Study On Cotton Growth Simulation And Decision Support System

The crop growth simulation model and decision support system (DSS) for cultural management are the core contents of informational and digital agriculture. The crop growth simulation model is mainly characterized with the functions of systemic integration and real-time prediction. The dynamic model on crop quality formation is much needed for quality prediction under various conditions. Also, the growth model-based DSS for crop management is helpful for conducting strategy analysis and practice evaluation through different simulation trials. On the basis of the cotton growth simulation model established by the author's laboratory, using the system analysis method and mathematical modeling technique, this study developed a growth period simulation model based on the physiological development time (PDT), morphogenesis and abscission models, a model of dry matter accumulation and distribution in boll shell, fiber and seed, and spatio-temporal distribution models of fiber quality indexes. By further adopting the object-oriented programming and component technology, a comprehensive and component-based cotton growth simulation system (CGSS) was constructed for simulating cotton growth and development, yield and quality formation. Then, a growth model-based decision support system for cotton management (GMDSSCM) was developed with the function of decisionmaking support by combining the techniques of simulation experiments and strategy analysis. This work has provided a basic platform for development of digital cotton farming system, and should be helpful for construction of DSS for other crop management. The main results of the present study are summarized as follows.Based on the dynamic relationships between cotton growth and development and environmental factors such as fertilizers, genotypes and sowing dates, cotton growth period models were developed with consideration of relative thermal effectiveness (RTE), relative photoperiod effectiveness (RPE), variety earliness (VE), and compensation efficiency (CE). And in consideration of effects of the other main factors such as nitrogen nutrition, soil water stress and DPC regulation, cotton morphogenesis and abscission models were established too. Validation of the growth period model with data sets from different years, ecological zones, genotypes and sowing dates indicates a high goodness of fit between the simulated and observed values. The value of the root mean square error (RMSE) is 0.80 tol.65 for different growth stages, and the RMSE for main stem leaves, plant height, fruiting branches, branch nodes, abscission number and open bolls was 0.36, 1.11, 0.57, 2.87, 1.00, 0.59 respectively, and the error to observed values was less than 5% too. In the general, these models are accurate, applicable and predictable for predicting cotton growth period, morphogenesis and bud and boll abscission under different conditions.Cotton dry matter accumulation models were developed on the basis of RTE, in consideration of effects of the other main factors such as nitrogen nutrition, water stress and the ratio of supply and demand of assimilation product on the development of cotton bolls, with different genotypes, amount of applied fertilizer and different flowering stages. The results showed that the value of RMSE between simulated and absorbed values for dry matter accumulation and distribution of a single boll, seed cotton and fiber of cotton bolls of pre-July 20, July 21 to Aug.15, Aug.16 to Aug.31 and post-Aug.31 was 0.1767-0.5659, 0.0725-0.5279, and 0.0613-0.2634 g, respectively. RMSE of a single collective model for dry matter accumulation and distribution was higher or vicinaler than that of the models simulated by stages. This system of analysis methodology for cotton boll dry matter accumulation and distribution is precise and reliable.Based on the field experiments, effects of genotypes, different cotton fruit branches and nodes, mean daily temperature, sunlight length, soil moisture and nitrogen content of cotton plant on cotton fiber quality indexes were quantified at flowering and boll-forming stages, and spatio-temporal distribution models of cotton fiber length, strength, micronaire value and length uniformity were constructed based on present cotton fiber quality eco-models. These models were validated using the experiment date sets of different eco-sites, varieties, N fertilizer and water conditions too. The results showed that the value of RMSE between simulated and observed values for temporal distribution models of cotton fiber quality indexes of fiber length, strength, micronaire value and length uniformity was 0.15 ram, 0.29cN·tex^-1, 0.18 and 0.36, and the RMSE of spatial distribution models was 0.22 ram, 0.60 cN·tex^-1, 0.15 and 0.86, respectively, the error to observed values was less than 5% too. In the general, these models are accurate, applicable and predictable for predicting spatio-temporal cotton quality indexes under different conditions.By analyzing the cotton growth and development, using object-oriented program design and software component technology, the more mechanical and comprehensive cotton growth model (CottonGrow) were established. And based on CottonGrow, meteorological generation model and variety parameter generation model, CGSS with powerfully suitable, accurate and easily useful properties was programmed by Visual C++ 6.0 on Windows 2000. The CGSS includes major models of cotton growth systems, viz, development process, organ establishment, photosynthetic assimilation, biomass partitioning, growth respiration, yield and quality formation, nitrogen dynamics and water balance, and it can quantificationally forecast the cotton growth and development progress and the dynamic development of cotton productivity. The system realized the effective integration of the yield and quality prediction for overall cotton productivity, and provided the key techniques and framework for development of crop growth simulation system and growth model-based decision support system for crop management.Based on CottonGrow, weather generation model, variety parameterization model, and strategy evaluation model which was developed for analyzing the simulation results from different simulation experiments, GMDSSCM was established on the platforms of Visual C++ and Visual Basic by adopting the characteristics of object-oriented and componentbased software as non language relevance, re-utilization and portable system maintenance. The GMDSSCM can be used for realizing data management, growth model prediction, real-time regulation, temporal and spatial analyses, and individual and comprehensive management strategies valuating. The established system can be used for cotton growth simulation under various environments, growing conditions and different genotypes.