Sampling Strategy And Management Zone Delineation Based On Soil Spatial Variability In Coastal Saline Land

There are very abundant coastal land resources in Zhejiang province. As main mothball land resources, these coastal lands are very important to alleviate the conflict between sparse land resource and high population. The provincial government paid much attention to the reclamation, exploration and application for the coastal saline land, and the reclaimed lands have produced plenty of primary products and create greatly social and economic benefit. Because the soils in coastal region developed mainly from modern marine or fluvial deposits, this made them have the characteristics with the low nutrient and high salinity, which had become main limiting factors of continuably agricultural development in coastal saline land. The phenomena of high salinity content, low organic matter, available nitrogen and phosphorus is very pervasive in the reclaimed region, plus frequent typhoon attacks and lack of freshwater resources, which made the soil have very low output and agricultural benefit. As is reported, 70% of reclaimed lands are graded as middle or low production level. In order to make coastal land develop from conventional agriculture to modern agriculture, to realize effective production of agricultural land and soil resources, and high yiled of primary products, it is necessary to manage coastal soil resources scientifically and rationally by applying the principal ideas of precision agriculture and by adopting modern spatial information tenichque and means.Soil was a natural synthesis with complicated configuration and evolvement process. The soils with same texture were not completely homogenous in the same time and in the different spatial position. In the same way, the variable values of soil properties were not equivalent in different time and in the same spatial position. This change was named as temporal and spatial variability of soil properties. The study on soil spatio-temporal variability is very important to promote the development of soil science, hydrology, geology and so on. So, the study of spatio-temporal variability for soil properties was the study hotspot in soil science since 1970s.Recently, the spatial or temporal variability of soil properties become theoretic basis of precise agriculture. Understanding the spatio-temporal variability of soil properties and crop yield in farm scale, finding out the causes of yield variability, putting specific management measures into practice and assessing their economic benefit, accumulating spatial resource information to optimize management strategies, all these were the main ideas of precise agriculture. Soil is the basis of crop growth, the variability of soil properties will arouse the variability of crop growth. Only the variability of soil properties was admitted, controlled and assessed effectively, which was then combined with yield and quality of the crop, and the response mechanism of crop growth to soil spatio-temporal variability was understood, then precise agriculture can be carried out successfully. The study on spatio-temporal variability of soil properties not only can improve the precision of soil investigation, soil mapping and field experimentation, but also can ascertain rational sampling scale and sampling number in the farmland and can provide effective approaches for estimating and simulating soil process accurately. It can present vividly the effect of temporal and spatial variability on soil properties and soil-crop system, hence enhance the understanding of the relationship between soil and crop, whereas these quantificational information are the indispensably basic data and theoretic foundation for developing precise agriculture.Today, more and more soil scientists began to attach importance to the spatial variability and spatial distribution of soil properties. In this study, conventional statistics, Geostatistics and the spatial analysis technique of geographic information system were used to analyze spatial variability of soil properties in the study area, based on which, to design and optimize soil sampling strategies, classify precise management zones, finally, set up spatial variability analysis software package for soil properties. Some satisfied results were obtained as follows.(1) Establishment of optimal soil sampling strategy and study on methods of improving soil estimation precision.Three methods were utilized to establish and optimize soil sampling strategy, and to improve estimaion precision of soil salinity. ? Ordinary kriging and ordinary cokriging were used to study the estimation precision of profile soil electrical conductivity in coastal saline soil. The results indicated that cokriging, which used electrical conductivity data of surface layer as secondary variable, not only can reduce the sampling number of soil electrical conductivity in deeper layer soil, but also can improve their estimation precision. ? Regression kriging was provided to estimate soil salinity with reduced sample number in latter period by using the spatial variability of soil salinity data in previous period as secondary information. The results indicated that, compared with ordinary kriging and ordinary cokriging, regression kriging can greatly improve estimation precision of soil salinity in latter period and can save the samplingcost. (3) Four quad-tree method combined with semivariogram function (variance quad-tree method) was used to design the optimal sampling scheme of soil salinity in the study field. Ordinary kriging was used to interpolate the unsamplng positions by using the samples from regular grid sampling and variance quad-tree sampling. The results revealed that with same size of samples, the estimation error from variance quad-tree sampling scheme was much smaller than that from the grid sampling, and the sampling efficiency was improved up to 16% to 25%. The advantage of variance quad-tree method was that this scheme samples sparsely in uniform areas and more intensively where variation is large. Thus the maximum sampling efficiency was achieved while ensuring the variability within the sampling area is characterized effectively.(2) Study on classification of precise management zone in the coastal field.Precise management zones were defined by classifying the study field into sub-regions according to heterogenous combinations of the factors influencing crop yield, then uniform management measures were carried out in the same management zones and variable management schemes were implemented in defferent management zones to achieve the goal of precise soil management. This study used the most limitingfactor in coastal saline region------the spatio-temporal variability of soil salinity to classifythe precise management zone. At the same time, SPOT remote sensing image was combined with soil salinity data, nutrient data as multiple data sources, then correlation analysis, stepwise regression analysis and principal component analysis were used to select, compress and deal with the main limiting factors affecting soil quality and crop yield, finally fuzzy clustering arithmetic was utilized to classify the management zones, fuzzy performance index and normalized classification entropy were provided as indices to ascertain the optimal clusters number. It was found that the management zones classified by the two methods had good similarity. The results not only provide new approach for soil and agronomy parameters sampling, but also provide effective approach and foundation for soil precise management and basic information for precise agriculture.(3) Seting up spaial variability analysis software package for soil properties base on basic geographic information system platform.Based on the previous studies, soil spatial variability analysis software package was set up by using visual C++ software combined with database and modeling technique,. This software not only can analysis spatial variability of soil properties, but also can used to design sampling pattern and classify managements zones, which provide desion-making base for effective management and exploitation of soil resources in coastal region.The innovations or new developments were made as follows:(1)Variance four quad-tree was used to design effective soil sampling pattern in coastal saline land by using prior spatial variability of soil salinity. Cokriging was used to optimize the profile soil sampling strategy, and regression kriging was provided to design sequential soil sampling by using the similarity of spatial structure between previous and later samples. The study provided reference and effective means for methodology of reducing sampling number, saving sampling cost and impoving estimation precision.(2) Soil spatial variability analysis software pakage was set up by using visual C++ programme language and system integration technique in windows environment. The software can be used to analyse spatial variability of soil properties in the field, to design economic and effective sampling scheme, to classify precise management zones and mapping, thus can provide technical supports for precise soil management, consultation and decision-making. The software has many good characteristics such as powerful function, friendly interface, simple manipulation procedure and comprehensive information, so has good extending value. It can be applied widely to analysis the spatial correlations of spatial data with geographic properties, and interpolate and model their spatial distributions.(3)The spatial variabilities of soil salinity, nutrient and crop yield were used to classify management zones by using fuzzy cluster analysis. The classified management zones can be used to sample soil and agronomy parameters by treating management zones as sampling units, and to carrying out variable input and precise management decision-making according to spatial variability of different management units, and to provide basic information and effective means for precise field management.