The main objective of this paper is to provide researchers that investigate fossil phytolith assemblages and model/data comparisons a new tool for estimating C3/C4 grass composition over time. We tested the reliability of modern soil phytolith assemblages and phytolith indices for tracing the dominance of different grass subfamilies and tree cover density. We analyzed modern soil phytolith assemblages from sites over elevation gradients on Mount Kenya (Kenya), Mount Rungwe and around Lake Masoko (southern Tanzania). These data were compared with available botanical data. A phytolith index named Ic, proved to be an effective proxy of the proportions of Pooideae, Arundinoideae and Bambusoideae grasses (mainly C3 grasses) versus Panicoideae grasses (mainly C4 grasses), increasing with elevation in East-Africa. When tropical mountains are covered by open habitats (e.g. grasses and shrublands), Ic should be a reliable proxy of the C3/C4 grass composition. These results highlight the value of the phytolith index Ic, when interpreting paleo-environmental records from tropical mountains, to: 1) better understand past local and regional C3/C4 grass distributions and associated climatic changes and 2) increase the set of C3/C4 data available for model/data comparisons. 相似文献
We investigate trajectory forecasting as an application of ocean circulation ensemble modeling. The ensemble simulations are performed weekly, starting with assimilation of data for various variables from multiple sensors on a range of observational platforms. The ensemble is constructed from 100 members, and member no. 1 is designed as a standard (deterministic) simulation, providing us with a benchmark for the study. We demonstrate the value of the ensemble approach by validating simulated trajectories using data from ocean surface drifting buoys. We find that the ensemble average trajectories are generally closer to the observed trajectories than the corresponding results from a deterministic forecast. We also investigate an alternative model in which velocity perturbations are added to the deterministic results and ensemble mean results, by a first-order stochastic process. The parameters of the stochastic model are tuned to match the dispersion of the ensemble approach. Search areas from the stochastic model give a higher hit ratio of the observations than the results based on the ensemble. However, we find that this is a consequence of a positive skew of the area distribution of the convex hulls of the ensemble trajectory end points. 相似文献
Structural uncertainty exists when associating sparse fault interpretations made from two-dimensional seismic lines or limited outcrop observations. Here, a graph formalism is proposed that describes the problem of associating spatial fault evidence. A combinatorial analysis, relying on this formalism, shows that the number of association scenarios is given by the Bell number, and increases exponentially with the number of pieces of evidence. As a result, the complete exploration of uncertainties is computationally highly challenging. The available prior geological knowledge is expressed by numerical rules to reduce the number of scenarios, and the graph formalism makes structural interpretation easier to reproduce than manual interpretation. The Bron–Kerbosch algorithm, which finds maximal cliques in undirected graphs, is used to detect major possible structures. This framework opens the way to a numerically assisted exploration of uncertainties during structural interpretation.
On 8–9 September 2002, an extreme rainfall event caused by a stationary mesoscale convective system (MCS) occurred in the Gard region, France. Distributed hydrologic and hydraulic modelling has been carried out to assess and compare the various sources of data collected operationally and during the post-event field surveys. Distributed hydrological modelling was performed with n-TOPMODELs and assessed for ungauged basins with the discharge estimates of the post-event surveys. A careful examination of the occurrence in time and space of the flash floods over the head watersheds indicates that flooding was controlled by the trajectory of the convective part of the MCS. Stationarity of the MCS over the Gardon watershed (1858 km2 at Remoulins) for 28 h was responsible for the exceptional magnitude of the flood at this scale. The flood dynamics were characterized by an extensive inundation of the Gardonnenque plain upstream of the Gardon Gorges resulting in a significant peak flow reduction downstream. One-dimensional unsteady-flow hydraulic modelling was found to be required to reproduce these dynamics. Hydraulic modelling also proved to be potentially useful for the critical analysis and extrapolation of operational discharge rating curves. 相似文献
Over the last two decades, models of the Earth’s magnetospheric magnetic field have been continuously improved to describe
more precisely the different magnetospheric current systems (magnetopause current, symmetric and partial ring currents, tail
currents and field aligned currents). In this paper we compare the different Tsyganenko models and the Alexeev and Feldstein
model in the context of cosmic ray physics. We compare the vertical cutoff rigidity and asymptotic direction of vertical incidence
obtained with these models for the January 20, 2005, ground level enhancement and for the big magnetic storm of April 6, 2000.
For the event of January 20, 2005, we study the impact of the differences in asymptotic direction obtained with the models
on the radiation dose computation at aircraft altitude. For the magnetic storm of April 6, 2000, we discuss the importance
of the different magnetospheric current systems in causing cutoff rigidity variations. Finally we summarise the advantages
and drawbacks of the different models in the context of space weather. 相似文献