Performance analysis of radial basis function networks and multi-layer perceptron networks in modeling urban change: a case study

The majority of cities are rapidly growing. This makes the monitoring and modeling of urban change’s spatial patterns critical to urban planners, decision makers, and environment protection activists. Although a wide range of methods exists for modeling and simulating urban growth, machine learning (ML) techniques have received less attention despite their potential for producing highly accurate predictions of future urban extents. The aim of this study is to investigate two ML techniques, namely radial basis function network (RBFN) and multi-layer perceptron (MLP) networks, for modeling urban change. By predicting urban change for 2010, the models’ performance is evaluated by comparing results with a reference map and by using a set of pertinent statistical measures, such as average spatial distance deviation and figure of merit. The application of these techniques employs the case study area of Mumbai, India. The results show that both models, which were tested using the same explanatory variables, produced promising results in terms of predicting the size and extent of future urban areas. Although a close match between RBFN and MLP is observed, RBFN demonstrates higher spatial accuracy of prediction. Accordingly, RBFN was utilized to simulate urban change for 2020 and 2030. Overall, the study provides evidence that RBFN is a robust and efficient ML technique and can therefore be recommended for land use change modeling.     

Seismic activity along the central america volcanic arc: is it related to subduction of the Cocos plate?

We determine seismic strain rate of tectonic earthquakes along the Central America Volcanic Arc. We then compare this result to those obtained from earthquakes related to the convergence of the Cocos and Caribbean plates and to earthquakes in the back-arc region of northern Central America.

The seismic strain-rate tensor for shallow-focus earthquakes along the Central America volcanic arc since 1700, has a compressive eigenvector with a magnitude of 0.7 × 10⁻⁸ year⁻¹, and oriented in a 357° azimuth. The extensive eigenvector is oriented in a 86° azimuth, with a magnitude of 0.82 × 10⁻⁸ year⁻¹. When only Centroid Moment-tensor solutions (CMT) are considered, the respective eigenvectors are 1.2 × 10⁻⁸ year⁻¹ and 1.0 × 10⁻⁸ year⁻¹.

The compressive eigenvector from the seismic strain-rate tensor for earthquakes along the Cocos-Caribbean convergent margin is 2.0 × 10⁻⁸ year⁻¹, plunging at 25°, and oriented in a 29° azimuth. Its magnitude and direction are similar to those of the compressive eigenvector for earthquakes along the volcanic arc. The extensive eigenvector along the convergent margin, on the other hand, has a large vertical component. The compressive and extensive eigevenvectors are 4.9 × 10⁻⁸ year⁻¹ and 4.6 × 10⁻⁸ year⁻¹, using only CMTs as the database.

Earthquakes along the grabens of northern Central America yield a seismic strain-rate tensor whose extensive eigenvector has a magnitude of 2.4 × 10⁻⁸ year⁻¹, oriented in a 109° azimuth. Magnitude and direction are similar to those of the extensive eigenvector for earthquakes along the volcanic arc. The compressive eigenvector along the grabens is practically vertical.

Similarities in magnitudes and directions for compressive and extensive eigenvectors suggest to us that the strain field along the Central America volcanic arc is the result of compression along the convergent Cocos-Caribbean margin, and extension in the back-arc region, along the grabens of northern Central America. This field is resolved as strike-slip faulting along the arc.     

The lunar gravity mission MAGIA: preliminary design and performances

The importance of an accurate model of the Moon gravity field has been assessed for future navigation missions orbiting and/or landing on the Moon, in order to use our natural satellite as an intermediate base for next solar system observations and exploration as well as for lunar resources mapping and exploitation. One of the main scientific goals of MAGIA mission, whose Phase A study has been recently funded by the Italian Space Agency (ASI), is the mapping of lunar gravitational anomalies, and in particular those on the hidden side of the Moon, with an accuracy of 1 mGal RMS at lunar surface in the global solution of the gravitational field up to degree and order 80. MAGIA gravimetric experiment is performed into two phases: the first one, along which the main satellite shall perform remote sensing of the Moon surface, foresees the use of Precise Orbit Determination (POD) data available from ground tracking of the main satellite for the determination of the long wavelength components of gravitational field. Improvement in the accuracy of POD results are expected by the use of ISA, the Italian accelerometer on board the main satellite. Additional gravitational data from recent missions, like Kaguya/Selene, could be used in order to enhance the accuracy of such results. In the second phase the medium/short wavelength components of gravitational field shall be obtained through a low-to-low (GRACE-like) Satellite-to-Satellite Tracking (SST) experiment. POD data shall be acquired during the whole mission duration, while the SST data shall be available after the remote sensing phase, when the sub-satellite shall be released from the main one and both satellites shall be left in a free-fall dynamics in the gravity field of the Moon. SST range-rate data between the two satellites shall be measured through an inter-satellite link with accuracy compliant with current state of art space qualified technology. SST processing and gravitational anomalies retrieval shall benefit from a second ISA accelerometer on the sub-satellite in order to decouple lunar gravitational signal from other accelerations. Experiment performance analysis shows that the stated scientific requirements can be achieved with a low mass and low cost sub-satellite, with a SST gravimetric mission of just few months.     

Evaluation of liquefaction potential in an intermountain Quaternary lacustrine basin (Fucino basin,central Italy)

In this study, we analyse the susceptibility to liquefaction of the Pozzone site, which is located on the northern side of the Fucino lacustrine basin in central Italy. In 1915, this region was struck by a M 7.0 earthquake, which produced widespread coseismic surface effects that were interpreted to be liquefaction-related. However, the interpretation of these phenomena at the Pozzone site is not straightforward. Furthermore, the site is characterized by an abundance of fine-grained sediments, which are not typically found in liquefiable soils. Therefore, in this study, we perform a number of detailed stratigraphic and geotechnical investigations (including continuous-coring borehole, CPTu, SDMT, SPT, and geotechnical laboratory tests) to better interpret these 1915 phenomena and to evaluate the liquefaction potential of a lacustrine environment dominated by fine-grained sedimentation. The upper 18.5 m of the stratigraphic succession comprises fine-grained sediments, including four strata of coarser sediments formed by interbedded layers of sand, silty sand and sandy silt. These strata, which are interpreted to represent the frontal lobes of an alluvial fan system within a lacustrine succession, are highly susceptible to liquefaction. We also find evidence of paleo-liquefaction, dated between 12.1–10.8 and 9.43–9.13 kyrs ago, occurring at depths of 2.1–2.3 m. These data, along with the aforementioned geotechnical analyses, indicate that this site would indeed be liquefiable in a 1915-like earthquake. Although we found a broad agreement among CPTu, DMT and shear wave velocity “simplified procedures” in detecting the liquefaction potential of the Pozzone soil, our results suggest that the use and comparison of different in situ techniques are highly recommended for reliable estimates of the cyclic liquefaction resistance in lacustrine sites characterized by high content of fine-grained soils. In geologic environments similar to the one analysed in this work, where it is difficult to detect liquefiable layers, one can identify sites that are susceptible to liquefaction only by using detailed stratigraphic reconstructions, in situ characterization, and laboratory analyses. This has implications for basic (Level 1) seismic microzonation mapping, which typically relies on the use of empirical evaluations based on geologic maps and pre-existing sub-surface data (i.e., age and type of deposits, prevailing grain size, with particular attention paid to clean sands, and depth of the water table).     

Assessing the impact of large-scale dewatering on fault-controlled aquifer systems: a case study in the Acque Albule basin (Tivoli, central Italy)

The development of large-scale bedrock quarry operations often requires high-volume and long-term groundwater extraction to maintain a sustainable working environment. These dewatering activities often influence groundwater levels and flow patterns regionally. In the present study, the influence of the dewatering of the travertine quarry operations near the city of Tivoli, Italy, are quantitatively investigated through an integrated analysis of field data and numerical modeling. Lowering of regional groundwater levels in the vicinity of the quarry has led to destructive land subsidence and alterations to the flow system sustaining a hot-spring area. The study employs a finite element numerical model (FEFLOW) to evaluate and quantify the impact of the extensive dewatering on fault-controlled regional groundwater flow in the Acque Albule basin. By incorporating the physical field data and historical hydrologic information, the numerical model was calibrated against three groundwater scenarios, reproducing the effects of different exploitation activities, coupled with natural changes over the course of the quarry operation. The results indicate that groundwater withdrawals by the mining industry and by “Terme di Roma” spa resulted in the cessation of flow from the primary thermal spring and a drop in the phreatic level in the area consequently affected by land subsidence.     

The winds of WC10 central stars of planetary nebulae

We present the results of an analysis of the winds of two WC10 central stars of planetary nebulae, CPD-56°8032 and He 2-113. These two stars have remarkably similar spectra, although the former exhibits somewhat broader emission line widths. High resolution spectra (up to R=50 000) were obtained in May 1993 for both objects at the 3.9 m AAT, using the UCL Echelle Spectrograph. The fluxes in individual Cii auto-ionising multiplet components, many of which were blended, were derived. Lines originating from auto-ionising resonance states situated in the C²⁺ continuum are very sensitive to the electron temperature, since the population of the these levels is close to LTE. The measured widths and profile shapes of these transitions are presented and are in excellent agreement with those predicted on the basis of their calculated auto-ionising lifetimes. The wind electron temperature is derived for both stars from the ratio of the fluxes in four such transitions (T _e =18 500 K±1 500 K for CPD-56° 8032 andT _e =13 600 K±800 K for He 2-113). Optical depth effects are investigated using normal recombination lines to obtain an independent wind electron temperature estimate in excellent agreement with the dielectronic line analysis.     

Flow and transport in channels with submerged vegetation

This paper reviews recent work on flow and transport in channels with submerged vegetation, including discussions of turbulence structure, mean velocity profiles, and dispersion. For submerged canopies of sufficient density, the dominant characteristic of the flow is the generation of a shear-layer at the top of the canopy. The shear-layer generates coherent vortices by Kelvin-Helmholtz (KH) instability. These vortices control the vertical exchange of mass and momentum, influencing both the mean velocity profile, as well as the turbulent diffusivity. For flexible canopies, the passage of the KH vortices generates a progressive wave along the canopy interface, termed monami. The KH vortices formed at the top of the canopy penetrate a distance δ _e into the canopy. This penetration scale segregates the canopy into an upper layer of rapid transport and a lower layer of slow transport. Flushing of the upper canopy is enhanced by the energetic shear-scale vortices. In the lower layer turbulence is limited to length-scales set by the stem geometry, and the resulting transport is significantly slower than that of the upper layer.