The Modeling Approach Used to Address These Questions
- Has 2012 weather from planting through July 2 affected corn yield potential?
The Hybrid-Maize uses historic weather data from automated weather stations. In this case, I used weather data from three of Iowa State University's Research and Demonstration Farms - NW Iowa near Sutherland, SE Iowa near Crawfordsville, and central Iowa using a combination of data from Ames and Gilbert.
The model allows users to compare yield potentials given the weather actually recorded up through the simulation date. In this case it included 2012 weather data through last Monday, July 2. The model generates real-time yield predictions for the current season. What that means is that actual weather conditions up to the date of the simulation are in a sense considered the base from which to start. That is what we have to work with; unfortunately, we can’t change what has happened so far in 2012. Then the model ‘asks’ a series of ‘what if’ questions. For example: What is yield potential if, from this day forward, we have weather conditions like those we had in the best possible year in the weather database for that location? What if the worst historical weather occurred? The weather record begins in either 1986 or 1988 for the three research and demonstration farms we’re working with here (Table 1).
Common inputs for all three sites modeled are provided in Table 1. Factors that varied across locations such as soil textures are shown in Table 2. Residue levels at planting, corn suitability ratings and other field-specific information are not factored into the analysis. However, some of the variability, especially in the early-season factors, is removed by using emergence date rather than planting date in the model.
- What effect will this week’s extremely hot and dry weather have on yield potential?
The five to seven day NOAA forecast on July 3 was for continued hot-dry weather. Using those forecast high and low temperatures, I generated a new – some might say fictitious - weather file for each location for the crop model (see high/low temperatures used in Table 2). The model then provided real-time yield predictions using actual weather data through July 2, 2012, and the forecast average temperatures through July 8. All other weather variables – i.e. solar radiation, precipitation (which was nil at all three sites) and relative humidity - in these model runs were the same as those of July 2; only high/low temperatures were changed. Because crop canopies were already completely developed or nearly so on July 2, I assume in this analysis that light interception, crop water use, etc., are stable during this time frame.
As mentioned earlier, the real-time yield predictions are based on weather data from previous years at that location and are summarized into five yield-level groups ranging from best yield to worst yield years. For example, what if the best previous weather occurred and what if the worst previous weather occurred from July 9 through the end of the season. Data from these model runs are summarized in Figure 1. Bars for the three locations are grouped together according to the five yield-level categories from best to worst.