Research On Spatialization Of Grain Yield And Error Analysis

Grain is the foundation of agriculture and the basic consumer goods of human life. The food problem of countries is the most important thing that is related to people's livelihood of the national economy and the security, stability and development of countries. Therefore, food production is an eternal theme related to survival and development of countries. As the most populous country of earth, China is also the largest food production and consumption country of earth, so its statistical data of grain yield have important research value.The statistical data of grain yield are generally recorded by administrative units. Therefore, grain yield can only be regarded as uniformly distributed in an administrative zone though the actual situation is not so. Statistical data of grain yield is not suitable to comprehensive analysis among different study areas. However, Spatialization of attribute data can contribute to comprehensive analysis of grain yield with other natural and cultural data. Since the concept of spatialization technology was proposed, domestic and foreign researchers have done a lot of research in the fields of spatialization of attribute data, and put forward various kinds of methods and models. But, we can know that the research on spatialization of grain yield is still relatively small in the existing research. Therefore, this paper is to analyze the relationship between different data sources, different sample scales, different partitioning schemes, different error correction methods and the accuracy of spatialization of grain yield by the multiple linear regression analysis method. This paper was divided into two parts to analyze the relationship between different data sources, different sample scales, different partitioning schemes, different error correction methods and the accuracy of spatialization of grain yield. The following work has been carried out and the corresponding conclusions are obtained:(1)In the third chapter, the relationship between different sample scales, different partitioning schemes and accuracy of spatialization was explored by using multiple linear regression analysis model. Spatialization of nationwide grain yield relates to sample scales and partitioning schemes. Different sample scales and partitioning schemes will inevitably account for different errors of spatialization. In this part, models considering farmland distribution and sample scales and partition schemes were proposed to estimate grain yield and its spatial distribution. The grain yield data were collected from 2005 Yellow Book of China. Data of paddy field, irrigated land, and dry land areas in each county or district were calculated. Four datasets of 3 scales were selected including total grain yields of counties, total grain yields of prefectures and their average grain yields. A total of 2321 county data and 349 prefecture-level data were obtained. 3 partitioning schemes(no partition of China, 7 partitions of China, partitions of China by province) were considered. A total of 15 kinds of multiple variable linear models were constructed with area of different types of farmland as independent variables, grain yields as dependent variables. The results showed that:(1) For models without constant term, accuracy of spatialization results increased first and then decreased with scaling down of partitioning scheme. For models with constant term, accuracy of spatialization results decreased with scaling down of partitioning scheme.(2) In the 2 partitioning schemes(no partition of China, 7 partitions of China), accuracy of spatialization results increased first and then decreased with scaling down of samples from prefectural level to county level and 1 km by 1 km level.(3) Compared with other models, in the case of county grain yields as samples, without constant term and 7 partitions of China, the accuracy of result was the highest(Coefficient of determination was 0.655). By using the proportional coefficient method, the accuracy of result is improved further(Coefficient of determination was 0.968).(2)In the fourth chapter, the relationship between different data sources and accuracy of spatialization was explored by using multiple linear regression analysis model. spatialization of grain yield of Henan Province was taken as an example to analyze the relationship between different data sources and the accuracy of spatialization of grain yield. 3 data sources of different farmland types were selected including land cover dataset, land use dataset and statistical data. A total of 126 county data were obtained. A total of 3 kinds of multiple variable linear models were constructed with area of different types of farmland as independent variables, grain yields as dependent variables. Then, 7 kinds of error correction methods were used to correct spatialization result of grain yield of Henan Province. The results showed that: The accuracy of spatialization results obtained by land cover dataset is the highest, the accuracy of spatialization results obtained by statistical data is second, the accuracy of spatialization results obtained by land use dataset is the lowest.(3)In the fourth chapter, different error correction methods were used to correct the preliminary results, and the influence of different error correction methods on the accuracy of spatialization was explored. 7 kinds of error correction methods were used to correct spatialization result of grain yield of Henan Province. The results showed that:(1)The average correction method, weight coefficient correction method?and weight coefficient correction method?cannot be used to correct preliminary results of spatialization.(2)Proportional coefficient correction method, weight coefficient correction method?, weight coefficient correction method ? and weight coefficient correction method ? can be used to correct preliminary results of spatialization. Weight coefficient correction method?is the best method to improve accuracy of spatialization result, weight coefficient correction method? is second, proportional coefficient correction method is third and weight coefficient method ?is worst.(3)Accuracy of spatialization after error modification based on error correction methods, which can improve the accuracy of preliminary spatialization results, are close.This research made up for the deficiency of spatial error analysis of grain yield, explored the relationship between different sample scales, partitioning schemes, data sources, error correction methods and spatial error. Although the conclusions were got by studying spatialization of grain yield, it was also applicable to other research of spatialization of social and economic statistical data. In next researches,in order to improve the accuracy of spatialization results, it is necessary to consider more factors and methods.