Identifying causes of crop yield variability with interpretive machine learning

AbstractCrop yield prediction is crucial for global food security yet notoriously challenging due to multitudinous factors that jointly determine the yield, including genotype, environment, management, and their complex interactions. Integrating the power of optimization, machine learning, and agronomic insight, we present a new predictive model (referred to as the interaction regression model) for crop yield prediction, which has three salient properties. First, it achieved a relative root mean square error of 8% or less in three Midwest states (Illinois, Indiana, and Iowa) in the US for both corn and soybean yield prediction, outperforming state-of-the-art machine learning algorithms. Second, it identified about a dozen environment by management interactions for corn and soybean yield, some of which are consistent with conventional agronomic knowledge whereas some others interactions require additional analysis or experiment to prove or disprove. Third, it quantitatively dissected crop yield into contributions from weather, soil, management, and their interactions, allowing agronomists to pinpoint the factors that favorably or unfavorably affect the yield of a given location under a given weather and management scenario. The most significant contribution of the new prediction model is its capability to produce accurate prediction and explainable insights simultaneously. This was achieved by training the algorithm to select features and interactions that are spatially and temporally robust to balance prediction accuracy for the training data and generalizability to the test data.

Download Full-text

Substitution of Inorganic Nitrogen Fertilizer with Green Manure (GM) Increased Yield Stability by Improving C Input and Nitrogen Recovery Efficiency in Rice Based Cropping System

Agronomy ◽

10.3390/agronomy9100609 ◽

2019 ◽

Vol 9 (10) ◽

pp. 609 ◽

Cited By ~ 3

Author(s):

Qaswar ◽

Jing ◽

Ahmed ◽

Shujun ◽

Dongchu ◽

...

Keyword(s):

Crop Yield ◽

Green Manure ◽

Inorganic Nitrogen ◽

Cropping System ◽

N Fertilization ◽

Yield Stability ◽

N Fertilizer ◽

Yield Variability ◽

Variability Index ◽

Double Rice

A long-term field experiment was carried out (since 2008) for evaluating the effects of different substitution rates of inorganic nitrogen (N) fertilizer by green manure (GM) on yield stability and N balance under double rice cropping system. Treatments included, (1) N0 (no N fertilizer and no green manure); (2) N100 (recommended rate of N fertilizer and no green manure); (3) N100-M (recommended rate of N fertilizer and green manure); (4) N80-M (80% of recommended N fertilizer and green manure); (5) N60-M (60% of recommended N fertilizer and green manure); and (6) M (green manure without N fertilization). Results showed that, among all treatments, annual crop yield under N80-M treatment was highest. Crop yield did not show significant differences between N100-M and N80-M treatments. Substitution of different N fertilizer rates by GM reduced the yield variability index. Compared to the N0 treatment, yield variability index of early rice under N100-M, N80-M, and N60-M treatments was decreased by 11%, 26%, and 36%, respectively. Compared to the N0 treatment, yield variability index of late rice was decreased by 12%, 38%, 49%, 47%, and 24% under the N100, N100-M, N80-M, N60-M, and M treatments, respectively. During period of 2009–2013 and 2014–2018, nitrogen recovery efficiency (NRE) was highest under N80-M treatment and N balance was highest under N100 treatment. NRE of all treatments with GM was increased over the time from 2009–2013 to 2014–2018. All treatments with GM showed increasing trend of SOC over the years. Substitution of N fertilizer by GM also increased C inputs and soil C:N ratio compared to the N100 and N0 treatments. Boosted regression model indicated that C input, N uptake and AN were most influencing factors of crop yield. Thus, we concluded that N fertilization rates should be reduced by 20% under GM rotation to attain high yield stability of double rice cropping system through increasing NRE and C inputs.

Download Full-text

Combining Optical, Fluorescence, Thermal Satellite, and Environmental Data to Predict County-Level Maize Yield in China Using Machine Learning Approaches

Remote Sensing ◽

10.3390/rs12010021 ◽

2019 ◽

Vol 12 (1) ◽

pp. 21 ◽

Cited By ~ 6

Author(s):

Liangliang Zhang ◽

Zhao Zhang ◽

Yuchuan Luo ◽

Juan Cao ◽

Fulu Tao

Keyword(s):

Machine Learning ◽

Crop Yield ◽

Satellite Data ◽

Maize Yield ◽

Environmental Data ◽

Yield Prediction ◽

County Level ◽

Climate Data ◽

Source Data ◽

Optical Fluorescence

Maize is an extremely important grain crop, and the demand has increased sharply throughout the world. China contributes nearly one-fifth of the total production alone with its decreasing arable land. Timely and accurate prediction of maize yield in China is critical for ensuring global food security. Previous studies primarily used either visible or near-infrared (NIR) based vegetation indices (VIs), or climate data, or both to predict crop yield. However, other satellite data from different spectral bands have been underutilized, which contain unique information on crop growth and yield. In addition, although a joint application of multi-source data significantly improves crop yield prediction, the combinations of input variables that could achieve the best results have not been well investigated. Here we integrated optical, fluorescence, thermal satellite, and environmental data to predict county-level maize yield across four agro-ecological zones (AEZs) in China using a regression-based method (LASSO), two machine learning (ML) methods (RF and XGBoost), and deep learning (DL) network (LSTM). The results showed that combining multi-source data explained more than 75% of yield variation. Satellite data at the silking stage contributed more information than other variables, and solar-induced chlorophyll fluorescence (SIF) had an almost equivalent performance with the enhanced vegetation index (EVI) largely due to the low signal to noise ratio and coarse spatial resolution. The extremely high temperature and vapor pressure deficit during the reproductive period were the most important climate variables affecting maize production in China. Soil properties and management factors contained extra information on crop growth conditions that cannot be fully captured by satellite and climate data. We found that ML and DL approaches definitely outperformed regression-based methods, and ML had more computational efficiency and easier generalizations relative to DL. Our study is an important effort to combine multi-source remote sensed and environmental data for large-scale yield prediction. The proposed methodology provides a paradigm for other crop yield predictions and in other regions.

Download Full-text