Skip to main content

Spatial variability of BCI soil classes

The nested dumbbell augerings (Table 6.2) show that soils at 1m spacing are highly likely to be of the same class.  The one class distinction that seems somewhat heterogeneous over very short range is topsoil colour.  This suggests that the differentiation of the dark from the brown loams is not valid for general soil classification.  Ignoring variations in topsoil colour, soil classes are also likely to be similar at 10 m range.  However, soils at 100 m spacing are likely to be of different classes.  This indicates that the soil classes and mapping units are appropriate for map scales of 1:10 000 – 1:15 000.

These variability-spacing findings refer to the soil morphological features on which the classes are defined. Other soil attributes probably have different spatial variability patterns, but dedicated geostatistical studies are needed to elucidate them. In general, the variability patterns for soil classes are likely to apply to mineral-related soil attributes such as texture and total contents of mineral nutrients.  Soil attributes that derive mainly from biotic processes are likely to be variable over shorter, plant-sized ranges, as in the short-range variability of topsoil colour. Yavitt (2000) found some systematic difference between geologies for mineral-sourced P but hardly any for the atmospherically-sourced and biotically incorporated nutrients N and S.

Soil class spatial variability



Soil class similar






Augers at 1 m spacing

28 pairs



3 of 5 class differences are due to variation in topsoil colour

Augers at 1 m spacing, ignoring topsoil colour variation

28 pairs




Homogeneous (ignoring topsoil colour variation) pairs of 1 m augerings compared with those at 10 m distance

12 pairs of pairs



Homogeneous (ignoring topsoil colour variation) quartets of 10 m augerings compared with those at 100 m distance

5 pairs of quartets




Updating and using the BCI soil map

There is a PDF version of the 2007 soil map in STRI and Potsdam and on-line.  Potsdam and STRI also hold a GIS BCI soil map master, which can be augmented or amended.  Changes should be made under the coordination of the STRI Webmaster in the form of new, clearly labeled and dated covers.  There should be separate covers for new data, for changes in the definitions of soil mapping units, and for alterations to soil boundaries.  Geo-referenced soils data can be entered as new covers, even if not mainly intended for soil mapping, as they will amplify the characterisation of BCI soils.

It is technically feasible to print the whole or parts of the 2007 map at any scale.  However, our augering was at an overall density sufficient for semi-detailed mapping.  International guidelines (FAO, 1979; Legros, 2006) indicate that the largest map scale supportable by our data is 1:15 000 for most of the island.  Printing at 1:10 000 is justifiable for the more intensively augered areas such as the 50 ha LTER plot and Lutz Creek catchment (see 7.1 and 7.2).  If more detail on soil spatial distribution is required for other areas, it is necessary to supplement the data and refine the map.