The effect of long-term soil management on the physical and biological resilience of a range of arable and grassland soils in England

Abstract

In agricultural systems, soils are subjected to a range of stresses which may affect soil physical and biological properties and compromise the ability of the soil to function. How the soil responds to stress both initially and over time defines resistance and resilience respectively and is key to sustainable soil management. As soil organic matter (OM) is a fundamental soil property, we hypothesised that resilience would be controlled, in part, by this. As soil OM is itself affected by long-term soil management, we therefore sought to establish a direct link between this and resilience to physical and biological stresses.

Fifteen soils were taken from a range of arable and grassland sites in England, including the Broadbalk and Highfield long-term experiments at Rothamsted. The soils were subjected to physical (uniaxial compaction stress) or biological (transient heat stress or persistent Cu stress) stresses in the laboratory and the effects on general physical (void ratio) and biological (substrate-induced respiration) functions were monitored over a recovery period of up to 28 days. Wet–dry and freeze–thaw cycles were also imposed on the soils following the initial recovery after compression.

The initial recovery of void ratio following compression was greater in grassland soils (0.28–0.80) than arable soils (0.16–0.58) which was probably related to the elastic effect of particulate OM and the redistribution of the greater water content in the former soils. Wet–dry and freeze–thaw cycles had an additional effect in soils with clay contents greater than 26%. The transient heat stress reduced respiration rates in the grassland soils by up to 45%. Grassland soils are less likely to have been subjected to heat stress in the field since they carry a greater vegetation covering than arable soils. The more persistent Cu stress reduced respiration rates in arable soils by up to 85% but only up to 30% in grassland soils. The difference was ascribed to the greater microbial biomass and diversity in grassland soils and buffering by the OM. We defined resistance and resilience indices as the absolute value of the measured soil functions at maximum stress and after the recovery period respectively. Reasonable linear regressions (adjusted R² up to 0.68) were established for resistance or resilience as a function of soil OM content, used as a proxy index of soil management. Expressed in this way, grassland soils were more resistant and resilient to the physical and biological stresses than the arable soils. Linear relationships were also found to link physical and biological resistance and resilience (adjusted R² up to 0.96). An arable soil with 65% clay was found to be highly resilient also.

We found large differences in physical and biological resilience to stress between arable and grassland soils. We attributed this primarily to the effect of soil management on the OM content, although we recognised the importance of clay content.

Introduction

Soil must be able to recover from stresses imposed by man and nature if it is to continue to function, and management cannot be said to be sustainable if this is not the case. In agricultural systems, a range of stresses can be imposed on soils, including vehicle traffic, cultivation, fertilisers and weather, which may damage soil physical and biological properties and compromise the ability of the soil to function. Soils may become polluted, eroded or flooded in extreme instances. There are two components of the response of a soil property or function to stress: the initial response and the subsequent recovery over time following removal of the stress. Following the work of other soil scientists in this area (e.g. Lal, 1993, Kay et al., 1994, Seybold et al., 1999), we define these two components as resistance and resilience respectively.

Soil resilience studies often monitor a particular soil property or function prior to, during and following imposition of a stress. For example, soil physical resilience has been expressed through monitoring pore volume (O'O'Sullivan et al., 1999, Zhang et al., 2005, Kuan et al., 2007) and strength (Munkholm and Schjønning, 2004, Gregory et al., 2007) during and after compaction, by measuring the stability (Denef et al., 2001) and size distribution (Grant et al., 1995) of soil aggregates to wet–dry cycles, and by measuring vertical soil movements after compaction (Tobias et al., 2001). Biological resilience in soil has been quantified by measuring changes in the short-term mineralization of plant residues (Griffiths et al., 2000, Kuan et al., 2007), dissolved organic carbon (Merckx et al., 2001), catabolic function (Degens and Harris, 1997) specific microbial function groups (Tobor-Kaplon et al., 2005, Wada and Toyota, 2007, Wertz et al., 2007) and the size and activity of the microbial biomass (Franco et al., 2004) in response to disturbance. Orwin and Wardle (2004) offer an interesting discussion on how to convert measured soil properties into resilience indices.

Recently, the links between biological and physical processes have been assessed using resilience assays. Griffiths et al. (2005) found that soils treated with sewage-sludge had greater biological and physical resilience, probably because of the increased soil organic matter (OM) content. Griffiths et al. (2008) later reported a link between the physico-chemical properties of contrasting soils and their biological resilience. However there have been studies of other soils that have failed to find a clear link between soil physical resilience and biological resilience (Runion et al., 2004, Shestak and Busse, 2005). Later work by Kuan et al. (2007) found only a limited relationship between physical and biological resilience in a range of soils from Scotland. The latter studies that failed to find links, however, either considered the short-term response (< 1 year) to compaction (Runion et al., 2004, Shestak and Busse, 2005) or studied a wide range of soil types. The impact of long-term soil management on both physical and biological resilience has not been studied using paired experiments on experimental fields.

We suggest that long-term land management, lasting decades rather than several years, may influence resilience considerably, with biological and physical resilience being closely linked. Grassland management, in comparison to arable management where the soil is annually cultivated, generally results in greater OM and microbial biomass contents. This is driven by greater organic returns to the soil through plant litter, animal manures and root-derived substances, and less carbon mineralization by mechanical disturbance. Grassland soils also tend to have a better-developed soil structure due to the more extensive root system and associated fungal hyphae (Tisdall and Oades, 1982). Zhang & Hartge (1990) found strong links between the physical resistance to compression and OM, which Zhang et al. (2005) later postulated was driven by particulate OM. So if a grassland soil were then to be subjected to a stress, processes such as root penetration and earthworm and microorganism activity may aid structural and biological recovery (Barzegar et al., 1995, Seybold et al., 1999, Larink et al., 2001, Radford et al., 2001, Langmaack et al., 2002). Correlations have been made between the biodiversity of grassland soils and resilience to biological stresses (Girvan et al., 2005, Tobor-Kaplon et al., 2005, Brussaard et al., 2007). In addition, particulate OM may enhance elastic rebound from physical stresses (Watts and Dexter, 1997, Zhang et al., 2005) and may adsorb substances toxic to microorganisms (Hund-Rinke and Kördel, 2003). The better-developed structure provides the habitat (Young and Ritz, 2000) and may protect microorganisms from perturbation. Recovery processes may be inhibited in arable soils due to annual cultivation practices and the harvesting of the vegetative cover.

In this study we explore the effect of long-term soil management in agricultural systems on physical and biological resilience following a stress, using our soil resilience assays (Griffiths et al., 2000, Gregory et al., 2006, Kuan et al., 2007). Included in this study are contrasting soil management plots in the well-known Broadbalk (arable and wilderness) and Highfield (arable, fallow, grass) long-term experiments at Rothamsted Research in England, as well as more recent experiments developed on different soil types. We hypothesised that the OM content of a soil, arising from the agricultural system for which it is used, is a key soil property determining resilience through its effect on soil structure and soil microbiology. We hypothesised that initial resistance and longer-term resilience, measured with our rapid stress-recovery indices, would be greater in soils with a greater OM content. An inherent soil property that also affects soil physical resilience in particular is texture, and we hypothesised that resilience would be greater in clayey soils than sandy soils.