Methodology for the use of DSSAT models for precision agriculture decision support

doi:10.1016/j.compag.2008.05.022

Computers and Electronics in Agriculture

Volume 64, Issue 2, December 2008, Pages 276-285

https://doi.org/10.1016/j.compag.2008.05.022 Get rights and content

Abstract

A prototype decision support system (DSS) called Apollo was developed to assist researchers in using the Decision Support System for Agrotechnology Transfer (DSSAT) crop growth models to analyze precision farming datasets. Because the DSSAT models are written to simulate crop growth and development within a homogenous unit of land, the Apollo DSS has specialized functions to manage running the DSSAT models to simulate and analyze spatially variable land and management. The DSS has modules that allow the user to build model input files for spatial simulations across predefined management zones, calibrate the models to simulate historic spatial yield variability, validate the models for seasons not used for calibration, and estimate the crop response and environmental impacts of nitrogen, plant population, cultivar, and irrigation prescriptions. This paper details the functionality of Apollo, and presents the results of an example application.

Introduction

Producers in the Midwestern US often use yield monitors to map spatial yield variability within their fields. These data have revealed that a tremendous amount of yield variability exists within most fields. Spatial yield variability results from complex interactions among factors including water stress, nutrients, rooting depth, soil physical properties, drainage, weather, pests, and management. Much of the variability in a field is caused by soil properties, a source of variability which naturally exists and cannot easily be altered. On the other hand, some variability exists due to factors that can be managed (e.g. pests, poor emergence, water logging).

Crop growth models integrate the effects of soils, weather, management, genetics, and pests on daily growth, and can be used to gain insight into spatial yield variability. Among the numerous crop growth models, the most widely used are the Decision Support for Agrotechnology Transfer (DSSAT) models, which were designed to simulate growth, development, and yield of a crop growing on a uniform area of land, as well as the changes in soil water, carbon, and nitrogen that take place under the cropping system over time (Jones et al., 2003). DSSAT has been in use for the past 15 years by researchers all over the world, for a variety of purposes, including crop management (Fetcher et al., 1991), climate change impact studies (Alexandrov and Hoogenboom, 2001), sustainability research (Quemada and Cabrera, 1995), and precision agriculture (Paz et al., 2001, Paz et al., 2003), and is well validated for a number of regions and crops. Included in the DSSAT family are modules which simulate the growth of 16 different crops, including maize, soybeans, wheat, rice, and others. DSSAT uses common modules for soil dynamics and soil–plant–atmosphere interactions regardless of the plant growth module selected. Data requirements include weather inputs (daily maximum and minimum temperature, rainfall and solar radiation), soils classification, and crop management practices (variety, row spacing, plant population, fertilizer and irrigation application dates and amounts).

While the DSSAT family of crop growth models provides many opportunities for critical analysis, it is tedious to use for precision agriculture studies and decision support because the model is built for simulation of a single homogeneous field unit. In order to facilitate the use of DSSAT for precision agriculture decision support, automated procedures and related tools are needed to implement crop growth simulations spatially across field-level management zones.

We have developed methods to use the DSSAT family of crop growth models to understand causes of spatial yield variability, conduct yield gap analysis for factors that limit yield, and estimate the economic consequences of moving from uniform to spatially variable management. We have also developed methodologies to automate the preparation of necessary input files for DSSAT to be used to simulate spatially variable crop development. These methods have been used successfully in several studies (Thorp et al., 2006, Thorp et al., 2007, DeJonge et al., 2007, Miao et al., 2006) but have not yet been fully explained.

In this paper we describe these methods, combined into a prototype precision farming decision support system (DSS) called Apollo, which integrates several precision farming applications developed for the DSSAT crop growth models into a single interface. Specifically, we detail the development and utilization of Apollo to allow researchers to (1) build model input files for spatially varying parameters, (2) calibrate input parameters for the crop growth models in order to simulate historical yield variability within a field, (3) validate the calibrations for seasons not used for calibration, and (4) estimate the crop response and environmental consequences of precision prescriptions or other management decisions.

Section snippets

Building model input files for spatially varying parameters

DSSAT simulations require a soil input file that details the physical and hydraulic properties of the soil. If comparison of model output yield to observed yield data is to be done, for example in order to calibrate the model to specific local conditions, formatted yield data is also required. When using DSSAT for precision agriculture applications, a challenge is to efficiently generate these files for each management zone. One function of Apollo, then, is to automate the generation of these

Case study: impacts of climate change on spatial yield variability

Previous investigations have detailed the use of Apollo for model calibration and validation (Thorp et al., 2007), development of precision nitrogen management strategies (Thorp et al., 2006), spatially variable-irrigation (DeJonge et al., 2007), etc. We will therefore conclude this paper by demonstrating the use of Apollo for a new application. We will use the system to simulate the potential effects of projected climate change spatial yield variability.

Future scenario climates for the mid to

Conclusions

Crop models provide a mechanistic way to estimate the interaction of spatial differences in soil properties and pest populations with temporal stresses on yield variability within a field. This is possible because the models compute daily growth processes as a function of weather, stress, and pest damage. Once calibrated to simulate the historical yield variability within a field, crop models are a powerful tool to develop risk management strategies that can balance economic risk incurred by

Acknowledgements

The authors wish to thank Dr. Eugene Takle and Dr. Eric Lu of the Iowa State University Department of Agronomy for their assistance with and providing of the simulated climate data used in this study.

References (22)

W.D. Batchelor et al.
Examples of strategies to analyze spatial and temporal yield variability using crop models
Eur. J. Agron.
(2002)
K.C. DeJonge et al.
Simulating the effects of spatially variable irrigation on corn yields, costs, and revenue in Iowa
Agric. Water Manage.
(2007)
W.L. Goffe et al.
Global optimization of statistical functions with simulated annealing
J. Econometrics
(1994)
J.W. Jones et al.
DSSAT cropping system model
Eur. J. Agron.
(2003)
J.O. Paz et al.
Model-based techniques to determine variable rate nitrogen for corn
Agric. Syst.
(1999)
K.R. Thorp et al.
Methodology to link production and environmental risks of precision nitrogen management strategies in corn
Agric. Syst.
(2006)
V.A. Alexandrov et al.
The impact of climate variability and change on crop yield in Bulgaria
Agric. Forest Meteorol.
(2001)
Christensen, J.H., Christensen, O.B., Lopez, P., van Meijgaard, E., Botzet, M., 1996. The HIRHAM4 regional atmospheric...
J. Fetcher et al.
An evaluation of the SWATRER and CERES-Millet models for southwest Niger
F. Giorgi et al.
Regional comiate change scenarios over the United States produced with a nested regional climate model
J. Climate
(1994)

Y. Miao et al.

Evaluating management zone optimal nitrogen rates with a crop growth model

Agron. J.

(2006)

Cited by (114)

High spatial resolution seasonal crop yield forecasting for heterogeneous maize environments in Oromia, Ethiopia
2023, Climate Services
Seasonal climate variability determines crop productivity in Ethiopia, where rainfed smallholder farming systems dominate in the agriculture production. Under such conditions, a functional and granular spatial yield forecasting system could provide risk management options for farmers and agricultural and policy experts, leading to greater economic and social benefits under highly variable environmental conditions. Yet, there are currently only a few forecasting systems to support early decision making for smallholder agriculture in developing countries such as Ethiopia. To address this challenge, a study was conducted to evaluate a seasonal crop yield forecast methodology implemented in the CCAFS Regional Agricultural Forecasting Toolbox (CRAFT). CRAFT is a software platform that can run pre-installed crop models and use the Climate Predictability Tool (CPT) to produce probabilistic crop yield forecasts with various lead times. Here we present data inputs, model calibration, evaluation, and yield forecast results, as well as limitations and assumptions made during forecasting maize yield. Simulations were conducted on a 0.083° or ∼ 10 km resolution grid using spatially variable soil, weather, maize hybrids, and crop management data as inputs for the Cropping System Model (CSM) of the Decision Support System for Agrotechnology Transfer (DSSAT). CRAFT combines gridded crop simulations and a multivariate statistical model to integrate the seasonal climate forecast for the crop yield forecasting. A statistical model was trained using 29 years (1991–2019) data on the Nino-3.4 Sea surface temperature anomalies (SSTA) as gridded predictors field and simulated maize yields as the predictand. After model calibration the regional aggregated hindcast simulation from 2015 to 2019 performed well (RMSE = 164 kg/ha). The yield forecasts in both the absolute and relative to the normal yield values were conducted for the 2020 season using different predictor fields and lead times from a grid cell to the national level. Yield forecast uncertainties were presented in terms of cumulative probability distributions. With reliable data and rigorous calibration, the study successfully demonstrated CRAFT’s ability and applicability in forecasting maize yield for smallholder farming systems. Future studies should re-evaluate and address the importance of the size of agricultural areas while comparing aggregated simulated yields with yield data collected from a fraction of the target area.
Modeling the potential impact of climate change on maize-maize cropping system in semi-arid environment and designing of adaptation options
2023, Agricultural and Forest Meteorology
Climate change is showing detrimental effect on crop productivity and food security. Adaptive strategies should be implemented to offset the negative impacts of climate change on the maize crop. There is concern that food insecurity will increase in Pakistan due to climate change. Hence, present study was conducted to quantify the response of maize yield to projected climate change and suggest adaptation options i.e., planting date, fertilizer use, fertigation, plant density and hybrid choice to counterbalance the impact of climate change using the CSM_CERES_Maize model. Field data obtained during spring and autumn seasons of 2016 was used to calibrate the model while evaluation was done using experimental data of 2016 and 2017. Afterwards, sensitivity analysis for carbon dioxide, temperature, water, and nitrogen (CTWN) variation was performed, and it was then applied to determine the impact of climate change on spring and autumn maize for mid-century. The results revealed that model performance was better for the spring season than for the autumn season. The model was sensitive to CTWN during the spring and autumn season of maize. Results showed that without adaptation, grain yield will be reduced in the spring (17 to 29%) and autumn (19 to 33%) seasons maize while adoption of adaptation strategies (e.g., Fertigation, sowing date adjustment (earlier sowing of spring maize by 15 days and delayed sowing of autumn maize by 20 days), higher thermal time requiring cultivars, increasing 10% plant population and 25 kg N ha⁻¹ during mid-century) could increase grain yield during both seasons. Therefore, farmers should adopt these adaptation strategies so that negative impact of climate on maize yield could be reduced.
County level calibration strategy to evaluate peanut irrigation water use under different climate change scenarios
2023, European Journal of Agronomy
Regional scale simulation of crop yield is challenging due to the spatial variability of soil properties, crop varieties, management practices and weather conditions. Point-based crop models are commonly used for spatial simulation with increased availability of high-resolution spatial datasets. However, it is still difficult to calibrate crop models well due to the spatial variability of model inputs. The focus of this work was to determine if a single set of cultivar and soil parameters could be calibrated to simulate county level peanut yield and to evaluate the effects of irrigation to mitigate potential climate change impacts on peanut yield. Model input data for fourteen seasons and five major peanut producing counties was assembled and used for model calibration. Three seasons of data were withheld and used for an independent evaluation. Overall, peanut growth duration and county level yields were simulated well with a set of optimum cultivar and soil parameters for each county. The model calibration showed that simulated maturity dates and yields were in good agreement with the observed county level values reported by NASS, giving an overall R² of 0.71 and 0.73 and RMSE values of 6 days and 333 kg/ha, respectively. The model also generated good simulations of maturity dates and yield for the three evaluation seasons, with an overall R² of 0.83 and 0.76 and RMSE values of 5 days and 429 kg/ha, respectively. The results from future climate simulations indicated that the rainfed yields will suffer from increasing daytime temperature and an irrigation strategy could potentially offset the heat and water stress to maintain higher peanut production in the Southeastern USA. This study can enhance the accuracy of simulating the impact of climate change on crop production by providing a calibration and evaluation strategy that aggregates spatial heterogeneity of model inputs at the regional scale. Further research should test this method for other crops and more model applications, such as nitrogen leaching, groundwater use and fertilizer management.
Applying agroclimatic seasonal forecasts to improve rainfed maize agronomic management in Colombia
2022, Climate Services
Climate variability affects crop production in multiple and often complex ways. The development and use hybrid crops with greater productivity and tolerance to climate shocks is one of the approaches to climate adaptation and agricultural intensification. Since hybrid crops are more expensive for the producer, risk management is of paramount importance. Here, we pose that there is high potential for the Colombian maize sector to use crop-specific climate services for risk reduction. We used the CERES-Maize crop model connected to seasonal climate forecasts developed via Canonical Correlation Analysis (CCA) across key maize growing areas in Colombia to assess the performance of a maize-specific agroclimatic forecast to inform two key decisions, namely, the choice of sowing dates and genotypes. We find that the agroclimatic models perform well at discriminating yield categories (above, below, and normal) with discrimination capacity of up to 70–80 % for the ‘below normal’ and ‘above + below normal’ categories. Consistent with this, agroclimatic forecasts typically predict the optimal planting date with an error of 3 pentads or less. They also predict the optimal choice of genotype correctly around 50–70 % of the time depending on the site or season of interest. Notably, we identify specific cases in which the agroclimatic forecast is misleading but argue that the overall value of the forecasts outweighs these cases. Future work should focus on expanding the scope of the agroclimatic prediction to include other relevant farming decisions that are influenced by climate, and on the improvement of climate forecast performance.
Plant available water capacity (PAWC) of soils predicted from crop yields better reflects within-field soil physicochemical variations
2022, Geoderma
Within field variations of plant available water capacity (PAWC) of soil is one of the major causes of spatial yield variability in dryland agriculture systems, as PAWC interacts with pre-season and in-season rainfall and other climatic variables to determine crop growth and final yield. Quantification of such variations helps to better understand the changes in soil texture and subsoil constraints to inform spatially explicit management practice. An inverse modelling approach to estimate PAWC from crop yields was developed as a more cost-effective alternative to traditional soil sampling methods. In this study, we further extend this approach to predict and map in-field variations of PAWC from yield maps of single and multiple crops. Soil PAWC maps were produced based on inversely predicted PAWC using crop yield maps together with in-field management information, and compared with: 1) available water capacity derived using laboratory-measured soil properties, and 2) soil types derived from proximally sensed soil spectra and ground geophysics for four representative farms in Australia. The results show that the predicted PAWC maps matched well with within-field spatial variation of soil types, and well reflected the impact of soil constraints (e.g. salinity), and soil classifications from soil survey and local experience. This demonstrates that the predicted PAWC from crop yield using inverse modelling can reflect the soil physicochemical variations within-field. The generated PAWC maps can be combined with process-based modelling to predict crop yield and yield zones and to inform spatial field management and soil sampling.
Crop yield prediction using MODIS LAI, TIGGE weather forecasts and WOFOST model: A case study for winter wheat in Hebei, China during 2009–2013
2022, International Journal of Applied Earth Observation and Geoinformation
Timely and reliable in-season forecasting of crop yield is crucial for regional and national agricultural management. To improve the winter wheat yield prediction accuracy at the regional scale, we developed a data assimilation scheme that assimilated the MODIS leaf area index (LAI) into the WOrld FOod STudies (WOFOST) model. The meteorological data of the WOFOST model include current weather data, 15-day THORPEX Interactive Grand Global Ensemble (TIGGE) forecast dataset and weather forecast data generated by TIGGE forecast and historical meteorological data (1979–2008) using the vectorial angle method. The WOFOST model was calibrated within 10 subregions of Hebei province based on field measured winter wheat growth data from each corresponding agrometeorological station to account for the spatial variability of crop and soil parameters to some extent, and the Savitzky-Golay (S-G) filtered MODIS LAI was then assimilated into the WOFOST model for regional winter wheat yield forecasting. We constructed a four-dimensional variational data assimilation (4DVar) cost function to account for the observations and model errors from Feb. 10th to Apr. 30th, and the Shuffled Complex Evolution-University of Arizona (SCE-UA) algorithm was used to minimize the cost function by reinitializing three WOFOST parameters. The winter wheat yield forecasting date was started from Apr. 30th, and the results showed that assimilating MODIS LAI into the WOFOST model substantially improved the accuracy of regional wheat yield predictions (R = 0.60, CCC = 0.53, RMSE = 619.73 kg/ha) compared with the unassimilated results (R = 0.35, CCC = 0.24, RMSE = 857.32 kg/ha), and the relative error (RE) between the averaged predicted yield and official statistics for most cities decreased after data assimilation. This demonstrated that assimilating MODIS LAI can optimize the simulation of LAI in the WOFOST model, thereby further reduce the uncertainty of yield forecasting. These promising results highlighted the potential of integrating remotely sensed data, crop model and weather forecasts for in-season prediction of crop yield at the regional scale.

View all citing articles on Scopus

View full text

Methodology for the use of DSSAT models for precision agriculture decision support

Abstract

Introduction

Section snippets

Building model input files for spatially varying parameters

Case study: impacts of climate change on spatial yield variability

Conclusions

Acknowledgements

Eur. J. Agron.

Agric. Water Manage.

J. Econometrics

Eur. J. Agron.

Agric. Syst.

Agric. Syst.

The impact of climate variability and change on crop yield in Bulgaria

Agric. Forest Meteorol.

An evaluation of the SWATRER and CERES-Millet models for southwest Niger

Regional comiate change scenarios over the United States produced with a nested regional climate model

J. Climate

Evaluating management zone optimal nitrogen rates with a crop growth model

Agron. J.