Sampling distribution method for monitoring quality of arable land in county area based on spatial balanced
Yang Jianyu,Yue Yanli,Song Hairong,Ye Sijing,Zhao Long,Zhu Dehai
DOI: https://doi.org/10.11975/j.issn.1002-6819.2015.24.042
2015-01-01
Abstract:As a large agricultural country, China has a large population but not enough cultivated land. In 2011, the cultivated land per capita was 0.09 hm2, only 40% of the world average level; and it is getting worse with the rapid development of economy, industrialization and urbanization. Through the monitoring network for cultivated land quality in county area, the distribution and change trend of the cultivated land quality can be reflected. Besides, the quality of non-sampled locations should also be estimated with the data of sampling points. Therefore, this paper proposes a new sampling method for monitoring the quality of arable land in county area based on spatial balanced sampling, which is a pre-processing method to determine the number of sampling points, including preprocessing the data of cultivated land quality before sampling, exploring the spatial correlation and spatial distribution pattern of cultivated land quality, and computing the appropriate quantity of sampling points by analyzing the change trend of sampling number and sampling precision. And the spatial balanced sampling method is aimed to optimize spatial sampling design for setting up the monitoring network. It is required for sampling of a population to understand the trends and patterns in natural resource management because of the financial and time constrains. Spatial balanced sampling provides the mathematical foundation for statistical inference, and is efficient but remains flexible to inevitable logistical or practical constrains during filed data collection. There are integrated factors that affect arable land quality inventory and monitoring, such as geomorphic conditions, altitude, gradient and transport cost. Factors are commonly used to modify sampling intensity; some factors, such as category, gradient, or accessibility, can be readily incorporated into the spatially balanced sampling design. In this paper, we take the distance between the sampling points and the main roads, the slope of terrain and the sample size of each grading according to stratification sampling method as primary factors to generate the raster layer containing probability, by considering the cost of monitoring and the precision of estimation; and on this basis, the monitoring samples are selected by spatial balanced sampling method. Taking the Kriging standard error and the transport cost as the optimization criterion, the experiments in Ji'an County are conducted to compare this method with traditional sampling method in cost (the average distance between the sampling points and the main roads) and estimation accuracy (the mean of Kriging standard error). Seventy-eight monitoring of reference sample units are finally deployed, and the average of ordinary Kriging standard error of the proposed method is 140.23, which is smaller than the simple random sampling (216.96), the stratified sampling (157.14) and the traditional grid random sampling (152.70); the transport cost of this method is 2 277.95 m, which is lower than the simple random sampling (2658.93), the stratified sampling (2726.59) and the traditional grid random sampling (3221.83) when the quantity of samples is the same. Therefore, the result illustrates that the estimation accuracy of this method is higher than the simple random sampling, the stratified sampling, or the traditional grids random sampling when the number of sampling points is 78. Besides, the transport cost of this method is significantly lower than the traditional methods. Therefore, this method can meet the need of montoring the classification of cultivated land in county area.