M. Van Meirvenne

Evaluating the probability of exceeding a site-specific soil cadmium contamination threshold.

A non-parametric approach for assessing the probability that heavy metal concentrations in soil exceed a location-specific environmental threshold is presented. The methodology is illustrated for an airborne Cd-contaminated area in Belgium. Non-stationary simple indicator kriging, using a soft indicator coding to account for analytical uncertainty, was used in combination with declustering weights to construct the local conditional cumulative distribution function (ccdf) of Cd. The regulatory Cd contamination threshold (CT) depends on soil organic matter and clay content, which entails that its value is not constant across the study area and also is uncertain. Therefore, soft indicator kriging was used to construct the ccdfs of organic matter and clay. Latin hypercube sampling of the ccdfs of Cd, soil organic matter and clay yielded a map of the probability that Cd concentrations exceed the site-specific CT. Cross-validation showed that the ccdfs provide accurate models of the uncertainty about these variables. At a probability level of 80% we found that the CT was exceeded at 27.3% of the interpolated locations, covering 3192 ha of the study area, illustrating the extent of the pollution. Additionally, a new methodology is proposed to sample preferentially the locations where the uncertainty about the probability of exceeding the CT, instead of the uncertainty about the pollutant itself, is at a maximum. This methodology was applied in a two-stage sampling campaign to identify locations where additional Cd samples should be collected in order to improve the classification into safe and contaminated locations.

Kriging soil texture under different types of nonstationarity

Abstract A geostatistical analysis assumes some form of stationarity of the variable under study, but different types of stationarity exist and often spatial data exhibit some form of nonstationarity. However, most studies assume one type of nonstationarity and consequently apply one type of interpolation method within the study area. A study area of 8×18-km area was selected because it was expected to contain complex nonstationary conditions in soil texture. Therefore, four geostatistical interpolation methods were evaluated in their ability to account for different types of nonstationarity in the topsoil silt content: two univariate interpolation methods, ordinary kriging (OK) and universal kriging (UK), and two bivariate methods, simple kriging with varying local means (SKlm) and ordinary cokriging (OCK). A digital elevation model (DEM) was used as the exhaustive secondary information for the bivariate methods. Two kinds of nonstationary conditions were identified inside the study area: (1) a large-scale trend in both the silt content and elevation, with a strong correlation between them, and (2) a very strong local fluctuation around a mean value, representing a local nonstationarity. Consequently, different techniques were applied in different parts inside the study area: the global trend was best accounted for by OCK and UK could best account for the local nonstationarity. After combining the results of the two prediction methods, it was found that the overall estimation of the silt content was more precise than when any single method was used over the entire study area.

Phosphate enrichment in the sandy loam soils of West-Flanders, Belgium

The last three decades, pig breeding has evolved towards a specialised, large scaled, land independent bio-industry in the province of West-Flanders. Subsequently, in certain regions, very high amounts of liquid pig manure are produced each year. This pig slurry is used as a fertilizer at a rate which very often exceeds normal agricultural practices. Because of the nonequilibrium between the phosphorus crop requirements and the P-inputs, phosphates accumulate in the soil. However, the phosphate sorption capacity of a soil is limited. Once the sorption capacity is exceeded, phosphates will start leaching through the soil profile. Since, during winter, in these areas, the groundwater table is situated at a depth of less than 1.0 m, phosphate breakthrough might take place. In the sandy loam soil region (± 1000 km2) of the province, an inventory of the P status of the soil was made. The region was sampled according to a regular grid with 2 km intervals. At random, some sample points were only 500 m apart. This resulted in a total of 296 samplings. In view of fertilizer recommendations, lactate extractable P of the plough layer (0–30 cm) was determined. A maximum value of 101 mg P 100 g-1 of air dry soil, a minimum value of 6 mg P 100 g-1 and a median value of 31 mg P 100 g-1 were found, indicating that for half of the spots monitored, the P status of the soil is high to very high. An oxalate extraction was done to investigate the phosphate saturation of the soil profile (0–90 cm). Based on a critical phosphate saturation degree of 30%, more than half of the soil profiles are phosphate saturated. Phosphate leaching at a rate higher than 0.1 mg ortho-P l-1 at a depth of 90 cm can be expected. Therefore, a restriction of the P fertilization should be highly recommended. The geostatistical processing of the data using block kriging resulted in a spatial continuous estimate of the phosphate saturation degree. A good agreement was found between the pig density and the phosphate saturation degree of the soil profile.

Spatial variability of soil nitrate nitrogen after potatoes and its change during winter

Knowledge of the frequency distribution and spatial structure of the soil NO3-N is required to develop an efficient sampling strategy. A 1 ha polder field was sampled after the harvest of potatoes in October 1987, and in February and April 1988, without being fertilitzed since March 1987. These data sets were examined by a classical statistical as well as a spatial structure analysis. The October and February data sets were found to be lognormally distributed, the April data showed a normal frequency distribution. All three data sets had a spatial structure, although the October data were anisotropic and needed removal of a trend. The spatial variability of soil NO3−N decreased, became isotropic and evolved towards a larger range of spatial dependence during the winter. Knowledge of this structure permitted to krige or cokrige the data. The number of samples required to estimate the mean NO3−N content with an acceptable precision was found to be 39, 43 and 17 in October, February and April respectively.

Within-field variability of mineral nitrogen in grassland

Abstract The within-field variability of soil mineral nitrogen (Nmin) in a grazed grassland of 8000 m2 was examined. NO3–-N concentrations were characterized by a high spatial variability. This can be explained by the uneven deposition of animal excreta. All NH4+-N as well as NO3–-N values were lognormally distributed, before and after the grazing season. At the end of the grazing season the largest part of the variability of NO3–-N was found for NO3–-N concentrations measured within a distance of a few metres. A high variability for NO3–-N over very short distances was also indicated by a large nugget variance. During the grazing season, observed mean Nmin values increased from 22 to 132 kg N ha–1. Regions with clearly higher NO3–-N concentrations could be identified. These zones matched with the drinking place and the entrance of the pasture, places which were more frequently visited than others. High residual N levels in autumn led to relatively high losses of N, mostly by leaching, during the subsequent drainage period. Knowing the variability of Nmin, the number of samples needed to estimate the average Nmin in a field could be calculated for different probabilities and various degrees of precision. From the spatial distribution of the Nmin concentrations and the restrictions imposed by the new European decree, adapted fertilizer strategies can be proposed at least for places where systematically higher Nmin concentrations can be expected.

Temporal Stability of Spatial Patterns of Soil Salinity Determined from Laboratory and Field Electrolytic Conductivity

ABSTRACT We elaborated a procedure for the assessment of the temporal stability of soil salinity and the optimization of the sampling effort. Soil electrolytic conductivity data obtained from field electrode probes and laboratory analysis were compared and analyzed to check the temporal stability of salinity patterns. Sampling of 20 locations at different depths was repeated 19 times over a period from November 1994 to June 2001. Both methods showed a strong temporal stability. The Spearman rank correlation confirmed the persistence of the ranking of the different locations. Additionally, using the technique of relative differences, we identified three classes: (1) low saline locations, (2) locations which are representative of the average field soil salinity, and (3) high saline locations. Low saline locations were associated with the zones of waterlogging and/or salt leaching. High saline locations were exclusively in the zone of salt accumulation. Locations representative of the average soil salinity were found in all three possible zones. We investigated how precise the selected locations representing the average soil salinity can estimate this average. We found that using only two locations from the 20 available, the average was adequately estimated with a difference < 0.3 dS m−1. This representativeness was also checked by splitting the measurements into two temporal subsamples. For both subsamples the same locations were representative of the average soil salinity compared to when all measurement dates were considered.

Combining multiple signals of an electromagnetic induction sensor to prospect land for metal objects

Buried unexploded ammunition is a major problem on arable land in former battle areas. Many battlefields of the First World War (WWI) still contain a lot of unexploded shells just below the plough layer, posing serious threats to soil editors and trenchers. Electromagnetic induction (EMI) sensors have been used for a variety of agricultural and archaeological purposes to map the natural soil variability and to locate buried archaeological remains. Besides its sensitivity to variations in soil texture and anthropogenic disturbances, EMI proves to respond strongly to metal objects in the soil. Most EMI sensors rely on a single signal, with magnitude and sign of the metal anomalies differing according to the instruments coil distance and separation. The multi-coil EMI sensor, the DUALEM-21S, provides four simultaneous apparent electrical conductivity ( ECa ) signals enhancing significantly the possibilities for signal processing. To calibrate our instrument, we buried different masses of metal at different depths. The four ( ECa measurements showed a response to the metal objects down to 1.2 m. The measurements were subtracted by their gradual trend to obtain the local anomalies (Δ( ECa ). A combination of these four Δ( ECa ’s was used to amplify the signal response to metal, influenced by both depth and mass of the buried objects. At an intensively shelled former WWI battle field near Ypres (Belgium), a detailed prospection was conducted with the DUALEM-21S. Based on our multi-signal procedure, we located 40 positions, 20 where we predicted buried metal and 20 where we expected that no metal was present within 1.2 m depth. There were no false negative predictions and at the 20 locations where we expected metal, shells up to 90 kg were excavated. As a final outcome we produced a map with predictions of the mass of metal objects in the soil assuming a fixed depth and alternatively a map with predictions of the depth of metal objects assuming a given mass. Apart from their potential for agricultural and archaeological investigations, multi-( ECa signals were shown to be useful for locating metal objects, like unexploded WWI shells, in the top 1.2 m of soil.

Geostatistical modeling of sedimentological parameters using multi-scale terrain variables: application along the Belgian Part of the North Sea

In the nowadays highly pressurized marine environment, a science‐based approach to management becomes increasingly important. In many cases, the sediment nature and processes are the key to the understanding of the marine ecosystem, and can explain particularly the presence of soft‐substrata habitats. For predictions of the occurrence of species and habitats, detailed sedimentological information is required. This paper presents a methodology to create high quality sedimentological data grids of grain‐size fractions and the percentage of silt‐clay. Based on a multibeam bathymetry terrain model, multiple sources of secondary information (multi‐scale terrain variables) were derived. Through the use of the geostatistical technique, Kriging with an external drift (KED), this secondary information was used to assist in the interpolation of the sedimentological data. For comparison purposes, the more commonly used Ordinary Kriging technique was also applied. Validation indices indicated that KED gave better results for all of the maps.

Quantification of soil textural fractions of Bas-Zaïre using soil map polygons and/or point observations

Abstract Reconnaissance soil surveys typically produce qualitative choropleth maps and descriptions of sampled soil profiles. These can be converted to quantitative information directly. Alternatively quantitative predictors can be made by interpolating between the point observations, or the two can be combined. Three methods to predict the sand and clay content of Bas-Zaire using a reconnaissance soil map and 151 soil profiles are compared. A cross-validation showed that the combination method produced the most precise predictions with an acceptably small bias. This result was obtained only after verifying and correcting the soil surveyors estimations of the soil map prediction variances and stratifying the area to acquire a relative pooled within-stratum variogram.

Spatio-Temporal Kriging of Soil Salinity Rescaled from Bulk Soil Electrical Conductivity

Our spatial data consist of 413 measurements of the apparent electrical conductivity (ECa) obtained with electrical probes in the east of Hungary. Additionally, a limited subset of the locations (15 to 20) was sampled for laboratory analysis of soil electrical conductivity of 1:2.5 soil:water suspension (EC2.5), a simple proxy for the electrical conductivity of soil saturation extract (ECe). The latter formed our calibration data set. This procedure was repeated 17 times between November 1994 and December 2000 yielding a large spatio-temporal database. The first step was to rescale EC2.5 from ECa, based on the calibration data sets, using classical and spatial regression models. The residuals of the ordinary least squares model were tested for the absence of spatial dependence using the Moran’s I test. This hypothesis was accepted, the EC2.5 was rescaled using the classical regression model. The next step was to identify the structure of the variability of the rescaled EC2.5 by computing and modeling the spatial, the temporal, and the spatio-temporal covariograms. Finally, soil salinity maps were produced for the study area and for any time instant using spatio-temporal kriging. The estimates were more precise compared to the ones obtained using only the spatial covariogram computed and modeled separately for each time instant.