Study On Theory And Application Of Statistics To Groundwater Pollution Assessment

The evaluations of Groundwater pollution require both objective depict and forecast of the space distribution and the space-time's change about the contamination. On the basic of which to make evaluation of risk in water pollution. So we need not only numerical simulation and forecast model but also the more proper statistics analyses and depict methods.The research is carried out with the objective to study the law of time-space change of concentration of soluble substance during its diffusion and migration process. The research can provide scientific basis for effective prevention of water pollution and preferred plan for recovery.The mathematic model of the groundwater can be divided in determinacy model and randomness model. The determinacy model is on the theory basic of groundwater dynamics, and use PDE to calculate discharge and simulate the shift and change of the contamination. The randomness model always used for estimate of water parameter. From the long history of this study, the mathematic model is concentrate in determinacy model, the other is in the stage of developing, but they all require statistics analyses about results and describe the space distribution also the developing of contamination. So the describe method of the above problem is very important. The main content of this article is the statistics method of evaluate forecast and risk analyses about the groundwater assess.This article concentrates upon the assessment of water quality, which is an issue within environmental science, by introducing geological statistic methods into time-space domain since the target variable is not only a spatial one but also a time variable. Geological statistics, which is also called Kriging method, is the best unbiased estimation method that could be applied to any estimation procedure related to spatial data. With the aim to enhance the accuracy and stability of estimation and strengthen the estimating function, many new models have been created so far, forming many new branches. However the application of it in environmental science is still to be improved with very few examples of successful application. Based on refined mathematic models using Kriging method, this article point at assess, forecast and risk analyses of the groundwater pollution evaluate, then do several jobs as following:(1)After analyzing the statistical characteristic of the monitoring data, several features could be seen. First, the data is closely related to time and space. Different time, place and scope of monitoring can affect the data structure and probability distribution greatly. Therefore, both the spatial distribution and the timing changes should be considered in the research. Second, the data has significant randomness and continuity. No matter if it is atmosphere rainfall, surface runoff or migration of underground water solute that we are tracing; the spatial distribution and calculation of the parameters are have both randomness and continuity. Third, the data has a non-normal distribution. The structure of the data is not normally distributed. And the skewed distribution of it, which is widely seen,could not be analyzed by classical statistics. Fourth, exceptional values are frequently detected , so the process of calculation and analysis must be exceptional-value-adjusted. Fifth, qualitative variable and fuzziness are also seen in the qualitative analysis procedure; therefore useful information should be maintained to ensure the accuracy of the analysis result. Sixth, missing data can be frequently seen; therefore reasonable estimation of the missing data is necessary. These characteristics are defined the use of statistic analysis methods, so need select some can reflect the characteristics. In the article analyses the statistics methods which suit for measuring data of the groundwate, then set forth the necessary statistics analyses method.(2)The article illustrates the application of Kriging model in water resource assessment by demonstrating examples. The mathematical form of manifold Kriging models in time-space domain analyzes the development, application and problems of the models.(3)The application spatial Kriging model has been introduced into the space-time domain. The article advances the general theory of Kriging model in time-space domain.(4)Mainly focuses on the time-domain and universal Kriging model. By theorem derivation, the time-domain Kriging model has been found. The new form of universal Kriging model could be found by using drifting model, which is one original form of universal Kriging model, to meet the requirement of unbiased estimation.(5)This article discuss the definition of the groundwater risk theory and use Makov chain to determine matrix transfer, then give the risk analyses on groundwater pollution.In order to prove the effectiveness of the method, we study the application of Kriging model in water quality assessment taking pollution of surface and underground water as examples. We've got several conclusions as listed below.(1)Theory based on assumptions derived second-order stable model of a new form of pan-Krieger, this model is different from the airspace of the pan-Kriging model, it mainly in the time domain on the estimated drift, also taking into account the relatively fixed position of regionalized variable spatial variability of circumstances presents the characteristics of a more stable state, it suitable for detecting relatively fined position of the problem of estimating. Compared with the original model, reflecting the temporal and spatial characteristics of the data and the models forms to be simplified.(2)According to the observation data of nitrogen-based compound of underground water in Changchun area for the last three years, we draw a distribution diagram of the year 2006, and by using time domain Kriging model, we forecast the trend of the future two years. From the result Kriging method in time-space and variance in water,and provide a new thought about water resources assess. But it is sensitive about the space domain of data,so it has no advantage in forecast the trend of pollution. The research result shows that there won't be drastic change of water quality in Changchun area in short-term, and the three ion concentrations will have relatively great change (rise),therefore reasonable prevention and recovery is highly needed.(3)Based on the feature that the test data does not meet the classical statistical theory, and introduce the overall distribution without demand. When the condition does not meet the study requirements, through statistical analysis of test books, and reasonable, still be able to reach a more satisfactory assessment of the spatial distribution of figure.(4)Based on the definition of risk analysis, the risk features of underground water pollution have been elaborated.Category of underground water body from 2004 to 2008 has been figured out by using fuzzy integration comment and the water body pollution trend from 2009 to 2010 has be forecasted by introducing Markov chain risk analysis model. It is concluded that Changchun underground water has a high risk of being polluted and there won't be a favorable change in a short term.The article also points out the influence that the completeness of hydrology data have on the application of Kriging method. Meanwhile, since the Kriging method in time-space domain requires rather complicated mathematical derivation, related professional software need to be developed in order to realize wider application and better estimation. Therefore this is also the bottleneck of the further development of Kriging method in time-space domain.