Portability of semantic and spatial–temporal machine learning methods to analyse social media for near-real-time disaster monitoring

Natural Hazards - Up-to-date information about an emergency is crucial for effective disaster management. However, severe restrictions impede the creation of spatiotemporal information by current...     

Effective fetch and non-linear four-wave interactions during wave growth in slanting fetch conditions

Wave growth in slanting fetch (with wind blowing obliquely off a coast) is investigated with 7 years worth of routine wave measurements in Lake IJssel in The Netherlands and with the SWAN wave model. Two aspects are considered in particular for this case: the validity of the concept of effective fetch and the role of the non-linear four-wave interactions. For slanting and parallel fetch conditions, we found some significant deviations from the effective fetch assumption, leading to 20–35% mismatch in either the peak period T_p or the significant wave height H_m0 respectively. However, the effect of discrepancies between various widely accepted wave growth formulas turned out to be even more important. The wave directions during slanting fetch are significantly ‘steered’ by the coastline, especially in the first kilometre(s) off the coast. The role of the non-linear four-wave interactions is investigated by running the SWAN (version 40.41) wave model with three different quadruplet formulations. Exact quadruplet methods (Xnl) yielded relatively strong wave steering, despite the four-wave interactions being relatively weak. Application of Xnl did not lead to better overall agreement with measurements — improvements for the mean wave period T_m01 were offset by some deterioration for the wave height H_m0.     

Uncertainty in the application of the parabolic bay shape equation: Part 2

This paper attempts to highlight issues that are relevant in the application of the Parabolic Bay Shape Equation (PBSE) to a non-equilibrium bay. For this case, the bay of Imbituba in southern Brazil was chosen. The construction of a breakwater to shelter the port of Imbituba in the south of the bay was accompanied by an increase in sedimentation to the port area from an eroding downdrift beach. Superimposed plots of the coastline of the Bay of Imbituba from different years confirm a general trend of accretion of the southern part of the bay accompanied with a retreat of the coastline in the northern part. After the application of the PBSE it became clear that the breakwater caused a change in the equilibrium state of the bay. Between 1947 and 2001 the Bay of Imbituba has changed from a dynamic equilibrium to a natural beach reshaping or self-reshaping. The tendency of the sedimentation of the southern part of the bay can be explained by the SEP associated with the new updrift diffraction point (tip of the breakwater): The seaward position of the SEP predicts a need for additional sediment in order to achieve a stable plan form.     

Aspects of the cycle of inorganic nitrogen compounds in the tropical rain forest of the Ivory Coast

A partial balance of mineral N is given for the basins of two coastal rivers in a forest zone in the Ivory Coast. The dry and wet depositions on the basin surfaces is given for particulate matter (NO₃ ^–, NH₄ ⁺). The quantity of mineral N washed away in the rivers is evaluated. The losses from leaching of the soils by rainwater are about 0.33 to 1.0% of the atmospheric depositions for NH₄ ⁺–N and 2.2 to 5.8% for NO₃ ^––N. The yearly atmospheric input of N compounds to the ecosystem, about 1.4 g N m^–2 y^–1, is at least 14% of mineral N formed in the soils and is therefore quite significant.     

Etude des Déformations et Inclinaisons du Sous-sol: Développements Instrumentaux et Applications dans la Région Sismique d’Arette,France

We present a modified version of a tilt and strain meter conceived to have great sensibility and good stability. These instruments were deployed in a cave at the seismic Pyrenean region of Arette (France). We set in evidence several types of phenomena, of which deformations brought about by water recharge filling up fissures on the massif are of importance. Owing to the redundance in the data, the study of the co-seismic deformations allows one to separate instrumental displacements and local readjustments from regional motions.     

A numerically calibrated reference level (MP28) for the terrestrial mammal-based biozonation of the European Upper Oligocene

The fauna of the Enspel (Westerwald) and the neighbouring Kärlich (Neuwied basin) fossil deposits correspond to the Upper Oligocene Mammal Paleogene (MP) reference level 28 and 28–30, respectively. Basaltic flows and a trachyte tuff terminating and predating the fossil deposit sedimentation allow to numerically calibrate the MP reference levels by radioisotope dating. Laser fusion ⁴⁰Ar/³⁹Ar step heating on volcanic feldspars yield a time interval of 24.9–24.5 Ma for reference level MP28 at Enspel and a maximum age of 25.5 Ma for the time interval MP28–MP30 at Kärlich. Interpolation between the time intervals determined for the Enspel reference level MP28 and the age of the global Oligocene/Miocene boundary of 24.0 ± 0.1 Ma taken from literature results in time intervals of 24.5–24.2 Ma and 24.2–23.9 Ma for the younger reference levels MP29 and MP30, respectively. These intervals of ≤ 0.4 m.y. for MP reference levels of the latest Oligocene are short relative to older Oligocene MP reference levels 21–27 between 34 and 25 Ma. Since subdivision into MP reference levels essentially is based on assemblages of mammal taxa and on evolutionary changes in tooth morphology of mammals short MP time intervals during the latest Oligocene indicate a rapid evolutionary change relative to the early Oligocene.     

Field‐based and spectral indicators for soil erosion mapping in semi‐arid mediterranean environments (Coastal Cordillera of central Chile)

The Coastal Cordillera of central Chile is naturally sensitive to soil erosion due to moderate to steep slopes, intense winter rains when the vegetation cover is scarce, and deeply weathered granitic rocks. In 1965, 60 per cent of its surface was moderately to very severely eroded. Today this process is still largely active, but no data are currently available to evaluate the real extent, distribution and severity of soil degradation on a regional scale. This information is vital to support efficient soil conservation plans. A multi‐scale approach was implemented to produce regional land degradation maps based on remote sensing technologies. Fieldwork has shown that the surface colour or ‘redness’ and the density of coarse fragments are pertinent erosion indicators to describe a typical sequence of soil degradation in the context of mediterranean soil developed on granitic materials and micaschists. Field radiometric experiments concluded that both factors influence the reflectance of natural surfaces and can be modelled using radiometric indices accessible from most satellites operating in the optical domain, i.e. redness index and brightness index. Finally the radiometric indices were successfully applied to SPOT images to produce land degradation maps. Only broad classes of erosion status were discriminated and the detection of the degradation processes was only possible when most of the fertile layer had already been removed. This technology provides decision‐making information required to develop regional soil conservation plans and to prioritize actions between catchment areas, especially in vast inter‐tropical regions where spatialized data are not always readily available. Copyright © 2006 John Wiley & Sons, Ltd.     

Petrochemical Investigation of the Antique Ophiolite (Philippines): Implications on Volcanogenic Massive Sulfide and Podiform Chromitite Deposits

Abstract: The Antique ophiolite, located in Panay island (west‐central Philippines), corresponds to several tectonic slices within the suture zone between the Philippine Mobile Belt (PMB) and the North Palawan Block (NPB). It includes dismembered fragments of a basaltic sequence, dominantly pillow‐lavas with minor sheet flows, rare exposures of sheeted dikes, isotropic gabbros, subordinate layered mafic and ultramafic rock sequences and serpentinites. Most of the ophiolite units commonly occur as clasts and blocks within the serpentinites, which intrude the whole ophiolitic body, as well as, the basal conglomerate of the overlying Middle Miocene sedimentary formation. The volcanic rock sequence is characterized by chemical compositions ranging from transitional (T)‐MORB, normal (N)‐MORB and to chemistry intermediate between those of MORB and island arc basalt (IAB). The residual upper mantle sequence is harzburgitic and generally more depleted than the upper mantle underlying modern mid‐oceanic ridges. Calculations using whole‐rock and mineral compositions show that they can represent the residue of a fertile mantle source, which have undergone degrees of partial melting ranging from 9‐22.5 %. Some of the mantle samples display chondrite‐nor‐malized REE and extended multi‐element patterns suggesting enrichments in LREE, Rb, Sr and Zr, which are comparable to those found in fore‐arc peridotites from the Izu‐Bonin‐Mariana (IBM) arc system. The Antique ultramafic rocks also record relatively oxidizing mantle conditions (Δlog f_O2 (FMQ)=0.9‐3.5). As a whole, the ophiolite probably represents an agglomeration of oceanic ridge and fore‐arc crust fragments, which were juxtaposed during the Miocene collision of the PMB and the NPB. The intrusion of the serpentinites might be either coeval or subsequent to the accretion of the oceanic crust onto the fore‐arc. Volcanogenic massive sulfide (VMS) deposits occur either in or near the contact between the pillow basalts and the overlying sediments or interbedded with the sediments. The morphology of the deposits, type of metals, ore texture and the nature of the host rocks suggest that the formation of the VMS bodies was similar to the accumulation of metals around and in the subsurface of hydrothermal vents observed in modern mid‐oceanic ridge and back‐arc basin rift settings. The podiform chromitites occur as pods and subordinate layers within totally serpentinized dunite in the residual upper mantle sequence. No large coherent chromitite deposit was found since the host dunitic rocks often occur as blocks within the serpentinites. It is difficult to evaluate the original geodynamic setting of the mineralized bodies since the chemistry of the host rocks were considerably modified by alteration during their tectonic emplacement. A preliminary conclusion for Antique is that the VMS is apparently associated with a primitive tholeiitic intermediate MORB‐IAB volcanic suite, the chemistry of which is close to the calculated composition of the liquid that coexisted with the podiform chromitites.     

Feasibility of AFC models for the petrogenesis of calc-alkaline magma series

 One of the most popular petrogenetic concepts is that of simultaneous assimilation and fractional crystallisation (AFC). The equations governing these processes are well known and widely applied. The power and attraction of the model are that, with the minimum of constraints, petrogenetic schemes can be erected for essentially any group of rocks. These models are usually based on the observed variations in a small number of trace elements and/or isotopic systems. In this paper we have generated well-constrained AFC models for the Variscan, calc-alkaline Quérigut complex in the French Pyrénées. Our models are based on geological constraints on the identities of the parent magma and candidate assimilants, and the Sm-Nd and Rb-Sr isotope systematics of the rocks. The modelling is successful in mimicking the isotope array for the actual rocks. However, the results are massively inconsistent with those produced using the major- and trace-element concentrations. Our major conclusion is that AFC modelling should be undertaken only with robust initial constraints on the compositions of the parent magma and the proposed assimilant, with due respect for crystal-liquid partition coefficients, in so far as they are known. The results should then be accepted only if the isotope-based model parameters are consistent with those derived from major- and trace-element variations. Received: 19 August 1994/Accepted: 23 March 1995