A field comparison of three neutron moisture meters

A study was undertaken to compare 3 neutron moisture meters (NMMs) over a range of counting times on 2 soil types in north-westem Victoria. The meters were a Pitman Wallingford 225 (WAL225), a Campbell Pacific Nuclear 503 (CPN503), and a Campbell Pacific Nuclear 501DR (CPN501). The soil types were a grey self-mulching clay and a solonised brown soil (sandy loam). The 3 NMMs exhibited different count rates in water, from 242 to 2645 counts per second. It was necessary to increase the counting time of CPN501 to 64 s to achieve a comparable performance to WAL225 at 16 s. Over the 15-month experiment, standard counts remained relatively constant for WAL225 and CPN501, but CPN503 showed pronounced variation. In field calibrations, the inclusion of depth (as a separate variable to account for neutron absorbing gradients in the soil profile) made significant improvement to the regression of all NMMs irrespective of counting time. No improvements in the calibration were achieved using density corrections from core samples for individual layers. On both soil types, WAL225 and CPN503 performed similarly to each other, with very similar residual mean squares and coefficients of determination. On the grey clay, WAL225 and CPN503 explained 77 and 88% of the variation in soil water content, respectively. CPN501 performed poorly, even with a 64 s counting time, and explained only 52% of the variation when depth was included. When depth was excluded this fell to 25%. On the sandy loam, WAL225 and CPN503 explained 76 and 74%, respectively, of the variation in soil water content. CPN501 again performed poorly with a 64 s counting time, accounting for 65% of the variation when depth was included and 33% with depth excluded. Examples of poor calibrations or none at all in water use studies are discussed and typical errors are demonstrated.