On-farm sampling density and correction requirements for soil moisture determination in irrigated heavy clay soils in the Gezira,central Sudan

Using the neutron scattering technique, with separate calibration for each measuring depth and temperature corrections, an over-sampling experiment with a worst case analysis was conducted in tenant irrigated fields under arid conditions. The purpose was to better understand actual on-farm soil moisture distribution as well as to determine minimum sampling density requirements for water use efficiency calculations in the heavy cracking clay soils of the Gezira irrigation scheme, central Sudan, under inhomogeneous watering conditions. Results show that actual soil moisture inhomogeneities can seriously distort the moisture distribution and water use pictures if the sampling density is too low. In a 2.1 ha end field under Gezira conditions 20 equally spaced neutron probe samples had to be collected from the 30 cm soil depth if the total experimental errors were to be kept within 12.5% of the average moisture content being measured. Sampling density requirements increased to 24, 28 and 33 samples for worst case error limits of 10%, 7.5% and 5% at 30 cm depth. At the agronomically more important lower depths, at or below 70 cm, less than 10 samples only could be afforded with an error of 10% at 70 cm, of 15% at 50 cm and of 20% at 30 cm, the errors typically becoming smaller at larger depths throughout. Credible soil water averages were obtained with this sampling. Field moisture patterns were well recognized when averaging several days of measurements.