Characterization of Historical Precipitation Data in Relation to Production Risk of Fall-stockpiled Tall Fescue

Abstract

Weather data collected during field experiments are not often compared to long-term weather records in meaningful ways. We examine the results of a study of N application on yield of fall-stockpiled tall fescue in relation to concomitant and historical late-season precipitation in the Southern Piedmont Region of Virginia. Previous analysis indicated that splitting applications of 40 or 80 lb N/acre increased the probability of achieving designated yield goals. However, yields were also affected by precipitation during the stockpiling period. Monthly precipitation during the study was compared to the 58-year historical record via percentile distributions, principal component analysis, and Monte Carlo simulation. Total precipitation during the August through November period of the seven-year study averaged 4 inches above the long-term mean. Yields increased with N rate as expected. A dry matter production model that incorporated concomitant and historical precipitation records with field data appears more suitable to assess yield risk than a model based on field data alone. Experimental results should be interpreted with regard to the weather conditions during the experimental period. Management recommendations that consider results in the context of historical weather records provide a more robust interpretation of the data and more accurate estimates of risk.