SEARCH · Search USGS

Results for “Complexity”

Search indexed USGS publications on groundwater, aquifers, geologic maps, mineral resources and earthquakes. Explore source records by subject and place.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 253 records · Page 14Linked to original sources

USGS2004

Complex proximal deposition during the Plinian eruptions of 1912 at Novarupta, Alaska

Proximal (<3 km) deposits from episodes II and III of the 60-h-long Novarupta 1912 eruption exhibit a very complex stratigraphy, the result of at least four transport regimes and diverse depositional mechanisms. They contrast with the relatively simple stratigraphy (and inferred emplacement mechanisms) for the previously documented, better known, medial-distal fall deposits and the Valley of Ten Thousand Smokes ignimbrite. The proximal products include alternations and mixtures of both locally and regionally dispersed fall ejecta, and numerous thin complex deposits of pyroclastic density currents (PDCs) with no regional analogs. The locally dispersed component of the fall deposits forms sector-confined wedges of material whose thicknesses halve radially from and concentrically about the vent over distances of 100-300 m (cf. several kilometers for the medial-distal fall deposits). This locally dispersed fall material (and many of the associated PDC deposits) is rich in andesitic and banded pumices and richer in shallow-derived wall-rock lithics in comparison with the coeval medial fall units of almost entirely dacitic composition. There are no marked contrasts in grain size in the near-vent deposits, however, between locally and widely dispersed beds, and all samples of the proximal fall deposits plot as a simple continuation of grain size trends for medial-distal samples. Associated PDC deposits form a spectrum of facies from fines-poor, avalanched beds through thin-bedded, landscape-mantling beds to channelized lobes of pumice-block-rich ignimbrite. The origins of the Novarupta near-vent deposits are considered within a spectrum of four transport regimes: (1) sustained buoyant plume, (2) fountaining with co-current flow, (3) fountaining with counter-current flow, and (4) direct lateral ejection. The Novarupta deposits suggest a model where buoyant, stable, regime-1 plumes characterized most of episodes II and III, but were accompanied by transient and variable partitioning of clasts into the other three regimes. Only one short period of vent blockage and cessation of the Plinian plume occurred, separating episodes II and III, which was followed by a single PDC interpreted as an overpressured "blast" involving direct lateral ejection. In contrast, regimes 2 and 3 were reflected by spasmodic sedimentation from the margins of the jet and perhaps lower plume, which were being strongly affected by short-lived instabilities. These instabilities in turn are inferred to be associated with heterogeneities in the mixture of gas and pyroclasts emerging from the vent. Of the parameters that control explosive eruptive behavior, only such sudden and asymmetrical changes in the particle concentration could operate on time scales sufficiently short to explain the rapid changes in the proximal 1912 products. Springer-Verlag 2003.

Alaska↗

USGS2025

Complex staged emplacement of a basaltic lava: The example of the July 1974 flow of Kīlauea

Basaltic lava flows can be highly destructive. Forecasting the future path and/or behavior of an active lava flow is challenging because topography is often poorly constrained and lava has a complex rheology and emplacement history. Preserved lavas are an important source of information which, combined with observations of active flows, underpins conceptual models of lava flow emplacement. However, the value of preserved lavas is limited because pre-eruptive topography and, thus, syn-eruptive lava flow geometry are usually not known. Here, we use tree-mold data to constrain pre-eruptive topography and syn-eruptive lava flow geometry of the July 1974 flow of Kīlauea (USA). Tree molds, which are formed after advancing lava encloses standing trees, preserve the lava inundation height and the final preserved thickness of lava. We used data from 282 tree molds to reconstruct the temporal and spatial evolution of the ~ 2.1 km-long July 1974 flow. The tree mold dataset yields a detailed dynamic picture of staged emplacement, separated by intervals of ponding. In some ponded areas, flow depth during emplacement (~ 5 m) was twice the preserved thickness of the final lava (2–3 m). Drainage of the ponds led to episodic surges in flow advancement, decoupled from fluctuations in vent discharge rate. We infer that the final breakout occurred after the cessation of fountaining. Such complex emplacement histories may be common for pāhoehoe lavas at Kīlauea and elsewhere in situations where the terrain is of variable slope, and/or where lava is temporarily perched and stored.

Hawaii↗

USGS2020

Landslide monitoring and runout hazard assessment by integrating multi-source remote sensing and numerical models: An application to the Gold Basin landslide complex, northern Washington

The landslide complex at Gold Basin, Washington, has been drawing considerable attention after a catastrophic runout of the nearby landslide at Oso, Washington, in 2014. To evaluate potential threats of the Gold Basin landslide to the campground down the slope, remote sensing and numerical modeling were integrated to monitor recent landslide activity and simulate hypothetical runout scenarios. Bare-earth LiDAR DEM (digital elevation model) differencing, InSAR (Interferometric Synthetic Aperture Radar), and offset tracking of SAR images reveal that localized collapses at the headscarps have been the primary type of landslide activity at Gold Basin from 2005 to 2019, and currently no signs indicative of movement of a large centralized block or a deep-seated main body were detected. The maximum horizontal deformation rate is 5 m/year occurring primarily from headscarp recession of the middle lobe, and the annual landsliding volume of the whole landslide complex averages 1.03 × 10 5 m 3 . From three-dimensional limit equilibrium analysis of generalized terrace structures, the maximum landslide volume is estimated as 2.0 × 10 6 m 3 . Simulations of hypothetical runout scenarios were carried out using the depth-averaged two-phase model D-claw with above-obtained landslide geometry constraints. The simulation results demonstrate that debris flows with volume less than 10 5 m 3 only pose limited threats to the campground, while volumes over 10 6 m 3 could cause severe damages. Consequently, the estimated maximum landslide volume of 2.0 × 10 6 m 3 suggests a potential risk to the campground nearby. Adaption of our methodology could prove useful for evaluating other similar landslides globally for hazards prevention and mitigation.

Washington↗

USGS2011

Improved mapping of National Atmospheric Deposition Program wet-deposition in complex terrain using PRISM-gridded data sets

High-elevation regions in the United States lack detailed atmospheric wet-deposition data. The National Atmospheric Deposition Program/National Trends Network (NADP/NTN) measures and reports precipitation amounts and chemical constituent concentration and deposition data for the United States on annual isopleth maps using inverse distance weighted (IDW) interpolation methods. This interpolation for unsampled areas does not account for topographic influences. Therefore, NADP/NTN isopleth maps lack detail and potentially underestimate wet deposition in high-elevation regions. The NADP/NTN wet-deposition maps may be improved using precipitation grids generated by other networks. The Parameter-elevation Regressions on Independent Slopes Model (PRISM) produces digital grids of precipitation estimates from many precipitation-monitoring networks and incorporates influences of topographical and geographical features. Because NADP/NTN ion concentrations do not vary with elevation as much as precipitation depths, PRISM is used with unadjusted NADP/NTN data in this paper to calculate ion wet deposition in complex terrain to yield more accurate and detailed isopleth deposition maps in complex terrain. PRISM precipitation estimates generally exceed NADP/NTN precipitation estimates for coastal and mountainous regions in the western United States. NADP/NTN precipitation estimates generally exceed PRISM precipitation estimates for leeward mountainous regions in Washington, Oregon, and Nevada, where abrupt changes in precipitation depths induced by topography are not depicted by IDW interpolation. PRISM-based deposition estimates for nitrate can exceed NADP/NTN estimates by more than 100% for mountainous regions in the western United States.

Environmental Monitoring and Assessment↗

USGS2010

An exploratory investigation of polar organic compounds in waters from a lead–zinc mine and mill complex

Surface water samples were collected in 2006 from a lead mine-mill complex in Missouri to investigate possible organic compounds coming from the milling process. Water samples contained relatively high concentrations of dissolved organic carbon (DOC; greater than 20 mg/l) for surface waters but were colorless, implying a lack of naturally occurring aquatic humic or fulvic acids. Samples were extracted by three different types of solid-phase extraction and analyzed by electrospray ionization/mass spectrometry. Because large amounts of xanthate complexation reagents are used in the milling process, techniques were developed to extract and analyze for sodium isopropyl xanthate and sodium ethyl xanthate. Although these xanthate reagents were not found, trace amounts of the degradates, isopropyl xanthyl thiosulfonate and isopropyl xanthyl sulfonate, were found in most locations sampled, including the tailings pond downstream. Dioctyl sulfosuccinate, a surfactant and process filtering aid, was found at concentrations estimated at 350 μg/l at one mill outlet, but not downstream. Release of these organic compounds downstream from lead-zinc mine and milling areas has not previously been reported. A majority of the DOC remains unidentified.

Missouri↗

USGS2012

Suspended-sediment flux and retention in a backwater tidal slough complex near the landward boundary of an estuary

Backwater tidal sloughs are commonly found at the landward boundary of estuaries. The Cache Slough complex is a backwater tidal region within the Upper Sacramento–San Joaquin Delta that includes two features that are relevant for resource managers: (1) relatively high abundance of the endangered fish, delta smelt ( Hypomesus transpacificus ), which prefers turbid water and (2) a recently flooded shallow island, Liberty Island, that is a prototype for habitat restoration. We characterized the turbidity around Liberty Island by measuring suspended-sediment flux at four locations from July 2008 through December 2010. An estuarine turbidity maximum in the backwater Cache Slough complex is created by tidal asymmetry, a limited tidal excursion, and wind-wave resuspension. During the study, there was a net export of sediment, though sediment accumulates within the region from landward tidal transport during the dry season. Sediment is continually resuspended by both wind waves and flood tide currents. The suspended-sediment mass oscillates within the region until winter freshwater flow pulses flush it seaward. The hydrodynamic characteristics within the backwater region such as low freshwater flow during the dry season, flood tide dominance, and a limited tidal excursion favor sediment retention.

California↗

USGS2020

Sediment dynamics of a divergent bay–marsh complex

Bay–marsh systems, composed of an embayment surrounded by fringing marsh incised by tidal channels, are widely distributed coastal environments. External sediment availability, marsh-edge erosion, and sea-level rise acting on such bay–marsh complexes may drive diverse sediment-flux regimes. These factors reinforce the ephemeral and dynamic nature of fringing marshes: material released by marsh-edge erosion becomes part of a bay–marsh exchange that fuels the geomorphic evolution of the coupled system. The dynamics of this sediment exchange determine the balance among seaward export, deposition on the embayment seabed, flux into tidal channels, and import to the marsh platform. In this work, we investigate the sediment dynamics of a transgressive bay–marsh complex and link them to larger-scale considerations of its geomorphic trajectory. Grand Bay, Alabama/Mississippi, is a shallow microtidal embayment surrounded by salt marshes with lateral erosion rates of up to 5 m year −1 . We collected 6 months of oceanographic data at four moorings within Grand Bay and its tidal channels to assess hydrographic conditions and net sediment-flux patterns and augmented the observations with numerical modeling. The observations imply a divergent sedimentary system in which a majority of the suspended sediment is exported seaward, while a smaller fraction is imported landward via tidal channels, assisting in vertical marsh-plain accumulation, maintenance of channel and intertidal-flat morphologies, and landward transgression. These results describe a dynamic system that is responsive to episodic atmospheric forcing in the absence of a strong tidal signal and the presence of severe lateral marsh loss.

Alabama, Mississippi↗

USGS1994

Geochemistry and hydrology of a calcareous fen within the Savage Fen wetlands complex, Minnesota, USA

Savage Fen is a wetlands complex at the base of north-facing bluffs in the Minnesota River Valley. The complex includes 27.8 hectares of calcareous fen that host rare calciphile plants whose populations are declining in Minnesota. Water and sediment compositions in the calcareous fen were studied to gain a better understanding of the hydrologie System that sustains the rare vegetation. Groundwater in the fen is a calcium-magnesium-bicarbonate type with circumneutral pH values. The groundwater composition is the resuit of interactions among water, dissolved and gaseous carbon species, carbonates, and ion exchangers. Shallow groundwater is distinguished from deep groundwater by smaller concentrations of chloride, sulfate, magnesium, and sodium, and larger concentrations of calcium, bicarbonate, hydrogen sulfide, and ammonium. Magnesian calcite is the prevalent carbonate in unconsolidated sedimentary fill beneath the fen and is an important source and sink for dissolved calcium, magnesium, and inorganic carbon. Calcite concentrations just below the water table are small because aerobic and anaerobic oxidation of organic matter increase the partial pressure of carbon dioxide ( P CO 2 ), decrease pH, and cause calcite to dissolve. Thick calcite accumulations just above the water table, in the root zone of calciphile plants, result from water table fluctuations and attendant changes in P CO 2 . Groundwater beneath Savage Fen recharges in lakes and ponds south of the fen and upwells to the surface within the fen. Water at the water table is a mixture of upwelling groundwater and water near the surface that flows downslope from higher elevations in the fen. Changes in oxygen and hydrogen isotope compositions of shallow groundwater indicate that the proportion of upwelling groundwater in shallow groundwater decreases downgradient in the calcareous fen. Encroachment of reed grasses into the calcareous fen may reflect human-caused disturbances in the recharge area.

Minnesota↗

USGS1989

Spectroscopic evidence for organic diacid complexation with dissolved silica in aqueous systems—I. Oxalic acid

Increased solubility of quartz and mobilization in contaminated groundwater due to the complexation with dissolved organic acids has been recently proposed [Bennett and Siegel, Nature 326 , 684–686 (1987)]. Using laser Raman and Fourier transform infrared spectroscopies, we have examined mixed solutions of oxalic and silicic acids at near neutral pH in the tenth molar concentration ranges in an attempt to directly observe the proposed organo-silicate complexes. In both laser Raman and infrared spectra, product bands were observed that indicate an oxalate/silicic acid ester is being formed in the reaction. These data support the observation that organic diacids can lead to enhanced solubility of quartz in hydrogeological systems.

Organic Geochemistry↗

USGS1989

Contrasting zircon morphology and UPb systematics in peralkaline and metaluminous post-orogenic granite complexes of the Arabian Shield, Kingdom of Saudi Arabia

Uzircon ages are reported for seven metaluminous-to-peralkaline post-orogenic granites from the Late Proterozoic Arabian Shield of Saudi Arabia. Zircons from the metaluminous rocks are prismatic, with length-to-width ratios of ∼ 2–4: 1 and small pyramidal terminations. In contrast, zircons from three of the four peralkaline complexes either lack well-developed prismatic faces (are pseudo-octahedral) or are anhedral. Some zircons from the peralkaline granites contain inherited radiogenic Pb and have very high common Pb contents ( 206 Pb/ 204 Pb < 150), making the UPb method poorly suited for determining the age of these rocks. Zircons in the metaluminous granites do not contain inheritance and yield well-defined concordia intercepts. The span of ages of the seven complexes (670-470 Ma) indicates that post-orogenic granitic magmatism was not a singular event in the Arabian Shield but rather occurred as multiple intrusive episodes from the Late Proterozoic to the Middle Ordovician.

Chemical Geology: Isotope Geoscience Section↗

USGS1985

The copper-nickel concentration log: A tool for stratigraphic interpretation within the ultramafic and basal zones of the stillwater complex, Montana

An analogue to the electric well log was devised for copper-nickel concentration drill-hole data from the Basal and lower part of the Ultramafic zones of the Stillwater Complex using automated data processing. The copper-nickel concentration logs graphically represent intensity (concentration) values that reflect the distribution of the elements in sulfide and silicate minerals. Four major patterns are recognized by their characteristic variations in copper and nickel intensity: (1) relatively flat, low-level copper-intensity signatures associated with arcuate nickel-intensity patterns that correlate with rocks in the Peridotite member of the Ultramafic zone; (2) arcuate or bulb-like patterns of copper and nickel intensity that correlate closely with the Basal bronzite cumulate member of the Basal zone; (3) complex patterns consisting of intervals of low-intensity copper and moderate-intensity nickel, spikes of high nickel and copper intensity, and high copper intensity associated with low nickel intensity that correlate respectively with cordierite-pyroxene hornfels, massive sulfide, norites and mineralized diabase dikes in the Basal norite member; and (4) large intervals of extremely low copper and nickel intensity that correlate with quartz-orthopyroxene hornfels. The recognition and interpretation of these patterns allow two- and three-dimensional stratigraphic and lithologic reconstructions to be done by means of concentration-log correlations instead of variable quality lithologic logging.

Journal of Geochemical Exploration↗

USGS2010

Using a composite grid approach in a complex coastal domain to estimate estuarine residence time

We investigate the processes that influence residence time in a partially mixed estuary using a three-dimensional circulation model. The complex geometry of the study region is not optimal for a structured grid model and so we developed a new method of grid connectivity. This involves a novel approach that allows an unlimited number of individual grids to be combined in an efficient manner to produce a composite grid. We then implemented this new method into the numerical Regional Ocean Modeling System (ROMS) and developed a composite grid of the Hudson River estuary region to investigate the residence time of a passive tracer. Results show that the residence time is a strong function of the time of release (spring vs. neap tide), the along-channel location, and the initial vertical placement. During neap tides there is a maximum in residence time near the bottom of the estuary at the mid-salt intrusion length. During spring tides the residence time is primarily a function of along-channel location and does not exhibit a strong vertical variability. This model study of residence time illustrates the utility of the grid connectivity method for circulation and dispersion studies in regions of complex geometry.

Computers & Geosciences↗

USGS2024

A new method for bioassessment of ecosystems with complex communities and environmental gradients

Bioassessment of complex and heterogeneous ecosystems is a challenge when there are multiple, strong, natural environmental gradients; unknown, or spatially varying, mixtures of stressors; and large numbers of taxa with unknown responses to both the environmental gradients and the stressors. Current methods of bioassessment are not designed for use under this set of constraints. To address this gap, we have developed an assessment method appropriate for well-sampled, heterogeneous systems with many taxa. In the bioassessment described below, we model taxa occurrence as a function of natural environmental gradients, then use residual covariance patterns between all pairs of taxa to estimate the impact of human disturbance across sites as a latent construct. The derivation of the method from an underlying causal model allows the metric value at each site and the associated taxa responses to be partitioned into contributions from a set of putative stressors. We apply this method as a case study to the subtidal benthic invertebrate community of Puget Sound, WA (USA) and demonstrate a partial decomposition of the metric values to a set of stressors including sediment organic carbon, nitrogen, metals, and organic pollutants. While this method provides new opportunities to estimate, communicate, and understand the ecological condition of complex, heterogeneous ecosystems, due to the requirement for broad, detailed data to inform its estimates, it will likely be most appropriate for monitoring programs.

Washington↗

USGS2005

Influence of topographic complexity on solar insolation estimates for the Colorado River, Grand Canyon, AZ

Rugged topography along the Colorado River in Glen and Grand Canyons, exemplifies features common to canyon-bound streams and rivers of the arid southwest. Physical relief influences regulated river systems, especially those that are altered, and have become partially reliant on aquatic primary production. We measured and modeled instantaneous solar flux in a topographically complex environment to determine where differences in daily, seasonal and annual solar insolation occurred in this river system. At a system-wide scale, topographic complexity generates a spatial and temporal mosaic of varying solar insolation. This solar variation is a predictable consequence of channel orientation, geomorphology, elevation angles and viewshed. Modeled estimates for clear conditions corresponded closely with observed measurements for both instantaneous photosynthetic photon flux density (PPFD: ??mol m-2 s-1) and daily insolation levels (relative error 2.3%, CI ??0.45, S.D. 0.3, n = 29,813). Mean annual daily insolation levels system-wide were estimated to be 36 mol m-2 d -1 (17.5 S.D.), and seasonally varied on average from 13.4-57.4 mol m-2 d-1, for winter and summer, respectively. In comparison to identical areas lacking topographic effect (idealized plane), mean daily insolation levels were reduced by 22% during summer, and as much as 53% during winter. Depending on outlying topography, canyon bound regions having east-west (EW) orientations had higher seasonal variation, averaging from 8.1 to 61.4 mol m-2 d-1, for winter and summer, respectively. For EW orientations, 70% of mid-channel sites were obscured from direct incidence during part of the year; and of these sites, average diffuse light conditions persisted for 19.3% of the year (70.5 days), and extended upwards to 194 days. This predictive model has provided an initial quantitative step to estimate and determine the importance of autotrophic production for this ecosystem, as well as a broader application for other canyon systems. ?? 2004 Published by Elsevier B.V.

Ecological Modelling↗

USGS2004

Approaches to surface complexation modeling of Uranium(VI) adsorption on aquifer sediments

Uranium(VI) adsorption onto aquifer sediments was studied in batch experiments as a function of pH and U(VI) and dissolved carbonate concentrations in artificial groundwater solutions. The sediments were collected from an alluvial aquifer at a location upgradient of contamination from a former uranium mill operation at Naturita, Colorado (USA). The ranges of aqueous chemical conditions used in the U(VI) adsorption experiments (pH 6.9 to 7.9; U(VI) concentration 2.5 · 10 −8 to 1 · 10 −5 M; partial pressure of carbon dioxide gas 0.05 to 6.8%) were based on the spatial variation in chemical conditions observed in 1999–2000 in the Naturita alluvial aquifer. The major minerals in the sediments were quartz, feldspars, and calcite, with minor amounts of magnetite and clay minerals. Quartz grains commonly exhibited coatings that were greater than 10 nm in thickness and composed of an illite-smectite clay with occluded ferrihydrite and goethite nanoparticles. Chemical extractions of quartz grains removed from the sediments were used to estimate the masses of iron and aluminum present in the coatings. Various surface complexation modeling approaches were compared in terms of the ability to describe the U(VI) experimental data and the data requirements for model application to the sediments. Published models for U(VI) adsorption on reference minerals were applied to predict U(VI) adsorption based on assumptions about the sediment surface composition and physical properties (e.g., surface area and electrical double layer). Predictions from these models were highly variable, with results overpredicting or underpredicting the experimental data, depending on the assumptions used to apply the model. Although the models for reference minerals are supported by detailed experimental studies (and in ideal cases, surface spectroscopy), the results suggest that errors are caused in applying the models directly to the sediments by uncertain knowledge of: 1) the proportion and types of surface functional groups available for adsorption in the surface coatings; 2) the electric field at the mineral-water interface; and 3) surface reactions of major ions in the aqueous phase, such as Ca 2+ , Mg 2+ , HCO 3 − , SO 4 2− , H 4 SiO 4 , and organic acids. In contrast, a semi-empirical surface complexation modeling approach can be used to describe the U(VI) experimental data more precisely as a function of aqueous chemical conditions. This approach is useful as a tool to describe the variation in U(VI) retardation as a function of chemical conditions in field-scale reactive transport simulations, and the approach can be used at other field sites. However, the semi-empirical approach is limited by the site-specific nature of the model parameters.

Geochimica et Cosmochimica Acta↗

USGS2006

Uranyl adsorption and surface speciation at the imogolite-water interface: Self-consistent spectroscopic and surface complexation models

Macro- and molecular-scale knowledge of uranyl (U(VI)) partitioning reactions with soil/sediment mineral components is important in predicting U(VI) transport processes in the vadose zone and aquifers. In this study, U(VI) reactivity and surface speciation on a poorly crystalline aluminosilicate mineral, synthetic imogolite, were investigated using batch adsorption experiments, X-ray absorption spectroscopy (XAS), and surface complexation modeling. U(VI) uptake on imogolite surfaces was greatest at pH ∼7–8 ( I = 0.1 M NaNO 3 solution, suspension density = 0.4 g/L [U(VI)] i = 0.01–30 μM, equilibration with air). Uranyl uptake decreased with increasing sodium nitrate concentration in the range from 0.02 to 0.5 M. XAS analyses show that two U(VI) inner-sphere (bidentate mononuclear coordination on outer-wall aluminol groups) and one outer-sphere surface species are present on the imogolite surface, and the distribution of the surface species is pH dependent. At pH 8.8, bis-carbonato inner-sphere and tris-carbonato outer-sphere surface species are present. At pH 7, bis- and non-carbonato inner-sphere surface species co-exist, and the fraction of bis-carbonato species increases slightly with increasing I (0.1–0.5 M). At pH 5.3, U(VI) non-carbonato bidentate mononuclear surface species predominate (69%). A triple layer surface complexation model was developed with surface species that are consistent with the XAS analyses and macroscopic adsorption data. The proton stoichiometry of surface reactions was determined from both the pH dependence of U(VI) adsorption data in pH regions of surface species predominance and from bond-valence calculations. The bis-carbonato species required a distribution of surface charge between the surface and β charge planes in order to be consistent with both the spectroscopic and macroscopic adsorption data. This research indicates that U(VI)-carbonato ternary species on poorly crystalline aluminosilicate mineral surfaces may be important in controlling U(VI) mobility in low-temperature geochemical environments over a wide pH range (∼5–9), even at the partial pressure of carbon dioxide of ambient air ( p CO2 = 10 −3.45 atm).

Geochimica et Cosmochimica Acta↗

USGS2017

Geochemical evidence for a complex origin for the Kelso dunes, Mojave National Preserve, California USA

The Kelso Dune field in southern California is intriguing because although it is of limited areal extent (~ 100 km 2 ), it has a wide variety of dune forms and contains many active dunes (~ 40 km 2 ), which is unusual in the Mojave Desert. Studies over the past eight decades have concluded that the dunes are derived primarily from a single source, Mojave River alluvium, under a dominant, westerly-to-northwesterly wind regime. The majority of these studies did not, however, present data to support the Mojave River as the only source. We conducted mineralogical and geochemical studies of most of the 14 geomorphically defined dune groups of the Kelso Dune field as well as potential sand sources, alluvial sediments from the surrounding mountain ranges. Results indicate that sands in the nine western dune groups have K/Rb and K/Ba (primarily from K-feldspar) compositions that are indistinguishable from Mojave River alluvium (westerly/northwesterly winds) and Budweiser Wash alluvium (southwesterly winds), permitting an interpretation of two sources. In contrast, sands from the five eastern dune groups have K/Rb and K/Ba values that indicate significant inputs from alluvial fan deposits of the Providence Mountains. This requires either rare winds from the east or southeast or, more likely, aeolian reworking of distal Providence Mountain fan sediments by winds from the west, at a rate greater than input from the Mojave River or other western sources. The results indicate that even a small dune field can have a complex origin, either from seasonally varying winds or complex alluvial-fan-dune interaction. Application of K/Rb and K/Ba in K-feldspar as a provenance indicator could be used in many of the world's ergs or sand seas, where dune origins are still not well understood or are controversial. Four examples are given from Africa and the Middle East where such an approach could yield useful new information about dune sand provenance.

Geomorphology↗

USGS2023

When less is more: How increasing the complexity of machine learning strategies for geothermal energy assessments may not lead toward better estimates

Previous moderate- and high-temperature geothermal resource assessments of the western United States utilized data-driven methods and expert decisions to estimate resource favorability. Although expert decisions can add confidence to the modeling process by ensuring reasonable models are employed, expert decisions also introduce human and, thereby, model bias. This bias can present a source of error that reduces the predictive performance of the models and confidence in the resulting resource estimates. Our study aims to develop robust data-driven methods with the goals of reducing bias and improving predictive ability. We present and compare nine favorability maps for geothermal resources in the western United States using data from the U.S. Geological Survey's 2008 geothermal resource assessment. Two favorability maps are created using the expert decision-dependent methods from the 2008 assessment ( i.e., weight-of-evidence and logistic regression). With the same data, we then create six different favorability maps using logistic regression (without underlying expert decisions), XGBoost, and support-vector machines paired with two training strategies. The training strategies are customized to address the inherent challenges of applying machine learning to the geothermal training data, which have no negative examples and severe class imbalance. We also create another favorability map using an artificial neural network. We demonstrate that modern machine learning approaches can improve upon systems built with expert decisions. We also find that XGBoost, a non-linear algorithm, produces greater agreement with the 2008 results than linear logistic regression without expert decisions, because the expert decisions in the 2008 assessment rendered the otherwise linear approaches non-linear despite the fact that the 2008 assessment used only linear methods. The F1 scores for all approaches appear low (F1 score < 0.10), do not improve with increasing model complexity, and, therefore, indicate the fundamental limitations of the input features ( i.e., training data). Until improved feature data are incorporated into the assessment process, simple non-linear algorithms ( e.g., XGBoost) perform equally well or better than more complex methods ( e.g., artificial neural networks) and remain easier to interpret.

Geothermics↗