Sources of bias in the field calibration of a neutron meter

Abstract

Field calibrations of a neutron moisture meter and a gamma density meter were made on the individual layers of a duplex soil consisting of a metre depth of sand overlying a deep cracking clay. Two sources of bias were encountered in the case of the clay layer: (1) differences in bulk density and composition between sites caused a bias of 20% in the calibration coefficient resulting in an underestimate of change in water content, and (2) non-representivity of the access tubes, due to failure to sample cracks, caused an additional negative bias of up to 30%. The first of these inaccuracies can be avoided by using a better design for the calibration experiment and by correcting neutron count rate for soil density. The second bias due to poor crack sampling can be corrected for on the assumption that only three-dimensional swelling occurs on a field-wide basis.