The Estimation of Regional Crop Yield Using Ensemble-Based Four-Dimensional Variational Data Assimilation

Abstract

To improve crop model performance for regional crop yield estimates, a new four-dimensional variational algorithm (POD4DVar) merging the Monte Carlo and proper orthogonal decomposition techniques was introduced to develop a data assimilation strategy using the Crop Environment Resource Synthesis (CERES)-Wheat model. Two winter wheat yield estimation procedures were conducted on a field plot and regional scale to test the feasibility and potential of the POD4DVar-based strategy. Winter wheat yield forecasts for the field plots showed a coefficient of determination (R 2 ) of 0.73, a root mean square error (RMSE) of 319 kg/ha, and a relative error (RE) of 3.49%. An acceptable yield at the regional scale was estimated with an R 2 of 0.997, RMSE of 7346 tons, and RE of 3.81%. The POD4DVar-based strategy was more accurate and efficient than the EnKF-based strategy. In addition to crop yield, other critical crop variables such as the biomass, harvest index, evapotranspiration, and soil organic carbon may also be estimated. The present study thus introduces a promising approach for operationally monitoring regional crop growth and predicting yield. Successful application of this assimilation model at regional scales must focus on uncertainties derived from the crop model, model inputs, data assimilation algorithm, and assimilated observations.

Improving Winter Wheat Yield Estimation from the CERES-Wheat Model to Assimilate Leaf Area Index with Different Assimilation Methods and Spatio-Temporal Scales

“…In the past few decades, conventional agro-meteorological models and empirical statistical regression models based on field-measured yields and related remotely-sensed spectral vegetation indices [5] were used to monitor crop growth and yields, with the disadvantages of high-cost, laboriousness, inefficiency and only being applicable for local regions [6][7][8]. Process-oriented crop growth models can offer powerful tools to simulate the crop growth and obtain the crop yield under various environmental and management conditions with the advantages of cost, timeliness, accuracy and suitability at the field scale [6,[9][10][11][12][13][14], while they were also dynamically accounting for several limiting factors (e.g., soil, weather, water, nitrogen and field management data) at the regional scale [3,15]. With the development of remote sensing technology, extensive remotely-sensed data provide a synopsis of timely and up-to-date crop growing conditions over large areas throughout the crop's growing season and can be employed to provide many potential data for crop model simulations [16].…”

“…With the development of remote sensing technology, extensive remotely-sensed data provide a synopsis of timely and up-to-date crop growing conditions over large areas throughout the crop's growing season and can be employed to provide many potential data for crop model simulations [16]. Therefore, data assimilation approaches that integrate crop growth models with remote sensing data have been proposed recently and recognized as important approaches for monitoring crop growth conditions and improving the accuracy of yield estimations at the regional scale [6,7,9,[17][18][19][20][21][22][23][24][25].…”

“…However, the disadvantages of these methods are that they could not consider uncertainties in observations, and the model itself, usually failing to capture and maintain the crop system dynamics, and the iterative process may require excessive computing time [4,7,34,35]. Thus, the state-updating methods carried out using assimilation algorithms (e.g., an ensemble Kalman filter (EnKF) [3,10,36,37], particle filters (PF) [22,38,39], traditional variational methods [40,41] or ensemble-based four-dimensional variation [9,42,43], etc.) were introduced into crop growth simulations to alleviate the shortcomings of the recalibration/re-initialization methods and can combine many types of observations at random times, and the state variables can also be continually updated and simulated accurately [4].…”

“…Especially EnKF, as one of the typical state-updating methods, has received much attention because of its simple conceptual formulation and relative ease of implementation [3,10,21,36,41]. However, the disadvantages of this method are that it may result in an obvious deficiency when solving complicated estimation issues in a nonlinear and non-Gaussian dynamic system, and it also has low computational efficiency since it is designed to incorporate sequential information only [9].…”

Winter Wheat Production Estimation Based on Environmental Stress Factors from Satellite Observations

“…The Hengshui City was a part of the North China Plain, which was an alluvial plain of the Yellow River, Huaihe River, and Haihe River [40]. In this region, winter wheat was the dominant crop and the main crop planting patterns were the winter wheat and summer maize rotations.…”