ESTIMATING SOIL-SALINITY BY DISJUNCTIVE KRIGING

Abstract. Disjunctive kriging provides minimum variance estimates of properties from non‐linear combinations of spatially correlated sample data. In addition it can be used to estimate the conditional probability that some critical threshold is exceeded or that there is a deficit at unsampled points. The technique has been applied to estimate and map the salinity of the soil in the Bet Shean Valley of Israel from measurements of electrical conductivity. In November 1985 the estimated electrical conductivity of the soil exceeded 4 mS per centimetre throughout most of the region, and in only a small area was the probability of salinity less than 0.2. By March 1986 the electrical conductivity had declined everywhere to less than 4 mS per centimetre, and the conditional probability of exceeding this value nowhere exceeded 0.25. Despite the fluctuation in salinity farmers seem to have it under control. The results suggest that winter wheat is likely to germinate poorly in the saltier parts of the region and that lucerne (alfalfa, Medicago sativa) is unlikely to yield its maximum over most of it. Cotton, a summer crop sown in spring, should not suffer. Copyright © 1990, Wiley Blackwell. All rights reserved