Machine Learning Feature Selection Methods for Landslide Susceptibility Mapping

This paper explores the use of adaptive support vector machines, random forests and AdaBoost for landslide susceptibility mapping in three separated regions of Canton Vaud, Switzerland, based on a set of geological, hydrological and morphological features. The feature selection properties of the three algorithms are studied to analyze the relevance of features in controlling the spatial distribution of landslides. The elimination of irrelevant features gives simpler, lower dimensional models while keeping the classification performance high. An object-based sampling procedure is considered to reduce the spatial autocorrelation of data and to estimate more reliably generalization skills when applying the model to predict the occurrence of new unknown landslides. The accuracy of the models, the relevance of features and the quality of landslide susceptibility maps were found to be high in the regions characterized by shallow landslides and low in the ones with deep-seated landslides. Despite providing similar skill, random forests and AdaBoost were found to be more efficient in performing feature selection than adaptive support vector machines. The results of this study reveal the strengths of the classification algorithms, but evidence: (1) the need for relying on more than one method for the identification of relevant variables; (2) the weakness of the adaptive scaling algorithm when used with landslide data; and (3) the lack of additional features which characterize the spatial distribution of deep-seated landslides.     

The Density and Masses of Obscured Black Holes

Recent Chandra and XMM-Newton surveys have uncovered a large fraction of the obscured AGN responsible of the hard X-ray background. One of the most intriguing results of extensive programs of follow-up observations concerns the optical and near-infrared properties of the hard X-ray sources counterparts. More specifically, for a significant fraction of hard X-ray obscured sources the AGN responsible of the high X-ray luminosity remains elusive over a wide range of wavelengths from soft X-rays to near-infrared. This very observational result opens the possibility to investigate the host of bright obscured quasars in some detail. Here we briefly report on some preliminar results obtained for a small sample of elusive AGN in the HELLAS2XMM survey.     

A physical interpretation of the deterministic fractal–multifractal method as a realization of a generalized multiplicative cascade

In this study, we attempt to offer a solid physical basis for the deterministic fractal–multifractal (FM) approach in geophysics (Puente, Phys Let A 161:441–447, 1992; J Hydrol 187:65–80, 1996). We show how the geometric construction of derived measures, as Platonic projections of fractal interpolating functions transforming multinomial multifractal measures, naturally defines a non-trivial cascade process that may be interpreted as a particular realization of a random multiplicative cascade. In such a light, we argue that the FM approach is as “physical” as any other phenomenological approach based on Richardson’s eddies splitting, which indeed lead to well-accepted models of the intermittencies of nature, as it happens, for instance, when rainfall is interpreted as a quasi-passive tracer in a turbulent flow. Although neither a fractal interpolating function nor the specific multipliers of a random multiplicative cascade can be measured physically, we show how a fractal transformation “cuts through” plausible scenarios to produce a suitable realization that reflects specific arrangements of energies (masses) as seen in nature. This explains why the FM approach properly captures the spectrum of singularities and other statistical features of given data sets. As the FM approach faithfully encodes data sets with compression ratios typically exceeding 100:1, such a property further enhances its “physical simplicity.” We also provide a connection between the FM approach and advection–diffusion processes.     

Simultaneous identification of the pollutant release history and the source location in groundwater by means of a geostatistical approach

This paper describes an innovative procedure that is able to simultaneously identify the release history and the source location of a pollutant injection in a groundwater aquifer (simultaneous release function and source location identification, SRSI). The methodology follows a geostatistical approach: it develops starting from a data set and a reliable numerical flow and transport model of the aquifer. Observations can be concentration data detected at a given time in multiple locations or a time series of concentration measurements collected at multiple locations. The methodology requires a preliminary delineation of a probably source area and results in the identification of both the sub-area where the pollutant injection has most likely originated, and in the contaminant release history. Some weak hypotheses have to be defined about the statistical structure of the unknown release function such as the probability density function and correlation structure. Three case studies are discussed concerning two-dimensional, confined aquifers with strongly non-uniform flow fields. A transfer function approach has been adopted for the numerical definition of the sensitivity matrix and the recent step input function procedure has been successfully applied.     

Seismic behavior of single‐story asymmetric‐plan buildings under uniaxial excitation

The critical parameters that influence the nonlinear seismic response of asymmetric‐plan buildings are identified by evaluating the effects of different asymmetries that may characterize the structure of a building as well as exploring the influence of the ground motion features. First, the main findings reported in the literature on both the linear and nonlinear dynamic response of asymmetric‐plan buildings are presented. The common findings and the conflicting conclusions reached in different investigations are pointed out. Then, the results of comprehensive nonlinear dynamic analyses performed for evaluating the seismic response of systems characterized by different strength and stiffness configurations, representative of a large class of asymmetric‐plan buildings, are reported. Findings from the study indicate that the building response changes when moving from the linear to the nonlinear range, so that the seismic behavior of asymmetric‐plan buildings, apart from the source of asymmetry, can be always classified as irregular. Additionally, it was observed that as the seismic demands cause amplification of system nonlinearity with increasing earthquake intensity, the maximum displacement demand in the different resisting elements tends to be reached with the same deformed configuration of the system. The resultant of the seismic forces producing such a maximum demand is located at the center of resistance and corresponds to the collapse mechanism of the system that provides the maximum lateral strength in the exciting direction of the seismic action. Copyright © 2008 John Wiley & Sons, Ltd.     

Acoustic core logging in blast-damaged rock

Thill, R.E. and D' Andrea, D. V., 1975. Acoustic core logging in blast-damaged rock. Eng. Geol., 10: 13–36.The Bureau of Mines, in cooperation with the Duval Corp., conducted a blast-fragmentation experiment to determine the feasibility of preparing a porphyry copper-molybdenum deposit for in-situ leaching. The blast was designed with ten 9-inch-diameter blastholes to depths of 110 feet in an equilateral triangle configuration; spacings between blastholes were 15, 20, and 25 ft. One of the major problems in the experiment was in assessing blast damage. Acoustic core-logging equipment and methods were devised and used as one approach in solving this problem. Ultrasonic pulse travel-times were determined in four diametral directions at 2-ft intervals of depth to a final depth of 120 ft in three preblast and six postblast drill cores at the Duval test site. The acoustic logging program provided compressional wave travel-time at 0°, 45°, 90°, and 135° around the core circumference, maximum travel-time difference, mean compressional-wave velocity, and an anisotropy factor. Other acoustic parameters introduced in the analyses were stiffness modulus, seismic quality designation (SQD), and a compensated velocity to account for portions of the core that were nonrecoverable or too highly fractured to permit diametral travel-time measurements.The acoustic parameters all indicated the deterioration in structural quality from the preblast condition, in which the rock already was badly fractured and weathered, to the more highly fractured postblast condition. Because of the highly fragmented, poor structural condition of the rock after blasting, the rock was indicated to be suitable for in-situ leaching, at least at the 20- and 15-ft blasthole spacings, and even in some zones in the rock at the 25-ft blasthole spacing.     

Reporting negative results to stimulate experimental hydrology: discussion of “The role of experimental work in hydrological sciences – insights from a community survey”*

ABSTRACT

Experimental work in hydrology is in decline. Based on a community survey, Blume et al. showed that the hydrological community associates experimental work with greater risks. One of the main issues with experimental work is the higher chance of negative results (defined here as when the expected or wanted result was not observed despite careful experimental design, planning and execution), resulting in a longer and more difficult publishing process. Reporting on negative results would avoid putting time and resources into repeating experiments that lead to negative results, and give experimental hydrologists the scientific recognition they deserve. With this commentary, we propose four potential solutions to encourage reporting on negative results, which might contribute to a stimulation of experimental hydrology.     

Integration error over very long time spans

The main limit to the time span of a numerical integration of the planetary orbits is no longer set by the availability of computer resources, but rather by the accumulation of the integration error. By the latter we mean the difference between the computed orbit and the dynamical behaviour of the real physical system, whatever the causes. The analysis of these causes requires an interdisciplinary effort: there are physical model and parameters errors, algorithm and discretisation errors, rounding off errors and reliability problems in the computer hardware and system software, as well as instabilities in the dynamical system. We list all the sources of integration error we are aware of and discuss their relevance in determining the present limit to the time span of a meaningful integration of the orbit of the planets. At present this limit is of the order of 10⁸ years for the outer planets. We discuss in more detail the truncation error of multistep algorithms (when applied to eccentric orbits), the coefficient error, the method of Encke and the associated coordinate change error, the procedures used to test the numerical integration software and their limitations. Many problems remain open, including the one of a realistic statistical model of the rounding off error; at present, the latter can only be described by a semiempirical model based upon the simpleN ² formula (N=number of steps, =machine accuracy), with an unknown numerical coefficient which is determined only a posteriori.     

Engineering geomorphological characterisation of the Vajont Slide,Italy, and a new interpretation of the chronology and evolution of the landslide

Although the 1963 Vajont Slide in Italy has been extensively studied for over 50 years, its regional geological and geomorphological context has been neglected. In this paper, we use field observations and remote sensing data to elucidate the interaction between endogenic and exogenic processes that brought the north slope of Monte Toc to failure. We present the first detailed pre- and post-failure engineering geomorphology maps of the slide area. The maps delineate two main landslide blocks, several sub-blocks, compressional and extensional zones, and secondary failures in the deposit. The maps provide new insights into the kinematics, dynamics and evolution of the slide. Finally, we discuss the origin of Vajont Gorge and a prehistoric failure that occurred at the same location as the 1963 slide. We propose, as part of a newly developed multi-stage landscape evolution sequence, that the prehistoric failure was a deep-seated gravitational slope deformation (sackung) that initiated during deglaciation and continued to slowly move until the catastrophic failure in 1963. We argue that the gorge was created by these deep-seated slow movements.