A new and final crop yield forecast was released yesterday (Oct 12, 2017) by USDA-NASS reporting an increase in corn yield at the state-level (+1 bushels per acre, averaging 134 bushels per acre) and also in the final statewide crop acreage (+200,000 acres, averaging 5.2 million acres total). Check out the full report published by USDA-NASS for further details.
In a previous Agronomy eUpdate article in Issue 653, “Forecasting 2017 Kansas corn yields” on September 29, 2017, a new yield forecast tool was discussed for Kansas corn. This tool is primarily based on real-time satellite data, historical yield data at county-level, and prediction of current geo-location of corn fields across the state (based on satellite data and a field survey of +500 field locations across the state). To obtain more information about the K-State “Yield Forecasting Tool” (YFT), visit the previous article referenced above. The primary steps for this tool, presented as a simplified approach, are highlighted in Figure 1.
Figure 1. Theoretical framework portraying the main steps involved in the development of forecasting corn yields for Kansas. Steps: 1- Data collection; 2- Building and validating yield forecasting models (YFM); 3- Building and validating land layer for corn 2017; and 4- validation of previous years. Infographic developed by Ignacio Ciampitti, Rai Schwalbert, and Luciana Nieto, K-State Research and Extension.
This corn yield forecast tool will be evolving in the coming months with the goal of providing producers reliable yield predictions for improving the decision-making process.
An updated corn yield forecast was obtained for Kansas via utilization of satellite imagery from planting until beginning of September for this current growing season. Based on the satellite yield model developed by our team, the state-level yield prediction is 134.1 bushels per acre, which is right on target to the yield prediction released by USDA-NASS (134 bushels per acre; link).
A new step on the Yield Forecast Tool
Since our last news release, several farmers were also asking about information related to statewide total corn production for the 2017 growing season. The main challenges on any crop production estimates are to adequately account for the number of acres across the state. In order to move forward with this step of the forecasting tool, models were developed to estimate the number of corn acres for past growing seasons and to compare those values with final corn acreage. A validation analysis was done to test the precision our models in predicting corn acreage relative to the final number reported by USDA-NASS in each growing season. For this purpose, four past growing seasons (2009, 2012, 2014, and 2016) were selected to forecast total production benchmarking these values against the final reports published by USDA-NASS. As reflected in Figure 2, forecasted total corn production was close (1.8% deviation from mean) relative to the final values.
Figure 2. Validation for total corn production estimated via the yield forecasting model and the final value reported by USDA-NASS in past growing seasons (2009, 2012, 2014, and 2016) for the state of Kansas.
Final Forecast Total Corn Production
Implementing the previous step, implying the estimation of the total number of corn acres, the forecasted yield production based on our model was 698 million of bushels, comparable to the final value released by USDA-NASS (October 2017) of 697 million of bushels (Figure 3).
Figure 3. Forecasting corn yields (left panel) and total production (right panel) derived from satellite data for the state of Kansas.
In summary, the Yield Forecasting Tool (YFT) predicted an average state yield value of 134.1 bushels per acre (including satellite imagery data until corn maturity) and a total production of 698 million bushels for corn for the state. The predicted values are in agreement with the latest forecast released yesterday by USDA-NASS (Oct 12, 2017) for the overall corn production in Kansas (Figure 3).
Stay tuned for further details coming out about the YFT as we continue to work on other components, such as early-season crop classification (related to remote sensing and satellite imagery data), to be integrated to these complex crop yield forecast models.
Ignacio A. Ciampitti, Crop Production and Cropping Systems Specialist
Rai Schwalbert, KSUCROPS Production, Dr. Ciampitti’s Lab
Luciana Nieto, KSUCROPS Production, Dr. Ciampitti’s Lab