PLOTSIZE AND SAMPLE NUMBER FOR NEUTRON PROBE MEASUREMENTS IN SMALL-FIELD TRIALS

Soil water storage over a 2.85-m soil depth was measured from 200 alum inum access pipes, separated by 0.3 m in both directions, in a 1.2 x 1 5.0-m plot in dry and wet periods during 1988 and 1989 The objective w as to determine the minimum plot size and number of soil water content measurements if measured with a neutron probe for small field trials. The minimum plot size representing the 15-m plot was found to be depe ndent on water storage variance and on the distance over which soil wa ter storage measurements were spatially correlated. We concluded that a plot length of 5 m was needed to represent the mean and variance of the 15-m plot. Bootstrapping and temporal stability analysis were used to estimate the minimum number of observation tubes required to estim ate the mean and variance of 1.2 x 5.0-m plots. Bootstrapping showed t hat at least 10 soil water storage measurements were required in the 5 -m plot. Soil water storage distribution within the plot was found to be highly stable in time, especially for individual soil layers. Using temporal stability analysis, the number of required access pipes need ed to estimate a plot-average soil water storage was further reduced t o three. However, the variance of soil water storage was not conserved while reducing the number of measurement locations. We propose that a field study with small field trials should start with the maximum fea sible plot size and number of measurement locations. In this initial p hase, statistical techniques as proposed in this study can then be app lied to reduce the required number of observations, using predetermine d error limits.