Presentation: A New Look at Location Yield Data

Introduction

When multi-year and multi-location data are analyzed, they are most often summarized on a geographic basis over years. This combines sites and years that are in a geographic area despite production levels.

The fallacy of this becomes evident when drought, soil fertility or heavy rainfall makes a site or year significantly different than the norm. The use of this type of averaged data doesn’t pick varieties that are best adapted to low or high management conditions. In reality not all fields on a farm are capable of the same level of production, yet producers are given the impression that in their geographic area, varieties will respond in accordance with the long-term average.

We set out to determine if comparing varieties by level of production of the test compared with a designated check variety or the test mean would give new information. In particular we wanted to determine which varieties did best under low-yielding conditions and which varieties responded well to high input levels to give the best results under high-yielding conditions.

Materials and Methods

The Field Crop Development Centre (FCDC) has developed a SASŪ-based (SAS Institute, Cary, NC) program called the ‘FCDC Data Miner’ that allows the search of all data bases across years and locations and has the power to present the summary of the data in either geographic zonal format or by yield level of the test that the selected varieties are in.

We also developed a separation of data for malting varieties into locations which produce samples in the protein range for selectable malt.

The Data Miner program allows the breeder to compare yield data in a variety of formats in a matter of minutes. It will also graph the results. The program can perform analysis for yield, agronomic (test weight, 1000 KWT, maturity, height, whole plant biomass) and quality traits.

Production areas in Alberta
The Province of Alberta is traditionally broken into Agro Climatic Zones (ACZ) by soil type, precipitation, and length of growing season.

• Areas 1, 2, and 3 are dry areas of brown to transitional black soils.
• Area 4 has deep black soil with high levels of precipitation.
• Areas 5 and 6 have black to grey-wooded soils and have short growing seasons.

Results and Discussion

Traditional area analyses may combine normal and abnormal years (Table 1, Figure 1). Harrington and CDC Dolly are very stable-yielding over most Alberta production areas, while AC Metcalfe, Merit and Xena do especially well in Areas 2 and 4.

Table 1. Yield of 2-row barley varieties by production area

Figure 1. Yield of 2-row barley varieties by production area

Non-traditional area analyses, based on yield potential of the test, can reveal responses to high and low production systems (Table 2, Figure 2).
When analysis is done on a yield basis, yields of Harrington actually decline under high production, while yields of the other varieties continue to increase.

Table 2. Yield of 2-row barley varieties by yield class

Figure 2. Yield of 2-row barley varieties by yield class

• Non-traditional analyses based on protein classes can reveal the yield potential of 2-row malting varieties when grown to make malting grade (Table 3, Figure 3).
• When yield data are sorted by protein content, AC Metcalfe and Merit malting types actually have their highest yields at the 11% to 12.5% protein class, while Harrington yields best when protein is too low to make malting grade.
• Non-traditional yield analyses can also reveal the potential of varieties under production systems for both 6-row hulled (Figures 4 and 5) and hulless barley (Figures 6 and 7).

Figure 3. Yield of 2-row barley varieties by protein class


Figure 4. Yield of 6-row barley varieties by production area

Figure 5. Yield of 6-row barley varieties by yield class

Figure 6. Yield of hulless barley varieties by production area

Figure 7. Yield of hulless barley varieties by yield class

Conclusions

• While data analysis based on traditional areas of production do reveal cultivar differences in adaptation, there is a masking effect due to taking means over high and low yielding sites and years
• Analysis based on yield class allows a non-traditional look at yield data that reveals cultivar response to both high and low yielding production systems.
• The Data Miner also allows for other non-traditional analyses, such as by protein class in 2-row malting barley, that can be important in making breeding and production decisions.