Search

Retrieving additional information about a spatial distribution of chemical elements in topsoil is an important part of environmental protection. Artificial neural networks (ANN) such as generalized regression neural networks (GRNN) and multilayer perceptrons (MLP) are used to predict the spatial distribution of chemical element Copper (Cu) and Nickel (Ni) in topsoil. The initial data are soil survey data sets on surface content of Cu and Ni at a specific location in subarctic Novy Urengoy, Russia. The network structures were chosen during computer modeling, based on the minimization of the root mean squared error (RMSE). Each network has advantages and limitations for the content of the different elements in the various soil types. The hybrid models combining the techniques of ANN and geostatistics consists of several steps: 1) comparing of kriging estimates and network ones; 2) building residues; 3) residual kriging and combination with ANN estimates. The work confirms that trained ANN is suitable for modeling both the normal (Cu) and abnormal (Ni) spatial distribution of pollutants. The best prognostic hybrid model was model of the multilayer perceptron with residual kriging (MLPRK), which allows improving the predictive accuracy for the considered elements. [ABSTRACT FROM AUTHOR]

The paper deals with application of Multilayer Perceptron (MLP) to the prediction of Chromium (Cr) distribution features in different soil. The distributions of the element content in urban soils are intersectional and multifactorial in nature. The local features of soil formation are necessary taking into account to forecast the chemical elements content at the unmonitored sites. The raw data of elements content in the topsoil are the results of chemical analysis of the soil sample, selected in three cities of the Yamalo-Nenets Autonomous Okrug - Noyabrsk, Tarko-Sale and Novy Urengoy. During the soil surveys, 486 soil specimens were taken from the residual zones. According to spatial distribution of total chromium concentration of the spots with high-level concentration were recognized in Tarko-Sale and Novy Urengoy. Mean total concentration of chromium in specimens of the spot subgroups is by an order of magnitude higher than in subgroups of urban background. The MLP network structure is selected during the computer simulation based on the minimization of the RMSE individually for each area. The better accuracy is for Noyabrsk and for Tarko-Sale. MLP is suitable for modeling both the normal and abnormal spatial distribution of pollutants. For the distribution without marked features, like Novy Urengoy area, MLP model detects enough high error. [ABSTRACT FROM AUTHOR]

The study is based on the data obtained as a result of soil screening in the city of Noyabrsk, Russia. A comparison of two types of neural networks most commonly used in this type of research was carried out: multi-layer perceptron (MLP), generalized regression neural network (GRNN), and a combined MLP and ordinary kriging approach (MLPRK) for predicting the spatial distribution of the chemical element Chromium (Cr) in the surface layer of the urbanized territory. The model structures were developed using computer modeling, based on minimizing of a root mean squared error (RMSE). As input parameters, the spatial coordinates were used, and the concentration of Cr - as the output. The hybrid MLPRK approach showed the best prognostic accuracy. [ABSTRACT FROM AUTHOR]

Forecasting soil pollution is an important part of environmental protection around the world. The work is devoted to the use of artificial neural networks for spatial prediction of the anomalously distributed chemical element Chromium (Cr). Generalized regression neural networks (GRNN) and multilayer perceptrons (MLP) are classes of neural networks widely used for continuous mapping of functions. Each network has its pros and cons; however, both demonstrated quick preparation and good modeling capabilities. In the paper, we studied and compared two neural networks: GRNN and MLP. The present study is based on real data sets on surface contamination of Cr at a specific location in subarctic Novy Urengoy, Russia, obtained during previous screening. The network structures were chosen during computer modeling, based on the minimization of the root mean squared error (RMSE). GRNN (anisotropy) showed the best predictive accuracy in comparison with MLP and kriging. [ABSTRACT FROM AUTHOR]