Calibrating a neural network‐based urban change model for two metropolitan areas of the Upper Midwest of the United States

We parameterized neural net‐based models for the Detroit and Twin Cities metropolitan areas in the US and attempted to test whether they were transferable across both metropolitan areas. Three different types of models were developed. First, we trained and tested the neural nets within each region and compared them against observed change. Second, we used the training weights from one area and applied them to the other. Third, we selected a small subset (～1%) of the Twin Cities area where a lot of urban change occurred. Four model performance metrics are reported: (1) Kappa; (2) the scale which correct and paired omission/commission errors exceed 50%; (3) landscape pattern metrics; and (4) percentage of cells in agreement between model simulations. We found that the neural net model in most cases performed well on pattern but not location using Kappa. The model performed well only in one case where the neural net weights from one area were used to simulate the other. We suggest that landscape metrics are good to judge model performance of land use change models but that Kappa might not be reliable for situations where a small percentage of urban areas change.     

Hydrochemical constraints between the karst Tabular Middle Atlas Causses and the Saïs basin (Morocco): implications of groundwater circulation

The karst Tabular Middle Atlas Causses reservoir is the main drinking-water supply of Fez-Meknes region (Saïs Basin) in Morocco. Recent analyses showed a decline in associated groundwater chemical quality and increased turbidity. To understand this hydrosystem, four surveys were undertaken during fall and spring, 2009–2011. Hydrogeochemical studies coupled with isotopic analyses (δ¹⁸O, δD and ²²²Rn) showed that the aquifers between the causses (mountains) and the Saïs Basin are of Liassic origin and at the southern extremities are of Triassic origin. Five recharge zones of different altitudes have been defined, including two main mixing zones in the south. Deuterium excess results suggest local recharge, while a plot of δ¹⁸O versus δD characterizes a confined aquifer in the eastern sector. ²²²Rn results reveal areas of rapid exchanges with an upwelling time of less than 2 weeks. A schematic conceptual model is presented to explain the groundwater circulation system and the behavior of this karst system.     

Seismic intensity measures for above‐ground liquid storage tanks

A series of scalar and vector intensity measures is examined to determine their suitability within the seismic risk assessment of liquid storage tanks. Using a surrogate modelling approach on a squat tank that is examined under both anchored and unanchored support conditions, incremental dynamic analysis is adopted to generate the distributions of response parameters conditioned on each of the candidate intensity measures. Efficiency and sufficiency metrics are used in order to perform the intensity measure evaluation for individual failure modes, while a comparison in terms of mean annual frequency of exceedance is performed with respect to a damage state that is mutually governed by the impulsive and convective modes of the tank. The results reveal combinations of spectral acceleration ordinates as adequate predictors, among which the average spectral acceleration is singled out as the optimal solution. The sole exception is found for the sloshing‐controlled modes of failure, where mainly the convective period spectral acceleration is deemed adequate to represent the associated response due to their underlying linear relationship. A computationally efficient method in terms of site hazard analysis is finally proposed to serve in place of the vector‐valued intensity measures, providing a good match for the unanchored tank considered and a more conservative one for the corresponding anchored system.     

Regional ground‐motion prediction equations for amplitude‐, frequency response‐, and duration‐based parameters for Greece

Although all of the main properties of a ground motion cannot be captured through a single parameter, a number of different engineering parameters has been proposed that are able to reflect either one or more ground‐motion characteristics concurrently. For many of these parameters, especially regarding Greece, there are relatively few or no predictive models. In this context, we present a set of new regionally‐calibrated equations for the prediction of the geometric mean of the horizontal components of 10 amplitude‐, frequency response‐, and duration‐based parameters for shallow crustal earthquakes. These equations supersede previous empirical relationships for Greece since their applicability range for magnitude, and epicentral distance has been extended down to M_w 4 and up to 200 km, respectively, the incorporation of a term accounting for anelastic attenuation has been investigated, while their development was based on a ground‐motion dataset spanning from 1973 to 2014. For all ground‐motion parameters, we provide alternative optimal equations relative to the availability of information on the different explanatory variables. In all velocity‐based and contrary to the acceleration‐based parameters, the anelastic attenuation coefficient was found statistically insignificant when it was combined with the geometric decay and the coefficient accounting for saturation with distance. In the regressions where the geometric decay coefficient simultaneously incorporated the contribution of anelastic attenuation, its increase was found to be much less considerable in the velocity‐based than in the acceleration‐based parameters, implying a stronger effect of anelastic attenuation on the parameters that are defined via the acceleration time history.     

Linking sardine spawning dynamics to environmental variability

The present paper investigates the effect of environmental variability on the spawning dynamics of sardine, Sardina pilchardus, in the Eastern Mediterranean. To do so, variability in four different reproductive traits, i.e. reproductive activity, incidence of recent spawning, prevalence of early ovarian atresia and relative batch fecundity was tested along a gradient of habitat conditions like water temperature, chlorophyll a (Chl a) concentration and an index of zooplankton production. Apart from atretic prevalence all remaining traits were significantly and positively correlated with Chl a concentration which indicates that sardine in the E. Mediterranean tends to enhance egg production under conditions of increased primary productivity. On the other hand, prevalence of atresia together with spawning incidence were positively affected by water temperature whilst the only factor that correlated significantly with the index of zooplankton production was relative fecundity. These findings were interpreted as an indication of direct energy flow from phytoplankton filter-feeding to gonadal development and egg production, implying that besides capitalized energy sardine also uses current income for supporting reproduction.     

Local and regional sources of fine and coarse particulate matter based on traffic and background monitoring

The aim of this study was to identify local and exogenous sources affecting particulate matter (PM) levels in five major cities of Northern Europe namely: London, Paris, Hamburg, Copenhagen and Stockholm. Besides local emissions, PM profile at urban and suburban areas of the European Union (EU) is also influenced by regional PM sources due to atmospheric transport, thus geographical city distribution is of a great importance. At each city, PM₁₀, PM_2.5, NO₂, SO₂, CO and O₃ air pollution data from two air pollution monitoring stations of the EU network were used. Different background characteristics of the selected two sampling sites at each city facilitated comparisons, providing a more exact analysis of PM sources. Four source apportionment methods: Pearson correlations among the levels of particulates and gaseous pollutants, characterisation of primal component analysis components, long-range transport analysis and extrapolation of PM size distribution ratios were applied. In general, fine (PM_2.5) and coarse (PM₁₀) particles were highly correlated, thus common sources are suggested. Combustion-originated gaseous pollutants (CO, NO₂, SO₂) were strongly associated to PM₁₀ and PM_2.5, primarily at areas severely affected by traffic. On the contrary, at background stations neighbouring important natural sources of particles or situated in suburban areas with rural background, natural emissions of aerosols were indicated. Series of daily PM_2.5/PM₁₀ ratios showed that minimum fraction values were detected during warm periods, due to higher volumes of airborne biogenic PM coarse, mainly at stations with important natural sources of particles in their vicinity. Hybrid single-particle Lagrangian integrated trajectory model was used, in order to extract 4-day backward air mass trajectories that arrived in the five cities which are under study during days with recorded PM₁₀ exceedances. At all five cities, a significantly large fraction of those trajectories were classified in short- and medium-range clusters, thus transportation of particulates along with slow moving air masses was identified. A finding that supports the assumption of long-range transport is that, at background stations, long-range transportation effects were stronger, in comparison to traffic stations, due to less local particle emissions. Short-range trajectories associated to PM transport in Stockholm, Copenhagen and Hamburg were mainly of a continental origin. All three cities were approached by slow moving air masses originated from Poland and the Czech Republic, whereas Copenhagen and Stockholm were also influenced by short-range trajectories from Germany and France and from Jutland Peninsula and Scandinavian Peninsula, respectively. London and Paris are located to the north-west part of Europe. Trajectories of short and medium length arrived to these two megacities mainly through France, Germany, UK and North Atlantic.