A comparison of surface fitting algorithms for geophysical data

Many treatments in geophysics require regular grids of data. Since the data are generally recorded irregularly (e.g. gravity measurements along roads) or even on tracks (e.g. satellite measurements), it is necessary to grid the observed data. We present the result of a comparison of various surface fitting algorithms carried out in order to check their reliability. Two different types of sampling have been verified: (i) clouds of points and (ii) points along tracks. Five algorithms have been extensively tested: (1) polynomial fit, (2) algorithm using a combination of spline-laplace, (3) krigeage (4) least-squares fitting method, and (5) finite element method. The suitability of each method for different sets of data and the limitations (in terms of amplitude and gradient) are discussed.     

Failure of Layered Sandstone under Brazilian Test Conditions: Effect of Micro-Scale Parameters on Macro-Scale Behaviour

The experimental research in this paper focuses on the effect of micro-scale parameters on macro-scale behaviour of samples tested under Brazilian test conditions for one particular layered sandstone, i.e. sandstone from Modave in the South of Belgium. Five (visually) similar blocks are studied; however, they are different at micro-scale. Their differences on microscopic level (such as grain size, mineral contents and number of layer boundaries) affect the failure process. Fracture length parallel to the layers increases if more layer boundaries (e.g. per cm) are present and more weak minerals. The Brazilian tensile strength of studied layered Modave sandstone is larger for larger quartz grains, while fracture length parallel to the layers is smaller. Quartz grains in the studied layered sandstone are classified as fine sand.     

Efficiency Evaluating of Automatic Lineament Extraction by Means of Remote Sensing (Case Study: Venarch,Iran)

One of the main geological evidence for many ore-bearing deposits and mineralized regions is the existence of points in highly fractured zones and the fragmentation intensity resulted from development of hydrothermal alteration along the fractures, the process leading to the occurrence of ore deposits. In this paper, a new algorithm has been proposed including a set of image processing techniques for detection the lineaments in satellite images by means of programming in MATLAB environment. The set of utilized methods includes line segment extraction by EDLine algorithm, merging line segments by Tavares method and linking the resulting line segments based on the collinearity and proximity criterion. The tectonic structures were stabilized by B-Spline curve fitting. The proposed algorithms were implemented on the ASTER image of a structurally multiple fractured region located in the central Iran, and the lineament map of Venarch area has been depicted. The results obtained from the proposed algorithms indicate a high accuracy of the operations detection of 80% for the reference map lineaments and the overall accuracy of the method is effectively reported as 62%. Combination of the above algorithms proposes a new method that precisely resulted in obtaining image processing of geological evidence for increasing the accuracy and decreasing the risk, before any field operations.     

Groundwater Quality Modeling with a Small Data Set

Seventeen groundwater quality variables collected during an 8‐year period (2006 to 2013) in Andimeshk, Iran, were used to implement an artificial neural network (NN) with the purpose of constructing a water quality index (WQI). The method leading to the WQI avoids instabilities and overparameterization, two problems common when working with relatively small data sets. The groundwater quality variables used to construct the WQI were selected based on principal component analysis (PCA) by which the number of variables were decreased to six. To fulfill the goals of this study, the performance of three methods (1) bootstrap aggregation with early stopping; (2) noise injection; and (3) ensemble averaging with early stopping was compared. The criteria used for performance analysis was based on mean squared error (MSE) and coefficient of determination (R²) of the test data set and the correlation coefficients between WQI targets and NN predictions. This study confirmed the importance of PCA for variable selection and dimensionality reduction to reduce the risk of overfitting. Ensemble averaging with early stopping proved to be the best performed method. Owing to its high coefficient of determination (R² = 0.80) and correlation coefficient (r=0.91), we recommended ensemble averaging with early stopping as an accurate NN modeling procedure for water quality prediction in similar studies.