DISCRIMINANT PRINCIPAL COMPONENTS ANALYSIS

When the number of variables exceeds the number of samples, one method of multivariate discriminationis to use principal components analysis to reduce the dimensionality and then to perform canonicalvariates analysis (PC-CVA). This paper proposes an alternative approach in which discriminant analysisis carried out by a weighted principal component analysis of the group means (DPCA). This method doesnot require prior data reduction and produces discriminant factors that are orthogonal in the original dataspace. The theory and performance of the two methods are compared. Although the individual factors ofDPCA are found to be less discriminating than PC-CVA, the overall discrimination, calculated bymultivariate analysis of variance, and the predictive value, estimated by the leaving-one-out error rate,are broadly comparable.     

Seismic excitation of the polar motion, 1977–1993

The mass redistribution in the earth as a result of an earthquake faulting changes the earth's inertia tensor, and hence its rotation. Using the complete formulae developed byChao andGross (1987) based on the normal mode theory, we calculated the earthquake-induced polar motion excitation for the largest 11,015 earthquakes that occurred during 1977.0–1993.6. The seismic excitations in this period are found to be two orders of magnitude below the detection threshold even with today's high precision earth rotation measurements. However, it was calculated that an earthquake of only one tenth the size of the great 1960 Chile event, if happened today, could be comfortably detected in polar motion observations. Furthermore, collectively these seismic excitations have a strong statistical tendency to nudge the pole towards 140°E, away from the actually observed polar drift direction. This non-random behavior, similarly found in other earthquake-induced changes in earth rotation and low-degree gravitational field byChao andGross (1987), manifests some geodynamic behavior yet to be explored.     

Numerical investigation of the effects of upwelling on harmful algal blooms off the west Florida coast

The initiation of the toxic harmful algal bloom (HAB), Karenia brevis, along the west Florida coast has been associated with upwelling events. Upwelling processes may be responsible for the transport of nutrients or algae from deep offshore locations across the Florida shelf to the coast. The influence of coastal wind-driven upwelling on the onset and occurrences of K. brevis in this region was numerically investigated using Rutgers University's Regional Ocean Modeling System. Computations were carried out in an idealized model domain, a two-dimensional slice in the cross-shore and vertical directions. The surface forcing data used was from several offshore meteorological buoys. The motion of the algae was simulated using Lagrangian particles and a passive tracer. The numerical simulations of three K. brevis events in 2000–2002 showed that the particles respond (with upwelling/downwelling) to the along-shore wind stresses as expected and some upwelling was present during the events. Comparison of the passive tracer fields with measured fluorescence data exposed the model's sensitivity to the particular surface forcing data employed and the relatively more significant role played by surface forcing over initial conditions. The present model set-up constitutes a useful predictive tool for conditions conducive to the onset of HABs. It is planned to be used in a real-time mode to aid the NOAA HAB monitoring and forecasting system.     

Effect of chlorine on near-liquidus phase equilibria of an Fe–Mg-rich tholeiitic basalt

The importance of Cl in basalt petrogenesis has been recognized, yet constraints on its effect on liquidus crystallization of basalts are scarce. In order to quantify the role of Cl in basaltic systems, we have experimentally determined near-liquidus phase relations of a synthetic Fe–Mg-rich basalt, doped with 0.0–2.5 wt% dissolved Cl, at 0.7, 1.1, and 1.5 GPa. Results have been parameterized and compared with previous data from literature. The effect of Cl on mineral chemistry and liquidus depression is dependent on the starting basaltic composition. The liquidus depression measured for a SiO₂-rich, Al₂O₃-poor basalt is smaller than that observed for a basaltic melt depleted in silica and enriched in FeO_T and Al₂O₃. The effect of Cl on depression of the olivine–orthopyroxene–liquid multiple saturation pressure does not seem to vary with the starting composition of the basaltic liquid. This suggests that Cl may significantly promote the generation of silica-poor, Fe–Al-rich magmas in the Earth, Mars, and the Moon.     

Effect of chlorine on near‐liquidus crystallization of olivine‐phyric shergottite NWA 6234 at 1 GPa: Implication for volatile‐induced melting of the Martian mantle

Martian magmas are thought to be rich in chlorine compared with their terrestrial counterparts. Here, we experimentally investigate the effect of chlorine on liquidus depression and near‐liquidus crystallization of olivine‐phyric shergottite NWA 6234 and compare these results with previous experimental results on the effect of chlorine on near‐liquidus crystallization of the surface basalts Humphrey and Fastball. Previous experimental results showed that the change in liquidus temperature is dependent on the bulk composition of the basalt. The effect of chlorine on liquidus depression is greater for lower SiO₂ and higher Al₂O₃ magmas than for higher SiO₂ and lower Al₂O₃ magmas. The bulk composition for this study has lower Al₂O₃ and higher FeO contents than previous work; therefore, we provide additional constraints on the effect of the bulk composition on the influence of chlorine on near‐liquidus crystallization. High pressure and temperature crystallization experiments were performed at 1 GPa on a synthetic basalt, of the bulk composition of NWA 6234, with 0–4 wt% Cl added to the sample as AgCl. The results are consistent with previous notions that with increasing wt% Cl in the melt, the crystallization temperature decreases. Importantly, our results have a liquidus depression ?T (°C) from added chlorine that is consistent with the difference in bulk composition and suggest a dependence on both the bulk Al₂O₃ and FeO content. Our results suggest that the addition of chlorine to the Martian mantle may lower magma genesis temperatures and potentially aid in the petrogenesis of Martian magmas.     

On the Parametrization of Urban Land Use in Mesoscale Models

The effects of urban structures on the distribution of meteorological variables can be included in mesoscale models by an appropriate parametrization. The different approaches are conventionally tested against wind profiles in the centre of the urban area while flow distortions around are not considered. In this study, the quality of different parametrizations in capturing the main wind-field modifications in, as well as around, a complex obstacle is investigated. The method applied consists of a building resolved microscale model and a mesoscale model including a suitable parametrization. The results demonstrate that a drag or a porosity approach can reproduce very satisfactorily the main characteristics of the airflow completely, while a simpler roughness length concept in general approximates the mean flow unsatisfactorily.     

Simulating damage evolution and fracture propagation in sandstone containing a preexisting 3-D surface flaw under uniaxial compression

Preexisting flaws and rock heterogeneity have important ramifications on the process of rock fracturing and on rock stability in many applications. Therefore, there is great interest in numerical modelling of rock fracture and the underlying mechanisms. We simulated damage evolution and fracture propagation in sandstone specimens containing a preexisting 3-D surface flaw under uniaxial compression. We applied the linear elastic damage model based on the unified strength theory following the rock failure process analysis code. However, in contrast to the rock failure process analysis code, we used the finite element method with tetrahedron elements on unstructured meshes. It provided higher geometrical flexibility and allowed for a more accurate representation of the disk-shaped flaw with various flaw depths, angles, and lengths through locally adapted meshes. The rock heterogeneity was modelled by sampling the initial local Young's modulus from a Weibull distribution over a cubic grid. The values were then interpolated to the computational finite element method mesh. This method introduced an additional length scale for the rock heterogeneity represented by the cell size in the sampling grid. The generation of three typical surface cracking patterns, called wing cracks, anti-wing cracks, and far-field cracks, were identified in the simulation results. These depend on the geometry of the preexisting surface flaw. The simulated fracture propagation, coalescence types, and failure modes for the specimens with preexisting surface flaw show good agreement with recent experimental studies.     

Storage of carbon in U.S. forests predicted from satellite data,ecosystem modeling,and inventory summaries

A plant and soil simulation model based on satellite observations of vegetation and climate data was used to estimate the potential carbon pools in standing wood biomass across all forest ecosystems of the conterminous United States up to the year 1997. These modeled estimates of vegetative carbon potential were compared to aggregated measurements of standing wood biomass from the U. S. Forest Service’s national Forest Inventory and Analysis (FIA) data set and the Carbon Online Estimator (COLE) to understand: 1) predominant geographic variations in tree growth rate and 2) local land cover and land use history including the time since the last stand-replacing disturbance (e.g., from wildfire or harvest). Results suggest that although wood appears to be accumulating at high rates in many areas of the U.S. (Northwest and Southeast), there are still extensive areas of relatively low biomass forest in the late 1990s according to FIA records. We attribute these low biomass accumulation levels to the high frequency of disturbances, which can be observed even in high production areas such as the Southeast due to frequent forest harvests. Ecosystem models like the one presented in this study have been coupled with satellite observations of land cover and green plant density to uniquely differentiate areas with a high potential for vegetative carbon storage at relatively fine spatial resolution.     

Reflection on the first five years of South Africa’s Acoustic Tracking Array Platform (ATAP): status,challenges and opportunities

The Acoustic Tracking Array Platform (ATAP) is a marine science programme that monitors the movements and migrations of inshore marine animals along the South African coastline. Acoustically tagged animals are monitored by an expanded network of approximately 100 automated data-logging acoustic receivers moored at strategic node sites, from Cape Point in the west to the South Africa–Mozambique border in the east. During five years since its inception, in 2011, the ATAP has achieved outstanding progress in terms of the numbers of animals and the variety of species tagged. To date, the ATAP has yielded over 2.6 million detections from more than 700 acoustically tagged animals, representing 27 different species from 20 families, including the African penguin Spheniscus demersus, fishery-at-risk species (e.g. dusky kob Argyrosomus japonicus) and iconic elasmobranchs (e.g. white shark Carcharodon carcharias). Following a period of considerable equipment loss in 2014, refinement of the receiver network was required and the deeper receivers at each site were decommissioned without influencing the integrity of the nationwide array. The platform, managed by the South African Institute for Aquatic Biodiversity, ultimately represents a low-cost method of collecting long-term data that currently benefits approximately 25 researchers from 14 organisations. Case studies are presented to expose the opportunities provided by the ATAP, which will undoubtedly yield new discoveries and provide a greater understanding of the movement patterns and migrations of a wide variety of inshore marine and estuary-associated species.