Hostname: page-component-7b7567c8c-wxvwp Total loading time: 0 Render date: 2026-11-15T14:17:38.401Z Has data issue: false hasContentIssue false
  • English
  • Français

Quantification of C and N stocks in grassland topsoils in a Dutch region dominated by dairy farming

Published online by Cambridge University Press:  27 July 2010

M. P. W. SONNEVELD and
J. J. H. VAN DEN AKKER
M. P. W. SONNEVELD*
Affiliation:
Land Dynamics Group, Wageningen University, PO Box 47, 6700 AA Wageningen, The Netherlands
J. J. H. VAN DEN AKKER
Affiliation:
Alterra, Wageningen University and Research Centre, PO Box 47, 6700 AA Wageningen, The Netherlands
*
*To whom all correspondence should be addressed. Email: marthijn.sonneveld@wur.nl
Rights & Permissions [Opens in a new window]

Summary

Estimates on soil organic carbon (SOC) and nitrogen (N) stocks in soils cannot be directly calculated from routine soil analyses, since these often lack measurements on soil bulk density (Bd). Hence, flexible pedotransfer functions are required that allow the calculation of SOC stocks from gravimetrically determined SOC contents. The present paper aimed to: (1) quantify SOC and N stocks in grassland topsoils for a Northern Dutch region dominated by dairy farming and (2) analyse the relationships between SOC and bulk density at the field level. As estimates of SOC and N stocks are potentially affected by soil compaction, the combined measurements on soil bulk density and soil organic matter (SOM) were also evaluated with respect to critical limits for soil compaction using soil density (Sd) for sandy soils and packing density (Pd) for clay soils. The SOC and Bd measurements were done in the upper 0·1–0·2 m of grasslands at 18 dairy farms, distributed across sandy, clay and peat soils. Both farm data and grassland management data were collected. Non-linear regressions were used to analyse relationships between Bd and SOM. Significant non-linear relationships were found between gravimetric SOC contents and bulk density for the 0–0·1 m layer (R2=0·80) and the 0·1–0·2 m layer (R2=0·86). None of the fields on sandy soils or clay soils indicated signs for limited rooting in the topsoil although some fields appear to approach the critical limit for compaction for the 0·1–0·2 m layer. Stocks of SOC in the top 0·2 m at farm level were highest in the peat soils (21·7 kg/m2) and lowest in the sandy soils (9·0 kg/m2). Similarly, N stocks were highest for farms on peat soil (1·30 kg/m2) and lowest for farms on sandy soil (0·60 kg/m2). For the sandy soils, the mean SOC stock was significantly higher in fields with shallow groundwater tables.

Information

Type
Crops and Soils
Information
The Journal of Agricultural Science , Volume 149 , Issue 1 , February 2011, pp. 63 - 71
Copyright
Copyright © Cambridge University Press 2010

INTRODUCTION

Drivers and sizes of soil organic carbon (SOC) stocks in agricultural soils have received increased attention over the past decade because of their contribution to the overall global carbon (C) budget (Jones et al. Reference Jones, Hiederer, Rusco and Montanarella2005). Grassland soils, in particular, have the capacity to sequester or to release substantial amounts of C (Freibauer et al. Reference Freibauer, Rounsevell, Smith and Verhagen2004; Soussana et al. Reference Soussana, Loiseau, Vuichard, Ceschia, Balesdent, Chevallier and Arrouays2004; Kätterer et al. Reference Kätterer, Andersson, Andren and Persson2008). In order to calculate actual volumetric SOC stocks from gravimetrically determined SOC contents, it is critical to know the bulk density of the soil (Batjes Reference Batjes1996; Lettens et al. Reference Lettens, Van Orshoven, Van Wesemael, De Vos and Muys2005; Frogbrook et al. Reference Frogbrook, Bell, Bradley, Evans, Lark, Reynolds, Smith and Towers2009). Because gravimetric SOC contents are often negatively correlated with soil bulk density (De Vos et al. Reference De Vos, Van Meirvenne, Quataert, Deckers and Muys2005), changes in SOC contents cannot be directly translated into changes in SOC stocks in grassland soils.

SOC contents in grassland topsoils are related to land use history (Hoogerkamp Reference Hoogerkamp1984; Sonneveld et al. Reference Sonneveld, Bouma and Veldkamp2002; Schulp & Veldkamp Reference Schulp and Veldkamp2008), soil and landscape properties (Landi et al. Reference Landi, Mermut and Anderson2004) and grassland management. Higher SOC levels are, for example, found in grazed pastures compared with pastures receiving only mowing regimes (Hassink & Neeteson Reference Hassink and Neeteson1991; Hassink Reference Hassink1994; Mestdagh et al. Reference Mestdagh, Lootens, Van Cleemput and Carlier2006). Additionally, soil bulk density in grassland topsoils can also be directly related to grassland management such as stocking rates and use of machinery (Wopereis Reference Wopereis1994; Hansen Reference Hansen1996; Hamza & Anderson Reference Hamza and Anderson2005). Thus, the quantification of changes in SOC stocks in grasslands needs to take into account both SOC content as well as bulk density.

Routine soil analyses from farmers' fields provide an opportunity to monitor SOC development in agricultural topsoils (Reijneveld et al. Reference Reijneveld, Van Wensum and Oenema2009), but soil bulk density is mostly not determined in these analyses (Batjes Reference Batjes1996). In order to calculate SOC stocks from SOC contents, use has been made of representative data (Frogbrook et al. Reference Frogbrook, Bell, Bradley, Evans, Lark, Reynolds, Smith and Towers2009) or of continuous pedotransfer functions (Rawls Reference Rawls1983; Batjes Reference Batjes1996; Springob et al. Reference Springob, Brinkmann, Engel, Kirchman and Bottcher2001). Existing functions, however, need to be validated for specific studies, since pedotransfer functions for estimating soil bulk density may result in serious underestimations (De Vos et al. Reference De Vos, Van Meirvenne, Quataert, Deckers and Muys2005). For the Netherlands, the continuous pedotransfer functions developed are restricted to mineral soils and need additional data such as the median sand particle size, which are also not determined in routine analyses (Wösten Reference Wösten, Gregorich and Carter1997). More flexible pedotransfer functions are therefore not only needed to quantify SOC but also nitrogen (N) stocks in grassland soils based on data from routine soil analyses at farmers' fields.

The aim of the present paper was to quantify SOC and N stocks in grassland soils for a Northern Dutch region dominated by dairy farming. As the current international standard depth for SOC analyses for grasslands is 0·1 m, which is also adopted by Dutch standard laboratories (Hanegraaf et al. Reference Hanegraaf, Hoffland, Kuikman and Brussaard2009), the analysis is limited to the upper 0·1–0·2 m of the soil. The second aim was to analyse the relationships between soil organic matter (SOM) and bulk density at the field level for this area. The combined measurements on soil bulk density and SOM are also evaluated with respect to soil compaction as this affects the quantification of SOC and N stocks in soils.

MATERIALS AND METHODS

Study area

The study area is located in the north of the Netherlands in the province of Friesland and covers a total area of 60 000 ha (Sonneveld et al. Reference Sonneveld, De Vos, De Vries, Knotters, Kros, Roelsma, Bleeker, Hensen and Frumau2009). The dominant land use is grassland for dairy farming (41 500 ha). Field sizes are generally small, averaging c. 1–2 ha. Sandy soils developed in late-glacial cover sands (FAO: Gleyic Podzols) are dominant and most have an anthropogenic topsoil of 0·3–0·5 m as a result of peat reclamations and historic plaggen applications (plaggen soils have been created by cutting turves of peat from an outfield area, using them as bedding for cattle and then spreading the slurry-soaked bedding on the arable fields as fertilizer) (Sonneveld et al. Reference Sonneveld, Bouma and Veldkamp2002). Soils developed in peat deposits (FAO: Terric Histosols) and marine clay sediments (FAO: Calcaric Fluvisols) are also found, but mainly in the western and northern parts of the study area (Fig. 1). Data on soils and groundwater tables were available in GIS format (1:50 000 scale; Stiboka 1981). No significant spatial trends in climate are expected for this region of the Netherlands.

Fig. 1.

Study area in the North of the Netherlands, distribution of major soil types and location of the dairy farms in this study (indicated with dots).

Selected farms

In 2007, about 60 dairy farms were monitored with respect to their nutrient management. These participated in a large research project where relationships between farm management and environmental quality were investigated. Using the 1:50 000 scale soil map (Stiboka 1981), 18 dairy farms were randomly selected from this larger group, of which six were located on sandy soils, six on clay soils and six on peat soils (Fig. 1). For all soil types, half of the group of farms applied animal slurry through above-ground spreading, whereas the other farms used conventional techniques to apply slurry to the grassland, such as shallow injection. It was later found that for one farm mapped as a farm on peat soil, samples had actually been taken from a sandy soil. This is most likely caused by the fact that this farm is located on the transition from sandy soils to peat soils and peat may have disappeared through mineralization after the 1:50 000 soil map was constructed in 1981 (Kempen et al. Reference Kempen, Brus, Heuvelink and Stoorvogel2009). Therefore, this farm was reclassified as a farm on sandy soil and was treated as such in the statistical analyses. Besides differences in manure application techniques, farms also varied in size, milk production and animal density. Characteristics for the farms are given in Table 1.

Table 1.

Mean values and standard errors (s.e.m.) for farm characteristics stratified by the soil type

On average, dairy farms on sandy soils are largest in size and animal density, but considerable variation was found. Cultivation of silage maize was only found for farms on sandy soils and was not found for farms on clay soils or peat soils. Mean sizes for all farm groups were higher compared with the average size (44 ha) of a dairy farm in the Netherlands in 2007 (LEI 2008). For the whole country, the proportion of maize on dairy farms is higher (0·15) than that observed in the study area (0·07).

Soil sampling

For each farm, topographical maps were available in GIS which indicated the location of the individual fields. The fields were numbered and three grassland fields were randomly selected at each farm for field sampling. Sampling was carried out in spring 2007 and followed a systematic w-shaped diagonal transect. For each field, bulk samples were taken for SOM and total N from 25 selected points along this transect. Bulk samples were taken from the 0–0·1 m and 0·1–0·2 m layers. The SOM contents and total N were determined using near infrared spectrometry; SOC contents were calculated from this assuming a C content in soil organic matter of 0·58 (Hanegraaf et al. Reference Hanegraaf, Hoffland, Kuikman and Brussaard2009). The clay content (fraction <2 μm) was determined using the pipette method. Clay contents are generally very low for the sandy soils (c. 4% by volume) and were only determined for marine clay soils. Through standardized queries, grassland management on the fields in the preceding year was recorded, distinguishing between only mowing, only grazing or a combination of the two. Fields were also classified as ‘grazing’ if only the latest cut was mown. Within each field, core samples (0·1 litre) were taken from two depths (0·05–0·10 m and 0·15–0·20 m) at three locations on the sampling transect. In total, 18 core samples were collected for each farm, resulting in a total of 324 samples. Samples were weighed and dried at 105°C. For each field and depth, the average bulk density was calculated from the individual samples.

Analyses

The relationships between SOM and bulk density were analysed in two ways: first by using non-linear regression and secondly by analysing the relationship with soil density (S d) and packing density (P d). Soil density is defined following Rawls (Reference Rawls1983) and Van den Akker (Reference Van Den Akker, Horn, Fleige, Peth and Peng2006):

(1)
$$S_{d,\,j} = {{100} \over {\left\{ {{{{\rm{SOM}}_j /(d_s ) + (100 - {\rm{SOM}}_j )} \mathord{\left/ {\vphantom {{{\rm{SOM}}_j /(d_s ) + (100 - {\rm{SOM}}_j )} {(d_m )}}} \right. \kern-\nulldelimiterspace} {(d_m )}}} \right\}}} $$

where S d,j is the soil density (kg/m3) for layer j when saturated, and SOMj is the gravimetric SOM content for layer j. The constants d s and d m refer to the material density of SOM (1470 kg/m3) and the density of the mineral fraction (2660 kg/m3), respectively (Van den Akker Reference Van Den Akker, Horn, Fleige, Peth and Peng2006).

Using eqn (1), the total pore (TP) volume can also be calculated as

(2)
$${\rm{TP}}_j = 1 - (B_{d,\,j} /S_{d,\,j} ) $$

where TPj=total pore volume for layer j and B d,j=bulk density for layer j (kg/m3). For sandy and loamy soils (clay content <17·5%), Van den Akker (Reference Van Den Akker, Horn, Fleige, Peth and Peng2006) indicates that limitations for root growth occur if TP drops below a critical threshold value of 0·40. From rewriting eqn (2), it follows that the ratio of B d and S d should be smaller than 0·60 if compaction is absent. For clay soils in the Netherlands (clay content >17·5% by volume), a critical TP value of 0·40 cannot be used as a criterion for rooting (Van den Akker Reference Van Den Akker, Horn, Fleige, Peth and Peng2006). To also allow an evaluation of compaction within clay soils, packing densities were calculated. Following (Jones et al. Reference Jones, Spoor and Thomasson2003), P d can be calculated as

(3)
$$P_{d,\,j} = B_{d,\,j} + 9 \times C_j $$

where P d,j is the packing density for layer j (kg/m3) and C j is the clay content for layer j. As a threshold, a critical limit of PD of 1750 kg/m3 is adopted here (Van den Akker Reference Van Den Akker, Horn, Fleige, Peth and Peng2006).

Estimated stocks of SOC in the top 0·2 m at field and the farm level were calculated according to eqns (4) and (5), respectively:

(4)
$${\rm{SOC}}_{{\rm{\,fld}}} = \sum\limits_j^2 {B_{d,\,j} \times {\rm{SOC}}_j } $$
(5)
$${\rm{SOC}}_{{\rm{frm}}} = {1 \over 3}\sum\limits_i^3 {\sum\limits_j^2 {B_{d,\,j} } } \times {\rm{SOC}}_j $$

where SOCfld is the estimated total SOC content in the upper 0·2 m of grassland soils at the field level and SOCfrm is the estimated total SOC content in the upper 0·2 m of grassland soils at the farm level based on individual samples per field i. Similar to eqns (4) and (5), total N stocks at field and farm levels were also calculated. All statistical analyses were performed in SPSS. Significant differences between groups were tested using a T-test.

RESULTS

SOM and bulk density

Results for gravimetrically determined SOM contents and dry bulk density for both the 0–0·1 m layer and the 0–0·2 m are given in Table 2. As expected, average SOM values of peat soils were highest and SOM values of sandy soils were lowest for both the 0–0·1 m layer as well as the 0–0·2 m layer. The range in SOM contents was considerable for all soils but especially for the peat soils. Using the SOM data, the proportion of SOM content at 0·1–0·2 m depth with respect to the 0–0·1 m depth was also calculated. These relative proportions were 0·94, 0·78 and 0·66 for peat, sandy and clay soils, respectively. This indicates that the relative distribution of soil organic matter within the soil profile is different for the different soil types. The largest difference found between 0–0·1 m and 0·1–0·2 m for the clay soils may be due to less biological homogenization as a result of shallow groundwater tables and swelling and shrinking of the soil. In general, average bulk densities from sandy soils are highest and average bulk densities from peat soils are lowest. It can also be observed that bulk densities for the 0·1–0·2 m layer were always higher compared with the 0–0·1 m layer. For the sandy soils and clay soils, standard errors (s.e.m.) were, on an average, also lower for the 0·1–0·2 m layer compared to s.e.m. values from the top 0–0·1 m. As with the SOM contents, the range in B d values for the peat soils were largest.

Table 2.

Gravimetric SOM contents and bulk density among the different soil types for the 0–0·1 m layer and the 0·1–0·2 m layer at the field level

For the marine clay soils, the fine earth fraction (<2 um) proved to be uncorrelated with SOM for both the 0–0·1 m layer as well as the 0·1–0·2 m layer. As indicated, SOC contents are calculated from SOM contents assuming a C content of 0·58. In Figs 2 and 3, the data on bulk density, averaged per field, are plotted against SOC observations at field level for sand, clay and peat soils. From regression analysis, significant non-linear relationships were found between bulk density and SOC for both the 0–0·1 m layer (R 2=0·80, P <0·001) and the 0·1–0·2 m layer (R 2=0·86, P<0·001). This relationship could be described by

(6)
$$B_d = a - b \times {\rm{Ln}}({\rm{SOC}}) $$

where a=1622 with a 95% confidence interval [1524; 1719] and b=340 with a 95% confidence interval [292; 388] for the top 0–0·1 m layer, and a=1687 with a 95% confidence interval [1617; 1757] and b=339 with a 95% confidence interval [301; 378] for the 0·1–0·2 m layer.

Fig. 2.

Bulk density in relation to SOCfld content for 0–0·1 m depth for sand, peat and clay soils.

Fig. 3.

Bulk density in relation to SOCfld content for 0·1–0·2 m depth for sandy, peat and clay soils.

The relationship between bulk density (B d) and soil density (S d) for sandy soils at depths of 0–0·1 m and 0·1–0·2 m is given in Fig. 4. As can be seen in Fig. 4, none of the fields on sandy soils show B d/S d ratios higher than 0·6. Thus, limitations for root growth as a result of compaction are not expected in these fields. However, some grassland fields on sandy soil do approach this critical threshold for the 0·1–0·2 m layer and are close to being defined as compacted. For fields on sandy soils, the mean pore volume (TP) was 0·55 for the 0–0·1 m layer and 0·47 for the 0·1–0·2 m layer. For both layers, there were no significant differences in B d/S d ratios between farms that applied manure through above-ground spreading and farms that used low-emission techniques. The relationships between B d and P d for clay soils for both layers are given in Fig. 5. Also for the clay soils, there are no signs of compaction as the critical P d threshold of 1750 kg/m3 is not exceeded. Again, some fields approach this threshold value at greater depth, but limitations for rooting are not to be expected. Thus, estimations of SOC stocks and N stocks in the 0–0·1 m layer and 0·1–0·2 m layer of grasslands are not affected by strong levels of soil compaction in the study area.

Fig. 4.

Relationship between S d and B d for 0–0·1 m depth and 0·1–0·2 m depth for individual grassland fields for sandy soils (clay % by volume <8). The threshold of 0·6 is indicated by the dashed line.

Fig. 5.

Relationship between P d and B d for 0–0·1 m depth and 0·1–0·2 m depth for individual grassland fields for clay soils (clay % by volume >17·5). The threshold is indicated by the dashed line.

SOC and total N stocks at field and farm level

Most fields on sandy soils (0·81 of all sandy fields) were used for a combination of mowing and grazing. SOCfld stocks were higher (P=0·058) for fields that received only a grazing regime (SOCfld=10·3 kg/m2) compared with fields that received a combination of grazing and mowing (SOCfld=8·7 kg/m2). Also total N stocks were higher (P=0·068) for fields with pure grazing regimes (Nfld=0·39 kg/m2) compared with fields that received a combination of grazing and mowing (Nfld=0·31 kg/m2). For clay and peat soils, no differences between management regimes were observed. For the sandy soils, the effect of hydrological regime proved to be even more significant compared with field management (Table 3). Stocks of SOCfld under fields with very shallow groundwater tables (‘wet soils’, Dutch groundwater class III) were significantly higher (11·4 kg/m2, P=0·003) compared with soils that are better drained (‘dry soils’, Dutch groundwater class VI) which had a mean SOCfld stock in the upper 0·2 m of 8·6 kg/m2. The mean difference between total Nfld stocks were also higher for these wet soils (Nfld=0·71 kg/m2, P=0·081) compared with drier soils (Nfld= 0·57 kg/m2).

Table 3.

SOCfld stocks and total Nfld stocks in the upper 0·2 m of grasslands on sandy soils with different hydrological regimes at the field level. Groundwater classes, following the Dutch classification scheme, are given between brackets. Standard error of differences (s.e.d.) and P values are indicated

As with gravimetric SOM contents, estimated SOCfrm stocks in the top 0·2 m were highest for the peat soils and lowest for the sandy soils (Table 4). The maximum SOCfrm stock for the upper 0·2 m of grasslands for farms on sandy soil was similar to the lowest level found for farms on clay soils. Estimated N stocks in the upper 0·2 m in grassland for farms on peat soils were more than twice as high compared with estimated N stocks in the upper 0·2 m in grassland for farms on sandy soils.

Table 4.

SOC stocks and N stocks at farm level in the top 0·2 m for the different soil types

Comparing the SOC stocks and N stocks between farms in relation to manure application technique did not yield significant differences. Although farms on sandy soils using above-ground spreading of manure had both a higher SOCfrm stock (1·28 kg/m2) as well as a higher Nfrm stock (726 kg N/ha) these were not significantly different from farms which used low-emission techniques to apply manure (P=0·20 and P=0·34 for SOCfrm and Nfrm, respectively).

DISCUSSION

For the area studied, significant non-linear relationships between SOC content and B d were found, which may be used as pedotransfer functions to calculate SOC stocks from SOM data that originate from routine soil analyses. The established relationships closely match the relationship found by Springob et al. (Reference Springob, Brinkmann, Engel, Kirchman and Bottcher2001) for Ap horizons (i.e. soil surfaces disturbed by human activity through, e.g. ploughing) in North Pleistocene sands in Germany. Such pedotransfer functions are appealing because of the low volume of data needed for translating SOC contents into SOC stocks. It has not been validated to what extent these functions also hold for other parts of the country. In the present work, the established pedotransfer functions were developed for the topsoil of grassland, but these may perform differently with respect to the subsoil (De Vos et al. Reference De Vos, Van Meirvenne, Quataert, Deckers and Muys2005).

Grasslands on sandy soils in the northern part of the Netherlands generally have higher SOM contents compared with the southern parts. Relatively high SOM contents were found in the present study for sandy topsoils, in agreement with earlier findings (Sonneveld et al. Reference Sonneveld, Bouma and Veldkamp2002; Sonneveld Reference Sonneveld2008). Hanegraaf et al. (Reference Hanegraaf, Hoffland, Kuikman and Brussaard2009) reported a mean SOM content of 9·4 g/100 g for the top 0·05 m of grassland on sandy soils in a Northern province in 2003 and 5 g/100 g for the top 0·05 m for grassland on sandy soils in the southern province of Brabant.

Increases in SOC stocks can potentially be observed as a sole result of increases in bulk density. As this is the result of using fixed sampling depths (Gifford & Roderick Reference Gifford and Roderick2003), it may be better to express SOC contents on a constant soil dry mass per unit of area. This would additionally require assessments of SOC deeper in the soil profile (Lettens et al. Reference Lettens, Van Orshoven, Van Wesemael, De Vos and Muys2005). However, the present work was associated with routine soil analyses and these generally use a fixed depth for sampling. Calculations of elevated SOC stocks can, for example, be the result of soil compaction. Significant differences in bulk density for the top 0·2 m in grasslands as a result of soil management have been reported in several studies (Wopereis Reference Wopereis1994; Hansen Reference Hansen1996; Mestdagh et al. Reference Mestdagh, Lootens, Van Cleemput and Carlier2006). In grasslands, the application of manure is a likely potential contributor for soil compaction, because this often takes place early in the growing season under relatively wet conditions. High levels of soil compaction in grassland topsoils may even result in a reduction of dry matter yield (Douglas et al. Reference Douglas, Campbell and Crawford1992; Douglas & Crawford Reference Douglas and Crawford1998). In the present study, samples for the topsoil of grassland on sandy soils (0–0·2 m) did not indicate compaction as B d/S d ratios were always below 0·6. Also, samples for the topsoil of grassland on clay soils (0–0·2 m) did not indicate compaction, since P d values were always below 1750 kg/m3. Although grassland management may affect the observed density of the soils, this was not to the extent that critical thresholds were exceeded. In reality, the limitation for rooting is not dependent on a fixed value. On sandy soils, other crops such as maize or bulbous plants have limitations at B d/S d ratios smaller than 0·6 (Van den Akker Reference Van Den Akker, Horn, Fleige, Peth and Peng2006).

For the sandy soils, soil hydrology was significant in explaining part of the variation in SOC stocks observed at field level, similar to what has been observed elsewhere (Schulp & Veldkamp Reference Schulp and Veldkamp2008). Shallow groundwater levels in wet soils contribute to lower decomposition rates compared with drier sandy soils (Springob et al. Reference Springob, Brinkmann, Engel, Kirchman and Bottcher2001). Because groundwater levels have artificially been lowered in most parts of the landscape, high SOM contents in cultivated grasslands on relatively wet soils may also reflect locations which were previously covered with peat.

It was also found that most types of grassland management consisted of a combination of grazing and mowing. Still, the present data suggest that only grazing regimes result in higher SOC stocks for sandy soils compared with a combination of grazing and mowing. This supports earlier findings by Mestdagh et al. (Reference Mestdagh, Lootens, Van Cleemput and Carlier2006) and Hassink & Neeteson (Reference Hassink and Neeteson1991) and it has been suggested that pure grazing regimes give higher inputs of ungrazed grass returns and faeces to the field resulting in higher SOC stocks in fields that are used for grazing.

The conclusion of the current work is that significant non-linear relationships exist between gravimetric SOC contents and bulk density (B d) for the 0–0·1 m layer (R 2=0·80, P<0·001) and the 0·1–0·2 m layer (R 2=0·86, P<0·001). These relationships can be used as pedotransfer functions to calculate SOC and N stocks in the upper 0·2 m of grasslands using data available from routine soil analyses. Estimations of SOC and total N stocks for the top 0–0·2 m were not affected by strong levels of soil compaction although some fields on sandy soils and clay soils were found to approach a critical limit for the 0·1–0·2 m layer. Mean SOC stocks in the top 0·2 m at farm level were highest in the peat soils (21·7 kg/m2) and lowest in the sandy soils (9·0 kg/m2). Similarly, N stocks were highest for farms on peat soil (1·30 kg/m2) and lowest for farms on sandy soil (0·60 kg/m2). For the sandy soils, the mean SOC stock was significantly higher in fields with shallow groundwater tables.

Thanks go to Frans Bakker for his contribution in data collection in the field.

References

REFERENCES

Batjes, N. H. (1996). Total carbon and nitrogen in the soils of the world. European Journal of Soil Science 47, 151–163.CrossRefGoogle Scholar
De Vos, B., Van Meirvenne, M., Quataert, E., Deckers, J. & Muys, B. (2005). Predictive quality of pedotransfer functions for estimating bulk density of forest soils. Soil Science Society of America Journal 69, 500–510.CrossRefGoogle Scholar
Douglas, J. T. & Crawford, C. E. (1998). Soil compaction effects on utilization of nitrogen from livestock slurry applied to grassland. Grass and Forage Science 53, 31–40.CrossRefGoogle Scholar
Douglas, J. T., Campbell, D. J. & Crawford, C. E. (1992). Soil and crop responses to conventional, reduced ground pressure and zero traffic systems for grass silage production. Soil and Tillage Research 24, 421–439.CrossRefGoogle Scholar
Freibauer, A., Rounsevell, M. D. A., Smith, P. & Verhagen, J. (2004). Carbon sequestration in the agricultural soils of Europe. Geoderma 122, 1–23.CrossRefGoogle Scholar
Frogbrook, Z. L., Bell, J., Bradley, R. I., Evans, C., Lark, R. M., Reynolds, B., Smith, P. & Towers, W. (2009). Quantifying terrestrial carbon stocks: examining the spatial variation in two upland areas in the UK and a comparison to mapped estimates of soil carbon. Soil Use and Management 25, 320–332.CrossRefGoogle Scholar
Gifford, R. M. & Roderick, M. L. (2003). Soil carbon stocks and bulk density: Spatial or cumulative mass coordinates as a basis of expression? Global Change Biology 9, 1507–1514.CrossRefGoogle Scholar
Hamza, M. A. & Anderson, W. K. (2005). Soil compaction in cropping systems: A review of the nature, causes and possible solutions. Soil and Tillage Research 82, 121–145.CrossRefGoogle Scholar
Hanegraaf, M. C., Hoffland, E., Kuikman, P. J. & Brussaard, L. (2009). Trends in soil organic matter contents in Dutch grasslands and maize fields on sandy soils. European Journal of Soil Science 60, 213–222.CrossRefGoogle Scholar
Hansen, S. (1996). Effects of manure treatment and soil compaction on plant production of a dairy farm system converting to organic farming practice. Agriculture, Ecosystems and Environment 56, 173–186.CrossRefGoogle Scholar
Hassink, J. (1994). Effects of soil texture and grassland management on soil organic C and N and rates of C and N mineralization. Soil Biology and Biochemistry 26, 1221–1231.CrossRefGoogle Scholar
Hassink, J. & Neeteson, J. J. (1991). Effect of grassland management on the amounts of soil organic N and C. Netherlands Journal of Agricultural Science 39, 225–236.CrossRefGoogle Scholar
Hoogerkamp, M. (1984). Changes in productivity of grassland with ageing. PhD thesis, Wageningen University.Google Scholar
Jones, R. J. A., Spoor, G. & Thomasson, A. J. (2003). Vulnerability of subsoils in Europe to compaction: a preliminary analysis. Soil and Tillage Research 73, 131–141.CrossRefGoogle Scholar
Jones, R. J. A., Hiederer, R., Rusco, E. & Montanarella, L. (2005). Estimating organic carbon in the soils of Europe for policy support. European Journal of Soil Science 56, 655–671.CrossRefGoogle Scholar
Kätterer, T., Andersson, L., Andren, O. & Persson, J. (2008). Long-term impact of chronosequential land use change on soil carbon stocks on a Swedish farm. Nutrient Cycling in Agroecosystems 81, 145–155.CrossRefGoogle Scholar
Kempen, B., Brus, D. J., Heuvelink, G. B. M. & Stoorvogel, J. J. (2009). Updating the 1:50,000 Dutch soil map using legacy soil data: A multinomial logistic regression approach. Geoderma 151, 311–326.CrossRefGoogle Scholar
Landi, A., Mermut, A. R. & Anderson, D. W. (2004). Carbon distribution in a hummocky landscape from Saskatchewan, Canada. Soil Science Society of America Journal 68, 175–184.CrossRefGoogle Scholar
LEI (2008). Land-en Tuinbouwcijfers 2008. Gravenhage, The Netherlands: LEI.Google Scholar
Lettens, S., Van Orshoven, J., Van Wesemael, B., De Vos, B. & Muys, B. (2005). Stocks and fluxes of soil organic carbon for landscape units in Belgium derived from heterogeneous data sets for 1990 and 2000. Geoderma 127, 11–23.CrossRefGoogle Scholar
Mestdagh, I., Lootens, P., Van Cleemput, O. & Carlier, L. (2006). Variation in organic-carbon concentration and bulk density in Flemish grassland soils. Journal of Plant Nutrition and Soil Science-Zeitschrift Fur Pflanzenernahrung Und Bodenkunde 169, 616–622.CrossRefGoogle Scholar
Rawls, W. J. (1983). Estimating soil bulk density from particle size analysis and organic matter content. Soil Science 135, 123–125.CrossRefGoogle Scholar
Reijneveld, J. A., Van Wensum, J. & Oenema, O. (2009). Soil organic carbon contents of agricultural land in the Netherlands between 1984 and 2004. Geoderma 153, 231–238.CrossRefGoogle Scholar
Schulp, C. J. E. & Veldkamp, A. (2008). Long-term landscape-land use interactions as explaining factor for soil organic matter variability in Dutch agricultural landscapes. Geoderma 146, 457–465.CrossRefGoogle Scholar
Sonneveld, M. P. W. (2008). Soil water repellency in an old and young pasture in relation to N application. Soil Use and Management 24, 310–317.CrossRefGoogle Scholar
Sonneveld, M. P. W., Bouma, J. & Veldkamp, A. (2002). Refining soil survey information for a Dutch soil series using land use history. Soil Use and Management 18, 157–163.CrossRefGoogle Scholar
Sonneveld, M. P. W., De Vos, J. A., De Vries, W., Knotters, M., Kros, J., Roelsma, J., Bleeker, A., Hensen, A. & Frumau, A. (2009). 3MG: Meervoudige Milieumonitoring voor Gebiedssturing. Een Case Study voor de Noordelijke Friese Wouden. Working Paper 9. Zoetermeer: TransForum Agro & Groen.Google Scholar
Soussana, J. F., Loiseau, P., Vuichard, N., Ceschia, E., Balesdent, J., Chevallier, T. & Arrouays, D. (2004). Carbon cycling and sequestration opportunities in temperate grasslands. Soil Use and Management 20, 219–230.CrossRefGoogle Scholar
Springob, G., Brinkmann, S., Engel, N., Kirchman, H. & Bottcher, J. (2001). Organic C levels of Ap horizons in North German Pleistocene sands as influenced by climate, texture, and history of land-use. Journal of Plant Nutrition and Soil Science 164, 681–690.3.0.CO;2-V>CrossRefGoogle Scholar
Stiboka (1981). Bodemkaart van Nederland 1:50 000: toelichting bij de kaartbladen 6 West Leeuwarden 6 Oost Leeuwarden en het vaste land van de kaartbladen 2 West Schiermonnikoog en 2 Oost Schiermonnikoog. p. 181. Wageningen: Stiboka.Google Scholar
Van Den Akker, J. J. H. (2006). Evaluation of soil physical quality of Dutch subsoils in two databases with some threshold values. In Soil Management for Sustainability (Eds Horn, R., Fleige, H., Peth, S. & Peng, X.), pp. 490–497. Advances in GeoEcology 38. Reiskirchen, Germany: Catena Verlag GMBH.Google Scholar
Wopereis, F. A. (1994). Invloed van Bodemverdichting op de Wortelontwikkeling van Grasland op Zandgrond. Rapport 260. Wageningen, The Netherlands: DLO-Staring Centrum.Google Scholar
Wösten, J. H. M. (1997). Pedotransfer functions to evaluate soil quality. In Soil Quality for Crop Production and Ecosystem Health (Eds Gregorich, E. G. & Carter, M. R.), pp. 221–245. Amsterdam: Elsevier.CrossRefGoogle Scholar
Figure 0

Fig. 1.Study area in the North of the Netherlands, distribution of major soil types and location of the dairy farms in this study (indicated with dots).

Figure 1

Table 1.Mean values and standard errors (s.e.m.) for farm characteristics stratified by the soil type

Figure 2

Table 2.Gravimetric SOM contents and bulk density among the different soil types for the 0–0·1 m layer and the 0·1–0·2 m layer at the field level

Figure 3

Fig. 2.Bulk density in relation to SOCfld content for 0–0·1 m depth for sand, peat and clay soils.

Figure 4

Fig. 3.Bulk density in relation to SOCfld content for 0·1–0·2 m depth for sandy, peat and clay soils.

Figure 5

Fig. 4.Relationship between Sd and Bd for 0–0·1 m depth and 0·1–0·2 m depth for individual grassland fields for sandy soils (clay % by volume <8). The threshold of 0·6 is indicated by the dashed line.

Figure 6

Fig. 5.Relationship between Pd and Bd for 0–0·1 m depth and 0·1–0·2 m depth for individual grassland fields for clay soils (clay % by volume >17·5). The threshold is indicated by the dashed line.

Figure 7

Table 3.SOCfld stocks and total Nfld stocks in the upper 0·2 m of grasslands on sandy soils with different hydrological regimes at the field level. Groundwater classes, following the Dutch classification scheme, are given between brackets. Standard error of differences (s.e.d.) and P values are indicated

Figure 8

Table 4.SOC stocks and N stocks at farm level in the top 0·2 m for the different soil types