Assessing quality of volunteer crowdsourcing contributions: lessons from the Cropland Capture game

Volunteered geographic information (VGI) is the assembly of spatial information based on public input. While VGI has proliferated in recent years, assessing the quality of volunteer-contributed data has proven challenging, leading some to question the efficiency of such programs. In this paper, we compare several quality metrics for individual volunteers’ contributions. The data were the product of the ‘Cropland Capture’ game, in which several thousand volunteers assessed 165,000 images for the presence of cropland over the course of 6 months. We compared agreement between volunteer ratings and an image's majority classification with volunteer self-agreement on repeated images and expert evaluations. We also examined the impact of experience and learning on performance. Volunteer self-agreement was nearly always higher than agreement with majority classifications, and much greater than agreement with expert validations although these metrics were all positively correlated. Volunteer quality showed a broad trend toward improvement with experience, but the highest accuracies were achieved by a handful of moderately active contributors, not the most active volunteers. Our results emphasize the importance of a universal set of expert-validated tasks as a gold standard for evaluating VGI quality.     

Preprocessing of gravity gradients at the GOCE high-level processing facility

One of the products derived from the gravity field and steady-state ocean circulation explorer (GOCE) observations are the gravity gradients. These gravity gradients are provided in the gradiometer reference frame (GRF) and are calibrated in-flight using satellite shaking and star sensor data. To use these gravity gradients for application in Earth scienes and gravity field analysis, additional preprocessing needs to be done, including corrections for temporal gravity field signals to isolate the static gravity field part, screening for outliers, calibration by comparison with existing external gravity field information and error assessment. The temporal gravity gradient corrections consist of tidal and nontidal corrections. These are all generally below the gravity gradient error level, which is predicted to show a 1/f behaviour for low frequencies. In the outlier detection, the 1/f error is compensated for by subtracting a local median from the data, while the data error is assessed using the median absolute deviation. The local median acts as a high-pass filter and it is robust as is the median absolute deviation. Three different methods have been implemented for the calibration of the gravity gradients. All three methods use a high-pass filter to compensate for the 1/f gravity gradient error. The baseline method uses state-of-the-art global gravity field models and the most accurate results are obtained if star sensor misalignments are estimated along with the calibration parameters. A second calibration method uses GOCE GPS data to estimate a low-degree gravity field model as well as gravity gradient scale factors. Both methods allow to estimate gravity gradient scale factors down to the 10⁻³ level. The third calibration method uses high accurate terrestrial gravity data in selected regions to validate the gravity gradient scale factors, focussing on the measurement band. Gravity gradient scale factors may be estimated down to the 10⁻² level with this method.     

Relict silicate inclusions in extraterrestrial chromite and their use in the classification of fossil chondritic material

Chromite is the only common meteoritic mineral surviving long-term exposure on Earth, however, the present study of relict chromite from numerous Ordovician (470 Ma) fossil meteorites and micrometeorites from Sweden, reveals that when encapsulated in chromite, other minerals can survive for hundreds of millions of years maintaining their primary composition. The most common minerals identified, in the form of small (<1-10 μm) anhedral inclusions, are olivine and pyroxene. In addition, sporadic merrillite and plagioclase were found.Analyses of recent meteorites, holding both inclusions in chromite and corresponding matrix minerals, show that for olivine and pyroxene inclusions, sub-solidus re-equilibration between inclusion and host chromite during entrapment has led to an increase in chromium in the former. In the case of olivine, the re-equilibration has also affected the fayalite (Fa) content, lowering it with an average of 14% in inclusions. For Ca-poor pyroxene the ferrosilite (Fs) content is more or less identical in inclusions and matrix. By these studies an analogue to the commonly applied classification system for ordinary chondritic matrix, based on Fa in olivine and Fs in Ca-poor pyroxene, can be established also for inclusions in chromite. All olivine and Ca-poor pyroxene inclusions (>1.5 μm) in chromite from the Ordovician fossil chondritic material plot within the L-chondrite field, which is in accordance with previous classifications. The concordance in classification together with the fact that inclusions are relatively common makes them an accurate and useful tool in the classification of extraterrestrial material that lacks matrix silicates, such as fossil meteorites and sediment-dispersed chromite grains originating primarily from decomposed micrometeorites but also from larger impacts.     

Ice-core pollen record of climatic changes in the central Andes during the last 400 yr

This paper presents a high-resolution ice-core pollen record from the Sajama Ice Cap, Bolivia, that spans the last 400 yr. The pollen record corroborates the oxygen isotopic and ice accumulation records from the Quelccaya Ice Cap and supports the scenario that the Little Ice Age (LIA) consisted of two distinct phases—a wet period from AD 1500 to 1700, and a dry period from AD 1700 to 1880. During the dry period xerophytic shrubs expanded to replace puna grasses on the Altiplano, as suggested by a dramatic drop in the Poaceae/Asteraceae (P/A) pollen ratio. The environment around Sajama was probably similar to the desert-like shrublands of the Southern Bolivian Highlands and western Andean slopes today. The striking similarity between the Sajama and Quelccaya proxy records suggests that climatic changes during the Little Ice Age occurred synchronously across the Altiplano.     

Late Pleistocene baldcypress (Taxodium distichum) forest deposit on the continental shelf of the northern Gulf of Mexico

Approximately 13 km south of Gulf Shores, Alabama (United States), divers found in situ baldcypress (Taxodium distichum) stumps 18 m below the ocean surface. These trees could have only lived when sea level fell during the Pleistocene subaerially exposing the tectonically stable continental shelf. Here we investigate the geophysical properties along with microfossil and stratigraphical analyses of sediment cores to understand the factors that lead to this wood’s preservation. The stumps are exposed in an elongated depression (~100 m long, ~1 m deep) nested in a trough of the northwest–southeast trending Holocene sand ridges and troughs with 2–5 m vertical relief and ~0.5 km wavelength. Radiocarbon ages of the wood were infinite thus optically stimulated luminescence (OSL) dating was used to constrain the site’s age. Below the Holocene sands (~0.1–4 m thick), separated by a regional erosional unconformity, are Late Pleistocene mud-peat (72±8 ka OSL), mud-sand (63±5, 73±6 ka OSL), and palaeosol (56±5 ka OSL) facies that grade laterally from west to east, respectively. Foraminiferal analysis reveals the location of the terrestrial-marine transitional layer above the Pleistocene facies in an interbedded sand and mud facies (3940±30 (1σ) ¹⁴C a BP), which is part of a lower shoreface or marine-dominated estuarine environment. The occurrence of palaeosol and swamp facies of broadly similar ages and elevation suggests the glacial landscape possessed topographic relief that allowed wood, mud and peats to be preserved for ~50 ka of subaerial exposure before transitioning to the modern marine environment. We hypothesize that rapid sea-level rise occurring ~60 or ~40 ka ago provided opportunities for local flood-plain aggradation to bury the swamp thus preserving the stumps and that other sites may exist in the northern Gulf of Mexico shelf.     

The stability of a remediated bed in Hamilton Harbour,Lake Ontario,Canada

ABSTRACT In situ measurements of lakebed sediment erodibility were made on three sites in Hamilton Harbour, Lake Ontario, using the benthic flume Sea Carousel. Three methods of estimating the surface erosion threshold (τ_c(0)) from a Carousel time series were evaluated: the first method fits measures of bed strength to eroded depth (the failure envelope) and evaluates threshold as the surface intercept; the second method regresses mean erosion rate (E_m) with bed shear stress and solves for the floc erosion rate (E_f) to derive the threshold for E_m = E_f = 1 × 10^?5 kg m^?2 s^?1; the third method extrapolates a regression of suspended sediment concentration (S) and fluid transmitted bed shear stress (τ₀) to ambient concentrations. The first field site was undisturbed (C) and acted as a control; the second (W) was disturbed through ploughing and water injection as part of lakebed treatment, whereas the third site (OIP) was disturbed and injected with an oxidant used for remediation of contaminated sediment. The main objectives of this study were: (1) to evaluate the three different methods of deriving erosion threshold; (2) to compare the physical behaviour of lacustrine sediments with their marine estuarine counterparts; and (3) to examine the effects of ploughing and chemical treatment of contaminated sediment on bed stability. Five deployments of Sea Carousel were carried out at the control site. Mean erosion thresholds for the three methods were: τ_c(0) = 0·5 (±0·06), 0·27 (±0·01) and 0·34 (±0·03) Pa respectively. Method 1 overpredicted bed strength as it was insensitive to effects in the surface 1–2 mm, and the fit of the failure envelope was also highly subjective. Method 2 exhibited a wide scatter in the data (low correlation coefficients), and definition of the baseline erosion rate (E_f) is largely arbitrary in the literature. Method 3 yielded stable (high correlation coefficients), reproducible and objective results and is thus recommended for evaluation of the erosion threshold. The results of this method correlated well with sediment bulk density and followed the same trend as marine counterparts from widely varying sites. Mass settling rates, expressed as a decay constant, k, of S(t), were strongly related to the maximum turbidity at the onset of settling (S_max) and were also in continuity with marine counterparts. Thus, it appears that differences in salinity had little effect on mass settling rates in the examples presented, and that biological activity dominated any effects normally attributable to changes in salinity. Bedload transport of eroded aggregates (2–4 mm in diameter) took place by rolling below a mean tangential flow velocity (U_y) of 0·32 ms^?1 and by saltation at higher velocities. Mass transport as bedload was a maximum at U_y = 0·4 ms^?1, although bedload never exceeded 1% of the suspended load. The proportion of material moving as bedload was greatest at the onset of erosion but decreased as flow competence increased. Given the low bulk density and strength of the lakebed sediment, the presence of a bedload component is notable. Bedload transport over eroding cohesive substrates should be greater in estuaries, where both sediment density and strength are usually higher. Significant differences between the ploughed and control sites were apparent in both the erosion rate and the friction coefficient (φ), and suggest that bed recovery after disruption is rapid (< 24 h). τ_c(0) increased linearly with time after ploughing and recovered to the control mean value within 3 days. The friction coefficient was reduced to zero by ploughing (diagnostic of fluidization), but increased linearly with time, regaining control values within 6 days. No long‐term reduction in bed strength due to remediation was apparent.     

Petrogenesis of the Greenhills Complex,Southland, New Zealand: magmatic differentiation and cumulate formation at the roots of a Permian island-arc volcano

A Permian (~265 Ma) intrusive complex which formed as a magmatic feeder reservoir to an immature island-arc volcano is fortuitously exposed in southern New Zealand. Known as the Greenhills Complex, this intrusion was emplaced at shallow crustal levels and consists of two layered bodies which were later intruded by a variety of dykes. Cumulates, which include dunite, olivine clinopyroxenite, olivine gabbro, and hornblende gabbro-norite, are related products of parent-magma fractionation. Both primary (magmatic) and secondary platinum-group minerals occur within dunite at one locality. Using the composition of cumulus minerals, mafic dykes and melt inclusions, we have determined that the parent magmas of the complex were hydrous, low-K island-arc tholeiites of ankaramitic affinities. Progressive magmatic differentiation of this parent magma generated fractionated melt of high-alumina basalt composition which is now preserved only as dykes which cut the Complex. Field evidence and cumulus mineral profiles reveal that the magma chambers experienced turbulent magmatic conditions during cumulate-rock formation. Recharge of the chambers by primitive magma is likely to have coincided with eruption of residual melt at the surface. Similar processes are inferred to account for volcanic-rock compositions in other parts of this arc terrane and in modern island-arc systems.     

Electrical resistivity imaging of hydrologic flow through surface coal mine valley fills with comparison to other landforms

Surface coal mining has altered land cover, near‐surface geologic structure, and hydrologic processes of large areas in central Appalachia, USA. These alterations are associated with changes in water quality such as elevated total‐dissolved solids, which is usually measured via its surrogate, specific conductance (SC). The SC of valley fill effluent streams is a function of fill construction methods, materials, and age; yet hydrologic studies that relate these variables to water quality are sparse due to the difficulty of conducting traditional hydrologic studies in mined landscapes. We used electrical resistivity imaging (ERI) to visualize the subsurface geologic structure and hydrologic flow paths within a valley fill. ERI is a noninvasive geophysical technique that maps spatiotemporal changes in resistivity of the subsurface. We paired ERI with artificial rainfall experiments to track infiltrated water as it moved through the valley fill. Results indicate that ERI can be used to identify subsurface geologic structure and track advancing wetting fronts or preferential flow paths. Our results suggest that the upper portion of the fill contains significant fines, whereas the deeper profile is primarily large rocks and void spaces. Water tended to pond on the surface of compacted areas until it reached preferential flow paths, where it appeared to infiltrate quickly down to >15 m depth in 75 min. ERI applications can improve understanding of how fill construction techniques influence subsurface water movement, and in turn may aid in the development of valley fill construction methods to reduce water quality effects.