Conversion of Modified Morgan and Mehlich-III Soil Tests to Morgan Soil Test Values

In several states in the Northeastern U.S., the Morgan or Modified Morgan soil tests are used as the basis for both fertilizer recommendations and P runoff risk indices. However, private laboratories servicing these states typically use the Mehlich-III solution as their basic soil test extractant. To meet Natural Resources Conservation Service standards for nutrient management and to use land grant university research databases to derive recommendations, accurate conversions from Mehlich-III and Modified Morgan to Morgan are needed. A study was conducted in New York to: (i) develop models that convert Mehlich-III P, K, Ca, and Mg soil tests to Morgan equivalents; and (ii) evaluate the impact of the use of these prediction models on agronomic recommendations for corn and the New York P index. Soils from 235 locations (27 different New York soil types) were analyzed for pH, modified Morgan P, and Morgan-and Mehlich-III extractable P, K, Ca, Mg, Fe and Al. Multiple regression analysis was used to derive conversion models that fit the data. The models were validated, and the impact of their use on fertilizer recommendations was assessed using a set of >10,000 independently collected New York soil samples. Morgan and Mehlich-III extractable K, Ca, and Mg were correlated linearly with slopes approaching 1. For P, the best model fit was obtained using pH, Mehlich-III P, Ca, and Al as independent variables (r(2) = 0.88). A slightly less reliable prediction was obtained without Al (r(2) = 0.82). The use of Morgan equivalents did not alter the percentage of low, medium, high, and very high soils in the New York data set, nor did it affect the P index classification. Recommendations for corn derived from a database with independently collected Mehlich-III soil test data (without Al) were identical to those obtained using measured Morgan values for 57% of all samples. An additional 32% showed <10 kg P(2)O(5) ha(−1) difference. Eight percent predicted P applications that were 15-20 kg P(2)O(5) ha(−1) larger or smaller, whereas 3% showed deviations >20 P(2)O(5) ha(−1). Inclusion of Mehlich-III Al data is expected to improve the accuracy of the recommendations and assessment of the P index. We conclude that conversion equations can be used to derive accurate Morgan equivalents based on pH, Mehlich-III P, Ca, and Al, but the results need to be verified based on land use history.