The influence of multivariate analysis methods and target grain size on the accuracy of remote quantitative chemical analysis of rocks using laser induced breakdown spectroscopy

Laser-induced breakdown spectroscopy (LIBS) was used to quantitatively analyze 195 rock slab samples with known bulk chemical compositions, 90 pressed-powder samples derived from a subset of those rocks, and 31 pressed-powder geostandards under conditions that simulate the ChemCam instrument on the Mars Science Laboratory Rover (MSL), Curiosity. The low-volatile (<2 wt.%) silicate samples (90 rock slabs, corresponding powders, and 22 geostandards) were split into training, validation, and test sets. The LIBS spectra and chemical compositions of the training set were used with three multivariate methods to predict the chemical compositions of the test set. The methods were partial least squares (PLS), multilayer perceptron artificial neural networks (MLP ANNs) and cascade correlation (CC) ANNs. Both the full LIBS spectrum and the intensity at five pre-selected spectral channels per major element (feature selection) were used as input data for the multivariate calculations. The training spectra were supplied to the algorithms without averaging (i.e. five spectra per target) and with averaging (i.e. all spectra from the same target averaged and treated as one spectrum). In most cases neural networks did not perform better than PLS for our samples. PLS2 without spectral averaging outperformed all other procedures on the basis of lowest quadrature root mean squared error (RMSE) for both the full test set and the igneous rocks test set. The RMSE for PLS2 using the igneous rock slab test set is: 3.07 wt.% SiO₂, 0.87 wt.% TiO₂, 2.36 wt.% Al₂O₃, 2.20 wt.% Fe₂O₃, 0.08 wt.% MnO, 1.74 wt.% MgO, 1.14 wt.% CaO, 0.85 wt.% Na₂O, 0.81 wt.% K₂O. PLS1 with feature selection and averaging had a higher quadrature RMSE than PLS2, but merits further investigation as a method of reducing data volume and computation time and potentially improving prediction accuracy, particularly for samples that differ significantly from the training set. Precision and accuracy were influenced by the ratio of laser beam diameter (∼490 μm) to grain size, with coarse-grained rocks often resulting in lower accuracy and precision than analyses of fine-grained rocks and powders. The number of analysis spots that were normally required to produce a chemical analysis within one standard deviation of the true bulk composition ranged from ∼10 for fine-grained rocks to >20 for some coarse-grained rocks.     

African climate variability in the satellite era

Interannual fluctuations in rainfall and ocean-atmosphere fields over and around Africa were studied in the satellite era of 1979–2007 using singular value decomposition. The leading modes of rainfall variability in GPCP satellite-gauge merged fields include a leading mode over central Africa, two modes of marine origin in the Gulf of Guinea and Eastern Africa, and two sub-tropical modes over the Sahel and Southern Africa. This differs from earlier gauge-based studies that tend to isolate three leading modes over western, eastern, and southern Africa. In the sea-surface temperature, sea-level pressure and upper wind fields, ENSO signals dominate the leading modes. However, for the low-level wind field, a trough circulation over the southeast Atlantic – Kalahari is the leading mode. It demonstrates predictive potential when cross-correlated with rainfall at 6- to 12-month lead time. Based on continuous filtered data, the value of various indices and the predictability of different zones are examined. The Sahel achieves the highest rank followed by the Congo and southern zones in the next tier. The Guinea and East African rains, which peak in the March-to-May season, appear least predictable. The seasonal rainfall is shown to modulate economic growth rate, and multi-variate predictive algorithms are tested at 6-month lead time.     

Beyond the circle of poison: Significant shifts in the global pesticide complex, 1976–2008

Almost 30 years after its introduction, the “circle of poison” remains a common conceptualization of the global pesticide complex among scholars and especially in popular understanding. The circle of poison describes a situation in which, pesticides banned in industrialized countries continue to be manufactured there and exported to developing countries, are then used in developing countries almost entirely on export crops, and return to industrialized countries as pesticide residues on food. Using secondary data and a case study of pesticide use in Costa Rica, I review the applicability of the circle of poison conceptualization to the current global pesticide complex. I argue that (1) the circle of poison is no longer accurate due to important global changes in pesticide regulation, production, trade, sales, and use driven by a number of dynamic economic, social, and ecological processes; (2) using industrialized countries’ pesticide regulations as proxies for safety should be replaced by multi-characteristic risk assessments; and (3) revisions of the circle of poison conceptualization should be updated because of export farmers’ adoption of newer classes of pesticides. The paper concludes by offering a new characterization of the global pesticide complex vis-à-vis pesticide use in developing countries: pesticide divergence by market orientation.     

Gravitational waves from scattering of stellar-mass black holes in galactic nuclei

Stellar-mass black holes (BHs) are expected to segregate and form a steep density cusp around supermassive black holes (SMBHs) in galactic nuclei. We follow the evolution of a multimass system of BHs and stars by numerically integrating the Fokker–Planck energy diffusion equations for a variety of BH mass distributions. We find that the BHs 'self-segregate', and that the rarest, most massive BHs dominate the scattering rate closest to the SMBH  (≲10⁻¹ pc)  . BH–BH binaries form out of gravitational wave emission during BH encounters. We find that the expected rate of BH coalescence events detectable by Advanced LIGO is  ∼1–10² yr⁻¹  , depending on the initial mass function of stars in galactic nuclei and the mass of the most massive BHs. We find that the actual merger rate is likely ∼10 times larger than this due to the intrinsic scatter of stellar densities in many different galaxies. The BH binaries that form this way in galactic nuclei have significant eccentricities as they enter the LIGO band (90 per cent with   e > 0.9  ), and are therefore distinguishable from other binaries, which circularize before becoming detectable. We also show that eccentric mergers can be detected to larger distances and greater BH masses than circular mergers, up to  ∼700 M_⊙  . Future ground-based gravitational wave observatories will be able to constrain both the mass function of BHs and stars in galactic nuclei.     

Compositional and lithological diversity among brecciated lunar meteorites of intermediate iron concentration

Abstract— We present new compositional data for 30 lunar stones representing about 19 meteorites. Most have iron concentrations intermediate to those of the numerous feldspathic lunar meteorites (3–7% FeO) and the basaltic lunar meteorites (17–23% FeO). All but one are polymict breccias. Some, as implied by their intermediate composition, are mainly mixtures of brecciated anorthosite and mare basalt, with low concentrations of incompatible elements such as Sm (1–3 μg/g). These breccias likely originate from points on the Moon where mare basalt has mixed with material of the FHT (Feldspathic Highlands Terrane). Others, however, are not anorthosite‐basalt mixtures. Three (17–75 μ/g Sm) consist mainly of nonmare mafic material from the nearside PKT (Procellarum KREEP Terrane) and a few are ternary mixtures of material from the FHT, PKT, and maria. Some contain mafic, nonmare lithologies like anorthositic norites, norites, gabbronorites, and troctolite. These breccias are largely unlike breccias of the Apollo collection in that they are poor in Sm as well as highly feldspathic anorthosite such as that common at the Apollo 16 site. Several have high Th/Sm compared to Apollo breccias. Dhofar 961, which is olivine gabbronoritic and moderately rich in Sm, has lower Eu/Sm than Apollo samples of similar Sm concentration. This difference indicates that the carrier of rare earth elements is not KREEP, as known from the Apollo missions. On the basis of our present knowledge from remote sensing, among lunar meteorites Dhofar 961 is the one most likely to have originated from South Pole‐Aitken basin on the lunar far side.     

Star clusters around recoiled black holes in the Milky Way halo

Gravitational wave emission by coalescing black holes (BHs) kicks the remnant BH with a typical velocity of hundreds of  km s⁻¹  . This velocity is sufficiently large to remove the remnant BH from a low-mass galaxy but is below the escape velocity from the Milky Way (MW) galaxy. If central BHs were common in the galactic building blocks that merged to make the MW, then numerous BHs that were kicked out of low-mass galaxies should be freely floating in the MW halo today. We use a large statistical sample of possible merger tree histories for the MW to estimate the expected number of recoiled BH remnants present in the MW halo today. We find that hundreds of BHs should remain bound to the MW halo after leaving their parent low-mass galaxies. Each BH carries a compact cluster of old stars that populated the core of its original host galaxy. Using the time-dependent Fokker–Planck equation, we find that the present-day clusters are  ≲1 pc  in size, and their central bright regions should be unresolved in most existing sky surveys. These compact systems are distinguishable from globular clusters by their internal (Keplerian) velocity dispersion greater than 100 km s⁻¹ and their high mass-to-light ratio owing to the central BH. An observational discovery of this relic population of star clusters in the MW halo would constrain the formation history of the MW and the dynamics of BH mergers in the early Universe. A similar population should exist around other galaxies and may potentially be detectable in M31 and M33.     

Particle propagation effects on solar flare X- and γ-ray time profiles

Much of the evidence for second stage particle acceleration in solar flares lies in the temporal variation of solar X- and -ray emissions. However, the solar flare X- and -ray burst time-intensity profiles are governed not only by the production or acceleration of electrons and protons but by the propagation of these particles in the solar atmosphere. The effects of particle propagation on X-ray and -ray time profiles are illustrated and compared through the use of three models with the result that a variety of particle propagation schemes reproduce effects commonly associated with second stage acceleration. The first model is that of a closed uniform density trap. The other two models employ particle diffusion from a trap to denser regions of the solar atmosphere to produce the high energy radiation. These calculations show that delayed peaking of the photon flux with respect to particle production and reduction in the impulsiveness of the high energy emission is to be expected, effects commonly associated with second stage acceleration. Thus, well understood physical processes are capable of producing so-called time delays in the high energy emission independent of any delays produced by additional particle acceleration processes. Diagnostic differences between these models are also discussed.     

Side-looking sonar backscatter response at dual frequencies

Dual-frequency side-looking sonars have the potential to be used as remote sensing tools to characterize subaqueous terrains. In one case study of the carbonate-ooze-coated Blake Plateau off-shore of Georgia, U.S.A., the difference in acoustic attenuation for 50 and 20 mm wavelengths (30 and 72 kHz frequency) permits the discrimination of sub-bottom scatterers from seabed surface textural features to reveal patchy regions where a buried hard ground had been pock-marked by karst-like depressions. In a second study of the Upper Hudson River in New York, U.S.A., related to environmental contaminates, the backscatter response at 15 and 3 mm acoustic wavelengths (100 and 500 kHz frequency) serves as a useful proxy for sediment grain size with coarser detritus distinguished from finer sediments. Sand and gravel regions inferred from the backscatter were confirmed by ground truth sampling.     

Sea level fluctuations in central California at subtidal to decadal and longer time scales with implications for San Francisco Bay, California

Sea level elevations from near the mouth of San Francisco Bay are used to describe the low-frequency variability of forcing of the coastal ocean on the Bay at a variety of temporal scales. About 90% of subtidal fluctuations in sea level in San Francisco Bay are driven by the sea level variations in the coastal ocean that propagate into the Bay at the estuary mouth. We use the 100-year sea level record available at San Francisco to document a 1.9 mm/yr mean sea level rise, and to determine fluctuations related to El Nino-Southern Oscillation (ENSO) and other climatic events. At time scales greater than 1 year, ENSO dominates the sea level signal and can result in fluctuations in sea level of 10–15 cm. Alongshore wind stress data from central California are also analyzed to determine the impact of changes in coastal elevation at the mouth of San Francisco Bay within the synoptic wind band of 2–30 days. At least 40% of the subtidal fluctuations in sea level of the Bay are tied to the large-scale regional wind field affecting sea level variations in the coastal ocean, with little local, direct wind forcing of the Bay itself. The majority of the subtidal sea level fluctuations within the Bay that are not related to the coastal ocean sea level signal are forced by an east–west sea level gradient resulting from tidally induced variations in sea level at specific beat frequencies that are enhanced in the northern reach of the Bay. River discharge into the Bay through the Sacramento and San Joaquin River Delta also contributes to the east–west gradient, but to a lesser degree.