Soil organic matter dynamics in semiarid agroecosystems transitioning to dryland

Introduction

The rapid expansion of dryland areas in recent years significantly depleted SOC and increased greenhouse gas (GHG) emissions (Huang et al., 2016). With projected changes in climate, global population, and land-use, it is anticipated that dryland will cover half the world’s land surface by the end of the century and directly impact the livelihoods of ∼250 million people through degradation of soil and environmental quality (Huang et al., 2016; Reynolds et al., 2007). The High Plains region of USA, where the Ogallala Aquifer underlies about 450,660 km² of the area, produces 30% of total crop and animal products in the country (Guerrero et al., 2010; NASS, 2016). Agricultural intensification using irrigation water pumped from the Ogallala Aquifer has regulated SOC and nutrient cycling, crop yields, and the rural economy. Land production values have increased by more than $12 billion annually after irrigation systems were developed (Hornbeck & Keskin, 2014). However, the saturated thickness of the Ogallala Aquifer has declined by more than 50% since its pre-development state in the Southern High Plains (Haacker, Kendall & Hyndman, 2016), with projected increase in dryland area. The extent of dryland expansion and the associated economic impacts remains unknown due to a lack of spatially explicit data on the Aquifer depletion in the region (Deines et al., 2020). Improved knowledge of SOC and N cycling during dryland transition will help in the development of a farming strategy that sustains agriculture in the regions facing challenges due to water limitation.

Management practices that increase SOC storage and improve soil health have been increasingly promoted for maximizing the economic potential of the land while enhancing environmental quality. Soil disturbance associated with agriculture around the globe has led to a loss of ∼133 Gt of soil C as CO₂ to the atmosphere in the past century and affected crop production and the environment (Sanderman, Hengl & Fiske, 2017). The depletion of SOC and N is associated with increased soil erosion and decreased water storage, nutrient cycling, and soil biological activity, ultimately leading to reduced crop yields and profitability (Baumhardt, Stewart & Sainju, 2015; Ghimire, Machado & Bista, 2018). In contrast, increased biomass C input through crop rotations along with conservation tillage systems, nutrients, and soil amendment additions can increase SOC sequestration and nutrient cycling (Al-Kaisi, Yin & Licht, 2005; Sainju, Jabro & Stevens, 2008). Soils with no-tillage and cover cropping store more SOC and nutrients than conventional tillage soils without a cover crop (Mitchell et al., 2015). No-tillage with cover cropping also served as a net sink for GHG (De Gryze et al., 2011). No-tillage and cover crops increase microbial abundance and diversity, which is often associated with increased SOC accumulation, reduced GHG emissions, and improved soil health (Mondal et al., 2020; Schmidt, Mitchell & Scow, 2019).

Grassland restoration could be another strategy to increase SOC storage in agroecosystems transitioning from irrigated to dryland management. A recent study documented up to a 37% increase in SOC with grassland restoration (∼50 years), which was not significantly different from SOC in a lightly grazed native pasture under dryland condition (Ghimire et al., 2019). The increase in SOC and nutrients with grassland restoration in the previously cultivated soils have also been documented from other semiarid regions of the world, the magnitude of increase varying with management practices and environment (Liu et al., 2017; Yang et al., 2019). Native grasses often have denser and deeper root systems than annual crops, and increased SOC with grassland restoration often results from the high root biomass of perennial species and addition of root-derived C (Roumet, Urcelay & Díaz, 2006).

Early response to management changes can be evaluated by monitoring labile SOC and N fractions, including CO₂ and N₂O emissions, because the labile organic matter changes rapidly with alternative management systems (Franzluebbers & Stuedemann, 2008). Although physical, chemical, and biological transformation processes convert plant residues into organic matter, the biologically active fraction of SOC could be affected by management changes more rapidly than other fractions (Lehmann & Kleber, 2015), while the soil disturbance and long-fallow periods can lead to the rapid loss of biologically active SOC and N fractions conservation systems that reduce tillage and increase cropping intensity and diversity can improve SOC storage and nutrient cycling (Ghimire, Machado & Bista, 2019). The conservation systems may be important for the Southern High Plains agroecosystems, where hot, dry environment and intermittent irrigation favors rapid mineralization of organic matter and limits biomass C inputs, as well as microbial activities (Cano et al., 2018; Ghimire et al., 2019). Improved understanding of management practices that affect various SOC and N fractions will help in designing sustainable cropping systems in semiarid regions in the time of uncertainty due to decreasing irrigation capacity.

This study aimed to evaluate SOC and N components, including CO₂ and N₂O emissions during laboratory incubation of soil samples, in irrigated and dryland cropping systems with varying crop diversity and tillage intensity, and a restored grassland. We hypothesized that grassland restoration and adoption of conservation systems that reduce soil disturbance along with cover cropping could improve soil health in hot, dry, semiarid environments through increased SOC and N storage.

Materials & Methods

Results

Soil pH and EC varied significantly between treatments in both 0–15 cm and 15–30 cm depths (Table 1). In the 0–15 cm depth, soil pH was highest in limited-irrigation cropping systems (LINTF and LINTC), which was not significantly higher than NG but higher than dryland cropping systems (DLCTF and DLNTF). In the 15–30 cm depth, limited-irrigation cropping systems had significantly higher soil pH (7.9–8.0) than dryland systems and NG (7.3–7.6).

Soil EC was significantly lower in dryland cropping systems (0.11–0.12 ds m⁻¹) than in limited-irrigation systems and NG (0.20–0.22 ds m⁻¹) in the 0–15 cm depth. In the 15–30 cm depth, the EC was significantly higher in DLCTF and LINTF than in NG and DLNTF (Table 1). Soil EC in LINTC was intermediate between other treatments.

Soil CEC did not differ between treatments in either depth. Soil CEC was in the range of 15.9–22.7 cmol_c kg⁻¹ in the 0–15 cm depth and 19.0–23.9 cmol _c kg⁻¹ in the 15–30 cm depth. Similarly, soil Db was in the range of 1.06–1.25 g cm⁻³ in the 0–15 cm and 1.14–1.30 g cm⁻³ in the 15–30 cm depth and no significant difference between treatments in either depth.

The SOC content was significantly greater in NG than all cropping systems in both soil depths. In the 0–15 cm depth of cropland systems, LINTC had significantly greater SOC than dryland cropping systems (Fig. 1A). The SOC in LINTF was intermediate between dryland cropping systems and LINTC. In the 15–30 cm depth, there was no difference in SOC content between irrigated and dryland cropping systems. The SOC in the 0–30 cm depth was 14% less in dryland systems, while it was 51% more in grasslands than limited-irrigation systems.

Soil total N content followed a similar trend as SOC that NG had significantly greater total N than all cropping systems in both depths. Total N in the 0–15 cm depth of LINTC was significantly greater than dryland systems but not significantly greater than LINTF (Fig. 1B). The total N in the 15–30 cm depth was greater in NG than in dryland cropping systems and LINTC. In LINTF, total N was intermediate between NG and other systems. Total N content was 13% less in the 0–30 cm depth of dryland and 41% more in the same depth of NG compared to limited-irrigation cropping systems.

Soil inorganic N in the 0–15 cm depth was greater in LINTC than other cropping systems (Fig. 2A). The inorganic N was considerably lower in NG than all cropping systems. In the 15–30 cm, it was significantly greater in DLCTF than all other systems. The inorganic N in limited-irrigation cropping systems and DLNTF was significantly lower than DLCTF but greater than NG.

The SMBC was significantly greater in NG than all the cropping systems in both soil depths (Fig. 2B). The SMBC in the 0–15 cm depth was highest in NG, followed by limited-irrigation and dryland cropping systems. The SMBC in the 15–30 cm depth followed the trend of 0–15 cm depth, with no significant difference between DLNTF and DLCTF, as well as irrigated systems.

The microbial respiration measured as CO₂-C release during laboratory incubation varied among agroecosystems. The NG released the highest amount of CO₂-C, followed by limited-irrigation cropping systems and dryland cropping systems (Fig. 3A). Since these agroecosystems varied in SOC content, we reanalyzed the data by normalizing with SOC content in the system. The CO₂-C release was consistent among NG, limited-irrigation cropping systems, and DLNTF (Fig. 3B). Only DLCTF had a lower microbial respiration than other treatments. The first-order kinetic model showed k in the range of 0.49–0.57, while labile substrate (C ₀) was in the range of 1.41–5.1 mg kg⁻¹ soil (0.22–0.38 g kg⁻¹ SOC), indicating most of the variation in microbial respiration was related to labile substrate present in the soil. The regression analysis showed that microbial respiration was strongly related to the SOC content and microbial biomass (Fig. 4). The model fit-statistics (r, RMSE, NRMSE) revealed that the model explained well the variability in CO₂-C release from these systems (Table 2).

The N₂O-N release during laboratory incubation showed a higher rate of N release from limited-irrigation cropping systems than dryland cropping systems and NG (Fig. 5A). When the N₂O-N release was normalized to total N, the value was even smaller per unit N in NG than all the cropping systems (Fig. 5B). The N mineralization constant k was in the range of 0.47–0.60, and N ₀ was in the range of 0.85–2.24 mg kg⁻¹ soil (0.95–3.85 g kg⁻¹ TN), showing the broader range in N₂O-N release relative to CO₂-C release (Table 2). The model-fit statistics (r, RMSE, NRMSE) for observed versus modeled N₂O-N release revealed that the first-order kinetic model explained well the variability in N₂O-N release from these systems. Soil N₂O-N release followed a weak trend with inorganic N (p = 0.12), and it was not related to the total N content.

Discussion

This study demonstrated an increase in SOC and N with native grassland restoration in agroecosystems transitioning to dryland, with greater magnitude of increase in biologically active SOC and N. Grasslands had 100% more SMBC and 51% more SOC than irrigated cropping systems and 243% more SMBC and 66% more SOC than dryland cropping systems. Biologically active soil organic matter is derived from high plant biomass production along with deeper and denser root systems of perennial grasses than annual crops (DuPont et al., 2010; Guo & Gifford , (2002). The SOC under perennial grasses also increases from a higher root: shoot ratio and low decomposition rate of roots from perennial grasses than from annual crops (Yang et al., 2019). Perennial grasses also increase microbial biomass, thereby contributing to more SOC and N accumulation in grasslands compared to croplands (Ghimire et al., 2019). Our previous study shows two- to six-times greater biomass C inputs in grassland than in cropland soils (Ghimire, Machado & Rhinhart, 2015). In the current study, native grasslands accumulated more C in the soil profile with a net increase in SOC sequestration. The SOC in the NG was even greater than the cover crop-integrated cropping system (LINTC). Grassland treatment was in place for 14 years, while the cover cropping treatment was in place for only three years.

The SOC sequestration or loss in agricultural soils depends on management practices, biomass inputs, and quality (Lehmann & Kleber, 2015; Ghimire, Machado & Bista, 2019). Above- and belowground plant production decreases with a decrease in water availability for irrigation. Reduced C and nutrients inputs from plant biomass recycling and enhanced microbial respiration with repeated disturbance can deplete SOC and nutrients in cropping systems transitioning to dryland. We did not monitor plant biomass production in different cropping systems and grasslands in this study. However, a significantly lower SOC, total and inorganic N, and SMBC in drylands than irrigated cropping systems demonstrated the effects of this transition. The SMBC and SOC were 42% and 14% lower in dryland than in irrigated cropping systems, irrespective of tillage management or cover cropping. The rate of CO₂-C and N₂O-N release per unit SOC and N, respectively, were similar across all agroecosystems compared in this study, which suggests no change in the rate of SOC and N loss despite the difference in inputs during the transition to dryland production. Agroecosystems in the semiarid regions continuously change SOC and nutrient status until they reach a new equilibrium (Bista et al., 2016). It appears SOC in dryland cropping systems in this study will continue to decline until it reaches a lower equilibrium.

Adoption of no-tillage typically increases SOC in surface soil (Mitchell et al., 2015; Sainju, Jabro & Stevens, 2008). Three years of no-tillage after 11 years of dryland cropping and conventional tillage was not enough to detect increase in SOC in this hot, dry semiarid agroecosystem. Management systems that reduce soil disturbance along with cropping system intensification could increase SOC and maintain net positive ecosystem carbon balance compared to no-tillage alone in hot, dry environments of the Southern High Plains (Rosenzweig, Stromberger & Schipanski, 2018; Thapa et al., 2019). Cropping system intensification increases SOC through greater C inputs to soil, as well as through increased fungal biomass and soil aggregation (Rosenzweig, Stromberger & Schipanski, 2018). Our study did not compare crop intensification and diversification options in drylands, but cover crop-integrated no-tillage system (LINTC) had numerically higher SOC than without cover crops under limited irrigation. A greater magnitude of increase in biologically active organic matter was observed with cover cropping than without, which suggests a potential long-term improvement in soil health through soil C accumulation. Cover crop plots in this study used a mixture of legume, brassica, and grass species. The legume and brassica cover crops produce high-quality biomass due to their low carbon to nitrogen ratios (Finney, White & Kaye , 2016). A mixture of cover crops also diversify microbial substrates through diverse root exudates and aboveground plant parts, which alters decomposition dynamics of crop residue inputs, leading to improvements in soil nutrient cycling and SOC accumulation (Mc Daniel et al., 2016). Our study shows higher SMBC and microbial respiration during eight weeks of incubation with cover crops than without cover crops suggesting the possibility of improving SOC in long-term.

The first-order kinetic model was used to evaluate the labile SOC and N pools and decomposition kinetics under various management systems. The decomposition rate constant (K) was highest in conventionally tilled drylands, followed by limited-irrigation croplands, and native grasslands. It is not surprising that decomposition rate increases with increasing disturbance and annual cropping compared to perennial grasses. Previous studies show a positive relationship between labile SOC pools, CO₂-C production, and SOC sequestration (Ahlering, Fargione & Parton, 2016; Ghimire et al., 2019; Ghimire, Machado & Bista, 2019; Six, Elliott & Paustian, 2000). However, we expected a lower decomposition rate in dryland cropping systems because of low biomass recycling in dryland cropping systems. The CO₂-C release did not vary between agroecosystems when normalized to SOC content, except in DLCTF. Relatively small CO₂-C fluxes in DLCTF when normalized to SOC content may be due to loss of labile carbon during repeated tillage activities.

The release of N₂O-N was highest in limited-irrigation cropping systems, followed by native grassland and dryland cropping systems. The highest rate of N release in limited-irrigation systems may have been due to high N input through external sources (fertilizer and atmospheric N fixation). Limited-irrigation systems received more N (97 kg N ha⁻¹ for sorghum) than dryland systems (33.6 kg N ha⁻¹ for sorghum). Although the relationship was not statistically significant for both N pools (p ≤ 0.05), there was a better correlation with inorganic N than with total N. The lowest N₂O-N release and inorganic N despite highest total N storage in native grassland suggests N limit the grassland productivity. Grasslands are often low in soil inorganic N because they do not receive any external N inputs and solely depend on internal N cycling. High C inputs for several years from biomass recycling along with no external N inputs increased SOC in grasslands, but low soil inorganic N and N₂O-N release. This shows possibility of mitigating global warming through grassland restoration, considering 310 times higher global warming potential of N₂O than that of CO₂ (IPCC, 2007). Evaluating CO₂ and N₂O emissions under grasslands and croplands in the Northern High Plains, Ahlering, Fargione & Parton (2016) reported the potential to offset SOC depletion and contribution to climate mitigation by avoiding conversion of grasslands into agricultural lands.

Water level in the Ogallala Aquifer has been depleting and the decrease in SOC and N levels are underway (Cano et al., 2018). Areas currently under limited-irrigation crop production could benefit from cover cropping in long-term because we observed small but positive effects of cover crops on SOC and N storage within three years of study establishment. However, about 24% of currently irrigated acreage is projected to be converted to dryland by 2100 and 13% of these areas are not suitable for dryland crop production (Deines et al., 2020). In such areas, grassland establishment could restore SOC and N in the soil profile. There was a strong positive relationship between microbial respiration, SMBC, and SOC. Specifically, the response of labile SOC components, including SMBC and microbial respiration were very clear. Agroecosystems that reduce soil disturbance, integrate cover crops, and increase the residue inputs can increase SOC and N storage and improve the sustainability of agriculture in semiarid regions transitioning from limited-irrigation to dryland management.

Conclusions

A vast area of the High Plains of USA is facing challenges in maintaining irrigated crop production due to a continuous depletion of the water level in the Ogallala Aquifer. Management selection plays a crucial role on SOC and N cycling in semiarid agroecosystems transiting to dryland. Our study demonstrated loss of SOC and N in agroecosystems transitioned to dryland crop-fallow systems and greater magnitude of change was observed in the biologically active fraction of soil organic matter. Grassland restoration could sequester C and nutrients in the soil profile. Although no-tillage cover crops were planted for only three years, there was a small increase in SOC components. The large CO₂ fluxes from soils with higher SOC agroecosystem suggests a critical role of residue inputs in C cycling. The N cycling was more related to the external N input via fertilizer application. An in-depth and regional-scale study will reveal the role of various agroecological measures in mitigating SOC and N loss during transition to dryland production. This study demonstrated that no-tillage and cover cropping had a small but positive change in SOC storage and nutrient cycling. Grassland restoration can increase SOC and N storage and improve agroecosystem sustainability in the hot, dry, semiarid agroecosystems facing transition to dryland.

Supplemental Information