Evidence for a relationship between soil nitrate-N levels and topography and comparison of topography and grid soil sampling for soil nutrients.

D.W. Franzen, V.L. Hofman and L.J. Cihacek.
North Dakota State University


Sugarbeet tonnage and quality are affected by soil N availability. Currently, site-specific management of soil N levels is being conducted by sugarbeet producers using 4 to 5 acre grids and variable-rate N application. In many cases, additional nutrients, such as phosphates, are also being variably applied. Although previous studies have shown that 4 to 5 acre grids may increase economic return to sugarbeet growers through variable-rate N application (Smith and Rains, 1995, 1996; Cattanach, Franzen, Stanislowski and Sax 1996), it is feared that failure of sampling to describe soil fertility levels and their boundaries correctly may result in poor performance of variable rate application (Franzen and Peck, 1994). Also, too many samples needed to describe field variability adequately enough to ensure some level of variable-rate N application success for a producer may lead to lower economic returns than suggested presently by profitability research. This study is being conducted to investigate the ability of different grid sizes and topography sampling to reveal soil nutrient levels within fields, with the fewest samples possible.


Two, square forty acre fields were sampled in a 110 ft. grid. The first site is located southwest of Gardner, ND. The Gardner site is composed of Fargo/Hegne soil series, and has two surface drains running east to west in the north, and west to east in the south. Gardner was sampled in 1995 and 1996. The north 15 acres has been in alfalfa for the length of the study, while the south 25 acres was in spring wheat in 1995 and barley in 1996. The second site is located south of Colfax along the Galchutt blacktop, west of I-29 about 3 miles. This site is composed of Glyndon/Wyndmere silt loams in the lower ground and Embden/Tiffany sandy loam in the higher ground. The Colfax field was sampled in 1995. Both sites were sampled to 4 ft., separating the cores into 0-6 inch, 6-24 inch and 24-48 inch depths.

Each field was mapped for elevation using a laser-surveying device. Both elevation and soil sampling locations were recorded using a differential global positioning satellite receiver (DGPS) with about 3 meter accuracy.

Grid density was compared by deleting sampling locations from the original 110 ft. mapping until the desired density was reached. Maps were then developed of the less dense grids using inverse distance squared weighting. The estimates were then recorded at each of the original sampling locations. These estimates were compared to the original figures, leaving out common numbers used to make the less dense grid maps to avoid auto-correlation.

Topography sampled maps were developed first by selecting the boundaries of topographic differences between high and low elevations locally within each field to produce polygons representing similar topographic areas. A soil test value was then selected randomly from the original 110 ft. data pool within the polygon area and all 110 ft. sampling locations within that polygon were given that soil test value. Correlations were made between the 110 ft. sampling values and the estimates from the topographic sampled data set.


The first evidence of topographic relationship with soil N levels was found when comparing between years. Figure 1 shows soil N levels from Gardner between 1994 and 1995. If soil nitrate-N levels are randomly distributed in the field from year to year, then the probability of high areas being high and low N areas being low in different years, with different crops and different climate would be very low. However, at all of our locations there are strong similarities in soil N values from year to year. This suggests that some underlying soil features must help to control soil N levels. Since topography affects both soil water content and soil organic matter accumulation, then local elevation differences may be an important factor affecting soil N levels through soil N mineralization, crop removal, subsoil water movement, N denitrification patterns and nitrate leaching and movement.

The second piece of evidence of topographic influences in soil nitrate N levels is shown in the comparison of Gardner nitrate-N levels in 1995 with an enhanced topography map of elevation (Figure 2). Higher soil N levels generally follow areas of higher elevation in this wetter than normal year. Lower N levels follow lower elevation levels, especially in the surface drains, possibly due to increase denitrification during the year. The Colfax nitrate-N levels also follow trends in elevation. Low soil N levels are found generally in the lower landscapes, while higher soil N levels are found in the higher landscapes (Figure 3).

The third piece of evidence is the correlation of topographic sampling with the 110 ft. grid compared to different grid sizes, including 220 ft., 330 ft. and 5 acre grids (Table 1). Topography sampling at Gardner was determined using 6 sampling polygons, while the Colfax topography sampling was developed using 5. Compared to the sample per acre (220 ft. grid) or the sample per 2.5 acre (330 ft. grid) density, topography correlation was not as high in Gardner, 1994 or at Colfax in 1995. However, the Gardner 1995 topography was higher than the 330 ft. grid correlation. At all three site years, topography was better than a 5 acre grid.

Table 1. Correlation (r) of topography sampling and various grid size nitrate-N estimates with 110 ft. nitrate-N levels.
Site, Year Topography 220 ft. grid 330 ft. grid 5 acre grid
Gardner, 1994 0.28 0.51 0.35 0.16
Gardner, 1995 0.30 0.39 0.22 0.04
Colfax, 1995 0.32 0.62 0.45 0.06

Topography may not solve all nutrient variability questions. Figure 4 shows zinc levels at Colfax in 1995. The boundary between high and low zinc is a straight line running east and west across the middle of the field. This suggests that zinc fertilizer may have been applied to the south half of the field, minimizing the influence of topography in explaining field zinc levels. Caution should be used, including additional sampling, when entering a field the first time to sample for variable-rate application.


Topography sampling appears better correlated than a 5 acre grid at representing field nitrate-N levels. Soil nitrate-N levels appear similar between years, suggesting that once identified, landscape features may predict soil nitrate-N levels with some level of confidence. Caution should be used when first attempting topographic sampling because of the possible influence of merged fields, past manure application fertilizer history in affecting soil nutrient levels, especially if non-mobile nutrients such as phosphate are important to the producer. A combination of gridding and topography when a field is first sampled may help sort out these questions.

The following guidelines may be helpful in determining a field sampling strategy:

Candidates for field grid sampling -

- Field history is unknown.
- Soil fertility is high, or has been treated with high levels of fertilizer or lime recently.
- Has a history of manure application.
- Fields have been merged.
- Non-mobile nutrient levels are important to know.

Candidates for topography based field sampling

- Yield monitoring or remote imaging suggest landscape trends.
- There is no history of manure (includes old farmsteads).
- Soil fertility levels of non-mobile nutrients are generally low.
- Mobile nutrient levels, especially N, are most important to know.


Cattanach, A., D. Franzen, H. Stanislowski, and L. Sax. 1996. pp. 113-116. IN:1995 Sugarbeet Research and Extension Reports. NDSU Ext. Serv. Fargo, ND.

Franzen, D.W. and T.R. Peck, 1995. Field soil sampling density for variable rate fertilization. J. Prod. Agric. 8:568-574.

Smith, L. and D. Rains. 1995. Grid soil testing and variable rate spreading for profitable sugarbeet production-Fact of fiction? pp. 181-185. IN:1994 Sugarbeet Research and Extension Reports. NDSU Ext. Serv. Fargo, ND.

Smith, L. and D. Rains. 1996. Grid soil testing and variable rate fertilization for profitable sugarbeet production-Part 2. pp. 113-116. IN: 1995 Sugarbeet Research and Extension Reports. NDSU Ext. Serv. Fargo, ND.

1996 Sugarbeet Research and Extension Reports. Volume 27, pages 111-117.