Search

HighlightsSoftware was developed for estimation of DSSAT CSM-CROPGRO-Soybean cultivar coefficients.Phenology-related coefficients were estimated based on observed phenological events.Growth-related cultivar coefficients were estimated based on time-series observations.Cultivar coefficients were optimized based on single- and multiple-experiment data sets.Abstract. The Decision Support System for Agrotechnology Transfer (DSSAT) is one of the most popular software solutions for predicting crop growth and yield while capturing the effects of management practices and interactions between the crop and the environment. Accurate estimation of the crop cultivar coefficients that govern in-season growth and development is critical for correct yield estimates. The manual cultivar coefficient estimation process is time-consuming and results in user-dependent, subjective optimums that are difficult to reproduce. Typically, end-of-season observations (point-based) are used for estimating dynamic in-season biomass accumulation rates. The objective of this study was to develop a time-series estimator (TSE) capable of using multiple in-season observations for estimating the coefficients that define in-season growth and biomass partitioning. Using the TSE, cultivar coefficients were estimated based on multiple in-season observations of leaf area index (LAI) and shoot, leaf, and grain dry matter weights. The cultivar coefficients were estimated from single- and multiple-treatment (seasons and locations) in-season observations. This was done for two cultivars for six management × environment combinations. Estimated multiple-treatment based cultivar coefficients were evaluated with an independent data set and compared to DSSAT standard (manual) coefficients and the cultivar coefficients estimated with the GLUE method. The average normalized root mean squared error (nRSME) for LAI and shoot, leaf, and grain weights was 26% lower for one cultivar and about the same for the other cultivar when compared to the DSSAT standard. Because GLUE uses end-of-season point-based cultivar coefficient estimation, the grain weight over time was underestimated in earlier phases and more accurate toward harvest. The TSE-estimated cultivar coefficients based on 346 in-season observations across multiple target variables and six experiments more accurately reflected in-season growth and grain weight without compromising final grain weight predictions. Keywords: . CROPGRO-Soybean, DSSAT, Genetic coefficients, Normalized root mean square error minimization, Time-seris observations.

Parameter estimation is a critical step in successful application of dynamic crop models to simulate crop growth and yield under various climatic and management scenarios. Although inverse modeling parameterization techniques significantly improve the predictive capabilities of models, whether these approaches can recover the true parameter values of a specific genotype or cultivar is seldom investigated. In this study, we applied a Markov Chain Monte-Carlo (MCMC) method to the DSSAT dry bean model to estimate (recover) the genotype-specific parameters (GSPs) of 150 synthetic recombinant inbred lines (RILs) of dry bean. The synthetic parents of the population were assigned contrasting GSP values obtained from a database, and each of these GSPs was associated with several quantitative trait loci. A standard inverse modeling approach that simultaneously estimated all GSPs generated a set of values that could reproduce the original synthetic observations, but many of the estimated GSP values significantly differed from the original values. However, when parameter estimation was carried out sequentially in a stepwise manner, according to the genetically controlled plant development process, most of the estimated parameters had values similar to the original values. Developmental parameters were more accurately estimated than those related to dry mass accumulation. This new approach appears to reduce the problem of equifinality in parameter estimation, and it is especially relevant if attempts are made to relate parameter values to individual genes. Keywords: Crop models, Equifinality, Genotype-specific parameters, Markov chain Monte-Carlo, Parameterization.

The introduction of yield monitors, global positioning systems, and geographic information systems have provided new means to measure crop yield within a field, allowing very fine description of the spatial variability. However, only limited progress has been made on diagnosing reasons for yield variability and on identifying and managing areas of the field to maximize profit or reduce environmental impacts. Methods are needed to determine causes of yield variability and simulate spatial variability of yield across fields. The objective of this article was to quantify the effects of three yield-limiting factors (water stress, soybean cyst nematodes [SCN], and weeds) on soybean yield in a spatially variable field using a combined crop model-regression approach. The procedure was tested on 77 grid cells (0.2 ha) in the McGarvey field in Perry, Iowa, for 1995, 1997, and 1999. After estimating soil water and site parameters, predicted soybean yields were in good agreement with measured yield (r2 = 0.88). The root mean square of error (RMSE) was 243 kg ha-1, within 11% of mean actual yield for each year. The combined model was then used to simulate the effect of each factor separately on yield loss for each grid cell. Water stress had the biggest impact on soybean yield among the yield-limiting factors for all years. Soybean yields were reduced by an average of 1391 kg ha-1 as a result of water stress in 1997, while average yield reductions due to weeds and SCN were 167 and 109 kg ha-1, respectively. Average estimated yield loss in 1997 due to the combined effects of water stress, SCN, and weeds in each grid cell was 1667 kg ha-1. The regression coefficients for attributing yield losses to site factors were field-specific and may not be transferable to other fields and years, but they can be computed for fields where spatial layers include data on these factors.

Spatial variation in landscape and soil properties combined with temporal variations in weather can result in yield patterns that change annually within a field. The complexity of interactions between a number of yield-limiting factors makes it difficult to accurately attribute yield losses to conditions that occur within a field. In this research, a back-propagation neural network (BPNN) model was developed to predict the spatial distribution of soybean yields and to understand the causes of yield variability. First, we developed a BPNN model by relating soybean yield to topography, soil, weather, and site factors and evaluated model predictions for the same field for independent years. We also explored the potential use of BPNN for predicting yields in independent fields. Finally, we evaluated the ability of the BPNN to attribute yield losses due to soybean cyst nematodes (SCN), soil pH, and weeds. A total of 14 input datasets with combinations of four controlling factors (topographic, soil fertility, weather, and site) were used. For each objective, data from fields in Iowa were used for training the BPNN, while a portion of the data was withheld to verify the accuracy of yield predictions. All BPNN models had fully connected feed-forward architecture with a back-propagation weight adjustment algorithm. When tested for a particular field, the BPNN captured the major patterns of yield variability in independent years; the root mean square error of prediction (RMSEP) was 14.2% of actual yield. When the BPNN was trained with inputs from five fields, the RMSEP at test sites was 11.2% of actual yield. When the BPNN was used to attribute yield losses to soil pH, SCN, and weed populations, standard errors were 92, 262, and 171 kg ha-1, respectively. The technique showed that the BPNN could predict spatial yield variability with an RMSEP of about 14%.

The current soil water component of the CROPGRO-Soybean model was structured for deep, well-drained soils. It did not have mechanisms for restricting root growth and root water uptake in soil layers with high mechanical impedance or soil layers with non-uniform root distribution. In this research, the model was modified to consider root growth and root water uptake under those conditions. Experimental data sets from Castana, Iowa, and Gainesville, Florida, with two contrasting soils were used for model validation. An empirical soil impedance factor was incorporated to influence the ability of roots to grow and proliferate in a layer. A root hospitality or clumping factor was incorporated to limit root water uptake in soil layers where roots are not uniformly distributed. A sensitivity analysis was conducted to demonstrate the effects of these two factors and rate of root depth increase on root growth, soil water extraction, seed yield, and crop biomass. Root growth and soil factors were adjusted so that simulated soil water extraction from each soil layer correctly matched observed data. A maximum rate of vertical root growth was estimated to be 4.0 cm d–1 for both locations. Root distribution with depth, soil water extraction pattern, and simulated crop responses were highly sensitive to this parameter. In both soils, the root hospitality factor was 1.0 (implying uniform root distribution) from the soil surface to layers near the bottom of the root zone below which roots extracted water at lower rates. Both soils required a soil impedance factor at 30 to 45 cm of 0.2 or less to match the apparent vertical rate of root growth, root length distribution with depth, and observed soil water extraction patterns. After confirming that the model accurately simulated soil water extraction patterns in both locations, the model predicted aboveground growth and yield very well at both locations. In the Castana experiment predicted seed yield was 3127 kg ha–1 and mean measured seed yield was 2644 kg ha–1. In Gainesville final seed yield predicted by the model was 1299 kg ha–1 and measured seed yield was 1178 kg ha–1. This research confirms the importance of root growth and root proliferation on soil water extraction, crop growth and yield, and it demonstrates the need for additional research to develop relationships between soil properties that are routinely measured and relatively simple root growth and distribution characteristics.

Multi-location cultivar trials are a primary source of information for plant breeding and agronomic research. To use cultivar trial data to evaluate model performance, there are many issues that need to be resolved for determining model inputs (e.g., weather, soil characteristics, cultivar coefficients, and cultural practices) as well as appropriate methods of analysis for model evaluation. Our objectives were to better understand methodological issues, limitations, and potential uses of cultivar trial data for evaluation and improvement of crop models. Cultivar trial data were obtained from “The Soybean Uniform Tests: Northern States and Southern States” publications for over 60 locations from the years 1970 to 1990. The plant process model SOYGRO Version 5.42 was used to simulate the cultivar trial results. The original SOYGRO model had a large bias in simulation of yield with latitude. Different hypotheses of temperature effects on photosynthesis, pod addition rate, and seed growth were tested to explain bias in model simulations. Results showed that the mean square error of prediction could be decreased to 52% of its original value by changing the temperature functions. While calibration removed the model bias with latitude and temperature, the model underpredicted yield in high and overpredicted yield in low yielding environments. This suggests that other factors such as soil fertility and pest problems not included in the model are important. The use of a large number of widely-located cultivar trials succeeded in testing the model over a very broad range of weather conditions, was useful in refining relationships originally used in the model, and has led to improvements in a subsequent soybean model release, the CROPGRO-Soybean model.

There is increasing interest in using crop growth models to study soybean (Glycine max L. Merr.) seed yield and quality. Because the availability of nitrogen (N) during late season growth is a critical factor affecting both seed yield and quality of the soybean crop, accurate prediction of N balance is important for process oriented models. The objective of this study was to evaluate the accuracy of the N balance components of the CROPGRO soybean model for Iowa-grown soybean, and where appropriate to modify the model in order to improve its accuracy. Field samples of ‘Kenwood’ soybean, taken during reproductive growth in four consecutive seasons (1992-1995), were analyzed for leaf, stem, pod wall, and seed N concentration. Soil and weather data from these trials were used to evaluate the ability of CROPGRO to predict N on both land-area and concentration bases for leaf, stem, pod wall, and seed tissues. The original N2-fixation routine was too sensitive to cool soil temperature, which caused it to underpredict N accumulation in cool seasons. The original model also predicted remobilization of N from leaf tissue to occur earlier than was observed in the field. This reduced predicted photosynthesis which led to underprediction in the rate of N assimilation late in the season. Several simple modifications were introduced to enhance accuracy of the model: (1) decrease the lower limit of the range of optimum soil temperatures for nodule growth (from 28 to 22°C), and nitrogenase activity (from 23 to 20°C); (2) decrease the base temperature of pod set (from 14 to 10°C), and increase the base temperature for emergence and early vegetative development (from 7 to 9°C); (3) delay of N remobilization from leaves and stems to seeds; (4) increase the maximum N concentration of various tissues to reflect observed values for ‘Kenwood’ soybean, and set initiation of N remobilization to occur earlier in pods than in other tissues.

The value of a peanut crop is determined by yield and grade. The grade of peanuts is largely dependent on seed size distribution. Higher percentages of large seeds result in a higher grade and subsequently a higher profit for the grower. Peanut growth models currently predict yield and seed mass distribution, however, they do not predict seed size distribution which is critical for determining overall crop value. The objective of this work was to develop and evaluate a procedure to compute seed size distribution based upon the seed mass distribution predicted by a crop growth model. A linear relationship between individual seed mass and seed size was derived from field data and used to compute seed size distribution from predicted seed mass distribution. Mass of sound mature kernels (SMK) and extra large kernels (ELK) was predicted for 15 field experiments and compared to measured results. Three hypotheses were tested to improve the computed seed size distribution in the CROPGRO-Peanut crop growth model. Hypotheses relating to variable seed and shell growth rate reduced errors between predicted and observed mass of SMK and ELK. However, adjusting the average final weight per seed coefficient required by the crop growth model resulted in the best predictions of seed size distribution. The results of this work can provide a dynamic method of predicting mass of SMK and ELK required to determine crop grade, value, and harvest date when coupled with a dynamic crop growth model.

A model that predicts peanut pod detachment was developed and linked to the peanut component of the CROPGRO crop growth model. The model predicts initiation and subsequent rate of pod detachment on an individual cohort basis. It was tested using five different hypotheses for initiation of pod detachment. Mean square error between predicted and measured overall attached pod mass was used as the criterion for parameter estimation in the detachment model. The best detachment model gave good results in 18 experiments that included irrigated and nonirrigated, disease and control, and artificial shade conditions. The mean square error of prediction was used to select the best detachment hypothesis and the jackknife method was used to obtain confidence intervals of the detachment model parameters. The best model initiated detachment in a pod cohort when it experienced six days without carbohydrate, or when the ratio of leaf mass to pod mass and the slope of leaf mass were less than 0.396 and –0.859 g m–2 d–1, respectively.

Most crop growth models do not account for damage caused by pests. This limitation must be removed if the models are to provide useful predictions of production under real farm conditions. A generic framework was developed to couple pest damage of various types into the PNUTGRO and SOYGRO models for peanut and soybean, respectively. Coupling points were identified in the models for applying damage to leaves, stems, roots, pods, seeds, whole plants, and to the supply of assimilate. The resulting models were tested by simulating crops with measured pest damage levels for peanut (foliar disease) and soybean (foliar feeding insects) and comparing observed and simulated crop growth and yield results. This approach for coupling pests with crop models has potential for extending the practical applications of crop models to a broad range of problems.

The interactions between plants and their environment involve an elaborate collection of biological, physical, and chemical processes. To better understand the responses of crops to their environments, computer models are being used to study both the simple and complex aspects of this system.

Asoybean (Glycine max (L.) Merr.) crop growth simulation model (SOYGRO) was developed to aid farm managers in making irrigation and pest management decisions. Non-linear first order differential equations describe dry matter rates of change, accumulation and depletion of protein pools, and changes in shell and seed numbers. Two data sets from defoliation and irrigation experiments were used for calibration and validation of the model. The model responds well to drought and defoliation stresses for two test cases. Sensitivity analyses of SOYGRO revealed that simulated yield was most sensitive to changes in gross photosynthesis and growth respiration. The sensitivity of simulated yield to changes in model parameters was increased by the occurrence of either water or defoliation stress.