An analogy-based crop yield forecast scheme assessing similarity of the time series of leaf area index

Authors: LIU, YADONG - Seoul National University; KIM, JUNHWAN - National Institute Of Crop Science - Korea; Fleisher, David; KIM, KWANG SOO - Seoul National University

Submitted to: Remote Sensing
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 7/27/2021
Publication Date: 8/4/2021
Citation: Liu, Y., Kim, J., Fleisher, D.H., Kim, K. 2021. An analogy-based crop yield forecast scheme assessing similarity of the time series of leaf area index. Remote Sensing. 13(6):3069. https://doi.org/10.3390/rs13163069.
DOI: https://doi.org/10.3390/rs13163069

Interpretive Summary: An estimate of seasonal crop yields is useful for farmers and agricultural policy officials to plan and establish pricing decisions. There are many methods that have been used to make these forecasts, but they vary in terms of accuracy and the amount of data needed. To simplify this forecasting effort, a new algorithm was developed that links crop yield with plant canopy size using a combination of data from prior growing seasons and near real-time satellite observations. This simpler approach, called the ABC forecast scheme, was shown to accurately predict corn and soybean yields in the Midwestern United States. Yields can be predicted up to two months in advance of expected harvest dates. The research indicated that the ABC approach gave a similar performance compared with more sophisticated and data-intensive alternatives. This new approach is of particular use for agricultural producers who live in regions where environmental data is limited in availability.

Technical Abstract: Seasonal forecasts of crop yield are important components of agricultural policy decisions and farmer planning. A wide range of input data are often needed to forecast crop yield, particularly where sophisticated approaches such as machine learning and process-based models are used. This requires considerable effort for data preparation in addition to identifying data sources. In this study, a simpler approach called the Analogy Based Crop-yield (ABC) forecast scheme is proposed to make timely and accurate prediction of regional crop yield using a minimum set of inputs. In this ABC method, a similarity index is first used to identify a season within the past ten years that has a similar time-series leaf area index (LAI) pattern to the current season. Crop yield in the current growing season is then forecast using the weighted average of crop yield reported in the analogous seasons for each county of interest. The MOD15AH2, which is a satellite data product for LAI, was used to provide inputs. The ABC approach was applied in this study to predict yield of corn and soybean in the Midwestern U.S. for the period of 2017-2019. The mean absolute percentage error (MAPE) of crop yield forecasts was less than 10 percent for both crops in each growing season when the time series of LAI from the day of year 89 to 209 was used as input. This prediction error was comparable to a deep neural network model that relied on soil and weather data as well as satellite data in a previous study. These results indicate that the ABC approach allowed for crop yield forecast with a lead-time of at least two months before harvest. The ABC scheme would be particularly useful for regions where crop yield forecasts are limited by availability of reliable environmental data.