Managing Fields with Isolated Low Fertility Areas

 

Jessie Peone and J. Lowenberg-DeBoer

 

Introduction

            Several studies in the 1990s showed that that variable rate application (VRA) of phosphate (P) and potash (K) is profitable on some fields and not on others. Some researchers tried to link profitability to the variability of P and K soil tests in the field. Unfortunately, soil test variability alone does not explain VRA profitability because many Eastern Corn Belt fields have high soil tests that are highly variable above the critical levels, with only small, isolated areas of low soil test. For P the critical level is about 30 lbs/acre and depending on the CEC the K critical level in the range from 175 to 300 lbs/acre. The key to profitability of VRA P & K appears to be soil test variability in the responsive range below the critical level.

            The objective of a recently completed study at Purdue University was to estimate the proportion of a field with low soil tests required for VRA application to be profitable. In a spreadsheet analysis researchers tested if a tenth of an acre was enough justify VRA, or if it required 5%, 10% or even 20% of the field. This article summarizes the key results. More information is available in Peone (2004).

                       

Methods

            A spreadsheet analysis was developed for a 10 acre field in southwestern Indiana that had been soil tested at 9 samples per acre. This means each sample represents about 1/9th of an acre (0.11 acre). Ninety samples were taken from the 10 acre field. This field was chosen because it had P and K soil tests that were highly variable, but all above the critical levels. Also 9 samples per acre allowed a more detailed view of soil variability than would be available with the conventional 2.5 acre grids. While 9 samples per acre resolution is very costly with manual sampling the data can be understood as being similar to what would be available if soil sensors were developed for P & K.

            Results with 2.5 acre grids were estimated by grouping the 1/9th of an acre cells in to rectangular areas of about 2.5 acres and using the average soil test of the four center cells to approximate a center point grid sample.  A whole field composite sample was approximated by averaging the soil test in five cells in a “X” shape. 

            The average P soil test for this field was 106 lbs/acre and the minimum cell tested 31 lbs/acre. The average K soil test for the field was 380 lbs/acre, and the minimum cell tested 250 lbs/ac. With an average CEC of 6 and a CEC range of 2 to 12, the K critical level is from 175 to 200 lbs/acre. Corn and soybean response functions developed in Ohio were used to estimate yields at each P and K level.

            Two decision rules were used: the Tri-State Fertilizer Recommendations (Vitosh et al, 1995) and the economic rule of setting marginal value equal to marginal cost. The study was implemented with two decision rules to test the robustness of the result. Because farmers use these and a variety of other rules of thumb in making fertilizer decisions, it was important to know that the result did not depend on the decision rule being used. The strategies to be discussed here are:

 

1)      Whole field management (WFM) – a uniform application of P and K is done based on a single “X” shaped composite sample of the field.

2)      Site-specific management (SSM) – P and K applications are varied by grid based on a center point sample. Two grid sizes were tried; the original 1/9th of an acre and 2.5 acres.

 

            The test was done by systematically lowering the soil test of the lowest soil test cells in the field into the responsive range for P alone, K alone and P & K together to see if it affected the choice of WFM or SSM. In each case the test was done for one cell testing low, two cells, 5% of the field, 10% of the field and 20% of the field. For P, soil tests were lowered to 10 lbs/acre. For K, soil tests were lowered to 100 lbs./acre. For example, in the P alone test, the lowest P soil test cell, was 31 lbs/acre. The P soil test value of this cell was lowered to 10 lbs/acre, the spreadsheet recalculated and net returns recorded. Then the next lowest P test cell was identified and the process repeated.

            Whole field soil testing was costed at about $1.00/acre. The 2.5 acre grid soil samples were charged at $6.19/acre the average in the 2003 Purdue Ag Retailer Precision Ag Survey. Based on $10/hour labor and $7.50/sample lab fees, the 9 samples per acre  were estimated to cost $90.91/acre. Variable rate application was costed at $7.12/acre based on the Purdue Precision Ag Survey and uniform application was estimated at $4/acre.

 

Results

            The results show that the profitability of SSM for managing fields with isolated low soil fertility areas depends on:

 

1)      the soil sampling procedure and

2)      the cost of soil sampling and variable rate application.

 

Using the procedure outlined above, the WFM strategy did not apply any P or K fertilizer even when 20% of the cells tested low for either the agronomic or the economic rule. This was because the low soil test cells did not happen to fall in “X” shaped composite sample.

            For the 2.5 acre grids, no P & K fertilizer was applied when 5% of the field or less tested low because few of the low test cells were picked up in the center point samples. Above 5%, some fertilizer was applied by the SSM strategy, but at baseline soil sample and VRA costs SSM was never the most profitable choice.

            With 9 samples per acre at current soil testing and VRA costs, SSM becomes profitable at around 20% of the field testing low (Fig. 1 & 2). The figures show the difference between SSM and WFM for the agronomic decision rule. Results are similar for both the agronomic and economic decision rules.

            If 9 samples per acre soil tests could be done for the current cost of 2.5 acre grid sampling, then SSM would become the most profitable alternative at about 5% of the field testing low (Fig. 1 and Fig. 2). When both P and K test low in some areas, the field percentage required is only about 3%. If soil testing and variable rate application were free, SSM would be the preferred choice even if only one 1/9th of an acre cell in the field tested low for either P or K.

 

 

Conclusions:

            This study indicates two key factors in determining whether SSM for P and K will be profitable for a field: 1) whether the soil sampling is dense enough to pickup the low test areas, and 2) the cost of soil sampling. Soil sampling on a coarse grid (e.g. 2.5 acre) may not reveal the low soil test areas. If soil sensors or other technologies could provide dense soil testing with multiple samples per acre for the current price of 2.5 acre grid sampling, then SSM for P and K would be a profitable alternative when about 5% of the field tested low for either P or K.

            The next step in this research would be to test the robustness of these conclusions with soil test data from other fields and with other response curves. Ultimately, the real test will be to determine if rules of thumb (e.g. VRA is profitable if 20% of the field tests below the critical level) actually produce profitable results. A critical issue is the density of soil testing. To specify the rule of thumb in terms of 20% of 2.5 acre grid samples is different from specifying the rule of thumb in terms of soil sensor data or multiple samples per acre.

 

For more information:

 

Peone, Jessie, “Management Strategies for Corn-Soybean Fields with Isolated Low Phosphorous and Potassium Areas,” MS Thesis, Department of Ag. Economics, Purdue University, West Lafayette, IN, 2004.

 

Vitosh, M., J. Johnson, and D. Mengel, “Tri-State Fertilizer Recommendations for Corn, Soybeans, Wheat, and Alfalfa,” Michigan State University, Ohio State University, and Purdue University, Extension Bulletin E-2567, July, 1995 (http://www.agry.purdue.edu/ext/forages/publications/ay9-32.htm).