Pre-plant vs Post Harvest Soil Tests

One of the other things I was interested in was to look at pre-plant versus post-harvest soil test data. We sampled the soil on October 23, 2014 (Planting was in Nov 2014) and then after harvest on June 4, 2015. I did a paired t-test to compare a few key parameters from October 2014 (pre-plant) to June 2015 (post harvest) and this is what I found:

 

Conventional thinking says that since we have removed nutrients from the soils, they should be depleted by the amount of removal a crop takes out. For winter wheat, and 50 bu/ac yield, we'd expect to remove 24 lb/ac of P2O5 (~10 lb/ac P) and 14.5 lb/ac K2O (12 lbs K) in the grain alone. So how can we explain an increase in both P and K, that doesn't make any sense if we follow the simple law of the conservation of mass. What are we missing?

Hidden Assumption 1: Soils are only 6" deep, that why we only sample in the "plow layer", right? When we sampled, we only sampled the the top 6" of soils - this is standard and I suppose it goes back to the "plow layer" idea. However, the soil is deeper than this and if a plant's roots go deeper than 6", they are accessing reserves of soil that we are not measuring. The assumption therefore that we should be sampling as if the plants are in a flower pot (a sterile one at that) are flat wrong.

False Assumption 2: Soils are inert mineral media. The second hidden assumption that we operationally live under is that soils are static and that if we measure P in January, they will be the same as P in June (all other things being equal). The fact is that soils are dynamic and micorhizzal relationships between plant and soil are happening all the time. Rather than pretend i know much about it, let's just step back and realize that our soils are living, dynamic ecosyetems that affect the plant and are affected by the plant at least in the root zone. Soils are NOT sterile flower pots 6" deep.

False Assumption 3: We can safely ignore the nutrient value of cover crops - mainly because it's difficult to measure how much will be plant available. We took the soil tests in detail, but hardly bothered (except for two random samples) to look at the cover crop. Fact is we had between 4,000 and 6,000 lb (1 - 1.5 tons of biomass) that measured an average of 0.3% P and 2.5% K (we'll set aside the %N in cover crop for a separate discussion). At face value, that's between 9 and 14 lbs/ac P and 100-150 lb/ac K. Literature shows that around 85% of P and >90% K is plant available from crop (cover crop) residues to the next plant. Well, since we saw increases in the soil test P and K, maybe some of that extra P and K came from the cover crop?

I am using a little sarcasm here, but this whole experiment has taught me and Carl so much an it's of necessity made us re-examine our assumptions about soils. If we didn't believe there was something funky with our assumptions about soils before, this experiment has moved the ground beneath our feet and shown us how little we know. As for people who continue to operationally consider soils as 6" deep inert flower pots? Well, they will continue to pay for inputs they don't need. At least now we know what lies at the root of the problem, now let's see if we can provide a viable alternative.