A Study On Ecological Zoning Of Wheat Grain Quality Based On GIS And Modeling

Spatial analysis and zoning on the similarities and differences of crop ecological factors and responses are of vital significance for regional crop production and food security in China. It helps to design high-quality, high-yield and high-efficiency crop production patterns, drive its regionalization and industrialization, as well as make full use of natural resources and excavate species' genetic potentials. Geostatistic theory serves as one of effective means to study spatial distribution and variation of production factors; however, it has rarely been used in quantified spatial analysis and zoning of crop quality. In this research, firstly, spatial data sources were built as input variables of a wheat grain quality estimation model, spatial scaling-up methods for model prediction were studied, and spatial variability of grain quality in wheat was quantified to the granularity of grid based on the geostatistical theory. Then, a preliminary digital ecological zoning scheme of wheat grain quality was proposed based on the unsupervised classification algorithm. Finally, a comprehensive digital wheat farming support system is established through effective coupling and integration of database, wheat models and WebGIS platform. The results will help promote applications of crop models at regional scale, evaluate the spatial variation of wheat quality and make ecological zoning, and provide the technical basis for development of digital wheat farming and modern agriculture.First, based on the parameter demand of the wheat grain quality estimation model and the change characteristics of regional climate, the method of Ordinary Kriging was determined as the appropriate spatial weather data interpolation method by comparing the precision of three different spatial interpolation methods. Then, the actual daily maximum temperature, minimum temperature, rainfall and sunlight hours, obtained from different weather stations within Jiangsu Province and across main winter wheat growing regions in China, were interpolated using this method at the 5km×5km resolution to generate the spatial grid map for two study areas of these four climate elements on a daily basis, as well as the spatial grid map of the four climatic factors (average daily temperature, difference of diurnal temperatures, total sunlight hours, and total rainfall) for six winter wheat varieties from anthesis to maturity. The results provide the basis regional data platform with high spatial-temporal resolution for scaling-up of wheat grain quality estimation model, and for analyzing spatial quality variation and making spatial ecology zoning.By analyzing the estimation mechanism of wheat grain quality model and technical approach of model scaling-up, the scaling-up method was explored based on the regional data platform. First, the regional performance of the wheat grain quality estimation model was simulated in two different ways. One way was to run the model at each location and then interpolate its results on a grid (i.e., "calculate first, interpolate later", denoted by "CI"). The other way was to interpolate the model inputs on a grid firstly, and then run the model at each grid nodes (i.e., "interpolate first, calculate later", denoted by "IC"). The IC technology turns out to be suitable for scaling up the wheat grain quality estimation model. Then, the model was applied at regional scale under various climate conditions, and the results were largely consistent with the results obtained by applying wheat grain quality estimation model under point scale scenarios.Based on the geostatistical theory and the spatial pattern of grain quality variation, quantitative methods to explore the characteristics of wheat grain quality variation were studied with semi-variance function theory and the suitable model scale-up method. First, regional simulation results were analyzed with semi-variance function; secondly, based on the characteristic parameters of the fitting function, the grain quality was quantitatively analyzed for spatial correlations, the maximum correlation scope, as well as spatial variability in different directions. The results show that based on geostatistics and scale-up model, quality indices based on grid can be effectively integrated with geo-spatial coordinates. The spatial variation of regional grain quality indices can be quantified, and the fine mapping based on grid can intuitively quantify the spatial distribution and variation trend of regional grain quality.The digital wheat quality zoning in China's main winter wheat growing regions was established based on the up-scaling wheat grain quality estimation model and the quantitative analysis of spatial variability. First, the climate environment of China's main winter wheat growing areas and the predicted quality results were quantitatively analyzed. Then, the quality indices simulated through up-scaling model were integrated with grid and an established spatial calculation model. The main quality indices were clustered spatially based on the non-supervision classification algorithm. Finally, according to the national standards of wheat grain quality, the statistical characteristics of main quality indices from the same region were analyzed and evaluated, resulting in a digital wheat grain quality ecological zoning method. Based on the above method, China's main winter wheat growing regions were divided into five primary ecological zones, with a fine mapping based on grid for a visual, quantitative and spatial expression.Based on the above-mentioned results, the mechanism and methodology for integrating the GIS technology with wheat grain quality estimation model, management knowledge model and growth simulation model were further explored. By designing a reasonable framework of system architecture with the XML technology plus a custom map engine, a model and WebGIS-based digital wheat farming support system (MGDWFSS) was developed. The system realized the comprehensive functions as suitable cultivation plan design, dynamic growth simulation, yield and quality forecasting, quality ecological zoning, and management decision-making in a network environment. Overall, the present study has provided quantitative methods and key technologies for up-scaling crop system models. It has also constructed a digital platform for regional productivity analysis and evaluation as well as management decision-making in wheat production.