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.
Increasing concentrations of atmospheric carbon dioxide are causing oceanic pH to decline worldwide, a phenomenon termed ocean acidification. Mounting experimental evidence indicates that near-future levels of CO2 will affect calcareous invertebrates such as corals, molluscs and gastropods, by reducing their scope for calcification. Despite extensive research into ocean acidification in recent years, the effects on non-calcifying anthozoans, such as sea anemones, remain little explored. In Western Europe, intertidal anemones such as Actinia equina are abundant, lower trophic-level organisms that function as important ecosystem engineers. Changes to behaviours of these simple predators could have implications for intertidal assemblages. This investigation identified the effects of reduced seawater pH on feeding and contest behaviour by A. equina. Video footage was recorded for A. equina feeding at current-day seawater (pH 8.1), and the least (pH 7.9) and most (pH 7.6) severe end-of-century predictions. Footage was also taken of contests over ownership of space between anemones exposed to reduced pH and those that were not. No statistically significant differences were identified in feeding duration or various aspects of contest behaviour including initiating, winning, inflating acrorhagi, inflicting acrorhagial peels and contest duration. Multivariate analyses showed no effect of pH on a combination of these variables. This provides contrast with other studies where anemones with symbiotic algae thrive in areas of natural increased acidity. Thus, novel experiments using intraspecific contests and resource-holding potential may prove an effective approach to understand sub-lethal consequences of ocean acidification for A. equina, other sea anemones and more broadly for marine ecosystems. 相似文献
The 1991 Pinatubo eruption left 5–6 km3 of debris on the volcano slopes, much of which has been mobilized into large lahars in the following rainy seasons. Also
during the eruption, collapse, localized in part along preexisting faults, left a caldera 2.5 km in diameter that almost immediately
began to accumulate a 1.6 × 108 m3 lake. By 2001, the water had risen to the fault-controlled Maraunot Notch, the lowest, northwestern portion of the caldera
rim comprising the physiographic sill of the Caldera Lake. That year, a narrow artificial canal dug into an old volcanic breccia
underlying the outlet channel failed to induce a deliberate lake breakout, but discharge from heavy rains in July 2002 rapidly
deepened the notch by 23 m, releasing an estimated 6.5 × 107 m3 of lake water that bulked up into lahars with a volume well in excess of 1.6 × 108 m3. Lakes in other volcanoes have experienced multiple breakouts, providing practical motivation for this study. Fieldwork and
high-resolution digital elevation models reveal andesites and ancient lacustrine deposits, strongly fractured and deformed
along a segment of the Maraunot Fault, a prominent, steeply dipping, left-lateral fault zone that trends N35°–40°W within
and parallel to the notch. Seismicity in 1991 demonstrated that the Maraunot Fault is still active. The fault zone appears
to have previously been the erosional locus for a large channel, filled with avalanche or landslide deposits of an earlier
eruption that were exhumed by the 2002 breakout floods. The deformed lacustrine sediments, with an uncalibrated 14C age of 14,760 ± 40 year BP from a single charcoal sample, attest to the existence of an earlier lake, possibly within the
Tayawan Caldera, rim remnants of which survive as arcuate escarpments. That lake may well have experienced one or more ancient
breakouts as well. The 2002 event greatly reduced the possibility of another such event by scouring away the erodible breccia,
leaving less erodible fractured andesites and lacustrine rocks, and by enlarging the outlet channel and its discharge capacity.
Several lines of evidence indicate, however, that future lahar-generating lake breakouts at the notch may keep populations
of Botolan municipality downstream at risk: (1) a volume of 9.5 × 107 m3 of lake water remains perched 0.8 km above sea level; (2) seismicity in 1991 demonstrated that the Maraunot Fault is still
active and movements of sufficient magnitude could enlarge the outlet and the discharge through it; (3) more likely, however,
with or without earthquake activity, landslides from the steep to overhanging channel walls could block the channel again,
and a major rainstorm could then cause a rise in lake level and sudden breakouts; (4) intrusion of a new dome into the bottom
of the lake, possibly accompanied by phreatic explosions, could expel large volumes of lahar-generating water. 相似文献
The influence of large bedrock landslides (“rock avalanches”) on the behaviour of glaciers is incompletely recognised. Here we present an example from an active tectonic margin in South Island, New Zealand where large earthquakes leave a significant imprint on glacial records. We demonstrate that terminal moraines on the western side of the Southern Alps record both ‘ordinary’ (i.e. climate-driven) and landslide-initiated glacial advances. Following consideration of the processes involved in rock avalanche-initiated moraine construction we suggest ways of determining the nature of the advance that built the terminal moraine. The implications of these observations are important in breaking the conventional linkage of individual terminal moraines with climate forcing. 相似文献
A total of 113 samples of waste and soil were collected from a site in the state of San Luis Potosi, Mexico, that was occupied for several years by the metallurgical industry. Specific magnetic susceptibility (MS), electrical conductivity (EC) and pH were determined, as well as the total and available concentrations of potentially toxic elements (PTEs) such as As, Cd, Cu, Fe, Pb and Zn, which may cause a health risk for humans, animals and ecosystems, and the concentrations of major ions in aqueous extracts of soils and wastes. The solid phases of the samples were also characterized. The results revealed that the soils and wastes exhibited elevated values of PTEs, MS and EC. For soils these values decreased with increasing distance from the waste storage sites. The MS values were elevated primarily due to the presence of Fe-oxyhydroxides, such as magnetite, hematite and goethite, which contain PTEs in their structure leading to a high correlation between the value of MS and the As, Cd, Fe and Pb contents (r = 0.57–0.91) as well as between the PTEs values (r = 0.68–0.92). The elevated EC values measured in the metallurgical wastes were the result of presence of the sulfate minerals of Ca, Mg and Fe. The pollution index, which indicates the levels of simultaneous toxicity from elements such as As, Cd and Pb, was determined, with extreme hazard zones corresponding to areas that exhibit high MS values (0.91 correlation). In conclusion, MS measurements can be used as an indirect indicator to evaluate the PTE contamination in metallurgical areas, and EC measurements can aid in the identification of pollution sources. 相似文献
Coral reefs and other coastal ecosystems such as seagrasses and mangroves are widely recognized to provide protection against
the devastating effects of strong waves associated with tsunamis and storms. The predicted warming climate brings to fore
the role of these ecosystems in providing protection against stronger typhoons that can result in more devastating waves of
greater amplitude. We performed a model simulation of storm generated waves on a Philippine reef, which is located along the
path of tropical storms, i.e., at least 10 typhoons on the average pass through the study site yearly. A model to simulate
wave propagation was developed using Simulating Waves Nearshore (SWAN) and DELFT3D-WAVE computer simulation software. Scenarios
involving local monsoonal wind forcing and storm conditions were simulated. In addition, as climate change may also result
to increased relative sea level, a 0.3 m and 1 m rise in sea level scenarios were also used in the wave model simulations.
Results showed that the extensive reef system in the site helped dissipate wave energy that in turn reduced wave run-up on
land. A significant reduction in wave energy was observed in both climate change, i.e., stronger wind and higher sea level,
and non-climate change scenarios. This present study was conducted in a reef whose coral cover is in excellent condition (i.e.,
50 to 80% coral cover). Estimates of coral reef growth are in the same order of magnitude as estimates of relative sea level
rise based on tide gauge and satellite altimeter data, thus it is possible that the role of reefs in attenuating wave energy
may be maintained if coral reef growth can keep up with the change in sea level. Nonetheless, to maintain reef growth, it
is imperative to manage coral reef ecosystems sustainably and to eliminate the stressors that are within human control. Minimizing
activities such as illegal and destructive blast and poison fishing methods, pollution and siltation, is crucial to minimize
the impacts of high-energy waves that may increase with climate change. 相似文献