| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Landcare Research, Private Bag 3127, Hamilton, New Zealand
* Corresponding author (SparlingG{at}LandcareResearch.co.nz)
Received for publication May 22, 2001.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: CEC, cation exchange capacity
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The principle legislative tool for environmental regulation in New Zealand is the Resource Management Act (New Zealand Government, 1991), which is administered by 17 autonomous Regional Authorities. The Resource Management Act requires effects-based criteria to determine whether a particular activity has a detrimental effect on the environment, but does not specify properties to be monitored or interpretative framework (Ministry for the Environment, 1997). To provide incentive for a unified approach to soil quality monitoring, the New Zealand Ministry for the Environment offered Regional Authorities a 60% subsidy toward monitoring costs during 1998–2001. The subsidy required that the participants use common methods and interpretation. Despite this financial incentive, only five of the 17 Regional Authorities participated in 1998–2000, with a commitment from a further five in 2000–2001. This paper reports on the data collected in the first two years of the project, which is popularly known as the 500 Soils Project. The data were supplemented with 56 sites from three other studies using comparable sampling and analytical methods (Schipper and Sparling, 2000; Sparling et al., 2000a, b).
Our data were collected primarily for state of the environment reporting at a regional scale, rather than for production purposes at a farm scale. Thus, our project differs from many in the USA, where much soil quality monitoring and reporting has a production focus (Karlen et al., 1999, 2001). Several large-scale studies in North America included a high proportion of arable farms and intensive land uses. Examples are Wander and Bollero (1999), who compared soil quality of 36 arable farms in Illinois, and Boehm and Anderson (1997), who measured soil quality on 20 farms under three crop management systems in Saskatchewan. Hellkamp et al. (2000) surveyed 293 sites in six U.S. Mid-Atlantic states with 12% of the sites on cropped land. In another large-scale study in the U.S. central and southern High Plains by Brejda et al. (2000b), 49% of the land supported intensive livestock production, and 31% was under some form of arable cropland.
Our survey is on a similar scale to some of the North American studies, but the 222 sites covered a wider range of soil orders and land uses, including arable, horticulture, pastoral, plantation forestry, and indigenous ecosystems. All sites other than indigenous ecosystems were on commercial farms or plantations. Twelve soil chemical and physical properties previously tested for variability and response to soil management (Schipper and Sparling, 2000) were used to characterize aspects of soil quality response to management.
Our objectives for this paper were to compare the soil orders and land-use frequency in the 500 Soils data set with the national distribution of soil orders and land uses, determine the contribution from land use or soil order to the variability in soil properties, provide justification for reducing the number of measured soil properties, and identify any items of concern for long-term sustainable land use.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Soil and Site Selection
Soil and land-use combinations were selected by the Regional Councils based on a combination of the area extent of the soils and land uses within their regions, and we used "targeted" sampling of soil and land use combinations perceived to be a risk to the environment. Soils were classified by soil order, using the New Zealand soil classification system (Hewitt, 1993). Where possible, matched sites on the same soil order but under different land uses were selected to produce a matrix of soil types and land uses (Schipper and Sparling, 2000). Soils were collected from 222 locations mainly from the North Island. The soils included examples from 13 different soil orders (Hewitt, 1993). Land use was divided into nine categories including indigenous forest; plantation forest (mainly Monterey pine, Pinus radiata D. Don); native tussock grassland used for sheep and beef farming; introduced European pasture species for drystock beef, sheep, or deer; intensively managed ryegrass–white clover pasture for dairy cows; orchards–horticulture; mixed arable cropland; and long-term use for intensive vegetable production. Not all land uses occurred on all soil orders (Table 1).
|
|
We took core samples to a depth of 100 mm, but subsurface and subsoil condition below this depth can affect the overall soil quality characteristics of the site. Hence, at the midpoint of each transect a pit was dug and the soil profile fully described down to at least a 1-m depth (Milne et al., 1995). The full depth of A horizon, total potential rooting depth, and the nature and depth of any limiting horizons were specifically recorded. The profile descriptions are not reported here, nor the site characterizations, but we consider that both are necessary to combine with the topsoil chemical and physical data for meaningful soil quality assessment.
Soil Quality Data Set
The number of soil properties measured for soil quality monitoring was smaller than used in our earlier work (Schipper and Sparling, 2000). Chemical properties were total C, total N, cation exchange capacity (exchangeable Ca2+, Mg2+, K+, and Na+), base saturation, Olsen phosphate, and soil pH. Soil physical properties were water retention (at 
s = -5, -10, -100, and -1500 kPa), dry bulk density, and particle density, which allowed calculation of readily and total available water, macroporosity, and total porosity. The only biochemical property measured was anaerobically mineralizable N. Properties previously measured but not included in the present data set were unsaturated hydraulic conductivity, soil respiration, microbial biomass C, and particle size distribution. The rationale for discarding some soil properties from the soil quality measurements was given by Schipper and Sparling (2000). In brief, unsaturated hydraulic conductivity was dropped because high variability meant impractical levels of replication were required to detect significant changes. Particle-size distribution was dropped because, although this is a useful property to identify matched sites, it is not a dynamic soil property responsive to land use. The biological measures of soil microbial biomass and soil respiration were not included because we did not feel able to interpret the data (Carter et al., 1999), and because in many cases there is a reasonable correlation with anaerobically mineralizable N (Hart et al., 1986).
Analyses
Total C and N were determined by dry combustion. For New Zealand soils total C is an acceptable measure of total organic C, as the contribution from carbonates to C content is normally negligible (Metson et al., 1979). The cation exchange capacity was determined by flame spectrometer after leaching of <2 mm air-dry soil with 1 M CH3COONH4 at pH 7.0. Soil pH and concentrations of Ca, Mg, K, Na, NH4, and PO4 were determined with standard methods (Blakemore et al., 1987). Potentially mineralizable N was estimated by the anaerobic (waterlogged) incubation method with a 7-d incubation at 40°C, extraction in 2 M KCl, and measurement by Technicon (Tarrytown, NY) AutoAnalyzer.
Water release was calculated from drainage on pressure plates (Gradwell and Birrell, 1979). Dry bulk density (
b) was measured on a subsample core dried at 105°C (Gradwell and Birrell, 1979), and the remaining soil was analyzed for particle density (p) as described by Claydon (1989). Readily available water and total available water were measured as the difference in volumetric water (
) between -10 kPa and -100 kPa s, and -10 kPa and -1500 kPa s, respectively. The total porosity (St) was calculated from the formula St = 100[1 - (b/p)], where p is the particle density, and macroporosity was calculated from St - at -5 kPa s (Klute, 1986).
Statistics
To allow valid land-use comparisons, gravimetric data were recalculated to a soil volume basis using the dry bulk density (Reganold and Palmer, 1995; Doran and Parkin, 1994). The length of each box shows the range within which the central 50% of the values fall, with the box hinge (borders) at the first and third quantiles; the midpoint line shows the median (Fig. 1–3)
. The whiskers show the range of values that fall within the inner fences. Values between the inner and outer fences are plotted with asterisks. Limits for the inner upper or lower fences were set as hinge minus (1.5 times the upper or lower hinge median, respectively). Those for the upper inner or lower fence were set as the hinge plus (1.5 times the upper or lower hinge median, respectively). Values outside the outer fences are plotted with empty circles and used a similar calculation, but were 3 times rather than 1.5 times the upper or lower hinge median.
|
|
|
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The most frequently sampled land use was pasture for dairy production, followed by pasture for drystock sheep and beef (Table 1). Satellite and aerial photography were not able to distinguish between different uses of pastures for dairy, sheep, and beef, or deer production, and had to be combined to a single category for Table 2. Crop production and horticulture could also not be separated by photography and are combined to a single class. Sixteen percent of the samples in the 500 Soils dataset were from soils used for crop production and horticulture, whereas from aerial and satellite photography this land use was <1% of the current land cover (even when combined). Conversely, low-risk soils are underrepresented; indigenous forest comprised 10% of the samples in the 500 Soils data set despite being 23% of the national vegetation cover (Table 2).
Within the 500 Soils dataset some soil order and land-use combinations are not represented at all (Table 1). This pattern reflects the pragmatic reality of land use within New Zealand, and that the sampling sites were on commercially operated farms rather than research farms with balanced experimental designs. It also reflects the geographic distribution of the participating regions. The strategy of targeted sampling to prioritize at-risk soils is efficient use of monitoring resources, but has led to the over- and underrepresentation of various soil and land use classes when considered on a wider basis. The sites selected provided a good range of soil and land-use comparisons, but caution is needed in aggregating such data into a national overview because of the sampling bias. Such bias can be partially overcome by applying weighting factors based on the area extent of the soil or land use relative to the sampling frequency (Hellkamp et al., 2000).
Land Use and Soil Order
Organic Resources
Total C content was best explained by the soil order (50% of the variability) with a comparatively small (27%) contribution from the land use (Table 3). Total C content (to a 100-mm depth) ranged from 7.8 to 154 Mg ha-1. There were few significant differences in total C between land uses because of high variability. Overall, there was a tendency for total C to be greater under pastures and indigenous vegetation than cropland and plantation forests (Fig. 1a). Land use had a greater effect on total N (43%) with an equivalent contribution from soil order (43%). Total N contents ranged from 0.39 to 10.0 Mg ha-1. Overall, the N content of pastures was greater than of other land uses (Fig. 1b,c). This pattern reflected the substantial organic matter content of permanent pastures, and the build up of organic N through the introduction of exotic legume species, principally white clover (During, 1984). The pattern of organic N accumulation was even more distinct when mineralizable N was considered; the highest values occurred under pastures for dairy and drystock production (Fig. 1d). Values ranged between 0.18 and 380 µg N cm-3, with 45% of the variability explained by land use and 22% by soil order (Table 3). The enhanced N status of the more intensively used soils was shown by the much lower C to N ratio of between 10:1 and 15:1 in managed pastures compared with between 15:1 and 20:1 for the plantation forests, tussock grassland, and indigenous forest (Fig. 1c). The low C to N ratio of managed pasture soils in New Zealand is a consequence of the widespread use of long-term productive ryegrass–white clover leys that maintain high organic matter levels with a high N content (Ledgard and Steele, 1992).
|
The Olsen P status of New Zealand soils showed a large range in values, ranging from undetectable on some forest sites, through to >200 µg cm-3 in a soil used for intensive vegetable production (Fig. 2b). However, a surprisingly small proportion of variability in Olsen P was attributed to land use (28%), with a small effect of soil order (13%). A combination of the two categorical variables explained 57% of the variability (see Table 3). There was a clear trend for Olsen P to be substantially greater under cropland and intensive pastures. Olsen P contents under forestry and tussock grassland were lower, although there were numerous outlier points in the range (Fig. 2b). Many unmodified New Zealand soils have a low available P status (Cornforth and Sinclair, 1984). The occurrence of high Olsen P contents in indigenous ecosystems predominantly occurred at sites adjacent to agricultural land without fencing; stock transfer and fertilizer drift could have contributed to the high P status.
The cation exchange capacity ranged from 4 to 69 cmol cm-3 (Fig. 2c). The greater part of the variation (56%) was explained by the soil order, and only 21% by land use (Table 3). Generally, the cation exchange capacity (CEC) in New Zealand topsoils is highly correlated with the organic matter content (Parfitt et al., 1995). However, amount and types of clay minerals also influence CEC. In the 500 Soils data set, the soils with the highest CEC (Fig. 2c) were not those that had the greatest total C contents. The data set included several soils on the Waiareka clay soil, much used for arable crops, which has exceptionally high CEC because of a large proportion of smectite clay (Sparling et al., 2000a). Inclusion of this particular soil means that the correlation between total C and CEC was not as strong as in the larger data set used by Parfitt et al. (1995). Base saturation ranged from 5 to 100%, being markedly lower under tussock grassland, pines, and indigenous forest than other land uses (Fig. 2d).
Physical Characteristics
Bulk density ranged between 0.27 and 1.44 Mg m-3, with 55% of the variability explained by soil order and 30% by land use (Fig. 3a, Table 3). Mixed cropland soils had the highest bulk density. The cropped soils were all sampled at the end of the harvesting operations, but before any further tillage, and probably represent the maximum bulk density for nontrafficked soil. In the majority of cases, bulk densities were well below values that are of concern for root growth, or for adequate drainage and aeration (Drewry et al., 2000; Pabin et al., 1998), although Houlbrooke et al. (1997) caution than on volcanic ash soils, adverse effects on ryegrass (Lolium perenne L.) root and shoot growth can occur at bulk densities < 1.3 Mg m-3.
Total porosity was generally a reverse of the bulk density pattern, with the lowest porosity on the mixed cropland soils (data not shown). Soil order explained 45% of the variability, and land use 29% (Table 3). The maximum porosity was 86%, and the lowest 46%. Macroporosity in these soils ranged from a minimum of 0.95% v/v to a maximum of 46.3% (Fig. 3b). Variability was partly explained by soil order (32%) and partly by land use (36%) (Table 3). Macroporosity is generally considered a more sensitive measure than total porosity because macropores are the dominant pores responsible for aeration and drainage (Dasilva and Kay, 1997), and are preferentially destroyed by compaction (Drewry et al., 2000). In our data set, the soils under mixed cropland, horticulture, and pastures for dairy and drystock production all had a substantial number of examples with macroporosity < 10%. While critical values for macroporosity are not well defined, and differ for different plant species, a precautionary value of >10% macroporosity has been suggested as advisable in New Zealand to maintain pasture production near optimum (Drewry et al., 1999, 2000). Several soils in the data set were below this value, raising concern about compaction by stock treading on dairy pastures and by tillage and vehicle traffic under cropland.
The total available water (Fig. 3d) showed a wide range in values (4–49% v/v), with little differentiation between land uses (21% of the variation) and soil order (25% of the variation, Table 3). Readily available water (Fig. 3c) also showed a wide range (2.6–23.1% v/v) but tended to be lower (median < 8%) under mixed and arable cropland than under other land uses. Even so, land use explained only 19% of the variability (Table 3), soil order explained 26%, and there was strong interaction (60%).
Interpretation
Most authors have found that organizing data by land use category has been a useful approach to show differences in soil quality properties despite spatial variability. Boehm and Anderson (1997) and Wander and Bollero (1999) found differences in soil properties due to tillage management on farms under different crop management systems, despite a range of soil classes and geographic distribution. Hellkamp et al. (2000) grouped sites by land use and management classes, and reported proportions of sites meeting target criteria. The proportion of cropland (12–31%) in the North American states (Brejda et al., 2000a,b; Hellkamp et al., 2000) was very much greater than in New Zealand (<1%). The latter authors weighted their data to adjust for differing sampling frequency and intensity across states and land uses. We also favor reporting soil quality using land use categories, but have also found soil order to be important for interpretation, particularly of organic resources, bulk density, and total porosity. The proportion of sites meeting target criteria (with weighting for area if necessary) is also our intended way to summarize data once the three-year data collection is completed. Established and scientifically justifiable target criteria are a prerequisite for this approach, but setting of such targets has proved contentious (Sojka and Upchurch, 1999; Karlen et al., 2001).
A Minimum Data Set
Principal components analysis using all 12 soil properties in the data set extracted four factors with eigenvalues > 1 that explained 75% of the variability (Table 4). Factor 1 (18.5% explanation) was primarily physical properties (bulk density and porosity), Factor 2 (24.4%) was organic resources (total C and N and mineralizable N), Factor 3 (16.6%) was soil water characteristics (readily and total available water), and Factor 4 (16.5%) was soil chemical characteristics (pH, base saturation, and Olsen P). We appraised the data set to determine whether all soil properties warranted retention. Some items such as pH and base saturation were highly correlated (r = 0.64, P < 0.001, n = 221) and there was no advantage in retaining both. We retained pH in favor of base saturation, as it was easier to measure. The CEC was also discarded because within a soil order much topsoil CEC was explained by total C content (for all soil orders r = 0.45, P < 0.001, n = 221), as was also reported by Parfitt et al. (1995). A further reason for not retaining CEC was that it was not highly responsive to soil management within a soil order (it is not a highly dynamic soil characteristic). Total porosity was dropped because, within soil orders, it was generally inversely related to total C, and was much less responsive than macroporosity. We also did not retain readily and total available water because it is the water storage in the whole soil profile, and not just the 0- to 100-mm depth, that is of importance when assessing water availability for plants (Hargrove, 1988; Rhoton and Lindbo, 1997). To assist with estimating water storage, we recommend recording the profile characteristics of the subsoil and subsurface horizons.
|
Soil Quality and Sustainability
The soil quality properties of concern were those associated with the more intensive land uses of dairy production and arable cropland. Soils under pasture for dairy production and arable cropland showed evidence of soil compaction (decreased macroporosity) relative to other land uses. In more than half the cases, macroporosity was <10% v/v, suggested as a limit below which some deterioration in pasture growth could be expected (Drewry and Paton, 2000; Drewry et al., 1999, 2000; Singleton et al., 2000). We have no information on the consequences of the reduced macroporosity in our study. High levels of inorganic nutrients and substantial organic matter loss were often found under arable cropland, in common with other studies (Boehm and Anderson, 1997; Wander and Bollero, 1999). While organic matter loss is often associated with a decline in desirable soil physical attributes (Soane, 1990; Reeves 1997), high inorganic fertility is undesirable because of the increased risk of leaching losses to the wider environment. The risk is much modified by the low proportion (<1%) of New Zealand soils under cropland, but will be of local concern. There was no evidence in our study that pine plantations had resulted in excessive soil acidification (a perceived risk by land managers in New Zealand). Some pine sites under the third rotation of pine had similar acidity to those under the first rotation, and indigenous vegetation frequently had a lower pH than forest plantations.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The pragmatic, simple, and low-cost approach of the project has been successful in obtaining the participation of the majority of the regions, and has provided quantitative information where none previously existed. We found similar patterns in soil quality characteristics within the same land use across all regions despite differences in soils and climate. Examples were the high total C and N contents under virtually all pastures, loss of macroporosity under pastures, and high chemical fertility under cropped soils. In no region did we find evidence that soil acidity under plantation pine was any lower than that under indigenous forest. The participating regions are waiting until the end of the project to fully assess the implications for future monitoring policy. At present, the data are being used to raise awareness of soil quality issues among regional council staff, scientists, and the general public, and to justify future monitoring. Aspects of the 500 Soils project findings on soil quality under pasture for dairy production have already been incorporated into guidelines for best management practices for the dairy industry (Roberts et al., 2001).
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. C. R. de Lima, W. Hoogmoed, and L. Brussaard Soil quality assessment in rice production systems: establishing a minimum data set. J. Environ. Qual., March 1, 2008; 37(2): 623 - 630. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. P. Sparling, D. Wheeler, E.-T. Vesely, and L. A. Schipper What is Soil Organic Matter Worth? J. Environ. Qual., March 1, 2006; 35(2): 548 - 557. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Sparling, R. L. Parfitt, A. E. Hewitt, and L. A. Schipper Three Approaches to Define Desired Soil Organic Matter Contents J. Environ. Qual., May 1, 2003; 32(3): 760 - 766. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. R. Lilburne, A. E. Hewitt, G. P. Sparling, and N. Selvarajah Soil Quality in New Zealand: Policy and the Science Response J. Environ. Qual., November 1, 2002; 31(6): 1768 - 1773. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Vadose Zone Journal | Journal of Plant Registrations | ||||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal |
