The Statistics Method Of Information In Space-time Domain And Its Application In Assessment Of Surface Water

The Statistics Method of Information in Space-time Domain and its Application in Assessment of Surface WaterThe research work of environmental science realm involves to many spatial datas and variables, and these variables exhibit in space-time area and present a certain random and structure. For example, in the research of the space distributing structure of a certain pollutant in the atmosphere, space variation of the pollutant in the soil, design of pollution monitoring net of groundwater, and the variety trend of surface water quality with time, etc, we want to analyzes the characteristic of the space distribution of these variables and make a superior estimate to them, or we want to imitate the long-lost and motion or other properties of the researched object. We will choose a suitable statistics method. The geology statistics method also be called a statistics method of spatial information. It takes regionalized variable(z(x),x∈G) as basic theory, and take variogram as basic tool to study the statistics method which is structural and stochastic. It has obvious advantage in optimal sampling strategies, analyzing to spatial data, handling irregular sample and optimal calculus of interpolation, etc. It is extensively applied in the realms, such as geography, ecology and environmental science, etc. Especially, the development of GIS(Geographic Information System) brings to the biggest abundant spatial datas and need to use statistics method of spatial information to analyze spatial data. In a lot of natural science researches and the problem of engineering technique involed to the sample which often takes its space position as a characteristic and to describe some one phenomenon or process with the single or multivariate.Information provided by this sample, should be related from the position or variable to study these informations. Under this kind of circumstance, to establish the best mathematics model of a certain phenomenon or process, we also adopt the statistics method of multivariate or single information(the geology statistics). However, in another research realm of natural science and the problem of the engineering technique, involved of variables which not only have a spatial characteristic, but also have the characteristic in time, then must take the variables studied as random functions of time-space. Study an useful information problem in this kind of time-space area, we can learn the theories and method of the spatial statistics (the geology statistics) by extending them to time-space area. Provide a new kind of method for analyzing the datas which depend on time and space, and provide a new method of establishing model for forecasting effectively the pollution events researched and the distribution of polluted material in time-space area, thus provide quantitative methods and basis to the pollution treatment, ecosystem repair, and play an important role in the management, programming development and protection of the environment resources and the people's health.The main work and result of my paper are as follows: First, summarize the basic concept and basic theories of statistics method of space information. For example, introduce the basic theories of statistics method of multivariate information in space-time domain(regionalized variable(z(x),x∈G),that is random function, basic tool variogram (γ(r)) ,the relation between variogram and Covariance and their properties, and while handle actual problem, how to make their approximate estimatation. After that, introduce the technique method of information in space, that is Kriging. Kriging method is the special form of the traditional regression analysis. It is divided into Kriging method of single variable and CoKriging method of multivariable, and both of them contain simple Kriging estimate, Ordinary Kriging estimate and universal Kriging estimate. These basic theories are pre-knowledge. In chapter 2, introduce the statistics method of multivariate information in a pure time area and discuss some theory of myself. The statistics method of multivariate information in a pure time area is a small branch of the statistics method of information in space-time domain. Its basic theories method is similar to the spatial statistics method. Just studied an area to be not space area any more, but see the space area as a point, and consider the problem about time series. For instance, the observation of some cosmos bodies, such as sunspot, the change of linar surface. etc. And monitor the safe scope of harmful composition in the atmosphere at some monitoring point; give a monitoring analysis of the tallest water level line of some important marine hydraulic engineering, such as reservoir, river and so on; and make successively sampling at some key oil well when produce and develop in the oil-field. etc. All of these researches can be formulated as statistics method of multivariate information in a pure time area. In this part of contentses, some properties of variogram are discussed. By thinking it over, discover that the variogram is similar to which about spatial statidtics, in fact. Just their denotion is different. So this chapter contents is put at the back of chapter 1. And then in chapter 3, summarize some of basic theory and method of the statistics of information in space-time domain. It includes random variables(Z(u,t))and random function of the space-time area, second order stationary hypothesis and the intrinsic hypothesis, stationary linear optimal estimation and non-stationary linear optimal estimation of information in space-time domain. In chapter 4, discuss some new theories basis on chapter 3. Regionalized variable has some different characterizations between in time area and in space area. The quantity of space information and time information usually is unbalant. So sometimes we solve the problem in two case, that is considering which have larger information. Analyzing specifically of specific matters. What my paper discuss is the cace of the quantity of time information larger than space information of former. And the other case can see [9]. In chapter 5, Kriging method is applied to solve specific problem and is verified the possibility of this method. Here, it use the statistics method of multivariate information in a pure time area to estimate and appraise water quality of some wetland. From corresponding charts to see the effect still is feasible. Under the condition of minimum variance, this kind of method can make the estimator is unbiased. It imply this method is a good interpolation method. Also we can use the Kriging method in chapter of this paper to handle this problem. I will continue to study this topic in the future.