Soil Sampling as a Basis for Fertilizer Application (continued)

SF-990 (Revised) August 1998

Soil Sample Collecting, Where and How (continued)

Directed sampling

Landscape/topography sampling

A more practical approach for North Dakota producers that combines low cost with a high degree of meaningful nutrient information is directed sampling. Directed sampling is based on some prior knowledge of the field, or some logical basis. The basis of most North Dakota directed sampling is the effect of landscape position on soil nutrient levels, particularly nitrogen. Soil pH, P, K, and Zn are non-mobile factors or nutrients in soil. The levels and patterns of non-mobile nutrients within fields are similar from year to year. North Dakota research has also shown that patterns of NO₃-N, S and Cl, which are mobile soil nutrients, are also stable between years because the patterns are affected by the landscape (Figure 8). Directed sampling based on landscape, or topography, has been shown to be similar to a one-sample-per-acre grid in providing within-field nutrient levels while requiring only a fraction of the sampling time and expense. Topography sampling of several fields across North Dakota only required four to seven samples per 40-acre field, compared to 36 for the one-sample-per-acre grid approach.

Figure 8. Valley City NO₃-N levels, 1994-1995.

Additional methods for directed sampling

Directed sampling should be considered an iterative process (a process that takes more than one attempt) in which information is added progressively to the general knowledge of the field. Producers will not have the 110- foot grid sample research base researchers at NDSU have to back up assumptions on where important management zones are located and where the boundaries might be. Several methods of determining management zones should be used in addition to topography to help the producer judge what areas are important.

Aerial photography and the use of satellite imagery can be used to show differences in soil color and differences in crop growth patterns and crop color. In years that are very dry or very wet, these areas will probably be related to topography. Aerial photography and satellite imagery has been shown to reveal patterns in sugarbeet leaf color which is especially useful to soil samplers.

Old FSA (ASCS) aerial photographs in slide format are available for most fields in North Dakota because of verification photography taken over the last 20 years of federal farm programs. These photographs may not only reveal past field boundaries and long-gone building sites, but provide patterns from past crops that align with present-day information. This information is inexpensive to acquire and can be scanned into computer software for use in decision making.

Yield monitor data may be useful to define some boundaries However, so many factors affect yields that unless yield is mainly affected by a single nutrient of interest in one year, most yield patterns may cross over several soil fertility levels. Yield monitor data has been most useful in recent North Dakota studies by identifying particularly poor yielding areas. These areas may have abnormally low fertility levels causing the poor yields, or they may have unusually high fertility levels if another factor is limiting yields, resulting in accumulation of excess nutrients in that location.

On-the-go soil electrical/electromagnetic soil conduc-tivity sensors may help define management zones. It is not possible to determine directly what the different levels of conductivity mean without sampling to ground-truth the areas, but they can reveal patterns that initially direct, reinforce or redefine existing layers of information regarding zonal boundaries.

Digitized GIS soil survey maps should be used with caution and not without the aid of other layers of information. Although soil surveys give generally reliable information useful in determining the general productivity of farms, the information is usually not fine enough in scale to direct site-specific decisions.

Pros and cons of different within-field sampling methods

The following are criteria for choosing grid sampling over a directed sampling approach:

• The field history is unknown
• Fertility levels are high due to high rates of fertilizer application.
• There is a history of manure application.
• Small fields have been merged into large fields.
• Non-mobile nutrient levels are of primary importance (P, K, Zn).

The following are criteria for choosing directed sampling methods over grid sampling:

• Yield monitor data or remote imaging show a relationship with landscape.
• There is no history of manure application.
• Relatively low fertility levels are present, or low fertilizer rates of non-mobile nutrients (less than maintenance) have been applied over the most recent years.
• Mobile nutrients, especially N, are important to map.

Another strength of the grid approach is that the procedure requires a lower level of interpretive skills by the sampler. Grid locations are imposed on a field map by the computer with a prompt to drive to the next location. Anyone who can drive and read a map can sample a field in a grid. The drawback is the expense of sampling and analysis, which may result in a less than adequate grid size needed to represent a field.

Directed sampling requires a much more intelligent approach. By using the zone method, either the sampler or the sampling supervisor who provides the sample location map to the sampler must have a high degree of agronomic savvy. It takes time to review aerial photography, satellite imagery, topography maps, and other layers of information, manipulate the maps to look for complementary patterns between different layers, and decide where the best management zone boundaries are located. Although the sampling and analysis expense of a directed approach to soil sampling is far less than a one-sample-per-acre grid approach, the expense of interpretation is considerably higher.

The value of determining within-field nutrient levels

Determining within-field nutrient levels allows the variable-rate application of fertilizers. When considerable variability is present, immediate economic returns are possible, provided the variability is on a portion of the yield/nutrient curve which allows increased yield or quality if application rates are varied. The rapid movement toward variable-rate N application in sugarbeets has been driven by the relationship between N levels and crop value.

Another important reason for determining within-field nutrient levels is to reveal the range of levels and location of the levels. In determining soil pH, for example, composite tests from 95% of North Dakota fields show a pH level of greater than 7. This led one author to announce that North Dakota "does not have an acid soil problem." However, in site-specific studies on five fields, three in the Red River Valley and two outside, the three fields in the Red River Valley all contained small areas (2-3% of total area) with values less than 7, and the fields outside the Red River Valley contained over one-half of each field area with pH levels less than 6. In one field, pH varied from 4.9 to 7.8. The pH ranges have implications on herbicide carryover, herbicide activity and the performance of some major crops with pH sensitivity. So what is the level of pH on the 20 million acres of cropland west of the Red River Valley? We really do not know, but perhaps up to half of these acres may have pH values lower than 7.

When a soil test shows high levels of nutrients in a composite soil test, does that mean that the whole field does not require fertilizer? Many producers have run on-farm tests in the past and have found that applying nutrients when composite soil test show that none are needed results in yield increases. Some producers simply do not trust a composite soil test. Sampling in a more intelligent manner using a directed approach should provide more accurate soil test results. The high soil test areas will be separated from the rest of the field, and areas needing fertilizer will be revealed. Whether or not variable-rate fertilization is used, more confidence in the soil test will result.

Sample core number and confidence in the sample value

Cell sampling or point sampling can be used to gather soil from a grid or management zone. Cell sampling (Figure 9) is a method where samples are gathered randomly from the grid or zone area, while point sampling limits the sample collection area to a 10-20 foot radius from a central area location. Point sampling is most often used in grid sampling, whereas cell sampling appears to better represent zone levels. Both methods require multiple soil cores. There is enough small-scale variability in most areas of fields that single cores are not likely to represent a grid or zone well (Table 3). Research on small-scale variability suggests that eight to 12 soil cores may be required to represent a grid or zone.

Figure 9. Cell sampling and point sampling.

Table 3. The percentage of composite NO₃-N values falling into a range of the mean � 20% with the cores taken in a random manner throughout a 60 foot X 60 foot plot area with individual sample cores obtained in a ten-foot grid. (Franzen and Dennis Berglund, 1997).

Once a sample value is obtained through careful sampling and analysis, what does that value mean and how much of the grid or zone is represented by that value? The lower the sample value is for NO₃-N, the more confidence there is in the value. For example, in a 10-acre zone, if a value is 20 lb NO₃-N/acre, then it would be expected that 9.9 acres of the zone test 10-30 lb NO₃-N/acre. However, if the value in the 10 acre zone were 100 lb NO₃-N, then only about 6.5 acres would test between 70 lb and 130 lb NO₃-N and the remaining 3.5 acres would have values above or below that range.

Some producers have become disillusioned with determining within-field nutrient levels because on closer inspection some areas have small-scale variability as great as the variability in the entire field. However, careful analysis of a field shows that even though some areas have extreme variability, most others do not.

Consider a 100-acre field with a NO₃-N composite test of 80 lb/acre (Figure 10) with a range from 10 to 200 lb NO₃-N/acre. Forty acres has a test level of 30 lb, 20 acres tests 50 lb, 20 acres tests 80 lbs, and 20 acres tests 120 lbs. Using the 80 lb/acre composite test over-fertilizes 20 acres while under-fertilizing 60 acres. The within-field method fertilizes 40 acres in the lowest category correctly, fertilizes 18 out of 20 acres testing 50 lb correctly, 15 of 20 acres at the 80-lb level correctly and 10 of 20 acres at the 120 lb level are fertilized correctly. The composite sample only fertilized 20 acres out of 100 correctly, while the within-field sampling method fertilized 83 acres out of 100 correctly. Even though some areas of the within-field approach were highly variable, the majority of the field benefits from revealing within-field variability.

Figure 10. An example of the properties of a field correctly and incorrectly fertilized with N using either a composite soil test-based or site-specific soil test based N approach.

Summary

Soil testing is the basis for fertilizer recommendations in North Dakota. A composite soil sample is a good first step in understanding relative levels among fields. Within-field management of nutrients based on grid sampling or directed sampling may inspire more confidence in soil test recommendations and provide more accurate field nutrient level information.

A composite field test requires from 20 to 30 cores to represent a field. By sampling three to four zones in the field, each with eight soil cores, the time spent sampling in the field and the cost of analysis is only increased a small amount, while the information gathered about the field is greatly increased.

Sampling should be considered seriously and soil samples handled properly to provide consistent results. Producers would not dare go to the field without checking the oil in their tractor engines. One should approach soil testing in a similar manner.

|  index  |

SF-990 (Revised) August 1998