Design And Implementation Of Maize Growth Simulation Model For Northwest Region In China

The global food supply is mainly depending on the cereals as a staple food and maize is one of the most important cereals in worldwide. Since it can be grown a wide range of agro-ecological environments all over the world and high value of economic importance, it is identified that maize has a large potential for a high yield production. Improved management practices aimed to increase the production in regional level are in demand due to this situation. Therefore, the regional level accurate yield estimation of the maize has become more important and, in fact, a big challenge. Consequently, improvement strategies to increase the yield are to be introduced. Therefore, several methods to expand the production through better solutions for the problems arise during the seeding to the harvesting have been used. Therefore, maize growth simulation models have become more popular replacing traditional yield monitoring methods. However, the regional yield simulation is not in a strong position, thus the accurate crop yields cannot be estimated in regional level. Therefore a region-specific maize model will have a greater value in yield simulation in a particular region. Meanwhile, the weather often plays an important role in crop growth due to specific variations and uncertainty of weather occurrences specially in regional level. Crop growth responses cannot accurately be estimated using historical weather averages, however, crops respond to the interactions which occurred between daily sequences of weather and other abiotic factors such as soil, water and nutrient balance. Reliable seasonal forecasting of weather at the starting of the cropping season could be a better solution to accurately estimate the production of a particular crop. Thus, there is a requirement to generate weather records on daily basis matching the statistical characteristics as much as closer to the actual records in particular region. As a result, this research was aimed to develop a maize growth simulation model integrating a weather sub model. In order to achieve this target, the research was conducted in two phases. In the first phase, a weather generator(CMWSim) was developed and evaluated for performances. In the second phase, a maize growth model(MAIZESim) was developed integrating the weather generator and evaluated for performances. Design and implementation of both models and evaluation results are presented.The proposed weather generator, CMWSim is a stochastic type software developed as a sub module of MAIZESim. The model generates daily weather records including precipitation and maximum and minimum temperature. The probabilistic simulation approach(explained by Markov chain analysis) was applied to the design of the CMWSim. It simulates(i) the precipitation occurrence,(ii) the precipitation amounts and(iii) the temperature values. Transitional probability matrices explained in firstorder two-state Markov chain model was based to calculate precipitation occurrence. The calculation of precipitation amounts was based on two-parameter gamma distribution function. The maximum and minimum temperatures were calculated using first-order auto-regressive model with applying a conditional scheme. The model was implemented as a user-friendly software which can run on Windows environment. As the first phase of the evaluation, only the weather sub model(CWMSim) was evaluated by statistically analyzing and comparing the observed and simulated data. The precipitation patterns were well preserved and all comparisons were statistically significant. Although there were some slightly over-estimated precipitation at higher intensity periods between July to September, the analysis was statistically acceptable. Maximum and minimum temperature values were also well fitted with observed values and no misleading results found. Thus, it is appeared that the simulated daily weather records by CMWSim were statistically acceptable and it could be used as inputs of the maize growth model(MAIZESim).The proposed maize growth model, MAIZESim, is a temperature driven maize growth simulation model which was developed by analyzing the quantitative growth of maize in daily basis and emphasizing the potential yield prediction with special focus on Northwest region in China. The model simulates daily maize growth and development, total accumulation of dry matter and final grain yield for one cropping cycle. The simulation runs through seven major stages for one cropping cycle including sowing date, germination and emergence, three-leaf unfolding, jointing, booting, spinning and harvesting. Primarily the model development was based on the Growing Degree Days(GDD). The model validation was done using five years real field observations(five cropping cycles in 2005, 2006, 2007, 2009 and 2011) collected from Yangling in Northwest region in China. Simulated and observed data were statistically analyzed and compared. The analyzed results have clearly shown a significant relationship between observed and simulated values. Based on the resultsand the comparisons of simulated and predicted values, it appears that the model can be used as a prediction tool in maize cultivation and also an economic management tool in long-term economic planning in regional level. Further the model provides positive evidence that it could be used as a research tool for maize growth and development. However, the MAIZESim is evaluated only for Xi’an in Northwest region in China. Further, field trials to be done for other locations of Northwest China to accurately validate the model. The model simulates through one cropping cycle, and therefore it should be expanded to simulate multiple cropping cycles. The model simulates the potential maize production, and therefore it has not been evaluated under nutrients and water limited conditions. Further improvements should be done including water and nutrient limited conditions to increase the accuracy and prediction capabilities. Integrated weather generator helps to consider multiple cropping years and water stress conditions as it predicts the weather records accurately including precipitation.