Corn as an Alternative Crop for Direct Seeding:

Evaluation through Crop Modeling

 

Tim Fiez, Dept. of Crop and Soil Sciences, Washington State University

Javier Marcos, Dept. of Crop and Soil Sciences, Washington State University

Gaylon Campbell, Dept. of Crop and Soil Sciences, Washington State University

Claudio Stockle, Dept. of Biological Systems Engineering, Washington State University

 

Time and cost efficient development of new cropping systems must be based on a fundamental knowledge of crop development and adaptation. With the 1996 Freedom to Farm Act, deficiency payments for program crops are no longer linked to actual planted acres. This new found flexibility coupled with an uncertain outlook of the direction of future commodity prices has spurred tremendous interest in developing new conservation tillage based cropping systems in the Pacific Northwest. Growers, agribusinesses, and researchers are considering a large range of "new" crops such as grain corn, yellow mustard, safflower, millet, and sunflowers. However, with each of these crops, we have a very limited knowledge base to determine if and where these new crops can be grown successfully. Not only do we face a large number of potential new crops, but we must also deal with a tremendous range in environment (precipitation, temperature, and soils) throughout the dryland cropping areas of the Pacific Northwest. While these new crops offer many potential advantages such as the breaking of disease cycles, increased economic security, and a greater compatibility with no-till and reduced tillage systems, a new crop must have a basic level of adaptation to a region before growers can incorporate it into a cropping sequence.

The consideration of corn as an alternative crop for the dryland farming areas is an excellent example of how we can determine if a new crop will be successful. In this paper, we will discuss 1) what information is needed to determine crop adaptation, 2) the use of crop simulation models to extend the results of field trials, and 3) the use of a crop model to evaluate the potential for growing dryland corn.

 

Evaluating New Crops

For any new crop (a non-traditional crop or a new approach or region for growing a traditional crop), we need the following knowledge:

 

Will it grow? For any new location, we need to know basic issues such as i) is the growing season long enough, ii) will the crop survive the winter if a winter annual, and iii) is there adequate soil water and precipitation? We might also want to know when the crop will flower, how deep will it root, and when will it be ready for harvest.

How well will it grow? Crop production decisions must be based on expected results; what will the crop yield over the long run? Acceptance or rejection of a new crop or system should not be based on one or even a couple years of data. Given year-to-year weather variability, it can take many years to know what a crop might yield on average and the range of yields to be expected. For example, to capture precipitation variability, climatologists recommend 30 years of record.

Where it will grow? In addition to year-to-year weather variability, there is tremendous location to location variability in the Pacific Northwest dryland cropping areas. Each location exhibits a unique combination of soils and climate. Because of these differences, the Washington State University Variety Testing Program tests winter wheat varieties at 18 locations annually throughout eastern Washington.

What are the best production practices? Final crop yield is the integration of a multitude of crop management factors such as soil fertility, weed control, seeding depth, rate, and date, row spacing, insect and disease control, tillage, and the previous crop.

Finding the Answers: The answers to will it grow, how well it will grow, where it will grow, and how best to grow a crop have traditionally been developed through grower and researcher experimentation and observation. We plant the crop, apply a set of production practices, and harvest the result. Most of what we know about crop production and crop adaptation have been determined through this approach. However, while effective, such methods are expensive and require many years of field experimentation and trials before confident answers can be deduced. Given the large number of potential new crops, the range in environmental conditions throughout the Pacific Northwest, and the desire to develop new systems quickly, it is time to consider the use of crop simulation models to extend the basic data gathered from limited field trials.

Use of Crop Simulation Models

Just as a growing crop responds to light, temperature, water, and other essential factors, crop models take as input the same driving variables, particularly temperature, solar radiation, and water, and predict the resultant crop response. A crop simulation model such as CropSyst developed at Washington State University can be adapted to model multiple crops. When properly parameterized and validated through comparison of model predictions and actual results, a crop simulation model can be used to assess many factors over time and location. By using historical weather records or programs that generate weather representative of historical trends of a specific site, a model can predict performance from fifty or one hundred different weather years. Averaging results from these multiple-year runs will produce a fair estimate of expected performance as wet, dry, and average years, for example, are included at the frequency indicated from past weather records. From such data, one can accurately estimate average yields, best and worst case performance and the probability of crop failures from events such as failing to reach maturity in cool years. To extend such analyses to different locations, such as Pullman, Grangeville, Ritzville, or Pendleton, one must simply use appropriate weather and soil data and crop production practices.

Through the ability to predict results over years and at different locations, models complement field trials. The field trial data provide basic data such as crop phenology, dry matter and leaf area development, and crop water use. From this basic data, we can then use the model to predict performance at sites where we do not have the resources to conduct field trials and to provide estimates of long-term average and frequency of yield performance accounting for weather variability.

In addition to testing the performance of a single crop, crop simulation models can be used to test the performance of a sequence of different crops (a rotation) over time and location. The consideration of new crops opens the possibility for many new crop rotations. In most instances, it will not be possible to test all possible rotations with long-term in-field cropping system trials. Given four crops, there are 6 possible two-year rotations, 8 three-year rotations, and 6 four-year rotations if you allow each crop to only occur once in a rotation. If you consider rotations where a crop repeats such as spring wheat followed by spring wheat, the number of possibilities is even greater. If you consider five potential crops, there are 84 possible two, three, four, and five-year rotations. While there are several important factors such as weed and disease levels for which no good models exist for our region, a rotation’s impact on soil water and compatibility with the climate can be predicted with models such as CropSyst. Questions such as should corn or safflower follow or precede winter wheat in terms of water use can be answered by current models.

 

Using a Crop Model to Evaluate the Potential for Growing Dryland Corn in Eastern Washington

Recent work with the crop simulation model, CropSyst is a good illustration of how crop models can be used to help evaluate new crop production and adaptation. Two years ago, a group of dryland growers became interested in the prospect of growing dryland corn. Many questions where raised such as would there be enough soil water and did we receive enough growing degree-days to grow corn to maturity. Through a combined effort of growers, industry and university personnel, and others, a series of trials were established in eastern Washington and part of northern Idaho. We were able to work with these individuals to collect the basic data of how corn grows under Pacific Northwest dryland conditions so that we could use the CropSyst model to extend the results in time, location, and over crop management.

Will Corn Grow? We collected data from two of the corn trials to measure water use, dry matter production, leaf area, and yield. When coupled with measurement of temperature, solar radiation, and other weather variables, we can use this basic data to develop the corn growth model. As discussed above, comparisons of actual observations and model predictions are used to validate the performance of the model. The following figure shows the ability of the model to predict soil water at the 1996 Dayton, Washington location.

 

How Well Will Corn Grow? After we develop the basic corn model, we can start to use the model to extend what we have learned from the field trials. The most fundamental questions concerning corn where with drought stress and the attainment of enough degree days to grow even short season corn hybrids to maturity. We know from the field trials how corn yielded during the 1996 and 1997 crop years, but it will take many more years to really determine a reliable yield expectation for these locations. Instead, we can predict a reliable expectation with the model by running it with many years of weather data. Illustrated below are the yield expectations from 50 years of model runs for Dayton and Pullman, Washington.

Note that corn will not reach maturity approximately 20% of the time in Pullman. It is impossible to predict this solely from two years of field trials.

Where Will it Grow? By running the model with weather and soil conditions representative of different locations, we can predict long term performance at locations where no trials were conducted. We simulated 50 years of corn production at four locations in eastern Washington, Dayton, Dusty, Pullman, and Harrington. From these data, we suggest that the Dayton area is the best suited for corn production, while the Pullman location may have the potential to produce the highest yields. However, corn in Pullman will not always reach maturity.

 

Location

Climate Data

Yield

Probability That Corn Will Reach Maturity

Avg. Ann. Precip.

Avg. Min. Temp.

Avg. Max. Temp.

Average When Reaching Maturity

Standard Dev.

Inches

º F

º F

Bu/a

Bu/a

%
Dayton

18.0

41.0

64.4

60

14

100
Pullman

22.0

35.6

55.4

85

19

76
Harrington

13.6

37.4

57.2

57

11

77
Lacrosse

12.8

39.2

60.8

43

11

100

 

What are the Best Production Practices? We can use the CropSyst model to predict the response to many basic agronomic questions. This can be extremely helpful in the early stages of learning about a new crop. To illustrate the use of the model in this way, we explored the effect of corn planting date at Pullman. Corn needs to be planted early enough to attain enough growing degree-days to reach maturity but it is possible that early planting may result in too much water use before the critical grain filling period. To assess the effect of planting date, the CropSyst model was run over 50 years of weather for planting dates ranging from April 24 to May 15 in one-week intervals. While yields rise with later planting dates, the chance of reaching maturity drops rapidly indicating that it is best to plant as early as possible.

 

 

While models will not eliminate field experimentation, they provide a way to extend the results of expensive and time consuming field research. With dryland corn, the model can help us decide where corn is most likely to succeed and what range of production practices should be proved out with actual field experimentation.

 

Conclusions

The goal of our modeling work with corn is to provide growers, agribusinesses, and researchers recommendations of how well corn is adapted to their particular location in the dryland region. With this information, they can start crop rotation studies involving corn, specific corn production research, and actual grower production in areas where the greatest chance of corn success is expected. We are currently in the process of extending our analysis of corn adaptation beyond the four sites presented earlier in the paper. For current information, call Tim Fiez at 509-335-2997 or check the http:\\drycrops.wsu.edu\crop_management\corn World Wide Web site.