Wheat harvest is well underway. USDA-NASS reported that the spring wheat harvest in Minnesota is 30% completed as of this weekend. South of US Hwy 2 that percentage is much higher. Most are pleasantly surprised by how much grain they haul off the field.
Before last week's widespread soaker, test weight was generally well above 60 lbs/bu. For many, the amount of visually scabby or tombstone kernels was low enough not to raise alarm bells. Growers are, however, complaining about the difficulty of combining. Despite having the concaves as tight as they dare, producers end up with white caps in the grain tank. The latter suggests that the initial scab infections caused immediate abortions of the developing kernels above the initial site of infection. This in turn means that there are yield losses because of scab despite the record-breaking yields.
West Central Minnesota and the southern Red River Valley are two areas where producers are more frequently running into problems with deoxynivalenol (DON), the toxin associated with scab that is also known as vomitoxin. Test results for DON from individual loads hauled into the elevator can be infuriating as they vary (widely). Much of this frustration has to do with how sensitive the testing is to sampling 'error'.
It is difficult to pull a representative sample from a truckload because one scabby kernel in the relatively small sample used for the testing can drive the results past the threshold of 2 ppm. The Federal Grain Inspection has, therefore, very defined rules and procedures on how elevators have to sample individual loads. Nevertheless, if you have any appreciation for statistics, probabilities, and sampling techniques you will understand the inevitability of the observed variability within a load and between loads from the same field or bin.
The single test is a single observation that falls somewhere on the distribution of the average DON concentration in that field or bin should you decide to test the DON concentration of every single kernel from that field or bin. Many kernels will have no DON at all. Some, especially the white chalky tombstone kernels, will have DON concentrations that can exceed 200 ppm. If you graph out all the test results of each individual kernel in a histogram you likely will end up with a distribution that is strongly skewed to the right.
It is an exercise in futility to test every single kernel as you would have no grain left to sell. Instead, elevators pull a relatively small sample from each load you haul in. Because of the underlying skewness in the DON concentration distribution of individual kernels, the test results from individual loads you haul in will vary. One scabby kernel can 'sour' a whole load as it has a disproportionate effect on the test. The FGIS sampling rules were developed with this understanding in mind.
Nevertheless, that doesn't make it any easier to accept. Try to think of that individual test result as a single observation. Not until you have replication, i.e. multiple loads/multiple tests, will you get an idea of the average DON concentration across the whole field. It will allow you to calculate an average and measure the variability between the loads. Those two metrics are much more informative than the individual observations. This is 'old hat' for those of you who grow sugarbeets. Every single load not only gets weighed but a quality sample is taken from each load to determine sucrose concentration, tare, and loss to molasses. The final payment per ton is not based on each load but on the field average and 'outliers' are removed from the calculation of the field average on which the payment per ton is based.
Given the underlying problems of sampling variability for DON, I think it is worthwhile to have a discussion about whether an approach similar to that used in sugarbeet production is feasible and fair.
Last week's widespread soaking and the hot and humid weather that has followed since are adding another problem. As of yesterday, low Hagberg Falling Numbers (HFN) are being reported in the Mahnomen area and up along the beach ridge. Low HFN numbers indicate the presence of alpha-amylase in the harvest grain. This in turn is an indication of sprout damage even if no sprouting is visually evident. I wrote about sprout damage and low HFN in this post from 2019. Once your HFN has dropped below 250 in a field, leave the field for later and harvest fields that have an HFN above 300.
Comments
Post a Comment