible to dry—wet stresses (Fierer and Schimel 2002). These seasonally inf luenced patterns are important to understanding the dynamics of soil C and other nutrients(Fierer and Schimel 2002), as belowground processes are controlled by these dry—wet cycles (Austin et al. 2004).

Since the beginning of the twentieth century, changes in rainfall amounts, temperatures, and frequent extreme climatic events have been observed (Milly et al. 2002; Peterson et al. 2002). India is highly sensitive to climatic changes due to long-term and inter-annual variability in monsoon observed over the country (Mall et al. 2006). The study of variabilities in carbon stocks due to moisture regimes becomes challenging as well as necessary. Variations in mean annual rainfall has a signif icant role in controlling aboveground biomass and SOC content. Mehta et al. ( 2014)showed that mean annual precipitation in forested regions of Gujarat, India was the driving factor inf luencing soil carbon and quality of biomass. Tropical Indian forests are likely to experience signif icant changes in cover and carbon stocks due to changes in rainfall patterns (Mehta et al. 2014).Year-to-year variability in monsoon rainfall has also been observed (Mall et al. 2006) and these changes will continue to develop as potential threats to tropical soils in the Indian subcontinent (Bhattacharyya et al. 2000). As carbon dynamics are inf luenced by rainfall patterns, many carbon stock inventories have been established globally. But a periodical assessment of soil carbon stock in different moisture regimes is also required. These data will help in understanding carbon stock dynamics and variability under moisture regimes and rainfall patterns. While several studies have shown the status of carbon stocks in India (Ravindranath et al. 1997,2001, 2008; Bhattacharyya et. al. 2000; Kaur et al. 2000;Chhabra and Dadhwal 2004; Lal 2004; Singh et al. 2007;Das et al. 2008; Saha et al. 2009; Gupta and Sharma 2012;Patil et al. 2012; Mehta et al. 2014; Venkanna et al. 2014;Hinge et al. 2018; Salunkhe et al. 2018), the variations in soil carbon stock under different moisture regimes have not been addressed specif ically.

Delhi has 13.0% of its geographical area under forest cover (FSI ( 2017). The state has scattered forest cover, and investigations on carbon stock and its dynamics in semi-arid conditions are very limited. However, estimation and variation in carbon stock in these forests is a necessity to create a complete dataset on Indian forests to prioritize areas for soil C sequestration. Therefore, the present study addresses the data gap and will be helpful in estimating variations in carbon stocks under different moisture regimes and C sequestration rates in semi-arid forests of Delhi, National Capital Region.

Material a nd methods

Study ar ea

This study was conducted in the ridge forests of Delhi which lies in the northern part of India between 28°24′17″to 28°53′00″ N and 76°45′30″ to 77°21′30″. The total geographical area is 1483.0 km 2 with 192.4 km 2 is under forest cover, constituting 13.0% of the geographical area of the state (FSI 2017). Ridge forests occupy 77.8 km2of cover and are the northern extension of the Aravalli Hill Range. The Delhi ridge is divided into four fragmented zones (Fig. 1):Southern ridge (6200 ha), Central ridge (864 ha), South-Central ridge (626 ha) and Northern ridge (87 ha). Temperatures in the study area June 2014 to May 2017 varied between 15.6 °C in January and 41.4 °C in May. Over the 3-year-study, annual rainfalls were 827.2 mm, 834.9 mm and 1285.2 mm June 2014—May 2015, June 2015—May 2016 and June 2016—May 2017, respectively. The highest rainfalls were in July and August in all 3 years (Fig. 2).

All the Delhi ridges are dominated by mesquite (Prosopis julifl ora(Sw.) DC.),an exotic invasive tree species introduced in 1877. The climate is semi-arid and the forests were classif ied as tropical thorn forest by Champion and Seth ( 1968). Soils are a sandy loam with pH 6.2—7.7.Delhi experiences four distinct moisture regimes in a year,pre-monsoon (March—May), monsoon (June—August),post-monsoon (September—November) and winter(December—February).

Fig. 1 Map of India showing Delhi and its four ridge forests

Fig. 2 Monthly variation in precipitation June 2014—May 2017) in Delhi ( Source: Agromet, IARI, Delhi)

Soil s ampling

Field visits and soil sampling were carried out over three years (June 2014—May 2017). Soil sampling was done after removing the litter and carried out on all ridges under four moisture regimes noted previously. On each ridge, three replicates of f ive, randomly selected sub-samples from two soil depths, 0—10 cm and ＞ 10—20 cm were collected and homogenized. Samples were transferred to the laboratory,plant remains and other debris removed, and kept at 4 °C for analysis.

Physical and c hemical anal ysis

Soil moisture was determined according to Allen et al.( 1974) within 48 h of collection. A 10 g sample was ovendried at 105 °C to constant weight. Soil bulk density was estimated by placing a steel core of known volume into the soil. Extra soil adhering to the core was removed from the end of the core. Soil from each depth was collected in separate polybags, and oven-dried at 105 °C until constant weight was achieved. Bulk density was estimated following Anderson and Ingram ( 1993).

Fig. 3 Annual carbon stock, mean ± SE (TC, SOC, and SIC) June 2014—May 2017 at two sampling depths on all Delhi ridges. a southern ridge (0—10), b southern ridge (＞10—20), c central ridge (0—10), d central ridge (＞10—20), e south-central ridge (0—10), f south-central ridge (＞10—20), g northern ridge (0—10), h northern ridge (＞10—20)

Soil texture was determined by hydrometric method(Allen et al. 1974). A 50-g air-dried 2-mm sieved sample was diluted with 25 ml of 5% sodium hexametaphosphate and 400 ml of distilled water for 15 min on a high-speed stirrer. The contents were transferred to a 1-l measuring cylinder, diluted to volume, stirred 60 s and timed for reading with a bouyoucos soil hygrometer. Initial reading was taken at 4.5 min and f inal reading at 5 h and temperature corrections applied, i.e., 0.3 units added or subtracted for every degree above or below 19.5 °C. Soil textural class was based on the USDA ( 2017) soil classif ication triangle.

Air-dried samples were used to estimate total carbon(TC), organic carbon (SOC), and inorganic carbon (SIC).TC was determined using the standard method of Liqui TOC II Analyzer (Elementar Analysissystems GmbH,Hanau, Germany). SOC was determined by digesting the sample with hydrochloric acid (HCl), oven-dried and the SOC content identif ied by subtracting SOC from TC. Carbon contents are in Megagrams of carbon per hectare (Mg C ha -1 ) (Mg = 10 6 g), derived by multiplying carbon concentration (C %) by bulk density (Mg m -3 ), soil depth (m),and forest area (m 2 ). Total carbon, SOC and SIC stock are in Mg C ha -1 using the following formula (Venkanna et al.2014):Query ID="Q1" Text="Equation are not sequential"

where C is carbon;Bbulk density;Ddepth.

Annual carbon stocks were calculated as the average of all moisture regimes. The rate of carbon sequestered was calculated by subtracting the mean annual stock of carbon in the 1st year from that of carbon in the 2nd year, and carbon in the 2nd year from that of the 3rd year.

Statistical anal ysis

Data were analyzed by one-way ANOVA to evaluate variations in moisture regimes on TC, SOC and SIC from June 2014 to May 2017 under two soil depths (0—10) cm and(＞10—20) cm in all four Delhi ridges. To evaluate the variation in yearly stocks, annual TC, SOC and SIC stocks were pooled. Similarly, one-way ANOVA was performed to analyze inter-annual variation under both soil depths on all four ridges. Pearson’s correlation tested the association between carbon stocks and soil moisture. All statistical analyses were carried out using IBM SPSS 16 software.

Results

Soil p hysical p arameters

Soil moisture ranged from 3.1% to 17.1% and 3.0% to 13.0%in the top 10 cm and ＞ 10—20 cm layers, respectively, in all four Delhi forests, and showed signif icant variations in moisture regimes on all ridges. It was signif icantly higher in the monsoon season than in the winter, pre-monsoon and postmonsoon seasons across all ridges.

Soil c arbon s tocks

SOC stocks in the 0—10 cm depth ranged between 4.69 and 12.75 Mg C ha -1 on the Southern ridge, 14.17 to 21.11 Mg C ha -1 on the Central ridge, 10.78 to 17.15 Mg C ha -1 on the South-Central ridge and 11.12 to 22.81 Mg C ha -1 on the Northern ridge. In the ＞ 10—20 cm depth, stocks ranged between 2.34 and 6.13 Mg C ha -1 on the Southern ridge, 12.01 to 20.20 Mg C ha -1 on the Central ridge, 7.37 to 13.86 Mg C ha -1 on the South-Central ridge and 5.26 to 14.33 Mg C ha-1on Northern ridge. SIC in the 0—10 cm depth ranged between 0.59 and 1.86 Mg C ha-1on the Southern ridge, 1.67 to 2.43 Mg C ha -1 on the Central ridge, 0.86 to 1.87 Mg C ha -1 on the South-Central ridge and 1.29 to 2.47 Mg C ha -1 on the Northern ridge. At the ＞ 10—20 cm depth, it ranged between 0.63 and 1.76 Mg C ha -1 on the Southern ridge, 1.69 to 2.94 Mg C ha -1 on the Central ridge, 1.08 to 1.75 Mg C ha -1 on the South-Central ridge and 1.21 to 2.71 Mg C ha -1 on the Northern ridge (Appendix S1).

Over the 3-year study period, mean annual SOC stock in the upper 10 cm layer increased from 8.87 to 9.21 Mg C ha -1 on the Southern ridge, 17.95 to 18.61 Mg C ha -1 on the Central ridge, 13.45 to 14.06 Mg C ha -1 on the South-Central ridge and 16.22 to 16.41 Mg C ha -1 on the Northern ridge. In the ＞ 10—20 cm depth, soil organic carbon increased from 4.27 to 4.60 Mg C ha -1 on the Southern ridge, 15.51 to 15.82 Mg C ha -1 on the Central ridge, 8.76 to 9.00 Mg C ha -1 on the South-Central ridge and 9.83 to 9.90 Mg C ha -1 on the Northern ridge. Mean annual SIC stock, on the other hand, showed a different trend, and in the 0—10 cm depth varied from 1.33 to 1.34 Mg C ha-1on the Southern ridge, 2.14 to 1.88 Mg C ha -1 on the Central ridge, 1.47 to 1.25 Mg C ha -1 on the South-Central ridge and remained same at 1.63 Mg C ha -1 on the Northern ridge. In the ＞ 10—20 cm depth, it decreased from 1.37 to 1.35 Mg C ha -1 on the Southern ridge, 2.21 to 2.17 Mg C ha -1 on the Central ridge, and remained the same 1.40 Mg C ha -1 on the South-Central ridge and 1.81 Mg C ha -1 on the Northern ridge (Fig. 3). The mean carbon stock values representing the actual carbon per total area of each of the ridges are shown in Table 1.

Table 1 Mean carbon stocks (Mg C) representing the actual carbon per total area in each of the ridges

Correlation anal ysis

The SOC and SIC stocks in the 0—10 and ＞ 10—20 cm layers showed signif icant correlation with soil moisture, considering the different moisture regimes on all ridges during the 3-year study (p＜ 0.05). The relationship between SOC and SIC stock with soil moisture under the two depths was negative, indicating that both SOC and SIC stock decreased with the sudden increase in soil moisture during the monsoon period (Fig. 4).

Discussion

SOC showed strong variations under different moisture regime and was highest during winter and lowest during the monsoon at both soil depths. This may be because of an increase in soil moisture in monsoon, resulting in a rapid increase in microbial activity involved in decomposition of organic matter, and is the main driver of soil carbon mineralization (Borken and Matzner 2009; Unger et al. 2010; Rey et al. 2016). Rainfall events lead to large soil CO 2 effluxes as the labile carbon and microbial biomass accumulated during the dry period becomes available for microbial decomposition (Harper et al. 2005; Rey et al. 2005, 2016; Sponseller 2007; Borken and Matzner 2009; Williams et al. 2009; Munson et al. 2010). During winter (dry conditions), low rates of soil respiration and decomposition result in relatively high soil carbon contents corresponding to the seasonal changes in soil moisture (Grunzweig et al. 2003). SIC stock did not follow a similar trend as it is mainly determined by leaching due to precipitation and thus creates a complicated soil prof ile (Schlesinger and Adrienne 1998; Wang et al. 2010).SIC dynamics are more complex than SOC as it may be a cumulative effect of factors other than decomposition and substrate availability, and may be mainly determined by weathering reactions and precipitation of secondary carbonates (Lal and Kimble 2000; Mi et al. 2008).

A signif icant negative correlation was observed between SOC and SIC with soil moisture under different moisture regimes at both the depths (Fig. 4). It is evident that for a region like Delhi, soil water limitations are a predominant issue and water availability may be regulating both aboveground and belowground processes. Precipitation is a strong driving force for C dynamics in arid and semi-arid regions(Liu et al. 2009). This suggests that large/sudden moisture changes result in a large amount of carbon loss from carbon accumulated during dry periods (Rey et al. 2016).

Annual variability in average SOC and SIC was not statistically signif icant over the 3-year period. However, there was an increased annual SOC stock and decreased (or constant levels in some cases) SIC stock in both soil depths in successive years. The reason could be higher annual precipitation during the 3rd year which resulted in increased SOC(Amundson et al. 1989; Burke et al. 1991; Quilchano et al.1995; Austin and Vitousek 1998; Wang et al. 2010). Several researchers have reported that aboveground and belowground carbon increases with greater annual rainfalls in terrestrial ecosystems (Schlesinger 1977; Roberts et al. 1989;Austin 2002). This affects aboveground biomass production by increasing plant nutrient uptake, thereby resulting in an accumulation of debris and soil organic carbon (Schlesinger 1977; Austin and Vitousek 1998; Austin 2002). Increases in the frequency of large rainfall events in semi-arid steppes in Inner Mongolia disproportionally increased primary productivity of higher plants compared to microbial activity, and thus the area could sequester more carbon in the soils (Chen et al. 2009). Similar f indings were reported by Mehta et al.( 2014) in tropical forests of Gujarat, India and Venkanna et al. ( 2014) in semi-arid tropical regions of southern India.In contrast, mean annual SIC stock decreases with increases in annual precipitation as leaching of pedogenic inorganic carbon (PIC) is enhanced by rainfall thus depleting accumulated PIC in soil (Nordt et al. 2000; Mi et. al. 2008; Wang et al. 2010; Venkanna et al. ( 2014).

Fig. 4 Relationship between SOC and SIC stock (Mg C ha -1 ) with soil moisture (%) in two depths

Soil carbon in a 30-cm soil layer of tropical thorn forest in India was 44 Mg C ha -1 (Ravindranath et al. 1997), and varied in tropical dry deciduous forests at 50-cm depth from 7.7 to 85.6 Mg C ha -1 (Chabbra et al. 2003). Both datasets compare with our mean soil carbon stock which ranged between 25.4 and 38.1 Mg C ha -1 in soil up to 20-cm depth. The total SOC stock estimated on the Delhi ridge ranged between 13.4 and 33.9 in the upper 20-cm layer, which is comparable with the total Delhi SOC stock of 34.0 Mg C ha -1 estimated by the Forest Survey of India (FSI 2017). Over the 3-year study, SOC sequestration ranged from 0.046 to 0.255 Mg C ha -1 y -1 and 0.208 to 0.741 Mg C ha -1 y -1 during the 1st—2nd years and 2nd—3rd years respectively in the 0—20 cm depth. Similarly, SIC sequestration ranged between 0.013 and 0.023 Mg C ha-1y -1 during the 1st—2nd years and was negative during 2nd—3rd years in the upper 20 cm (Table 2). SOC sequestration was higher during the 2nd—3rd years, and SIC sequestration rate, on the other hand, declined during the same period. This may be due to higher annual precipitation during this period, resulting in increased mean annual SOC and decreased mean annual SIC. A total rate of SOC sequestration for the Delhi ridge forests is estimated as 0.75—1.99 Mg C ha-1y -1 for the 0—20 cm depth. This is, however, higher than the estimates of 0.02—0.04 Mg C ha -1 y -1 SOC sequestration rate in semiarid regions of India by Lal ( 2004). The contrasting results are because all semi-arid regions of India were included,including degraded soils and ecosystems. In our study, the forests were natural and protected, hence higher sequestration rates were observed. At the national level, a total soil organic carbon sequestration potential of 2.33—4.66 Tg C y -1(Tg = Teragram = 10 12 g) has been estimated for 116.4 M ha of semi-arid regions of India (Lal 2004). Hence, the Delhi ridge forests, which comprise 0.007% of the country’s semiarid regions contributes 0.25—0.32% of the potential nationalSOC sequestration in the upper 20-cm soil layer. Most of the previous studies focused on SOC pool, although soil inorganic carbon might be similarly dynamic as SOC, and is also known to be an important factor for soil C sequestration(Jordan et al. 1999; Stone 2008; Wohlfahrt et al. 2008; Xie et al. 2009; Wang et al. 2010).

Table 2 Rate of carbon sequestered (Mg C ha -1 y -1 ) in all the Delhi ridges

Our study provides an overview of the current soil carbon stock and rates of carbon sequestration of semi-arid forests of Delhi. Variations in soil moisture should be considered when determining soil carbon estimates, with a focus on increasing the residence time of sequestered carbon. However, more data on aboveground and belowground processes,disturbances, soil dynamics and long-term effects of changes in precipitation are needed to improve the understanding of carbon dynamics in forest ecosystems.

Conclusion

In the semi-arid ridge forests of Delhi, variations were observed in soil organic and inorganic carbon stock under different moisture regimes. Higher soil organic carbon in winter suggests that this dry period during the year is important for carbon accumulation, as losses due to decomposition and runoffare less. Precipitation was a major controlling factor in carbon dynamics. Increased annual precipitation during the 3rd year accounted for a higher rate of SOC sequestration which could be due to an increase in primary productivity of higher plants, thereby enhancing organic carbon in litter and soil. Hence, we concluded that rainfall patterns would result in predominant changes in soil C storage and contribute to predicting the potential of carbon sequestration in semi-arid regions. The estimated carbon sequestration in our study may serve as the basis for policy decisions and management efforts for enhancing soil carbon sequestration and thus help to mitigate the effects of climate change.

AcknowledgementsThe research was fully funded by DST-SERB research Project NO. SB/YS/LS-88/2013. Minor Grants received through R&D Grants, University of Delhi is also highly acknowledged.