Should We Manage Small-Scale Variability?

How Extensive is Small-Scale Variability in Agricultural Fields?

How many times have you driven by this field on the right and wondered what was going on?  We seldom second guess these 'field chicken-pox' because we know that they are caused by animal deposits (both fecal and urine).  Do they contribute to long-term spatial variability?  If this field of winter wheat was topdressed, should all areas be treated the same?  At what scale (resolution) would we have to operate to account for SPOTS with N and areas without N?  

Should precision ag be working at a 1m2 resolution?

Yield Potential 222.jpg (193085 bytes)

On the left is a picture of one plot (yellow lines) in Experiment #222 that was started in 1969 by Dr. Billy Tucker.  This plot has received 80 lbs of N/acre, 60 lbs of P2O5/ac and 40 lbs of K2O per acre for 30 years.   Conventional tillage (2 shallow diskings prior to planting) has not changed for 30 years, nor have harvest or straw residue management practices.  What is important to note in this picture is that distinct differences in yield potential exist between the two squares (blue and red), less than 3 feet apart. The difference below is not due to planting density or planting error, but rather plant growth.  What has caused this difference has not been documented.  However, what is important is noting that large differences exist in what should be extremely homogenous areas.  The difference between the red and blue squares is one of 'yield potential.'  Increasing the topdress N rate in the red square will unlikely result in an equivalent yield as the blue square, even though need is greater.  In the future, fertilizing the red and blue squares based on their predicted yield potential will maximize yield and N use efficiency. 


Small-scale variability can be observed in a follow-up picture from Experiment #222 on the right.  This kind of small-scale variability can be found in farmer fields all over the world.  In the past we have ignored micro-variability in yield, opting to treat fields with one rate and one management practice.  In the developing world, treating each 1m2 area with specific rates is possible, largely because most (if not all) field work is accomplished by hand.  In the developing world where fertilizer N is applied for grain crop production, much of it is applied by hand, and more importantly, in-season.  This allows for visual inspection/calibration of the fertilization process since farmers can easily be taught how to fertilize plants based on predicted yield potential.  At this stage, it is unlikely that we can place sensors in the hands of developing nation farmers.  However, it would be very easy to place a color coded calibration chart (possibly taped to the farmers arm) for rapid use/reference concerning fertilizer application. 

 

222_99.jpg (114056 bytes)
Sludge_99.jpg (40582 bytes) An added example of small-scale variability is visualized in our sewage sludge experiment at Efaw, where fixed N rates have been applied annually since 1993.  

Much of this work is really common sense as related to small scale variability and how this relates to altered yield potential within the same plot.  If the plant is bigger, yield potential is higher.  If the plant is smaller and starved, yield potential is lower.  If color were the same for both large and small plants (e.g., dark green), N fertilization should be higher for the larger plants based on increased yield potential.  Programming the fertilization-decision into a mechanical-by-hand process for farmers will be cumbersome.  However, if developing nation farmers continue to apply fertilizers by-hand, this should be an achievable objective, and highly likely to result in increased yield and use efficiency.

One of our yeild potential trials at Morrison, OK, treats each 1m2 independently based on predicted yield potential .  This is one of two trials where N has been applied both based on need and expected yield using sensor-based estimates of in-season estimated yield (INSEY)

Only limited forage yield response to fertilizer N applied in early March, using INSEY was apparent at this site (right and bottom picture, Morrison, OK, April 8, 1999)

Morrison_99_2.jpg (63574 bytes)
Morrison_99.jpg (49617 bytes) This is another view of our INSEY work at Morrison.  Working with cooperators requires that we place electric fencing systems (solar powered unit on the left)  around our experiments, since most cooperators graze their wheat up until jointing (first hollow stem)
Small scale variability is apparent in the Magruder Plots (west of the OSU campus) that were first started in 1892.  Now 107 years after these native prairie soils were first tilled, small scale variablity continues to be present.  The plot on the right has received annually applied N and P since 1929 (33 lb N/ac and 30 lb P2O5/ac, 1929-1967 and 60 lb N/ac and 30 lb P2O5/ac, 1968-present).  The N rate was increased in 1968, consistent with increased yield potential of the dwarf wheat varieties grown in Oklahoma. Magruder_99.jpg (48284 bytes)
Efaw_NUE99.jpg (49351 bytes) One of two nitrogen use efficiency experiments, at the Efaw Experiment Station shows within plot variability consistent with that seen in the other photographs on this page.  This trial was initiated in 1992 and like all our other long-term experiments, each plot receives the same treatment year after year.
  222_99_flowering.jpg (43993 bytes)
Covington_2_99.jpg (54203 bytes) Wheat forage response to fertilizer N applied March 6, using INSEY (left and bottom pictures, Covington, OK, April 8, 1999).  Included in this work are treatments where N is applied topdress at flat rates and to every 1m2 based on predicted yield potential.
Tipton Waving Wheat.gif (58258 bytes) Covington.jpg (47331 bytes)
Is this small-scale variability or 'waving wheat in Oklahoma?' (picture on the right taken in long-term Experiment 502, near Lahoma, OK, May 21, 1999) Spatial Variability .jpg (56943 bytes)