Optimized Spatial Sampling Method For Complex Marine Environmental Monitoring Data

The establishment of monitoring network, covers space, sky, earth and underground, lays a good marine data foundation for the development of marine resources and economy. But restricted by hardware or software capability of marine data management, analysis and application, the phenomenon of "getting little knowledge from big data" emerged increasingly. Consequently, how to extract marine information from massive marine data rapidly and provide smart decision for ocean affairs is one of the most hot spots currently.Sampling survey is a good way to get key data from big marine data rapidly, applies to large scales data at low cost of investigation, and short periodicity is another main reason to use sampling survey for the rapid application on big data. Nevertheless, sampling survey methods mostly depend on traditional probability sampling theory, so the existing sampling methods will be challenged because how to get effective data to satisfy all kinds of requirements is still a big problem now.Marine environmental monitoring data is obtained by buoy, survey ship and manual monitoring, the data characteristics and the problems result from them can be concluded as:(1) Mass, marine environmental monitoring data is increasing rapidly, updating frequently and multi-source, the data is redundant in both space and time dimensions.(2) High spatial correlation, the attribute data in nearer place has higher similarity which is easy to cause space related failed and sample overlapped, sample precision drop.(3) Spatial heterogeneity, the data contains complex information and uneven spatial distribution, brings difficulties to data reuse and post process. Consequently, it is worth while studying an optimized spatial sampling method for marine environmental monitoring data, taking these characteristics into consideration, making good use of data.Reduce sampling cost will cause estimation precision deviated and the result of sampling distorted when design sampling and estimation, excessive numbers of sample will cause improving cost directly. Consequently, how to balance the precision of sampling, and taking cost saving into consideration is important. The main content of this article can be listed below:(1) Analyze marine environmental monitoring data which has multiple modal, high multiple dimensions and multi-attribute characteristics, analyze research status for existing sampling methods, discussing the issues by applying these sampling methods to marine environmental monitoring data.(2) An optimized system space sampling method is presented, considering spatial correlation of marine environmental monitoring. The semi-variable function is taken into this space sampling method design, giving the consideration of sample points uniform distribution in research sea area, and reducing information redundancy under ensuring sampling precision.(3) The optimized method meets the application requirements for this patch of marine environmental monitoring data is expanded by calculating the weight of each attribute of marine environmental monitoring data, considering the relationship between attributes. The optimized method can be used in the comprehensive evaluation by multi-attributes of marine environmental monitoring data, more comprehensive and more economic.(4) With the spatial data of a sea as an example, the optimized method for multi-attribute marine environmental monitoring data is arguably analyzed by variance, sample size and trend surface. The optimized method can reduces the number of sampling data while ensuring the accuracy of sampling compared with traditional systematic sampling.