Sources of Yield Variation Observed in Yield Maps

From year to year and field to field, the factors that impact crop yield and the degree of their impact will vary. As well the factors interact with each other in space and time. For example a sand soil which holds less water for the crops (reducing yield in a dry year) also has lower nutrient levels and often a lower pH. The interactions can either minimize an effect from a single factor or cause a more extreme impact. For example a sandy soil in a dry year has a much greater impact on crop yield than during a normal year. However if the spring was cold and wet, then the sandy soil will warm sooner ensuring a better germination. Remember that two low yielding areas might have low yields for completely different reasons - factors that limit yield will vary from place to place in a field. To further complicate the problem the effects of limiting factors may interact with weather and management. Possible sources of variation are given in Table 1.

Table 1: Sources of Yield Variation in a Field

Naturally occurring

weather  precipitation: amount , frequency and distribution, temperature, solar radiation: CHU units,  wind
soil- water relationships drainage, soil depth , water holding capacity
soil physical and chemical properties  texture: sand, silt, clay, structure and bulk density, topsoil depth, nutrient availability, pH, organic matter, cation exchange capacity,
slope and aspect (N,S,E or W) of a site  effects on soil erosion, soil temperature, machinery operations,
crop pest infestations  weeds, insects, diseases

Management induced 

crop inputs  hybrid or variety selection (yield potential),  plant population and uniformity
field history  herbicide/pesticide,  fertilizer/ manure inputs,
cultural practices and/or mistakes  crop rotation, tillage and compaction, previous practices, manure applications, land leveling , ditch cleaning, misapplication of nutrients or pesticides, planter, cultivator or harvesting problems

 

Recognizing Patterns on Yield Maps

It should be possible to divide a field into sub-regions or management zones in which the crop yield is affected by the same factor or combination of factors. In general, crop yield variations due to natural causes appear as irregularly shaped areas, management induced variations are often seen as regular geometric shapes and equipment problems often result in high or low yielding streaks or lines on the yield map. Table 2 gives a more complete description. Patterns in yield caused by natural variation in soil texture and by management are shown in Figures 1 and 2.

However, remember that the management zones will not necessarily perform consistently from one season to the next. For the producer to be able to use management zones, they have to show stability from year to year and from crop to crop. That is, zones should be equally high or low for all crops. As many producers do not have long term yield data to analyze and examine for consistent patterns the development of these management zones is limited by the lack of information. In some cases these patterns in crop yield may be influenced by the weather (rainfall distribution over the growing season) and management zones should be separated into wet and dry years.

Table 2: Possible Explanation of Patterns on Yield Maps

Pattern on Yield Map

Possible Explanations 

Natural patterns 

slope/aspect differences, topography, border shading from trees linear irregular areas
soil drainage, change in soil properties, pest infestations, nutrient availability irregular non-linear areas
depressions which affect drainage or organic matter round or elliptical areas
spreading pest infestations, soil quality changes, non- uniform plant population patches or spots

Management induced patterns 

different cropping history, hybrid or variety change, tillage or compaction effects, nutrient/pesticide changes rectangles or abrupt boundaries
irrigation problems circles, arcs or triangles
nutrient/pesticide skip or overlap, planter or cultivator malfunction, combining error, dead furrows, on-farm testing (nitrogen, hybrids etc), tillage, compaction or crop history streaks or lines

Yield monitor induced patterns 

inaccuracies in differential GPS signal the harvest pattern drifts away from the straight line 
loss of GPS signal missing data points
jamming (grain) when entering and exiting field "sawtooth" pattern

 

Analysis of Single Year Yield Maps

If only one year of a yield map is available for the field, it is still possible to evaluate the yield patterns and propose corrective measures. First, gather all the information together about that field - planting time, fertilization rates and times, tillage operations, pesticide applications, harvest situation (poor weather conditions), known problems (dry spring - poor germination, crusting), previous crop (crop rotation) etc. Additionally obtain the soil test reports, record weed observations, when was manure applied (uniformly or spot applications), land leveling, tile drainage and the rainfall distribution. If available, aerial photos of the field can help locate depressions, old ditches or changes in soil type etc. The more knowledge or information that is available the better the interpretation of the yield maps - at the end of the season.  Once all the information is at hand, look at the yield map and identify the patterns. Relate the patterns to the field data and history that was collected. The most difficult step is to determine the known or suspected factor that is influencing (limiting) the yield.

Finally, answer these questions.

  1. What is the nature of the yield variability?
  2. Can I identify the factors that limit yield?
  3. Does that variability correspond with any of the other factors?
  4. Are the key yield-limiting factors ones that can be managed?
  5. If so, is the management (correction action) cost effective?
  6. Would the correction action be environmentally beneficial or benign?

Figure 3  (Yield Variability Factors and Remedial Action, after T. Doerge, 2000) gives some of the possible causes of yield variability and the possibility of correction - not all problems can be corrected (for example the weather) and some problems are too expensive to be cost effective (spot drainage problems). The diagram should be used in conjunction with the producer's knowledge of the field and the field history. For example problems due to compaction or drainage may vary according to natural soil conditions and precipitation (heavier rainfall areas) and according to the field history (cannery operations). The factors in Figure 3  should be evaluated according to the region. Some of the factors that affect yield may have a greater impact in some regions compared to others, for example, the seeding date which is critical in short season climates but less so in areas with a longer growing season.

Identifying Yield Variability from Multiple Years

In the ideal situations, the main yield limiting factor or factors would be consistent from year to year and crop to crop. If that were truly the case, interpreting multiple year yield maps would be the same as a single year map. Unfortunately what may be a yield-limiting factor to one crop or in a given year does not necessarily have the same effect on a different crop or in the next year. Even the same crop grown two years in a row will often exhibit spatial instability (Figures and ).

One approach is to compare yields from either the same crop or different crops by using normalized yields. The normalized yield is obtained by dividing each yield sample point by the field average. Normalized yields are expressed as a percentage of the average yield of the field and can be used to compare spatial yield patterns across different crops and years. Thus a yield of 125% is actually 25% greater than the field average while any area less than a 100% normalized yield is not reaching its full yield potential. This allows different crops to be compared. Corn can have a maximum yield of 14 t/ha and a corn yield of 4 t/ha would be considered low. For wheat 6 t/ha would be high yield and 3 t/ha would be considered average. The normalized yield compares percentages of yields so that a 25% increase over the average would be the same for either corn or wheat for that field in a each year. Table 3 shows crop yield statistics for five years.

Table 3: Crop yield and Normalized yield for one field (Refer also to Case Study A)

year and crop

Field yield average, t/ha

Maximum yield, t/ha

Minimum yield, t/ha

normalized maximum yield % (as a percent of the field average)

normalized minimum yield % (as a percent of the field average)

1997 corn

8.3

14.9

2.5

180%

30%

1998 wheat

2.8

6.2

0.9

221%

32%

1999 corn

8.6

15.0

2.5

174%

29%

2000 corn

6.4

15.5

2.0

242%

31%

2001 corn

8.3

15.9

2.0

191%

24%

 

One methodology uses normalized yield data from multiple years and different crops to subdivide the fields into management zones or classes based on yield ranges (high, medium and low) and yield stability. The four classes are high yielding and stable, medium yielding and stable, low yielding and stable and the fourth class contains all areas that show no consistent pattern - they tend to increase or decrease differently from one year to the next. Each of these classes requires a different management approach - high to medium yielding and stable areas should be examined to determine if any input such as nutrients, seeding rate and pest control is restricting a potentially greater yield. In the low yielding and stable areas, the yield-limiting factor should be determined. If it can profitably be corrected then this is the best course of action otherwise, the producers may be able to reduce inputs without reducing yields. If a crop cannot use all of the nutrients then there is no benefit to applying the maximum amount. The unstable areas are the most difficult to interpret and manage. These areas should be examined according to the crop grown - are the areas unstable for all crops? Was it due to lodging, weed patches, poor germination etc. For example, sandy, well-drained areas in the field tend to yield well in seasons when wet conditions were present at seeding, and where summer rainfall was plentiful. Areas with heavier and/or poorly drained soils may have done poorly in these years. However, in a very dry year, or a year where soils were already extremely dry at seeding, the sandy areas would under-perform relative to the areas of heavier soil. These areas would show "unstable" yield ranges from year to year.

If an area of the field is lower yielding with different crops, it is likely a poor area and should be sampled to determine the cause. If an area is high yielding with one crop and low yielding with another, then consider what could reduce yield for one crop, but not affect the other?