Boosting for Mineral Prospectivity Modeling: A New GIS Toolbox

With an increasing demand for raw materials, predictive models that support successful mineral exploration targeting are of great importance. We evaluated different machine learning techniques with an emphasis on boosting algorithms and implemented them in an ArcGIS toolbox. Performance was tested on an exploration dataset from the Iberian Pyrite Belt (IPB) with respect to accuracy, performance, stability, and robustness. Boosting algorithms are ensemble methods used in supervised learning for regression and classification. They combine weak classifiers, i.e., classifiers that perform slightly better than random guessing to obtain robust classifiers. Each time a weak learner is added; the learning set is reweighted to give more importance to misclassified samples. Our test area, the IPB, is one of the oldest mining districts in the world and hosts giant volcanic-hosted massive sulfide (VMS) deposits. The spatial density of ore deposits, as well as the size and tonnage, makes the area unique, and due to the high data availability and number of known deposits, well-suited for testing machine learning algorithms. We combined several geophysical datasets, as well as layers derived from geological maps as predictors of the presence or absence of VMS deposits. Boosting algorithms such as BrownBoost and Adaboost were tested and compared to Logistic Regression (LR), Random Forests (RF) and Support Vector machines (SVM) in several experiments. We found performance results relatively similar, especially to BrownBoost, which slightly outperformed LR and SVM with respective accuracies of 0.96 compared to 0.89 and 0.93. Data augmentation by perturbing deposit location led to a 7% improvement in results. Variations in the split ratio of training and test data led to a reduction in the accuracy of the prediction result with relative stability occurring at a critical point at around 26 training samples out of 130 total samples. When lower numbers of training data were introduced accuracy dropped significantly. In comparison with other machine learning methods, Adaboost is user-friendly due to relatively short training and prediction times, the low likelihood of overfitting and the reduced number of hyperparameters for optimization. Boosting algorithms gave high predictive accuracies, making them a potential data-driven alternative for regional scale and/or brownfields mineral exploration.

     

Experimental impact cratering: A summary of the major results of the MEMIN research unit

This paper reviews major findings of the Multidisciplinary Experimental and Modeling Impact Crater Research Network (MEMIN). MEMIN is a consortium, funded from 2009 till 2017 by the German Research Foundation, and is aimed at investigating impact cratering processes by experimental and modeling approaches. The vision of this network has been to comprehensively quantify impact processes by conducting a strictly controlled experimental campaign at the laboratory scale, together with a multidisciplinary analytical approach. Central to MEMIN has been the use of powerful two-stage light-gas accelerators capable of producing impact craters in the decimeter size range in solid rocks that allowed detailed spatial analyses of petrophysical, structural, and geochemical changes in target rocks and ejecta. In addition, explosive setups, membrane-driven diamond anvil cells, as well as laser irradiation and split Hopkinson pressure bar technologies have been used to study the response of minerals and rocks to shock and dynamic loading as well as high-temperature conditions. We used Seeberger sandstone, Taunus quartzite, Carrara marble, and Weibern tuff as major target rock types. In concert with the experiments we conducted mesoscale numerical simulations of shock wave propagation in heterogeneous rocks resolving the complex response of grains and pores to compressive, shear, and tensile loading and macroscale modeling of crater formation and fracturing. Major results comprise (1) projectile–target interaction, (2) various aspects of shock metamorphism with special focus on low shock pressures and effects of target porosity and water saturation, (3) crater morphologies and cratering efficiencies in various nonporous and porous lithologies, (4) in situ target damage, (5) ejecta dynamics, and (6) geophysical survey of experimental craters.     

Hydraulic roughness over simple subaqueous dunes

Detailed studies of flow over subaqueous dunes in laboratory flumes were used to suggest a virtual near-bed layer of twice the dune height in which the mean velocity is accelerated towards the crest by contraction. The mean flow velocity in this layer above the crest, transformed into friction velocity by means of the surface skin roughness, is shown to give values consistent with measured values. The resulting dimensionless shear stress due to skin friction is depth-independent, in contrast to that derived by means of often cited traditional methods. As a result of the relationship between dune height and the thickness of the near-bed layer, an expression for the expansion loss behind dunes was formulated and used to relate form resistance directly to dune height.     

Thermophysical properties of Almahata Sitta meteorites (asteroid 2008 TC3) for high‐fidelity entry modeling

Asteroid 2008 TC₃ was characterized in a unique manner prior to impacting Earth's atmosphere, making its October 7, 2008, impact a suitable field test for or validating the application of high‐fidelity re‐entry modeling to asteroid entry. The accurate modeling of the behavior of 2008 TC₃ during its entry in Earth's atmosphere requires detailed information about the thermophysical properties of the asteroid's meteoritic materials at temperatures ranging from room temperature up to the point of ablation (T ~ 1400 K). Here, we present measurements of the thermophysical properties up to these temperatures (in a 1 atm. pressure of argon) for two samples of the Almahata Sitta meteorites from asteroid 2008 TC₃: a thick flat‐faced ureilite suitably shaped for emissivity measurements and a thin flat‐faced EL6 enstatite chondrite suitable for diffusivity measurements. Heat capacity was determined from the elemental composition and density from a 3‐D laser scan of the sample. We find that the thermal conductivity of the enstatite chondrite material decreases more gradually as a function of temperature than expected, while the emissivity of the ureilitic material decreases at a rate of 9.5 × 10^?5 K^?1 above 770 K. The entry scenario is the result of the actual flight path being the boundary to the load the meteorite will be affected with when entering. An accurate heat load prediction depends on the thermophysical properties. Finally, based on these data, the breakup can be calculated accurately leading to a risk assessment for ground damage.     

On the occurrence and origin of anthropogenic radionuclides found in a fragment of the Chelyabinsk (LL5) meteorite

A piece of the 2013 Chelyabinsk meteorite was investigated for its content of anthropogenic radionuclides. In addition to traces of cesium‐137 that had been previously reported for this particular fragment, we found an unusually high amount of strontium‐90, which indicates that the source of this contamination was the Kyshtym accident (1957). A high Sr‐90/Cs‐137 activity ratio is characteristic for Kyshtym‐derived contaminations. Based on the cesium‐137 content in the soil from the finding site, it is estimated that the fragment was contaminated with soil particles in the milligram range upon impact. Investigation of the soil revealed very unusual ferromagnetic characteristics and an iron‐rich chemical composition. Mössbauer spectroscopy indicated the presence of steel components in this soil, suggesting that the investigated meteorite fragment was found in an industrial dumping site rather than natural soil.     

Nonequilibrium Processes in the Solar Corona,Transition Region,Flares, and Solar Wind <Emphasis Type="Italic">(Invited Review)</Emphasis>

We review the presence and signatures of the non-equilibrium processes, both non-Maxwellian distributions and non-equilibrium ionization, in the solar transition region, corona, solar wind, and flares. Basic properties of the non-Maxwellian distributions are described together with their influence on the heat flux as well as on the rates of individual collisional processes and the resulting optically thin synthetic spectra. Constraints on the presence of high-energy electrons from observations are reviewed, including positive detection of non-Maxwellian distributions in the solar corona, transition region, flares, and wind. Occurrence of non-equilibrium ionization is reviewed as well, especially in connection to hydrodynamic and generalized collisional-radiative modeling. Predicted spectroscopic signatures of non-equilibrium ionization depending on the assumed plasma conditions are summarized. Finally, we discuss the future remote-sensing instrumentation that can be used for the detection of these non-equilibrium phenomena in various spectral ranges.     

Variability and propagation of Labrador Sea Water in the southern subpolar North Atlantic

The variability of two modes of Labrador Sea Water (LSW) (upper and deep Labrador Sea Water) and their respective spreading in the interior North Atlantic Ocean are investigated by means of repeated ship surveys carried out along the zonal WOCE line A2/AR19 located at 43–48°N (1993–2007) and along the GOOS line at about 48–51°N (1997–2002). Hydrographic section data are complemented by temperature, salinity, and velocity time series recorded by two moorings. They have been deployed at the western flank of the Mid-Atlantic Ridge (MAR) in the Newfoundland Basin during 1996–2004. The analysis of hydrographic anomalies at various longitudes points to a gradual eastward propagation of LSW-related signals, which happens on time scales of 3–6 years from the formation region towards the MAR. Interactions of the North Atlantic Current (NAC) with the Deep Western Boundary Current (DWBC) close to Flemish Cap point to the NAC being the main distributor of the different types of LSW into the interior of the Newfoundland Basin. Comparisons between the ship data and the mooring records revealed that the mooring sites are located in a region affected by highly variable flow. The mooring time series demonstrate an elevated level of variability with eddy activity and variability associated with the NAC considerably influencing the LSW signals in this region. Hydrographic data taken from Argo profiles from the vicinity of the mooring sites turned out to mimic quite well the temporal evolution captured by the moorings. There is some indication of occasional southward flow in the LSW layer near the MAR. If this can be considered as a hint to an interior LSW-route, it is at least of minor importance in comparison to the DWBC. It acts as an important supplier for the interior North Atlantic, distributing older and recently formed LSW modes southward along the MAR.     

Fe and O isotope composition of meteorite fusion crusts: Possible natural analogues to chondrule formation?

Meteorite fusion crust formation is a brief event in a high‐temperature (2000–12,000 K) and high‐pressure (2–5 MPa) regime. We studied fusion crusts and bulk samples of 10 ordinary chondrite falls and 10 ordinary chondrite finds. The fusion crusts show a typical layering and most contain vesicles. All fusion crusts are enriched in heavy Fe isotopes, with δ⁵⁶Fe values up to +0.35‰ relative to the solar system mean. On average, the δ⁵⁶Fe of fusion crusts from finds is +0.23‰, which is 0.08‰ higher than the average from falls (+0.15‰). Higher δ⁵⁶Fe in fusion crusts of finds correlate with bulk chondrite enrichments in mobile elements such as Ba and Sr. The δ⁵⁶Fe signature of meteorite fusion crusts was produced by two processes (1) evaporation during atmospheric entry and (2) terrestrial weathering. Fusion crusts have either the same or higher δ¹⁸O (0.9–1.5‰) than their host chondrites, and the same is true for Δ¹⁷O. The differences in bulk chondrite and fusion crust oxygen isotope composition are explained by exchange of oxygen between the molten surface of the meteorites with the atmosphere and weathering. Meteorite fusion crust formation is qualitatively similar to conditions of chondrule formation. Therefore, fusion crusts may, at least to some extent, serve as a natural analogue to chondrule formation processes. Meteorite fusion crust and chondrules exhibit a similar extent of Fe isotope fractionation, supporting the idea that the Fe isotope signature of chondrules was established in a high‐pressure environment that prevented large isotope fractionations. The exchange of O between a chondrule melt and an ¹⁶O‐poor nebula as the cause for the observed nonmass dependent O isotope compositions in chondrules is supported by the same process, although to a much lower extent, in meteorite fusion crusts.     

Phase center modeling for LEO GPS receiver antennas and its impact on precise orbit determination

Most satellites in a low-Earth orbit (LEO) with demanding requirements on precise orbit determination (POD) are equipped with on-board receivers to collect the observations from Global Navigation Satellite systems (GNSS), such as the Global Positioning System (GPS). Limiting factors for LEO POD are nowadays mainly encountered with the modeling of the carrier phase observations, where a precise knowledge of the phase center location of the GNSS antennas is a prerequisite for high-precision orbit analyses. Since 5 November 2006 (GPS week 1400), absolute instead of relative values for the phase center location of GNSS receiver and transmitter antennas are adopted in the processing standards of the International GNSS Service (IGS). The absolute phase center modeling is based on robot calibrations for a number of terrestrial receiver antennas, whereas compatible antenna models were subsequently derived for the remaining terrestrial receiver antennas by conversion (from relative corrections), and for the GNSS transmitter antennas by estimation. However, consistent receiver antenna models for space missions such as GRACE and TerraSAR-X, which are equipped with non-geodetic receiver antennas, are only available since a short time from robot calibrations. We use GPS data of the aforementioned LEOs of the year 2007 together with the absolute antenna modeling to assess the presently achieved accuracy from state-of-the-art reduced-dynamic LEO POD strategies for absolute and relative navigation. Near-field multipath and cross-talk with active GPS occultation antennas turn out to be important and significant sources for systematic carrier phase measurement errors that are encountered in the actual spacecraft environments. We assess different methodologies for the in-flight determination of empirical phase pattern corrections for LEO receiver antennas and discuss their impact on POD. By means of independent K-band measurements, we show that zero-difference GRACE orbits can be significantly improved from about 10 to 6 mm K-band standard deviation when taking empirical phase corrections into account, and assess the impact of the corrections on precise baseline estimates and further applications such as gravity field recovery from kinematic LEO positions.     

Ice resistance measurements in ridges with IB APU in the Baltic sea

Ice ridges form a difficult obstacle for ice navigation. Despite this fact, the resistance of ships in ridges has been investigated very little.A trial was performed with the Finnish icebreaker APU (propulsion machinery power of 8·8 MW) in April 1974 in the Baltic, in order to measure its resistance in ridges.Most of the 11 ridges were so massive that the ship could not penetrate through them by a continuous mode, the ramming mode had to be used.When the ship was penetrating a ridge, its speed and the propeller revolutions were registered for determining the resistance. A doppler radar was used for measuring ship spped.In case the ship was stopped in the ridge it was extracted, reversed and accelerated for the next ram. The time history was registered for determining the ship's speed of advance.Before starting a test, the ridge profile above the water level was measured. This was simply done by using a levelling instrument.The main object of the test was to determine the ship's speed of advance. The results, i.e. ice resistance, ship speed and ridge characteristics were analysed on three levels: momentary values average values for rams and average values for a series of rams in one ridge.The test series presented in the article was the first one in full scale in which the speed of the advance of a ship moving by ramming was determined and the ridge profiles were mapped. The measuring system developed worked well. As results the tests gave data of the ship's resistance and of the ship's speed of advance in ridges.