By Larry Dreiling

Over the past few weeks, two organizations have produced estimates of the Kansas Wheat crop. Both groups strive for objectivity and thoroughness. Each has different methods to achieve those estimates.

These estimates are important to producers and grain trade alike as they give an early season glimpse into the harvest season ahead.

When it comes to making Wheat harvest estimates, Eldon Thiessen should know how to do them.

As director of the Kansas Agricultural Statistics Service, Thiessen is charged with creating the state's official U.S. Department of Agriculture Wheat crop estimate through the National Agricultural Statistics Service. He also developed the procedure the Wheat Quality Council uses in its estimate of the Kansas Wheat crop.

While he may participate in creating both estimates, how the two are prepared are as different at Trego and TAM 107.

USDA estimates

The USDA estimate, Thiessen said, is to an exact computer model using highly trained enumerators working with hundreds of cooperating producers. The WQC estimate is done by participants of the tour, whose members include the trade, government officials, media and a few producers.

The method USDA-NASS uses for surveying the Wheat crop is similar to that of the many crops used by the professionals who perform the survey.

"We tell one of our enumerators to meet with a farmer. They then determine a given field that will be part of our survey," Thiessen said. "When we do that, they have to pick a starting corner so that from there, every area of that field would have a chance to be selected."

Using computer mapping, the enumerators are given a random set of points on the field from where to take samples. This grid-like structure increases the random selection factor that is needed when dealing with fields that may have small imperfections in them, but nothing so significant that if taken as a whole could be thought of as a bad field of wheat.

"We go into a driveway of a field, we have to find a corner to start, so that when you walk along the edge of the field and into the field, if we gave you a random number you'd have a potential of being in any area of that field. That way, we'd miss nothing," Thiessen said. "For the procedure the Wheat Quality Council uses, that's not practical."

WQC estimates

The WQC performs its estimate with volunteers who are given a one-hour long orientation session the evening before setting off into the fields.

"What I tell people is to walk out into a field, somewhere between 30 and 50 paces in a random direction. When a car empties out, no one should be following in the same area. They should be going in slightly different directions so that we can pick random spots within that small range that we have time to do," Thiessen said.

"The people on the tour might travel no more than 50 paces. Our enumerators might have 200 paces to 400 paces, depending on the acreage of the field. If it's a whole section, our computer may pick an entirely different end of the field from where you start. Our process is much more random."

Thiessen said he tells tour participants to use a constant number of paces when entering a field.

"I say: 'Don't watch where your feet are going per se. Instead, count your paces and then look down once you've reached your count'," Thiessen said. "From that point, we often suggest they throw their yardsticks out and where it lands is where you want to count."

The speed by which the WQC tour travels and--since it uses no cooperating producers--the type of fields surveyed makes the field selection process less random and therefore more prone to error.

"You really can't be very random the way the tour does it. The rule is we can't enter fenced-in fields--we can't climb over gates and fences--or enter fields with no trespassing signs on them," Thiessen said. "Plus, the tour moves fast. We're trying to get from Manhattan to Colby and then Wichita before finishing in Kansas City in less than four days."

Age of testing models

As one might guess, KASS has been surveying the Kansas Wheat crop for a long time. The earliest recorded survey data goes back to 1918, when the crop totaled 97.71 million bushels and yields averaged 13.5 bushels per acre on about the same number of acres as is the 2005 crop, about 10.1 million acres.

The first recorded May 1 KASS estimate was issued with the 1943 crop.

"Every May, we do a survey. If a field is in pre-jointing, or jointing with no head having emerged, or in late boot, then our enumerators code that field and then they count stalks. At harvest, they'll come back and won't count stalks. They'll count heads," Thiessen said.

WQC issued their first estimates in 1957. It's only been in recent years that WQC has used a solid system to determine their estimate.

"Before 1989, people would just go out into the fields and just make a eyeball guess," Thiessen said. "After I made my first tour in 1987, I said: 'There has to be a better way to do this.'

"That's why I put together this procedure and yield formula model we now use on the tour. In my mind, it gives a better view of what the Wheat crop's potential is plus it gives an education of how our process in at the Statistics Service works."

WQC methodology

Thiessen developed two yield formulas for WQC, one each for crops still in early growth stages and another for late season crops. The early season model requires tour participants to enter a field and count the number of stalks in a given foot of crop row.

The number of stalks is then multiplied to regression model number Thiessen has developed to serve for heads, then multiplied again to a formula number used to measure the average weight per head.

That number is then divided by the row space in inches and multiplied again by a ten-year moving average of kernels in each head. That standard deviation is used when analyzing poor or excellent looking fields.

"I have determined what each field counts as far as stalks in May are compared with the number heads produced," Thiessen said. "From that, I've created a regression model. I used a ten-year moving average on these fields. The model came from the data collected over the last ten years from the objective yield program operated by our agency. Every year the oldest year rolls off and the latest year rolls on. From there, we develop a standard deviation for the weights."

"If you reach a field that has already headed, use the stalk count model. Put the heads into the weight model to determine the yield," Thiessen said.

The formula works like this: Measure the number of heads in a foot of a row. Multiply that by the number of spikelets (The V-shaped cover where grain is produced and held until harvest.). Multiply that number by the number of kernels per spikelet. Divide that by the row width in inches. Multiply the total by a moving average of actual kernel yield.

"What's done on the tour is not as random as what we in Ag Statistics do in our operations."

What not to do

Thiessen also insists that no one should take the average yield the tour develops and place them next to a planted acreage estimate.

"We tell participants to develop a guess of their own field abandonment percentage to calculate into the acreage estimate to come up with a personal forecast for the state's Wheat production," Thiessen said. "This year, we looked like we were on the precipice of having problems because of drought. So you can't use the number of acres announced in November.

"I would call what is done the use of historic data by people to put forward their forecast of what production will be and nothing more than that. The benefit of the tour is allow the participants a better understanding of how the whole process of what the real estimators do. There's training here. Education."

System isn't perfect

Thiessen admits the system WQC uses isn't perfect, but it can offer a pretty snapshot of that moment in time when three to five persons are looking through an unfenced field of growing wheat.

"We aren't going to be perfect in our randomness. We also are not trying to put a number on an individual field," Thiessen said. "We are trying to put together a number that will represent the whole area we drive through. That can be accomplished with the over 400 stops the tour makes."

He also warns against making more out of the numbers than what they are construed to be--a snapshot in time of Wheat crop.

Date: 5/24/05

	Web hpj.com

Copyright/Privacy
Copyright 1995-2008.  High Plains Publishers, Inc.  All rights reserved.  Any republishing of these pages, including electronic reproduction of the editorial archives or classified advertising, is strictly prohibited. If you have questions or comments you can reach us at High Plains Journal 1500 E. Wyatt Earp Blvd., P.O. Box 760, Dodge City, KS 67801 or call 1-800-452-7171. Email: webmaster@hpj.com