Comparison of Strategies to Calculate Nitrate Leaching Load on Sandy Soils.

Assessing the impact of nitrogen [N] and irrigation [IRR] best management practices [BMPs] on nitrate leaching loads is difficult to directly measure. One strategy commonly used by researchers in sandy soils is to install suction cup lysimeters below the root zone to sample the soil solution. However, making comparisons between treatments is difficult to do becaues measurements of nitrate concentration have high inital spatial variabilty (29 – 153%) and the within treatment variance (46 – 56%) is much greater than the between treatment variance (11 – 17%). Based on this variabilty, it is very likely that large numerical differences in nitrate concentration can occur in the absence of a statistically significant difference for a between treatment comparisons. Current procedures to calculate nitrate leaching load do not account for the uncertainty in measured nitrate concentration values or whether or not numerical differences in nitrate concentration between treatments are statistically significant. This presentation proposes a new method to calculate nitrate leaching loads from suction cup lysimeter that accounts for variability in nitrate concentration using a protected multiple comparison procedure to identify significant between treatment differences on a given date. A statistical analysis using protected multiple comparisons should be used to adjust nitrate concentration. This adjustment procedure detects significant differences between treatments on a given date – if there are no significant differences found, the mean value rather than the treatment values are used to calculate nitrate leaching loads. This procedure reduces the likelihood of finding a false significant difference in nitrate load, but increases the likelihood of failing to find a true significant difference in nitrate load. Using data from a study on potato grown on a corse textured soil in Minnesota in 2010 and 2011 investigating the effect of IRR and N BMPs on nitrate leaching, a comparison of nitrate leaching loads using the unadjusted measurement and adjusted estimates from protected multiple comparison analysis was made. Of the 19 and 25 measurement dates in 2010 and 2011 respectively, only 3 and 13 of those dates has a significant effect of IRR, N, or IRR x N on nitrate concentration. By improving the methods to calculate nitrate leaching loads from suction cup lysimeters, the development of BMPs to reduce N-losses can be advanced.