Support vector machines and feed-forward neural networks for spatial modeling of groundwater qualitative parameters

The present study attempts to model the spatial variability of three groundwater qualitative parameters in Guilan Province, northern Iran, using artificial neural networks (ANNs) and support vector machines (SVMs). Data collected from 140 observation wells for the years 2002–2014 were used. Five variables, X and Y coordinates of the observation well, distance of the observation well from the shoreline, areal average 6-month rainfall depth, and groundwater level at the day of water quality sampling, were considered as primary input variables. In addition, nine qualitative variables were also considered as auxiliary input variables. Electrical conductivity (EC), sodium concentration (Na⁺), and sulfate concentration (SO₄ ^2?) of the groundwater in the region were estimated using ANNs and SVMs with different input combinations. The results showed that both ANNs and SVMs work well when the only primary input variable is the well location. The ANN yielded an RMSE of 1.03 mEq/l for SO₄ ^2?, 1.05 mEq/l for Na⁺, and 203.17 μS/cm for EC, using the X and Y coordinates of the observation wells in the study area. In the case of SVM, these values were, respectively, 0.87, 0.87, and 176.68. Considering the auxiliary input variables (pH, EC, and the concentrations of Na⁺, K⁺, Ca²⁺, Mg²⁺, Cl^?, SO₄ ^2?, and HCO₃ ^?) resulted in a significant decrease in the RMSE of both ANNs (0.22, 0.30, and 33.04) and SVMs (0.26, 0.34, and 36.23). Comparing these RMSE values with those of cokriging interpolation technique (0.59, 0.98, and 177.59) indicated that ANNs and SVMs produced more accurate estimates of the three qualitative parameters. The relative importance of auxiliary input variables was also determined using Gamma test. The output uncertainty of ANNs and SVMs were determined using p-factor and d-factor. The results showed that SVMs have less uncertainty than ANNs.     

Spatially-Weighted Factor Analysis for Extraction of Source-Oriented Mineralization Feature in 3D Coordinates of Surface Geochemical Signal

Natural Resources Research - This contribution proposes a spatially weighted factor analysis (SWFA) to recognize effectively the underlying mineralization-related feature(s) in geochemical signals....     

Estimation of Pore Volume Compressibility in Carbonate Reservoir Rocks Based on a Classification

Geotechnical and Geological Engineering - This study examined the petrographically classification, petrological and petrophysical characteristics by taking a vast range of carbonate reservoir rock...     

Implementation of online model updating in hybrid simulation

Hybrid simulation combines numerical and experimental methods for cost‐effective, large‐scale testing of structures under simulated earthquake loading. Structural system level response can be obtained by expressing the equation of motion for the combined experimental and numerical substructures, and solved using time‐stepping integration similar to pure numerical simulations. It is often assumed that a reliable model exists for the numerical substructures while the experimental substructures correspond to parts of the structure that are difficult to model. A wealth of data becomes available during the simulation from the measured experiment response that can be used to improve upon the numerical models, particularly if a component with similar structural configuration and material properties is being tested and subjected to a comparable load pattern. To take advantage of experimental measurements, a new hybrid test framework is proposed with an updating scheme to update the initial modeling parameters of the numerical model based on the instantaneously‐measured response of the experimental substructures as the test progresses. Numerical simulations are first conducted to evaluate key algorithms for the selection and calibration of modeling parameters that can be updated. The framework is then expanded to conduct actual hybrid simulations of a structural frame model including a physical substructure in the laboratory and a numerical substructure that is updated during the tests. The effectiveness of the proposed framework is demonstrated for a simple frame structure but is extendable to more complex structural behavior and models. Copyright © 2013 John Wiley & Sons, Ltd.     

Application of Discriminant Analysis and Support Vector Machine in Mapping Gold Potential Areas for Further Drilling in the Sari-Gunay Gold Deposit,NW Iran

In this contribution, we used discriminant analysis (DA) and support vector machine (SVM) to model subsurface gold mineralization by using a combination of the surface soil geochemical anomalies and earlier bore data for further drilling at the Sari-Gunay gold deposit, NW Iran. Seventy percent of the data were used as the training data and the remaining 30 % were used as the testing data. Sum of the block grades, obtained by kriging, above the cutoff grade (0.5 g/t) was multiplied by the thickness of the blocks and used as productivity index (PI). Then, the PI variable was classified into three classes of background, medium, and high by using fractal method. Four classification functions of SVM and DA methods were calculated by the training soil geochemical data. Also, by using all the geochemical data and classification functions, the general extension of the gold mineralized zones was predicted. The mineral prediction models at the Sari-Gunay hill were used to locate high and moderate potential areas for further infill systematic and reconnaissance drilling, respectively. These models at Agh-Dagh hill and the area between Sari-Gunay and Agh-Dagh hills were used to define the moderate and high potential areas for further reconnaissance drilling. The results showed that the nu-SVM method with 73.8 % accuracy and c-SVM with 72.3 % accuracy worked better than DA methods.     

APG:A novel python-based ArcGIS toolbox to generate absence-datasets for geospatial studies

One important step in binary modeling of environmental problems is the generation of absence-datasets that are traditionally generated by random sampling and can undermine the quality of outputs.To solve this problem,this study develops the Absence Point Generation(APG)toolbox which is a Python-based ArcGIS toolbox for automated construction of absence-datasets for geospatial studies.The APG employs a frequency ratio analysis of four commonly used and important driving factors such as altitude,slope degree,topographic wetness index,and distance from rivers,and considers the presence locations buffer and density layers to define the low potential or susceptibility zones where absence-datasets are gener-ated.To test the APG toolbox,we applied two benchmark algorithms of random forest(RF)and boosted regression trees(BRT)in a case study to investigate groundwater potential using three absence datasets i.e.,the APG,random,and selection of absence samples(SAS)toolbox.The BRT-APG and RF-APG had the area under receiver operating curve(AUC)values of 0.947 and 0.942,while BRT and RF had weaker per-formances with the SAS and Random datasets.This effect resulted in AUC improvements for BRT and RF by 7.2,and 9.7％from the Random dataset,and AUC improvements for BRT and RF by 6.1,and 5.4％from the SAS dataset,respectively.The APG also impacted the importance of the input factors and the pattern of the groundwater potential maps,which proves the importance of absence points in environmental bin-ary issues.The proposed APG toolbox could be easily applied in other environmental hazards such as landslides,floods,and gully erosion,and land subsidence.     

Artificial recharge efficiency assessment by soil water balance and modelling approaches in a multi-layered vadose zone in a dry region

To assess recharge through floodwater spreading, three wells, approx. 30 m deep, were dug in a 35-year-old basin in southern Iran. Hydraulic parameters of the layers were measured. One well was equipped with pre-calibrated time domain reflectometry (TDR) sensors. The soil moisture was measured continuously before and after events. Rainfall, ponding depth and the duration of the flooding events were also measured. Recharge was assessed by the soil water balance method, and by calibrated (inverse solution) HYDRUS-1D. The results show that the 15 wetting front was interrupted at a layer with fine soil accumulation over a coarse layer at the depth of approx. 4 m. This seemed to occur due to fingering flow. Estimation of recharge by the soil water balance and modelling approaches showed a downward water flux of 55 and 57% of impounded floodwater, respectively.