Three Dimensional Variability And Visualization Of Soil Electrical Conductivity In Coastal Saline Land

Soil salinization is a process of land degradation by which water-soluble salts accumulate in the soil, occurred naturally or because of conditions resulting from management practices. It leads to the change of soil chemical-physical properties, and the impairment and destruction of land biological productivity. Resulting in the broken of ecology balance, the change of environment, and threat the sustainable development of ecology, economy, and society. Soil salinization and soil secondary salinization is a worldwide problem, and becoming a bottle-neck in the sustainable development of irrigated farming. According to the incomplete statistics of UNESCO and FAO, there are 954 million hm~2 saline land all over the world, and amount to 99,133,000 hm~2 in the northwest, north, westward of northeast, and coastal area of China. Amelioration and utilization the saline land resource is an important approach for promoting the sustainable development of agriculture in China, and improving the development of the regional economy, society, and ecology. In order to pursuit high efficiency and benefit, it is necessary to introduce new technologies in soil resource utilization and management. In the saline soils, the highly precision field survey technology can be used to measure the soil salinity, then the advanced spatial information tools are adopted for decision making in the management of coastal soil resources scientifically and rationally.Spatial variability for soil properties was the study hotspot in soil science since 1970s. Soils are three-dimensional (3-D) bodies with properties that can vary greatly over small distances in every direction. However, soils are generally investigated in only the horizontal dimensions. Though several studies were mentioned the 3-D characterization of the spatial variability of soil properties is the aim, which only provide a set of horizontal layers for a number of soil depths. In this approach soil variation is modeled at every depth independently without considering the interaction of soil properties in the soil profile. It was said, saliferous groundwater under 3 meters can accumulate in the soil surface since the effort of capillarity. The phenomena of high salinity content, violent salt return, soil salinity changed frequently in vertical direction, which made the soil have very low output and agricultural benefit. So, the research of soil variability, especially in 3D, not only has theoretic value, which can enrich the 3D variability theory and method, but also has application value, whereas these quantificational information are the indispensably basic data and theoretic foundation for developing precise agriculture.Visualization is an important graphical communication of information. When exploring data, people usually look for structures, patterns and relationships between data elements. Such analysis is easier if the data are presented in a visual setting than a textural or a numerical form. Visualization provides an overview of complex and large dataset, shows a summary of data and helps in the identification of possible patterns and structures of the data. Therefore, the complex 3D soil EC variability can be described quantitatively and depicted accurately by the 3D visualization technology. Moreover, 3D visualization is a very important branch of 3D GIS and a research project by geographer and computer scientist all along.For 3D interpolation, 3D spatial scattered data for soil electrical conductivity (EC) must be collected firstly. But the traditional invasive sampling method for determining soil salinity requires considerable resources for field sampling and laboratory analysis. Moreover, the traditional method can't resample the same soil site since the soil bulk had been drilled or digged. The development of new technologies, however, such as electromagnetic (EM) induction sensors, has revolutionized the way in which soil profile EC is measured. Ground conductivity meters such as EM38 used in this study are relatively inexpensive and easy to use in obtaining EM measurements. At present, the EM method seems a first choice for measuring soil EC in a large scale fieldA coastal saline region selected as the study area, which located in the northern region of Shangyu City, Zhejiang Province, at the south shore of Hangzhou Gulf. In this study, soil profile EC data was predicted using the EM38 models firstly. Then, three dimensional spatial variability of soil EC, uncertainty analysis of stochastic simulation, and risk assessment of crop plant were done. Finally, new developed modules were integrated in the ARGIS platform. Some satisfied results were obtained as follows:(1) Prediction of soil profile electrical conductivity using EM38Verify experimentation results shown, the indices, such as air temperature, profile soil temperature, profile soil water content, has little influence on the measurement of Apparent Soil Electrical Conductivity (ECa) by EM38 in the study area. Soil salinity is the dominant factor influence the measurement of ECa by EM38. On soil profile, EM38 measurements ECa at the horizontal (EM_h) and vertical (EM_v) coil-mode configurations, EC1:5, and Bulk Electrical Conductivity (ECb) has high correlation mutually. The established experimental model can be used to predict well soil profile EC value. Although in the same study area, the experimental model must be re-calibrated while generalizing in different time. So, it is necessary to found a more universal soil profile EC prediction method. The method which detection of vertical ECa change in soil profiles from aboveground EM measurements can be used to solve the above problem. It uses the linear or non-linear model of the response of EM38 with second order Tikhonov regularization, that is, an inverse procedure, to predict ECa profiles. This method is based solely on electromagnetic physics, which requires no further field calibration when using in different soil type field. It was shown, the average prediction error of the linear model and non-linear model is 38.44% and 24.26%, respectively, and the non-linear model has higher precision.(2) Three dimensional variability and visualization of soil EC in a coastal saline landIn the present study, the apparent electrical conductivity data at ten different depth layers from the surface to 1.1m depth, which inversed by a procedure for estimating ECa profiles that combines the EM38 linear model with Tikhonov regularization from aboveground EM38 measurements were selected as the data source for the study of 3D variability. According to a certain horizontal and vertical search regulation, the horizontal and vertical variogram of soil EC were modelled, respectively. Combining both previous models into one 3D model, which can be used to depict the structure of spatial autocorrelation. It was shown, no directional effect was observed both in the horizontal and vertical planes. An isotropic six-parameter model can be established to depict the structure of spatial autocorrelation. By combining the horizontal with the vertical variogram, a 3D isotropic variogram was constructed which consisted of an isotropic nugget effect and three spherical models. 3D ordinary kriging method was used to predict the three-dimensional variability of soil EC. Plume model was used to present ECa spatial distribution, and presented a superior visualization of spatial distribution of ECa in 3D space directly that 2D interpolation can't achieve. Compared with the 2D ordinary kriging, the root-mean-square error, average standard error, and the correlation coefficient between the predicted value and the observed value produced by 3D ordinary kriging has great improved. Virtual reality model of EC was built by the VRML, and it is helpful for analyzing of the spatial distribution characteristic and the change trend for soil EC. At last, the EC virtual reality model was dispersed on the internet.(3) Spatial uncertainty of soil EC in 3D soil profile and risk assessment of crop plantFor instructing of soil management decision and making suggestion for suitable crops, the profile EC data predicted by the EM38 non-linear module, which has higher accuracy, was selected as the research data.â‘ The Sequence Gauss Simulation (SGS) method was used to simulate the spatial variability of ten different depth layers form surface to 1.1m depth. It was shown, the smoothing effect, commonly found in the maps generated by the ordinary kriging method, results in less variation in the estimated values than in the observed values. Fluctuation of the observed values was stood out in the simulation maps by the SGS method. Probability distribution and the histogram of the origin data were took into account in the process of SGS method. So, the specific spatial correlation structure of the observed values was kept in the prediction values. Moreover, comparing with the ordinary kriging method, the SGS method has higher prediction precision. Although the soil EC above or below the layer was not considered in this method, which combines the EC spatial variability maps of ten different depth layers together, it was a feasible approach to depict the 3D variability of soil EC. Deficiently, local area has obvious jump.â‘¡The value of 400mS/m of soil ECe was selected as the threshold of serious salinization. The uncertain of ECe exceed this threshold in ten different depth layers was simulated 1000 times by the SIS method. 3D probability map was produced by putting the ten probability maps together. The area, which has high probability and low probability, can be give explicit assessment result of whether it was salinized. The area, which its probability is close to 0.5, and its assessment result has great uncertainty. So, probability map is an appropriate approach for providing more information to the decision-maker. The relative error of simulation was analyzed. It was shown that the probability distribution result simulated 1000 times by SIS method is credible.â‘¢According to the salt tolerance capacity, risk assessment of crop plant was analyzed at ten different depth layers form surface to 1.1m depth by the disjunctive Kriging method. 3D probability map was produced by putting the ten probability maps together. Soil EC and exceed a certain threshold at deeper layers was took into account can instruct practical soil management and decision-making more accurately.(4) Extension of functional modules for Agricultural Resources Geographic Informational SystemFirstly, objective, principleand, and function of the Agricultural Resources Geographic Informational System (ARGIS) were analyzed. ARGIS was composed of five modules: GIS platform, soil sampling module, geostatistics analysis module, stochastic simulation module, and management zones module. Thereinto, GIS platform module is the base of the other modules, which support all the other modules can use map or grid layer in the data view of ARGIS as the operation object. According to the system framework, sampling design module of ARGIS Version1.0 (software register ID: 2006SR16131) was re-integrated into the soil sampling module in the new version of ARGIS. ARGIS was set up by using visual C++ language and system integration technique in windows environment. Soil electrical conductivity sampling module based on the EM38 and GPS device and stochastic simulation module were extended. At last, soil ECa was sampled in a coastal saline land with help of soil sampling module of ARGIS software. The spatial variability of EC was simulated and uncertain of exceeding a certain EC threshold was mapped by stochastic simulation module of ARGIS software.The innovations or new developments were made as follows:(1)The EM38 linear and non-linear model were used to inverse soil profile electrical conductivity(EC) in coastal saline land, and three dimensional ordinary Kriging method was used to predict three dimensional variability of soil EC. Virtual reality model of EC was built by the VRML, and then dispersed on the internet. Prediction, simulation, and internet dispersion of three dimensional soil EC in this study not only promoted the methodology for soil spatial variability, but also provided a new extension approach for soil research production.(2)The Sequence Gauss Simulation (SGS) and Sequence Indicator Simulation (SIS) methods were used to simulate the variability of soil EC and analyze the uncertain of simulation in ten different depth layers, respectively. Three dimensional probability of exceed a certain EC threshold was mapped. According to the salt tolerance capacity, risk assessment of crop plant was analyzed by the disjunctive Kriging method. The research of three dimensional uncertainty of stochastic simulation and risk assessment of crop plant was an important complementarity and development of the conventional works in two dimensions. Moreover, it can give more practical soil management and accurate decision-making.(3)Agricultural Resources Geographic Informational System (ARGIS, Version1.0) was set up by using visual C++ programme language and system integration technique in windows environment. The software has a universal Geographic Informational System (GIS) management platform, can be used to analyze spatial variability of soil properties in the field scale, to design economic and effective sampling scheme, to classify precise management zones and mapping. In this study, a new extension model for soil electrical conductivity sampling was developed, which integrated the device I/O read (i.e. EM38 and GPS) and stochastic simulation approach. This new model can provide technical supports for precise agricultural field management.