Landslide susceptibility mapping at Vaz Watershed (Iran) using an artificial neural network model: a comparison between multilayer perceptron (MLP) and radial basic function (RBF) algorithms

Landslide susceptibility and hazard assessments are the most important steps in landslide risk mapping. The main objective of this study was to investigate and compare the results of two artificial neural network (ANN) algorithms, i.e., multilayer perceptron (MLP) and radial basic function (RBF) for spatial prediction of landslide susceptibility in Vaz Watershed, Iran. At first, landslide locations were identified by aerial photographs and field surveys, and a total of 136 landside locations were constructed from various sources. Then the landslide inventory map was randomly split into a training dataset 70 % (95 landslide locations) for training the ANN model and the remaining 30 % (41 landslides locations) was used for validation purpose. Nine landslide conditioning factors such as slope, slope aspect, altitude, land use, lithology, distance from rivers, distance from roads, distance from faults, and rainfall were constructed in geographical information system. In this study, both MLP and RBF algorithms were used in artificial neural network model. The results showed that MLP with Broyden–Fletcher–Goldfarb–Shanno learning algorithm is more efficient than RBF in landslide susceptibility mapping for the study area. Finally the landslide susceptibility maps were validated using the validation data (i.e., 30 % landslide location data that was not used during the model construction) using area under the curve (AUC) method. The success rate curve showed that the area under the curve for RBF and MLP was 0.9085 (90.85 %) and 0.9193 (91.93 %) accuracy, respectively. Similarly, the validation result showed that the area under the curve for MLP and RBF models were 0.881 (88.1 %) and 0.8724 (87.24 %), respectively. The results of this study showed that landslide susceptibility mapping in the Vaz Watershed of Iran using the ANN approach is viable and can be used for land use planning.     

Modeling the strong ground motion and rupture characteristics of the March 31, 2006, Darb-e-Astane earthquake,Iran, using a hybrid of near-field SH-wave and empirical Green’s function method

We analyze the strong motion accelerograms of the moderate (M _w = 6.1), March 31, 2006, Darb-e-Astane earthquake of western Iran and also those of one of its prominently recorded, large (M _w = 5.1) foreshock and (M _w = 4.9) aftershock. (1) Using derived SH-wave spectral data, we first objectively estimate the parameters W _o\mathit{\Omega} _{\rm o} (long period spectral level), f _c (corner frequency) and Q(f) (frequency dependent, average shear wave quality factor), appropriate for the best-fit Brune ω ^− 2 spectrum of each of these three events. We then perform a non-linear least square analysis of the SH-wave spectral data to provide approximate near-field estimates of the strike, dip, and rake of the causative faults and also the seismic moment, moment magnitude, source size, and average stress drop of these three events. (2) In the next step, we use these approximate values and an empirical Green’s function approach, in an iterative manner, to optimally model the strong ground motion and rupture characteristics of the main event in terms of peak ground acceleration/velocity/displacement and duration of ground shaking and thereby provide improved, more reliable estimates of the causative fault parameters of the main event and its asperities. Our near-field estimates for both the main moderate event and the two smaller events are in good conformity with the corresponding far-field estimates reported by other studies.     

Estimation of coda and shear wave attenuation in the volcanic area in SE Sabalan Mountain,NW Iran

The quality factors of coda and shear waves have been estimated for the SE Sabalan Mountain, geothermal region in northwestern Iran. We have analyzed 65 local earthquakes with magnitude of 2.8 to 6.1 and 2.8 to 5 for shear and coda wave quality factor estimation, respectively. These events were recorded on five stations installed by Building and Housing Research Center Network. Coda normalization and Spectral decay methods have been used to estimate the frequency dependence attenuation relation for shear wave, and single back-scattering method for coda waves. We have observed that the coda normalization method has supplied significantly higher Q _S values as compared to the spectral method. The results show that, in general, Q values are significantly smaller for the entire frequency range as compared to tectonically active areas and are close to the values for volcanic areas.     

Coseismic slip model of the 2007 August Pisco earthquake (Peru) as constrained by Wide Swath radar observations

The Pisco earthquake ( M _w 8.0; 2007 August 15) occurred offshore of Peru's southern coast at the subduction interface between the Nazca and South American plates. It ruptured a previously identified seismic gap along the Peruvian margin. We use Wide Swath InSAR observations acquired by the Envisat satellite in descending and ascending orbits to constrain coseismic slip distribution of this subduction earthquake. The data show movement of the coastal regions by as much as 85 cm in the line-of-sight of the satellite. Distributed-slip model indicates that the coseismic slip reaches values of about 5.5 m at a depth of ∼18–20 km. The slip is confined to less than 40 km depth, with most of the moment release located on the shallow parts of the interface above 30 km depth. The region with maximum coseismic slip in the InSAR model is located offshore, close to the seismic moment centroid location. The geodetic estimate of seismic moment is 1.23 × 10²¹ Nm ( M _w 8.06), consistent with seismic estimates. The slip model inferred from the InSAR observations suggests that the Pisco earthquake ruptured only a portion of the seismic gap zone in Peru between 13.5° S and 14.5° S, hence there is still a significant seismic gap to the south of the 2007 event that has not experienced a large earthquake since at least 1687.     

Comparison of two-dimensional flood propagation models: SRH-2D and Hydro_AS-2D

This article presents a comparison between two two-dimensional finite volume flood propagation models: SRH-2D and Hydro_AS-2D. The models are compared using an experimental dam-break test case provided by Soares-Frazão (J Hydraul Res, 2007. doi: 10.1080/00221686.2007.9521829). Four progressively refined meshes are used, and both models react adequately to mesh and time step refinement. Hydro_AS-2D shows some unphysical oscillations with the finest mesh and a certain loss of accuracy. For that test case, Hydro_AS-2D is more accurate for all meshes and generally faster than SRH-2D. Hydro_AS-2D reacts well to automatic calibration with PEST, whereas SRH-2D has some difficulties in retrieving the suggested Manning’s roughness coefficient.     

Coastal Wave Height Prediction using Recurrent Neural Networks (RNNs) in the South Caspian Sea

The prediction of wave parameters has a great significance in the coastal and offshore engineering. For this purpose, several models and approaches have been proposed to predict wave parameters, such as empirical, soft computing, and numerical based approaches. Recently, soft computing techniques such as recurrent neural networks (RNN) have been used to develop sea wave prediction models. In this study, the RNN for wave prediction based on the data gathered and the measurement of the sea waves in the Caspian Sea, in the north of Iran is used for this study. The efficiency of RNNs for 3, 6, and 12 hourly and diurnal wave prediction using correlation coefficients is calculated to be 0.96, 0.90, 0.87, and 0.73, respectively. This indicates that wave prediction by using RNNs yields better results than the previous neural network approaches.     

Neuro-fuzzy modelling in mining geochemistry: Identification of geochemical anomalies

Local and mine scale exploration models for anomaly recognition within known ore fields are discussed. Traditional geochemical exploration methods are based on multivariate statistical analysis, metallometry, vertical geochemical zonality and criteria of natural field geochemical associations, which suffer several shortcomings, including lack of a geostatistical generalised approach for separating anomalies from background. These shortcomings make the interpretation process time consuming and costly. Fuzzy set theory, fuzzy logic and neural network techniques seem very well suited for typical mining geochemistry applications. The results, obtained from applying the proposed technique to a real scenario, reveals significant improvements, comparing the results obtained from applying multivariate statistical analysis. Computationally, the introduced technique makes possible, without exploration drilling, the distinction between blind mineralisation and zone of dispersed ore mineralisation. The methodology developed in this research study has been verified by testing it on various real-world mining geochemical projects.     

A new developed approach for the prediction of ground vibration using a hybrid PSO-optimized ANFIS-based model

Ground vibration is one of the common environmental effects of blasting operation in mining industry, and it may cause damage to the nearby structures and the surrounding residents. So, precise estimation of blast-produced ground vibration is necessary to identify blast-safety area and also to minimize environmental effects. In this research, a hybrid of adaptive neuro-fuzzy inference system (ANFIS) optimized by particle swarm optimization (PSO) was proposed to predict blast-produced ground vibration in Pengerang granite quarry, Malaysia. For this goal, 81 blasting were investigated, and the values of peak particle velocity, distance from the blast-face and maximum charge per delay were precisely measured. To demonstrate the performance of the hybrid PSO–ANFIS, ANFIS, and United States Bureau of Mines empirical models were also developed. Comparison of the predictive models was demonstrated that the PSO–ANFIS model [with root-mean-square error (RMSE) 0.48 and coefficient of determination (R ²) of 0.984] performed better than the ANFIS with RMSE of?1.61 and R ² of 0.965. The mentioned results prove the superiority of the newly developed PSO–ANFIS model in estimating blast-produced ground vibrations.