Search

Soil Quality or Soil Health are terms adopted by the scientific community as a metaphor for the effects of differing land management practices on the properties and functions of soil. Many other terms and metaphors are in use that defy neat quantification: human health, for example. Our challenge is to understand the importance of using such metaphors, but without compromising the underlying scientific understanding upon which they are based. We present here an approach based on expert elicitation in the field of soil quality and management, which offers a universal way of putting numbers to the metaphor. Like humans, soils differ and so do the ways in which they become unhealthy. We structure experts’ views of the extent to which soil delivers the functions expected of it within Bayesian Belief Networks anchored by measurable properties of soil. With these networks, we deduce the value of additional data to the precision of estimates of soil quality and health and infer the likely state of soil at locations in England & Wales. We conclude that the value of soil is best scored as its fitness for purpose or its utility. Our methodology has general applicability and could be deployed elsewhere or in other disciplines.

In the mid-19th century, John Bennet Lawes and Joseph Henry Gilbert established a series of field experiments at Rothamsted (Hertfordshire, in south-east UK) to examine the benefits of mineral fertilizers and organic manures on the production of several important agricultural crops. Seven of these “Classical” field experiments continue today. The Broadbalk Wheat, Hoosfield Spring Barley, and Park Grass Continuous Hay experiments are perhaps the most widely known of these long-term experiments. When the Classicals began, Lawes and Gilbert could not have envisaged the wide-ranging impact they would have on agricultural and other related sciences. They would probably also be surprised that the experiments have been maintained and that they have continued to be important platforms for research for such a long period. The experiments quickly demonstrated the importance of fertilizers and manures as a source of mineral nutrients, especially nitrogen (N) and phosphorus (P), for plant growth and development. Subsequent findings highlighted the need for good weed and disease control, the value of crop rotation, liming, and the use of modern short-strawed cereal varieties. Yield trends for wheat and barley on Broadbalk and Hoosfield, respectively, have followed the national averages, clearly demonstrating the capacity for arable soils in temperate Northwestern Europe to sustain good levels of crop production for well over a century provided exported nutrients are replaced, soil pH is kept around neutral, soil structure is maintained, and pests, weeds, and diseases are adequately controlled. In addition, the experiments generate information of value to agroecologists, farmers, and other land users on the management of seminatural systems to maintain and enhance biodiversity. They are an invaluable resource for monitoring the impact of agricultural management and environmental change, especially climate change, on the crop–soil system and other seminatural habitats. Provided the experiments, sample archive, and associated data are well-managed and used together with new analytical techniques and ideas, they will continue to provide a resource for agricultural research as relevant today as when the experiments started 175 years ago.

The management of optimal soil pH is fundamental to sustainable crop production. Understanding the lime requirement for arable crops has developed gradually over the last several decades. The aim of this study was to examine the yield-pH relationship for a range of arable crops to understand their response to liming, based on the Long-Term Liming experiments established in 1962 at Rothamsted Research, UK. The main treatments of four different rates of lime and, therefore, four distinctly different soil pH levels were maintained for 35 years at two sites (Rothamsted and Woburn). The pH ranged from 4.4 to 8.0. The lime response was tested on the following crops: spring barley, spring oats, spring beans, spring lupins, winter lupins, potatoes, linseed, winter oilseed rape, winter triticale and winter wheat. Relative yield (RY) was used for non-linear regression analysis to detect site, year and phosphorus (P) fertiliser effects on the relationship with pH. Liming had a highly significant positive effect on soil pH, but overall there was no consistent increase or decrease in soil extractable P (Olsen) or exchangeable K. There were significant site effects detected for RY for most crops which reflect differences in the two soil types. Spring oats and potatoes had very weak responses to lime within the pH range tested. For spring barley, winter triticale, winter wheat and winter oilseed rape significant effects of P fertiliser on the yield-pH relationship were found, although the nature of effects differed between crops and sites. Findings from the Long-Term Liming experiment are invaluable in improving the fundamental understanding on the yield-pH relationship for important arable crops and this has significant implications on selecting crops for rotations. The pH at 90% RY was calculated for selected crops and the beneficial effect of fertiliser P was detected in significantly reducing the critical pH value.

The mechanisms by which microbial communities maintain functions within the context of changing environments are key to a wide variety of environmental processes. In soil, these mechanisms support fertility. Genes associated with hydrolysis of organic phosphoesters represent an interesting set of genes with which to study maintenance of function in microbiomes. Here, we shown that the richness of ecotypes for each gene varies considerably in response to application of manure and various inorganic fertilizer combinations. We show, at unprecedented phylogenetic resolution, that phylogenetic diversity of phosphohydrolase genes are more responsive to soil management and edaphic factors than the taxonomic biomarker 16S rRNA gene. Available phosphorus – assessed by measuring Olsen-P - exerted some influence on alkaline phosphatase distribution: however, consistent and significant differences were observed in gene abundance between treatments that were inconsistent with bioavailable orthophosphate being the dominant factor determining gene abundance. Instead, we observed gene niche separation which was most strongly associated with soil exchangeable calcium. Our study suggests that the bioavailability of enzyme cofactors (exchangeable calcium in the case of phoD, phoX and βPPhy studied here) influence the abundance of genes in soil microbial communities; in the absence of cofactors, genes coding for alternative enzyme families that do not require the limiting cofactor (for example, non-specific acid phosphatases which require vanadate) become more abundant., Highlights • Metagenomics-based study of the longest running Long-Term Experiment in the World. • Studies 16SrRNA-contingent soil communities and nine phosphohydrolase genes. • Studies the effect of addition of farmyard manure or inorganic fertilizers. • Demonstrates an important role of bioavailable calcium in gene distribution. • The influence of bioavailable calcium is greater than bioavailable phosphorus.

This paper describes an agricultural model (Roth-CNP) that estimates carbon (C), nitrogen (N) and phosphorus (P) pools, pool changes, their balance and the nutrient fluxes exported from arable and grassland systems in the UK during 1800–2010. The Roth-CNP model was developed as part of an Integrated Model (IM) to simulate C, N and P cycling for the whole of UK, by loosely coupling terrestrial, hydrological and hydro-chemical models. The model was calibrated and tested using long term experiment (LTE) data from Broadbalk (1843) and Park Grass (1856) at Rothamsted. We estimated C, N and P balance and their fluxes exported from arable and grassland systems on a 5 km × 5 km grid across the whole of UK by using the area of arable of crops and livestock numbers in each grid and their management. The model estimated crop and grass yields, soil organic carbon (SOC) stocks and nutrient fluxes in the form of NH4-N, NO3-N and PO4-P. The simulated crop yields were compared to that reported by national agricultural statistics for the historical to the current period. Overall, arable land in the UK have lost SOC by −0.18, −0.25 and −0.08 Mg C ha−1 y−1 whereas land under improved grassland SOC stock has increased by 0.20, 0.47 and 0.24 Mg C ha−1 y−1 during 1800–1950, 1950–1970 and 1970–2010 simulated in this study. Simulated N loss (by leaching, runoff, soil erosion and denitrification) increased both under arable (−15, −18 and −53 kg N ha−1 y−1) and grass (−18, −22 and −36 kg N ha−1 y−1) during different time periods. Simulated P surplus increased from 2.6, 10.8 and 18.1 kg P ha−1 y−1 under arable and 2.8, 11.3 and 3.6 kg P ha−1 y−1 under grass lands 1800–1950, 1950–1970 and 1970–2010., Graphical abstract Unlabelled Image, Highlights • Roth-CNP model estimates C, N and P cycling within the UK agriculture for 1800–2010. • Simulated crop yields were comparable to the yields of UK's agricultural statistics. • Simulated SOC stock decreased under arable and increased under improved grassland. • Simulated N and P losses increased under both arable and grasslands. • Results shows the effect of local agriculture in a larger context of space and time.

The electronic Rothamsted Archive, e-RA (www.era.rothamsted.ac.uk) provides a permanent managed database to both securely store and disseminate data from Rothamsted Research’s long-term field experiments (since 1843) and meteorological stations (since 1853). Both historical and contemporary data are made available via this online database which provides the scientific community with access to a unique continuous record of agricultural experiments and weather measured since the mid-19th century. Qualitative information, such as treatment and management practices, plans and soil information, accompanies the data and are made available on the e-RA website. e-RA was released externally to the wider scientific community in 2013 and this paper describes its development, content, curation and the access process for data users. Case studies illustrate the diverse applications of the data, including its original intended purposes and recent unforeseen applications. Usage monitoring demonstrates the data are of increasing interest. Future developments, including adopting FAIR data principles, are proposed as the resource is increasingly recognised as a unique archive of data relevant to sustainable agriculture, agroecology and the environment.

We describe a model framework that simulates spatial and temporal interactions in agricultural landscapes and that can be used to explore trade-offs between production and environment so helping to determine solutions to the problems of sustainable food production. Here we focus on models of agricultural production, water movement and nutrient flow in a landscape. We validate these models against data from two long-term experiments, (the first a continuous wheat experiment and the other a permanent grass-land experiment) and an experiment where water and nutrient flow are measured from isolated catchments. The model simulated wheat yield (RMSE 20.3–28.6%), grain N (RMSE 21.3–42.5%) and P (RMSE 20.2–29% excluding the nil N plots), and total soil organic carbon particularly well (RMSE 3.1 − 13.8 %), the simulations of water flow were also reasonable (RMSE 180.36 and 226.02%). We illustrate the use of our model framework to explore trade-offs between production and nutrient losses., Graphical abstract Image 1, Highlights • Understanding trade-offs between yield and environment is essential for SI. • The Landscape Model aids the understanding of crop-soil-water interactions. • Model validated against 50 years of data from two long-term experiments. • Model validated against spatially-explicit data from the North Wyke farm platform. • The model simulated wheat yield, grain N and grain P particularly well.

The UK’s greenhouse gas inventory for agriculture uses a model based on the IPCC Tier 1 and Tier 2 26 methods to estimate the emissions of methane and nitrous oxide from agriculture. The inventory 27 calculations are disaggregated at country level (England, Wales, Scotland and Northern Ireland). 28 Before now, no detailed assessment of the uncertainties in the estimates of emissions had been 29 done. We used Monte Carlo simulation to do such an analysis. We collated information on the 30 uncertainties of each of the model inputs. The uncertainties propagate through the model and result 31 in uncertainties in the estimated emissions. Using a sensitivity analysis, we found that in England and 32 Scotland the uncertainty in the emission factor for emissions from N inputs (EF1) affected 33 uncertainty the most, but that in Wales and Northern Ireland, the emission factor for N leaching and 34 runoff (EF5) had greater influence. We showed that if the uncertainty in any one of these emission 35 factors is reduced by 50%, the uncertainty in emissions of nitrous oxide reduces by 10%. The 36 uncertainty in the estimate for the emissions of methane emission factors for enteric fermentation 37 in cows and sheep most affected the uncertainty in methane emissions. When inventories are 38 disaggregated (as that for the UK is) correlation between separate instances of each emission factor 39 will affect the uncertainty in emissions. As more countries move towards inventory models with 40 disaggregation, it is important that the IPCC give firm guidance on this topic.

World soil carbon (C) stocks are third only to those in the ocean and earth crust and represent twice the amount currently present in the atmosphere. Therefore, any small change in the amount of soil organic C (SOC) may affect carbon dioxide (CO2) concentrations in the atmosphere. Dynamic models of SOC help reveal the interaction among soil carbon systems, climate and land management and they are also frequently used to help assess SOC dynamics. Those models often use allometric functions to calculate soil C inputs in which the amount of C in both above and below ground crop residues are assumed to be proportional to crop harvest yield. Here we argue that simulating changes in SOC stocks based on C input that are proportional to crop yield is not supported by data from long-term experiments with measured SOC changes. Rather, there is evidence that root C inputs are largely independent of crop yield, but crop specific. We discuss implications of applying fixed belowground C input regardless of crop yield on agricultural greenhouse gas mitigation and accounting.

Bioenergy developments could lead to large-scale removal of cereal straw from fields, with consequences for soil organic carbon (SOC) and related properties. In 25 experiments of 6 to 56 yr duration there was a trend for SOC and total soil N content to increase where straw was incorporated annually. However the increases were only significant in six experiments and were

We have predicted the potential distribution of 26 bioenergy crops in the UK, based on the simple model described by Tuck et al. [1]. The model has been applied at a 5 km resolution using the UKCIP02 model for scenarios at Low, Medium-Low, Medium-High and High emissions. In the analysis of the results the limitations for crop growth are assigned to elevation, temperature, high and low rainfall. Most of the crops currently grown are predicted to remain prevalent in the UK. A number of crops are suitable for introduction to the UK under a changing climate, whereas others retreat to northern parts of the UK. The greatest changes are expected in England. The simplicity of the model means that it has a relatively high uncertainty, with minor modifications to the model leading to quite different results. Nevertheless, it is well suited for identifying areas and crops that are most likely to be affected by the greatest changes. It has been noted that Miscanthus and Short Rotation Coppice (SRC) willow and poplar, which are currently regarded as highly suitable for UK conditions, may be less suited to southern areas in the future, where, for example, kenaf could have a greater potential. Further investigations are required to reduce uncertainty associated with the projections based on this simple model and to make conclusions more firmly.

We compared alternate uses of cereal straw (4.25t dry matter ha(-1) containing 1.7t carbon (C)) for their effectiveness in relation to climate change mitigation. The scenarios were (1) incorporation into soil to increase soil organic carbon (SOC) content ("carbon sequestration") and (2) combustion to generate electricity. The Rothamsted Carbon Model was used to estimate SOC accumulation in a silty clay loam soil under the climatic conditions of north-west Europe. Using straw for electricity generation saved seven times more CO2 than from SOC accumulation. This comparison assumed that electricity from straw combustion displaced that generated from coal and used the mean annual accumulation of SOC over 100yr. SOC increased most rapidly in the early years, but then more slowly as a new equilibrium value was approached. We suggest that increased SOC from straw incorporation does not represent genuine climate change mitigation through carbon sequestration. In Europe, most straw not already incorporated in the field where it is grown is subsequently returned elsewhere, e.g., after use for animal bedding and production of manure. Only additional retention of C in soil compared to the alternative use represents sequestration. Maintenance of SOC for soil functioning is a more appropriate rationale for returning straw to soil than climate change mitigation. This analysis shows that considerably greater climate change mitigation is achieved through saved CO2 emissions by burning straw for electricity generation, replacing some use of fossil fuel.

Soil organic carbon (SOC) is a significant component of the global carbon (C) cycle. Changes in SOC storage affect atmospheric CO2 concentrations on decadal to centennial timescales. The C-TOOL model was developed to simulate farm- and regional-scale effects of management on medium- to long-term SOC storage in the profile of well-drained agricultural mineral soils. C-TOOL uses three SOC pools for both the topsoil (0–25 cm) and the subsoil (25–100 cm), and applies temperature-dependent first order kinetics to regulate C turnover. C-TOOL also enables the simulation of 14C turnover. The simple model structure facilitates calibration and requires few inputs (mean monthly air temperature, soil clay content, soil C/N ratio and C in organic inputs). The model was parameterised using data from 19 treatments drawn from seven long-term field experiments in the United Kingdom, Sweden and Denmark. It was found that the initial SOC content had to be optimised for each experiment, but also that one set of values for other model parameters could be applied at all sites. With this set of parameters, C-TOOL can be applied more widely to evaluate effects of management options on SOC storage in temperate agricultural soils. C-TOOL simulates observed losses of SOC in soils under intensive agricultural use and the gain in SOC derived from large inputs of animal manure and inclusion of perennial grassland. The model simulates changes in SOC for the entire profile, but lack of data on subsoil SOC storage hampers a proper model evaluation. Experimental verification of management effects on subsoil C storage, subsoil C inputs from roots, and vertical transport of C in the soil profile remains prioritised research areas.

In an earlier paper we presented data from an experiment in which nitrogen-15-labelled fertilizer was applied in spring to barley on the Rothamsted long-term Spring Barley Experiment, at rates of 48, 96 or 144 kg N ha−1. A substantial proportion (between 28 and 39%) of this 15N remained in the soil (0–70 cm) and stubble at harvest, mostly in organic form. The present paper follows the fate of this `residual' 15N over the following 2 years. Small amounts of `residual' 15N were recovered in the following two spring barley crops; 8% in the first and 3% in the second. The overall loss of `residual' 15N (i.e. `residual' 15N not recovered in crops and soil to a depth of 70 cm) over the 2 years was 23%. This is equivalent to just 8% of the total 15N originally applied. There was surprisingly little difference in the behaviour of the `residual' 15N in soils containing very different quantities of soil organic matter.

An experiment with 15N-labelled fertilizer was superimposed on the Rothamsted Hoosfield Spring Barley Experiment, started in 1852. Labelled 15NH4 15NO3 was applied in spring at (nominal) rates of 0, 48, 96 and 144 kg N ha-1. The labelled fertilizer was applied to microplots located within four treatments of the original experiment: that receiving farmyard manure (FYM) annually, that receiving inorganic nutrients (PK) annually and to two that were deficient in nutrients: applications were made in two successive years, but to different areas within these original treatments. Maximum yields in 1986 (7.1 t grain ha-1) were a little greater than in 1987. In 1987, microplots on the FYM and PK treatments gave similar yields, provided enough fertilizer N was applied, but in 1986 yields on the PK treatment were always less than those on the FYM treatment, no matter how much fertilizer N was applied. In plots with adequate crop nutrients, about 51% of the labelled N was present in above-ground crop and weed at harvest, about 30% remained in the top 70 cm of soil (mostly in the 0–23 cm layer) and about 19% was unaccounted for, all irrespective of the rate of N application and of the quantity of inorganic N in the soil at the time of application. Less than 4% of the added fertilizer N was present in inorganic form in the soil at harvest, confirming results from comparable experiments with autumn-sown cereals in south-east England. Thus, in this experiment there is no evidence that a spring-sown cereal is more likely to leave unused fertilizer in the soil than an autumn-sown one. With trace applications (ca. 2 kg N ha-1) more labelled N was retained in the soil and less was in the above-ground crop. Where P and K were deficient, yields were depressed, a smaller proportion of the labelled fertilizer N was present in the above-ground crop at harvest and more remained in the soil. Although the percentage uptake of labelled N was similar across the range of fertilizer N applications, the uptake of total N fell off at the higher N rates, particularly on the FYM treatment. This was reflected in the appearance of a negative Added Nitrogen Interaction (ANI) at the highest rate of application. Fertilizer N blocked the uptake of soil N, particularly from below 23 cm, once the capacity of the crop to take up N was exceeded. Denitrification and leaching were almost certainly insufficient to account for the 19% loss of spring-added N across the whole range of N applications and other loss processes must also have contributed.

Given that we must farm land in order to eat, the total environmental burden imposed by farming a crop, such as winter wheat in the UK, appears to be close to the minimum given current production techniques. The value of the services other than food production, such as flood water buffering, pollination, carbon storage and so on, that land can provide is relatively large compared with the value in reducing environmental burdens from pesticide use, nutrient pollution and greenhouse gas emissions that might arise by farming less intensively. More land will need to be brought into cultivation in order to provide the same amount of food if the intensity of farming is reduced and the resultant loss of ecosystem services (ES) outweighs the reduction in other burdens. Nevertheless, losses of nutrients, especially nitrogen (N), from agriculture are a serious concern and the current cost of the environmental footprint of agriculture is significant compared with the value of the food it produces. This article examines nutrient burdens and analyses the means by which the total environmental burden might be reduced relative to productivity. These include increasing the efficiency of farming, removing constraints to yield, and establishing multiple uses for land at the same time as farming. It concludes that agronomic measures which improve nutrient capture and which obtain more yield per unit area are valuable means to avoid degradation of environmental quality because both nutrient pollution and land consumption can be avoided.

Summary Realistic model representation of the dynamics of terrestrial ecosystems, including all relevant biophysical and biogeochemical interactions, is a prerequisite to accurate predictions of climate change. However, the current generation of models requires improvement. In particular, it is important to know the nitrogen status of soil for modelling global vegetation and the emission of pollutants such as nitrous oxide. For this reason, we collated several major datasets that record total nitrogen (TN) and other key properties of soils. From these data, we derived pedotransfer functions (PTFs) to predict TN of soils from other soil and site properties using stepwise multiple regressions. The most important predictor was found to be soil organic carbon (SOC). Other helpful soil and site properties for predicting TN included distance from the equator, soil texture class, soil group C:N ratio, pH, % clay and whether the soil contains more or less than 10% SOC. Several different PTFs are presented, giving the user a choice, depending on which soil and site properties are available. In all, we calculated that over 9000 observations were available, although not all datasets contained all the site properties listed above. The most useful PTF was found to be: TN = 0.0232 + 0.0250 ∗ C − 0.000534 ∗ D + 0.521 ∗ G + a(T ) ,f or C< 10%, and TN = 0.269 + 0.0449 ∗ C − 0.00575 ∗ D − 0.157 ∗ G + b(T ) ,f orC ≥ 10%, where C is % SOC, D is the degrees from the equator, G is a soil group factor defined as (C/Median C:N ratio of soil group), T is a factor based on texture classes (1, 2 or 3), depending on clay and sand content and a takes the value 0.0, 0.0228 or 0.0296 when T is 1, 2 or 3, respectively, and similarly b takes one of the values 0.0, 0.328 or 0.440, respectively. The PTFs were validated against data not used in their derivation, which were held back for this specific purpose. On the basis of the PTF above and the pre-existing digital soil map of the world, we publish a map of the global distribution of TN. It has been suggested that additional CO2 may be emitted from soils as a result of global warming, exacerbating warming still further. We calculate that the nitrous oxide formed as a result of the nitrogen that also mineralized could contribute a further 30% of additional warming.

Until recently, agricultural production was optimised almost exclusively for profit but now farming is under pressure to meet environmental targets. A method is presented and applied for optimising the sustainability of agricultural production systems in terms of both economics and the environment. Components of the agricultural production chain are analysed using environmental life-cycle assessment (LCA) and a financial value attributed to the resources consumed and burden imposed on the environment by agriculture, as well as to the products. The sum of the outputs is weighed against the inputs and the system considered sustainable if the value of the outputs exceeds those of the inputs. If this ratio is plotted against the sum of inputs for all levels of input, a diminishing returns curve should result and the optimum level of sustainability is located at the maximum of the curve. Data were taken from standard economic almanacs and from published LCA reports on the extent of consumption and environmental burdens resulting from farming in the UK. Land-use is valued using the concept of ecosystem services. Our analysis suggests that agricultural systems are sustainable at rates of production close to current levels practiced in the UK. Extensification of farming, which is thought to favour non-food ecosystem services, requires more land to produce the same amount of food. The loss of ecosystem services hitherto provided by natural land brought into production is greater than that which can be provided by land now under extensive farming. This loss of ecosystem service is large in comparison to the benefit of a reduction in emission of nutrients and pesticides. However, food production is essential, so the coupling of subsidies that represent a relatively large component of the economic output in EU farming, with measures to reduce pollution are well-aimed. Measures to ensure that as little extra land is brought into production as possible or that marginal land is allowed to revert to nature would seem to be equally well-aimed, even if this required more intensive use of productive areas. We conclude that current arable farming in the EU is sustainable with either realistic prices for products or some degree of subsidy and that productivity per unit area of land and greenhouse gas emission (subsuming primary energy consumption) are the most important pressures on the sustainability of farming.

Five white clover populations of Swiss origin and three bred varieties were grown in binary mixtures with two perennial ryegrass varieties, Aurora and S23. The seasonal yields of clover and grass plus clover were measured under a cutting regime during the second and third years after establishment. A series of destructive detailed sward measurements was made during the late autumn to spring period preceding each harvest year. In this way changes in the amounts of stolon, leaf plus petiole and numbers of growing points were monitored during the winter. There were large differences in clover yield between populations in both years. These were evident from the first (spring) harvest in each year. Higher-yielding clovers in spring tended to produce higher annual clover yields. No grass × clover interaction was evident at any harvest. Large differences between clovers were also apparent in the morphological characteristics measured, with the Swiss material generally having greater amounts of stolon, leaf plus petiole and numbers of growing points present in early spring. It is proposed that these factors contribute to the high spring yield in the Swiss populations. Loss of stolon length over each winter was less in the Swiss material, indicating that its good spring growth was not obtained at the expense of winter hardiness. Annual clover yield was found to be significantly positively correlated with the amount of stolon present in spring, exemplifying the importance of stolon survival over the winter.

Global change will alter the supply of ecosystem services that are vital for human well-being. To investigate ecosystem service supply during the 21st century, we used a range of ecosystem models and scenarios of climate and land-use change to conduct a Europe-wide assessment. Large changes in climate and land use typically resulted in large changes in ecosystem service supply. Some of these trends may be positive (for example, increases in forest area and productivity) or offer opportunities (for example, “surplus land” for agricultural extensification and bioenergy production). However, many changes increase vulnerability as a result of a decreasing supply of ecosystem services (for example, declining soil fertility, declining water availability, increasing risk of forest fires), especially in the Mediterranean and mountain regions.

A decision support system has been constructed around the nitrogen (N) turnover model SUNDIAL which allows farmers and policy makers to explore how arable rotations respond to practical strategies for reducing N losses. It automatically derives all crop rotations allowed within an imposed set of farming constraints and presents the N dynamics for each rotation. Total N losses (by leaching and gaseous losses) and crop N offtake were simulated for a six year arable rotation based on two winter wheat crops, spring barley, winter oilseed rape, winter beans and set-aside (cropped with industrial oilseed rape). The simulation was run using three basic soil types (sand, loam and clay), for three different productivity levels (low, medium and high), and using typical weather data from three different geographical regions in England (Southwest, Central and East Anglia), to look at the effects of changing the order of the crops in the rotation and incorporating straw on N losses and crop N offtake. The simulations suggest that straw incorporation decreases potential N losses to the environment, but that crops are generally unable to make use of the saved N. The average N offtake following straw incorporation is lower than when straw is not incorporated. Changing the sequence of crops in the rotation has a greater effect, significantly reducing N losses and increasing mean crop N offtake. On average, the best rotations lost 319 kg N ha −1 throughout the six year rotation, compared to the worst rotations, which lost 464 kg N ha −1 , a saving of 145 kg N ha −1 achieved merely by changing the order of the crops.

In this paper we estimate the European potential for carbon mitigation of no-till farming using results from European tillage experiments. Our calculations suggest some potential in terms of (a) reduced agricultural fossil fuel emissions, and (b) increased soil carbon sequestration. We estimate that 100% conversion to no-till farming would be likely to sequester about 23 Tg C y(-1) in the European Union or about 43 Tg C y(-1) in the wider Europe (excluding the former Soviet Union). In addition, up to 3.2 Tg C y(-1) could be saved in agricultural fossil fuel emissions. Compared to estimates of the potential for carbon sequestration of other carbon mitigation options, no-till agriculture shows nearly twice the potential of scenarios whereby soils are amended with organic materials. Our calculations suggest that 100% conversion to no-till agriculture in Europe could mitigate all fossil fuel-carbon emissions from agriculture in Europe. However, this is equivalent to only about 4.1% of total anthropogenic CO2-carbon produced annually in Europe (excluding the former Soviet Union) which in turn is equivalent to about 0.8% of global annual anthropogenic CO2-carbon emissions.

Many long-term experiments were not originally established to measure changes in soil organic matter (SOM) content. Thus, there are inevitably difficulties associated with interpreting SOM data from these sites. The general difficulties associated with long-term experiments mainly arise from the experimental design, sampling methods and record keeping. These include: little or no replication or randomisation; lack of time-zero samples; inappropriate control treatments; incomplete description of the experiment and sampling protocols; modifications to the experiment with time; soil movement between the plots. In addition, there are some difficulties of interpretation specifically associated with measuring long-term changes in SOM. These include: modifications to soil sampling protocol (depth, method, number and distribution of samples); changes in soil bulk density; methods of measuring soil C. In this paper we describe these potential problems, and identify the main sources of error when interpreting SOM data from long-term experiments.

Soil organic matter (SOM) is recognised as being of critical importance as a source and sink of carbon in the biosphere. As a result, research into predicting the effects of global environmental change on soil organic matter has been identified as a high priority within the Global Change and Terrestrial Ecosystems (GCTE) programme of the International Geosphere-Biosphere Programme (IGBP). The objectives of GCTE Task 3.3.1., “Soil Organic Matter” require that a global network of SOM modellers and experimenters be established. This global soil organic matter network (SOMNET) will comprise a number of regional networks, one of which will be the European GCTE SOMNET.

Three experiments are reported which examine the relative roles of host and Rhizobium genotypes as factors limiting clover (Trifolium repens L.) growth at low soil temperatures.

Clearing natural vegetation to establish arable agriculture (cropland) almost invariably causes a loss of soil organic carbon (SOC). Is it possible to restore soil that continues in arable agriculture to the pre-clearance SOC level through modified management practices? To address this question we reviewed evidence from long-term experiments at Rothamsted Research, UK, Bad Lauchstädt, Germany, Sanborn Field, USA and Brazil and both experiments and surveys of farmers’ fields in Ethiopia, Australia, Zimbabwe, UK and Chile. In most cases SOC content in soil under arable cropping was in the range 38–67% of pre-clearance values. Returning crop residues, adding manures or including periods of pasture within arable rotations increased this, often to 60–70% of initial values. Under tropical climatic conditions SOC loss after clearance was particularly rapid, e.g. a loss of >50% in less than 10 years in smallholder farmers’ fields in Zimbabwe. If larger yielding crops were grown, using fertilizers, and maize stover returned instead of being grazed by cattle, the loss was reduced. An important exception to the general trend of SOC loss after clearance was clearing Cerrado vegetation on highly weathered acidic soils in Brazil and conversion to cropping with maize and soybean. Other exceptions were unrealistically large annual applications of manure and including long periods of pasture in a highly SOC-retentive volcanic soil. Also, introducing irrigated agriculture in a low rainfall region can increase SOC beyond the natural value due to increased plant biomass production. For reasons of sustainability and soil health it is important to maintain SOC as high as practically possible in arable soils, but we conclude that in the vast majority of situations it is unrealistic to expect to maintain pre-clearance values. To maintain global SOC stocks at we consider it is more important to reduce current rates of land clearance and sustainably produce necessary food on existing agricultural land.

In this paper we describe the GCTE global Soil Organic Matter Network (SOMNET) before focusing on the European network of long-term sites. We then select two examples from the European site network and demonstrate how such data can be used to (a) track long-term changes in soil organic matter, (b) evaluate and compare soil organic matter models, and (c) make rough estimates of the potential for carbon (C) sequestration in soils at the regional (European) level. Our simple calculations based on only two long-term experiments suggest that amendment of arable soils with 10 Mg ha −1 of organic manure could lead to an increase in current total European soil C stock to 30 cm of about 4.8% over 90 yr, a scenario with limited potential for sequestering C. Similarly, afforestation through natural woodland regeneration of 30% of current arable land (surplus to requirements by 2010) could lead to an increase in current total European soil C stock of 12.4% over 100 yr. This is equivalent to 43 Tg C yr −1 or 3.8% of anthropogenic CO 2 –C emissions from Europe. If temporary C storage in standing woody biomass is included in the estimate, the amount of C sequestered is quadrupled and could account for 15.28% of Europe's annual CO 2 –C emissions. This is equivalent to 2.8% of annual global anthropogenic CO 2 –C emissions. These calculations are presented to demonstrate a simple technique for estimating rough C sequestration potentials but they do suggest some potential to sequester C in European agricultural soils. As a result, a more sophisticated approach using statistical relationships derived from a large number of long-term experiments was developed. The need for balancing the effects of these scenarios on soil C against other environmental considerations is discussed. Methods for improving estimates of the potential for soil C sequestration using the European site network are also discussed.

Aims Predictions of medium to long-term changes in soil organic carbon (SOC) stocks involve simulation models with plant carbon (C) inputs estimated from crop yields using crop harvest indices and allometric scaling involving fixed root-to-shoot ratios. The choice of allometric approach and initial distribution of SOC among model pools is crucial for model estimates. Methods We evaluated the impact of a modified allometric approach for grass leys and initial model set-up using the C-TOOL model. Four European agricultural long-term experiments were considered including three experiments with different frequency of grass leys in the crop rotation and one permanent grassland. Results The modified allometric function for ley relied on fixed below-ground C input regardless of mineral fertilizer inputs and modified pool initialization involved available site history. Including available, but insufficient, pre-experiment history to adjust the initial set-up of model SOC pools did not improve to the C-TOOL simulations. Changing the allometric approach for ley from fixed shoot-to-root ratios to fixed below-ground C input decreased the soil C input dramatically and improved the C-TOOL simulation of SOC stocks for fertilized treatments in all experiments when combined with standard model set-up. For unfertilized treatments, however, the efficiency of the standard allometric function was superior to the modified one. Conclusions Our results challenge the prevalent use of allometric approaches based on fixed root-to-shoot in SOC simulation models. The potential of modifying the initial distribution of SOC among model pools by accounting for historic land use remains in the dark.

Cumulative crop recovery of synthetic fertiliser nitrogen (N) over several cropping seasons (legacy effect) generally receives limited attention. The increment in crop N uptake after the first-season uptake from fertiliser can be expressed as a fraction (∆RE) of the annual N application rate. This study aims to quantify ∆RE using data from nine long-term experiments (LTEs). As such, ∆RE is the difference between first season (RE1st) and long-term (RELT) recovery of synthetic fertiliser N. In this study, RE1st was assessed either by the 15N isotope method or by a zero-N subplot freshly superimposed on a long-term fertilised LTE treatment plot. RELT was calculated by comparing N uptake in the total aboveground crop biomass between a long-term fertilised and long-term control (zero-N) treatment. Using a mixed linear effect model, the effects of climate, crop type, experiment duration, average N rate, and soil clay content on ∆RE were evaluated. Because the experimental setup required for the calculation of ∆RE is relatively rare, only nine suitable LTEs were found. Across these nine LTEs in Europe and North America, the mean ∆RE was 24.4% (±12.0%, 95% CI) of annual N application, with higher values for winter wheat than for maize. This result shows that fertiliser-N retained in the soil and stubble may contribute substantially to crop N uptake in subsequent years. Our results suggest that an initial recovery of 43.8% (±11%, 95% CI) of N application may increase to around 66.0% (±15%, 95% CI) on average over time. Furthermore, we found that ∆RE was not clearly related to long-term changes in topsoil total N stock. Our findings show that the—often used—first-year recovery of synthetic fertiliser N application does not express the full effect of fertiliser application on crop nutrition. The fertiliser contribution to soil N supply should be accounted for when exploring future scenarios on N cycling, including crop N requirements and N balance schemes. Highlights: Nine long-term cereal experiments in Europe and USA were analysed for long-term crop N recovery of synthetic N fertiliser. On average, and with application rates between 34 and 269 kg N/ha, crop N recovery increased from 43.8% in the first season to 66.0% in the long term. Delta recovery was larger for winter wheat than maize. Observed increases in crop N uptake were not explained by proportionate increases in topsoil total N stock.

An effective fertilizer recommendation system requires information on seasonal, soil-related and cultural variations in soil mineral nitrogen (N) and nutrient requirements of the crop. This can be provided by dynamic N turnover models, such as listed by Plentinger & Penning De Vries (1996). In this paper, we describe a survey of farmer opinion designed to ascertain what farmers want from such a decision support system. Over 100 farmers were surveyed. Surveyed farmers requested that default values be available for all model inputs. Inputs should be entered both by windows-based menu (for clarity) and tabular format (for speed), have user-selected units, and be fully supported by context-sensitive help. The system should have a hierarchical structure allowing access to fixed parameters, and be compatible with commonly used farm recording packages. Recommendations should be provided both for the field (single and optional application rates), and in tabular format across the whole farm. Simulations should be easily rerun using more recent crop and weather data. Turnover processes underlying recommendations should be illustrated by flow diagrams of flux between pools, pie charts of fertilizer fate, bar charts of movement down the soil profile and graphical plots of changes in N status against time.

SUMMARYThe Broadbalk Wheat Experiment at Rothamsted (UK) includes plots given the same annual applications of inorganic nitrogen (N) fertilizer each year since 1852 (48, 96 and 144 kg N/ha, termed N1 N2 and N3 respectively). These very long-term N treatments have increased total soil N content, relative to the plot never receiving fertilizer N (N0), due to the greater return of organic N to the soil in roots, root exudates, stubble, etc (the straw is not incorporated). The application of 144 kg N/ha for 135 years has increased total soil N content by 21%, or 570 kg/ha (0–23 cm). Other plots given smaller applications of N for the same time show smaller increases; these differences were established within 30 years. Increases in total soil N content have been detected after 20 years in the plot given 192 kg N/ha since 1968 (N4).There was a proportionally greater increase in N mineralization. Crop uptake of mineralized N was typically 12–30 kg N/ha greater from the N3 and N4 treatments than the uptake of c. 30 kg N/ha from the N0 treatment. Results from laboratory incubations show the importance of recently added residues (roots, stubble, etc) on N mineralization. In short-term (2–3 week) incubations, with soil sampled at harvest, N mineralization was up to 60% greater from the N3 treatment than from N0. In long-term incubations, or in soil without recently added residues, differences between long-term fertilizer treatments were much less marked. Inputs of organic N to the soil from weeds (principally Equisetum arvense L.) to the N0–N2 plots over the last few years may have partially obscured any underlying differences in mineralization.The long-term fertilizer treatments appeared to have had no effect on soil microbial biomass N or carbon (C) content, but have increased the specific mineralization rate of the biomass (defined as N mineralized per unit of biomass).Greater N mineralization will also increase losses of N from the system, via leaching and gaseous emissions. In December 1988 the N3 and N4 plots contained respectively 14 and 23 kg/ha more inorganic N in the profile (0–100 cm) than the N0 plot, due to greater N mineralization. These small differences are important as it only requires 23 kg N/ha to be leached from Broadbalk to increase the nitrate concentration of percolating water above the 1980 EC Drinking Water Quality Directive limit of 11·3mgN/l.The use of fertilizer N has increased N mineralization due to the build-up of soil organic N. In addition, much of the organic N in Broadbalk topsoil is now derived from fertilizer N. A computer model of N mineralization on Broadbalk estimated that after applying 144 kg N/ha for 140 years, up to half of the N mineralized each year was originally derived from fertilizer N.In the short-term, the amount of fertilizer N applied usually has little direct effect on losses of N over winter. In most years little fertilizer-derived N remains in Broadbalk soil in inorganic form at harvest from applications of up to 192 kg N/ha. However, in two very dry years (1989 and 1990) large inorganic N residues remained at harvest where 144 and 192 kg N/ha had been applied, even though the crop continued to respond to fertilizer N, up to at least 240 kg N/ha.

One of the main options for carbon mitigation identified by the IPCC is the sequestration of carbon in soils. In this paper we use statistical relationships derived from European long-term experiments to explore the potential for carbon sequestration in soils in the European Union. We examine five scenarios, namely (a) the amendment of arable soils with animal manure, (b) the amendment of arable soils with sewage sludge, (c) the incorporation of cereal straw into the soils in which it was grown, (d) the afforestation of surplus arable land through natural woodland regeneration, and (e) extensification of agriculture through ley-arable farming. Our calculations suggest only limited potential to increase soil carbon stocks over the next century by addition of animal manure, sewage sludge or straw (< 15 Tg C y(-1)), but greater potential through extensification of agriculture (approximate to 40 Tg C y(-1)) or through the afforestation of surplus arable land (approximate to 50 Tg C y(-1)). We estimate that extensification could increase the total soil carbon stock of the European Union by 17%. Afforestation of 30% of present arable land would increase soil carbon stocks by about 8% over a century and would substitute up to 30 Tg C y(-1) of fossil fuel carbon if the wood were used as biofuel. However, even the afforestation scenario, with the greatest potential for carbon mitigation, can sequester only 0.8% of annual global anthropogenic CO2-carbon. Our figures suggest that, although efforts in temperate agriculture can contribute to global carbon mitigation, the potential is small compared to that available through reducing anthropogenic CO2 emissions by halting tropical and sub-tropical deforestation or by reducing fossil fuel burning.

In an effort to mitigate anthropogenic effects on the global climate system, industrialised countries are required to quantify and report, for various economic sectors, the annual emissions of greenhouse gases from their several sources and the absorption of the same in different sinks. These estimates are uncertain, and this uncertainty must be communicated effectively, if government bodies, research scientists or members of the public are to draw sound conclusions. Our interest is in communicating the uncertainty in estimates of greenhouse gas emissions from agriculture to those who might directly use the results from the inventory. We tested six methods of communication. These were: a verbal scale using the IPCC calibrated phrases such as ‘likely’ and ‘very unlikely’; probabilities that emissions are within a defined range of values; confidence intervals for the expected value; histograms; box plots; and shaded arrays that depict the probability density of the uncertain quantity. In a formal trial we used these methods to communicate uncertainty about four specific inferences about greenhouse gas emissions in the UK. Sixty four individuals who use results from the greenhouse gas inventory professionally participated in the trial, and we tested how effectively the uncertainty about these inferences was communicated by means of a questionnaire. Our results showed differences in the efficacy of the methods of communication, and interactions with the nature of the target audience. We found that, although the verbal scale was thought to be a good method of communication it did not convey enough information and was open to misinterpretation. Shaded arrays were similarly criticised for being open to misinterpretation, but proved to give the best impression of uncertainty when participants were asked to interpret results from the greenhouse gas inventory. Box plots were most favoured by our participants largely because they were particularly favoured by those who worked in research or had a stronger mathematical background. We propose a combination of methods should be used to convey uncertainty in emissions and that this combination should be tailored to the professional group., Highlights • We tested 6 different methods for communicating uncertainty in greenhouse gases. • Choice of method of communication should be influenced by professional background. • The IPCC verbal scale has value but should be used alongside more numerate methods. • Boxplots communicate uncertainty effectively to those familiar with this method. • Shaded arrays communicate uncertainty effectively to most individuals.

The widely used ecosystem model DailyDayCent (DDC) was used to predict soil organic carbon (SOC) sequestration and yield under different fertilization treatments on a wheat (Triticum aestivum) monoculture on Broadbalk field, Rothamsted, UK. Over a period of about 170 years, the management practices on the test site, which is separated into plots with different treatments of fertilizer and manure application, are well documented. Four treatments consisting of control (no fertilizer), mineral nitrogen (N) fertilizer (MN), farm yard manure (FYM) and a combination of both (FYMN) have been selected for the study. DDC simulated the greatest increases in SOC (three times higher compared to simulation of the control plot) in the FYM plot. Overall a good agreement between modelled and measured yield was obtained (except in control plots). As observed in the experimental plots, the highest yield predicted by the model is with FYMN plot (more than three times higher relative to the control plot). A sensitivity test for the MN and FYM plots demonstrated that predicted SOC was relatively insensitive (1%) to bulk density, pH and field capacity. Higher sensitivity (4–7%) of predicted SOC changes were observed by changes in external C input. An increase of temperature by 1 °C or precipitation by 10% decreased predicted SOC by 2–4% and vice versa, and simulated a 1–4% variability in yield. DDC is a useful tool for simulating SOC in the long term plots, with a good fit to data, and is useful for examining alternate hypotheses of yield and SOC responses.