New South Wales (NSW) Soils Data

Description:

Soil is a key factor affecting agricultural production. Soil maps at high spatial resolution are beneficial for both sound land management and profitable crop production.

The NSW Soils Data consists of the total count for the full gamma ray spectrometer spectrum from three fields (F-Brigalow, 12-Brigalow, and Coda) covering about 2.66 km² of Nowley Farm (30.23°S, 150.24°E) in New South Wales, Australia.

A ground-based gamma ray spectrometer (GRS), or gamma radiometer, measures broad radioactive emissions from elements such as potassium, uranium, and throium in the top 40-60 cm of the soil. In this particular soil survey, the GRS was run on a single vehicle, which was driven along (approximately) evenly spaced swaths, 30-40 m apart, in the direction of tillage. Location data was received from a single GPS mounted on the vehicle. Total count (TC) in counts per second (cps) from the GRS is the variable of interest, it integrates the entire spectrum and often shows strong spatial patterns.

The original TC dataset is made up of 34,266 data. Exploratory analysis indicated that measurement-error variance on the original scale depended on the measured value, and hence Cressie and Kang (2010) made a shifted logarithmic transformation.

Define the adjusted count (AC) as AC ≡ TC + 160; a spatial analysis was carried out on the transformed variable, Z ≡ log(AC) = log(TC + 160). The measurement-error variance on the transformed scale was identified from an independent study to be σ² = 0.0016.

Format:

The file Soil_X.mat contains the X-coordinates, Soil_Y.mat the Y-coordinates, and Soil_TC.mat the total counts. The file basis_fns.m contains a script for generating the basis functions for Fixed Rank Kriging of the total counts.

Reference:

Cressie, N. and Kang, L. (2010). High-resolution digital soil mapping: Kriging for very large datasets, in Proximal Soil Sensing, eds R.A. Viscarra Rossel, A. B. McBratney, and B. Minasny. Springer, Dordrecht, NL, 49-63.