Limitations of Majority Agreement in Crowdsourced Image Interpretation

Crowdsourcing can efficiently complete tasks that are difficult to automate, but the quality of crowdsourced data is tricky to evaluate. Algorithms to grade volunteer work often assume that all tasks are similarly difficult, an assumption that is frequently false. We use a cropland identification game with over 2,600 participants and 165,000 unique tasks to investigate how best to evaluate the difficulty of crowdsourced tasks and to what extent this is possible based on volunteer responses alone. Inter‐volunteer agreement exceeded 90% for about 80% of the images and was negatively correlated with volunteer‐expressed uncertainty about image classification. A total of 343 relatively difficult images were independently classified as cropland, non‐cropland or impossible by two experts. The experts disagreed weakly (one said impossible while the other rated as cropland or non‐cropland) on 27% of the images, but disagreed strongly (cropland vs. non‐cropland) on only 7%. Inter‐volunteer disagreement increased significantly with inter‐expert disagreement. While volunteers agreed with expert classifications for most images, over 20% would have been mis‐categorized if only the volunteers’ majority vote was used. We end with a series of recommendations for managing the challenges posed by heterogeneous tasks in crowdsourcing campaigns.     

A Parallel Implementation of the Lattice Solid Model for the Simulation of Rock Mechanics and Earthquake Dynamics

— The Lattice Solid Model has been used successfully as a virtual laboratory to simulate fracturing of rocks, the dynamics of faults, earthquakes and gouge processes. However, results from those simulations show that in order to make the next step towards more realistic experiments it will be necessary to use models containing a significantly larger number of particles than current models. Thus, those simulations will require a greatly increased amount of computational resources. Whereas the computing power provided by single processors can be expected to increase according to “Moore’s law,” i.e., to double every 18–24 months, parallel computers can provide significantly larger computing power today. In order to make this computing power available for the simulation of the microphysics of earthquakes, a parallel version of the Lattice Solid Model has been implemented. Benchmarks using large models with several millions of particles have shown that the parallel implementation of the Lattice Solid Model can achieve a high parallel-efficiency of about 80% for large numbers of processors on different computer architectures.     

Glacier Retreat, Mass-Balance and Thinning: Sermilik Glacier, South Greenland

This study documents thinning and retreat of the South Greenland ice margin and discusses possible reasons in the light of mass‐balance and change of dynamic conditions. Analyses of satellite images have shown that the glacier tongue of Sermilik glacier disintegrated within the past 15 years. Furthermore, the observed thinning close to the Sermilik glacier front was as much as 120 m water equivalent during this period. This figure was derived by comparing surface elevation data from a digital elevation model (1985) and laser altimeter measurements from the year 2000, showing surface elevation changes along a flow line of Sermilik glacier. Mass‐balance data from in situ measurements performed at a centre flow line of the glacier are presented. These data are compared to results from remote sensing analyses of the study area. Net ablation reconstruction over the last 41 years from positive‐degree‐day modelling, at various locations along the Sermilik glacier massbalance transect, shows an increase during the past decades. These analyses indicate that only 55% of the total thinning in this area can be explained by mass‐balance changes. The remaining 45% of the thinning is attributed to changes in the dynamic behaviour of the glacier, such as an increase of creep towards the end of the twentieth century. The significant thinning along the Qagssimiut lobe can also be explained as a combination of mass‐balance changes and changes in ice dynamic behaviour.     

A matter of tolerance: Distribution potential of scyphozoan polyps in a changing environment

Jellyfish blooms are unpredictable, unsustainable events, frequently affecting aquatic ecosystems severely. Of particular interest are the consequences of environmental change for jellyfish populations, especially in semi‐enclosed habitats. Regional and seasonal changes in water chemistry and physics may control the distribution of sessile polyps in the Baltic Sea, hence potentially driving the population dynamics of the two abundant medusa species Aurelia aurita and Cyanea capillata (Scyphozoa, Cnidaria). In laboratory experiments, settlement, growth, survival and physiological condition of A. aurita polyps were investigated at different levels of water temperature, pH and salinity. Survival and physiological condition of C. capillata polyps were examined after exposure to low salinity levels. Increased settlement of A. aurita planula larvae was observed on substrate plates at low temperature (4°C), low pH (7.4) and low salinity (7.5 psu), whereas early polyp growth was constrained by salinity ≤10 psu. Aurelia aurita polyps were in good physiological condition over the whole temperature range, while exposure to pH <6.5 led to stepwise tissue degradation. Salinity reduction to ≤5 and ≤8 psu caused irreversible degeneration of A. aurita and C. capillata polyps, respectively. Observed physiological limits suggest distribution of polyp populations of A. aurita in central and of C. capillata in western parts of the Baltic Sea, while future climate changes may particularly restrict occurrence of the less tolerant C. capillata.     

Groundwater flow dynamics in the complex aquifer system of Gidabo River Basin (Ethiopian Rift): a multi-proxy approach

Hydrochemical and isotope data in conjunction with hydraulic head and spring discharge observations were used to characterize the regional groundwater flow dynamics and the role of the tectonic setting in the Gidabo River Basin, Ethiopian Rift. Both groundwater levels and hydrochemical and isotopic data indicate groundwater flow from the major recharge area in the highland and escarpment into deep rift floor aquifers, suggesting a deep regional flow system can be distinguished from the shallow local aquifers. The δ¹⁸O and δ²H values of deep thermal (≥30 °C) groundwater are depleted relative to the shallow (<60 m below ground level) groundwater in the rift floor. Based on the δ¹⁸O values, the thermal groundwater is found to be recharged in the highland around 2,600 m a.s.l. and on average mixed with a proportion of 30 % shallow groundwater. While most groundwater samples display diluted solutions, δ¹³C data of dissolved inorganic carbon reveal that locally the thermal groundwater near fault zones is loaded with mantle CO₂, which enhances silicate weathering and leads to anomalously high total dissolved solids (2,000–2,320 mg/l) and fluoride concentrations (6–15 mg/l) exceeding the recommended guideline value. The faults are generally found to act as complex conduit leaky barrier systems favoring vertical mixing processes. Normal faults dipping to the west appear to facilitate movement of groundwater into deeper aquifers and towards the rift floor, whereas those dipping to the east tend to act as leaky barriers perpendicular to the fault but enable preferential flow parallel to the fault plane.     

Effects of different boundary conditions on the simulation of groundwater flow in a multi-layered coastal aquifer system (Taranto Gulf,southern Italy)

The evaluation of the accuracy or reasonableness of numerical models of groundwater flow is a complex task, due to the uncertainties in hydrodynamic properties and boundary conditions and the scarcity of good-quality field data. To assess model reliability, different calibration techniques are joined to evaluate the effects of different kinds of boundary conditions on the groundwater flow in a coastal multi-layered aquifer in southern Italy. In particular, both direct and indirect approaches for inverse modeling were joined through the calibration of one of the most uncertain parameters, namely the hydraulic conductivity of the karst deep hydrostratigraphic unit. The methodology proposed here, and applied to a real case study, confirmed that the selection of boundary conditions is among the most critical and difficult aspects of the characterization of a groundwater system for conceptual analysis or numerical simulation. The practical tests conducted in this study show that incorrect specification of boundary conditions prevents an acceptable match between the model response to the hydraulic stresses and the behavior of the natural system. Such effects have a negative impact on the applicability of numerical modeling to simulate groundwater dynamics in complex hydrogeological situations. This is particularly important for management of the aquifer system investigated in this work, which represents the only available freshwater resource of the study area, and is threatened by overexploitation and saltwater intrusion.