Spatial and seasonal variability of P presents a challenge for accurate soil sampling for agronomic and environmental management of farm fields. Our objectives were to: (i) evaluate the accuracy of soil sampling protocols (sampling intensity) during the growing season and after harvest, with and without manure application; and (ii) evaluate the impact of temporal and spatial variability on the accuracy of conversion equations for Mehlich‐3 to Morgan P soil test P. Two 4‐ha corn (Zea mays L.) fields were sampled (150 samples per field; 64 in a regular grid, the remainder in a pattern that optimized distribution of lag distances) in July and after corn harvest. Semivariograms were constructed to investigate spatial dependence. Simulations were done using the regular grid samples to determine the 95% confidence interval (CI) and probability of obtaining a mean within the 95% CI as impacted by sampling intensity. Thirty or more samples per 4‐ha field were needed to obtain an estimate within 10% of the population mean with >90% probability independent of the timing of sampling. Manure application reduced spatial dependence and both summer sampling and sampling within a month after manure application greatly reduced the accuracy of Mehlich‐3 to Morgan conversions used in New York (P < 0.001). Additional research is needed to expand the database with more fields, both tilled and hay and pasture fields, but based on current research, we conclude that current guidelines for the timing of sampling and number of samples per hectare should be adjusted for most accurate P guidelines, especially if Mehlich‐3 to Morgan P conversion equations are used.

No abstract