COMPARING TOPOGRAPHY SOIL SAMPLING WITH OTHER KNOWN PRECISION AG METHODS

Kevin R. Hollands
Certified Professional Agronomist CPCC-I
CENTROL INC. Twin Valley, MN

INTRODUCTION

In 1996, the use of topography soil sampling greatly increased in the Red River Valley as another Precision Ag option for detecting nutrient variability in soils. Topography maps were generated on approximately 5,000 acres for the use in evaluating drainage and soil sampling by elevation. Topography maps were produced to soil sample by zones on 25 fields totaling over 2500 acres for 1997's sugarbeet crop. Results of elevation sampling in 1996, as in the past, have shown a direct correlation between the available nitrogen and the separate elevations of fields with varying topography. The lower areas of a field have trended lower in available nitrogen and the higher areas have trended higher in available nitrogen.

Satellite imagery recently became available to help visualize the vegetative vigor differences in a crop. These differences can sometimes appear extreme with only minor changes in final yield. It has been suggested that zones could also be derived from these satellite images for soil sampling.

OBJECTIVE

Objectives of the study were to determine how well nitrogen values, obtained from soil sampling using management zones derived from both satellite imagery and topography maps compare to normal grid sampling. Since topography sampling had been ground truthed using soil sampling, and satellite imagery has never been ground truthed in the Red River Valley, it was time to do so.

PROCEDURES / RESULTS

For both satellite image and topography sampling, fields were split into 3-5 different elevations called management zones. Each management zone was digitized by hand, with numerous mouse clicks using a powerful PC, and a GPS file was created. These files, along with GPS units, were then used in soil sampling pickups for locating and soil sampling within the desired zone. A representative sample was taken within each zone to a depth of 42". A range of 10-20 cores were used depending upon the size of the zone.

Please refer to an example of a topography map (Figure 1) produced for Scott Knutson of Fisher, MN. This field is located in the NW quarter of Fisher Township, Section 33. The field was a square 76.3 acre field and the previous crop was wheat. The total elevation difference was 5.04 feet. The legend for this field (Figure 2) describes the various differences in relative feet. The lower right portion of the field map was 0.86 feet above the bench mark while the upper left area was the lowest area of the field measuring -4.18 feet below the bench mark. By using elevations from the topography map, five different management zones were digitized along contour lines and the zones were sampled accordingly. Using the soil test results, a spreader map was created for a variable rate fertilizer applicator. The spreader map (Figure 3) shows the various zones created by the digitizing process. The soil test results table (Table 1) verifies that the adjusted nitrogens of this field were very high 126 lbs/A in the higher elevations and very low 10 lbs/A in the lowest elevations.

Table 1
Fisher 33 topography sampled

To compare most of our available technology, a 144 acre field in the NW quarter of Nesbit Township, Section 18 for Scott Love, Fisher, MN was chosen. This field was picked because of it's varying topography, extremely good yields of 100 Bu/A of barley in 1996, presumed low residual nitrogen levels, and some field drainage problems. A normal 42" soil test was taken across the field for comparison purposes. The results show very low nitrogen in all sample depths (Table 2).

Table 2
Nesbit 18 regular 4' sample

NDSU's deep nitrogen adjustment formula was used to establish total available nitrogen. The value for the 24-42" depth would be -19 lbs/A, and when added to the top 2 feet nitrogen of 21 lbs/A, the sum is +2 lbs/A.

To compare other Precision Ag methods in determining the nutrient variability of this field, it was decided to use regular grid soil sampling, satellite imagery digitized into zones and a topography map digitized into zones. The grid sample results are a good foundation to start from to compare the data since the field was point sampled. There were 42 grid points plotted in the Love field, an average of 3.43 acres per grid. The results of this work (N,P,K, Figures 3,4,5 respectively) illustrate the variability found in the field. Results on 32 grid points exhibited values of less than 0 lbs/A adjusted available nitrogen. The remaining 10 grids had positive nitrogen values with a high of 340 lbs/A adjusted available nitrogen. The average of all the adjusted nitrogen values was 6 lbs/A adjusted available nitrogen, only a 4 lbs/A adjusted nitrogen difference from the normal soil test. The phosphorus ranged from a low of 3 PPM to a high of 39 PPM, with an average of 10 PPM. The potassium ranged from a low of 53 PPM to a high of 808 PPM, with an average of 162 PPM.

Table 3
Nitrogen results from Nesbit 18 (Grid)

Table 4
Phosphorus results from Nesbit 18 (Grid)

Table 5
Potassium results from Nesbit 18 (Grid)

The nitrogen spread map (Figure 8) shows how varying amounts of nitrogen were applied.

Table 6
Nesbit 18 satellite image sample by zones

The satellite image (Figure 9), which is a picture of the various degrees of crop vigor at a certain crop stage, was then used to develop sample zones. This field was segregated into five different zones based on the crop reflections from a June photo. The nitrogen results in the table (Table 6) indicate values with low variability. The adjusted available nitrogens varied from -11 lbs/A to 19 lbs/A. The phosphorus varied from 7 PPM to 22 PPM and the potassium varied from 179 PPM to 280 PPM. The satellite image zone sample did not uncover much of the nitrogen variability in this field.

A topography map (Figure 11) was then produced with an elevation variance of 5.06 feet. The legend for this field (Figure 11) describes the various differences in relative feet. The lower right portion of the field map was 2.01 feet above the bench mark while the upper left area was the lowest area of the field measuring -3.05 feet below the bench mark.

It was decided, after a bird's eye view of this topography map, to sample by four different zones. The adjusted nitrogen values varied from 32 lbs/A in the lowest elevation zone to 138 lbs/A in the highest elevation zone. The phosphorus varied from 10 PPM to 18 PPM and the potassium varied from 236 PPM to 343 PPM. Organic matters were also tested. The results ranged from 5.3% in the high areas and declined with the elevation respectively to a low of 4%. The nitrogen spread map (Figure 13) shows how the nitrogen applications varied across the field. The nitrogen values in the table (Table 7) explains the nutrient values of each different zone.

Table 7
Nesbit 18 topography sample

If we compare the nitrogen spread map of the grid sampled field (Figure 4) and the nitrogen spread map of the topography sampled field (Figure 13) the nitrogen rates applied were very similar across the entire field.

OVERALL RESULTS

As mentioned earlier, 25 fields were sampled this past fall using zones derived from topography maps. The results (Table 8) show how the adjusted available nitrogen consistently followed the topography in 18 of 25 fields. The fields in which in the nitrogens did not follow the topography are underlined in the table. The separate elevations are categorized as low, medium low, medium, medium high, and high. The fields with 3 zones were always labeled low, medium, and high. The fields with 4 zones were labeled depending upon how the field was digitized with different elevation groupings. The fields with 5 zones obviously covered the entire spectrum of the zone designations.

The Phosphorus results did not show any particular trends or patterns across the different elevations. The Potassium did show a slight trend of declining levels with increased elevations. There was not enough organic matters tested to establish any definite trends.

Table 8
Nitrogen Levels across all 25 fields topography sampled in 1996

Of the 7 fields which had nitrogen values which did not follow the topography:

2 fields had their lowest elevations on sloping hillsides which tested high in adjusted nitrogen
1 field had 5 zones with all adjusted nitrogen values 0 lbs/A or less
1 field had 15 feet of fall from side to side, therefore all steeply sloped
3 fields had only 1 zone out of place as far as the nitrogen values following elevation

CONCLUSION

The satellite imagery soil sampling by zones did not show the variability as well as the other two methods. If satellite images are to used for zone sampling, more experimentation needs to be done.

Topography maps proved to be a very valuable tool in determining how to soil sample more precisely. Sloping hillsides (lowest elevations with high nitrogens) are areas normally not soil sampled. Since these areas do exist in some fields, topography sampling enables this method to discover the nitrogen differences and apply fertilizer rates accordingly. Even though there were seven fields sampled which the nitrogen values did not follow the topography, these fields suggest that the particular slope of an area is probably just as important as the elevation itself. Since elevation can not change without some degree of slope, elevation and slope are inter-related.

In fields with nitrogen variability and limited changes of elevation in the range of 2-6 feet, topography sampling appears to be working very well. When comparing fields which have been topography sampled and grid sampled the same year, topography results using 3-5 zones have almost mirrored grid results using 4 acre grids. Elevation sampling accuracy could undoubtedly be improved if the number of zones sampled were increased. Since no two fields are alike, choosing the number of management zones in each field to be sampled is one of the biggest challenges faced with this technology.

ACKNOWLEDGEMENTS

1. A special thanks to the growers Steve Williams, Scott Knutson, and Scott Love for their individual field information.
   
2. Soil-Teq's software was used for making the spreader maps.
   
3. A special thanks is due Mike Freeberg & CENTROL's Mapping Center for assistance in completing this research.

1996 Sugarbeet Research and Extension Reports. Volume 27, pages 98-110.

Questions/Comments