Scientists critical of soil health testing often stress absolute differences between soil labs performing tests. But what about ranking? Rank-Order statistics focus not on the absolute numbers but on whether labs are similarly distinguishing one type from others, an essential aspect. Now, recent farmer comparisons of soil health tests sent to differing labs give strong evidence that this is key. Despite numerical differences, the health score ranking provided by differing labs are actually very similar. “If you look at any of these reports you are likely to draw similar conclusions” says Dr. Will Brinton of Woods End Labs. “But if you listen to the national uproar of scientists criticizing precision of soil health tests, you’d reach a different conclusion”.
This finding is supported by at least two recently reported soil test projects conducted by farmers, one reported by s Crop Editor Bill Spiegel (Kansas) and the other by Dr Dan Davidson, CCA in Nebraska.
Spiegel used a soil contractor to sample two 40-acre blocks on differently managed farms (conventional vs No-Till). Davidson took samples from 6 Nebraska farms with varying management of conventional, CRP, and permanent pasture. Both sent single-blind samples to four 4 labs for soil health.
Bill Spiegel’s two farms are on Harney Silt Loam – the Kansas State Soil, and all labs’ results- Cornell, Woods End, Ward and OSU – similarly rated one farm soil somewhat healthier than the other. A portable OSU test for oxidizable carbon reached the same conclusion for the same pair of samples. There were differences in the magnitude of separation, but overall the labs agreed on the trend. It wasn’t however what Spiegel expected (see article).
Davidson reported his findings at the recent Northern Prairies Alliance Meeting in Minot SD. Dan compared the 6 farm soils at Ward, Woods End, Midwest and Cornell. The results were the same: the labs closely agreed on the best and worst soils and rank-ordering showed scoring of farms was highly equivalent.
Dan drilled down into test minutiae and found more interesting agreements. For example, Solvita soil health scores correlated closely with Ward labs PLFA test method, a complex test which estimates microbial biomass. An unusually close correlation was found when comparing Cornell’s overall soil quality ranking to the Woods End Fertility Score (r² 98%). Woods End original soil health test was offered for organic farmers in 1987 by combining the physical, chemical, and biological methods in one report. The newer fertility scoring combines 5 health factors and 4 nutrient factors.
Farmer-based studies are not what researchers consider validated data, but it’s how comparisons are commonly run, and these farmers footed the bills. The results underscore curiosity over what all the debate is about for soil health tests. If independent soil labs are reaching similar conclusions using differing tests, “that’s good” says Brinton: “it means we are observing the same population.” In a statistical trial recently reported at a CCA event by University of Connecticut concerning routine P, K, Ca, Mg, and OM tests compared to Solvita health tests, the latter were found to have less variability under the same sampling regimens.