Line-driven disc wind models with an improved line force

We describe an efficient method of calculating the radiation pressure resulting from spectral lines, including all the terms in the velocity gradient tensor. We apply this method to calculate the two-dimensional, time-dependent structure of winds from luminous discs. Qualitative features of our new models are very similar to those we calculated including only the dominant terms in the tensor. In particular, we find that models which displayed unsteady behaviour in our earlier paper are also unsteady with the new method, and gross properties of the winds, such as mass-loss rate and characteristic velocity, are not changed by the more accurate approach. The largest change caused by the new method is in the disc-wind opening angle: winds driven only by the disc radiation are more polar with the new method, whilst winds driven by the disc and central object radiation are typically more equatorial. In the closing discussion, we provide further insight into the way the geometry of the radiation field and consequent flow determine the time properties of the flow.     

A spatially resolved 'inside-out' outburst of IP Pegasi

We present a comprehensive photometric data set taken over the entire outburst of the eclipsing dwarf nova IP Peg in 1997 September/October. Analysis of the light curves taken over the long rise to the peak-of-outburst shows conclusively that the outburst started near the centre of the disc and moved outwards. This is the first data set that spatially resolves such an outburst. The data set is consistent with the idea that long rise times are indicative of such 'inside-out' outbursts. We show how the thickness and the radius of the disc, along with the mass transfer rate, change over the whole outburst. In addition, we show evidence of the secondary and the irradiation thereof. We discuss the possibility of spiral shocks in the disc; however, we find no conclusive evidence of their existence in this data set.     

Measurements of the vertical correlation in turbulence under broken waves

Turbulence measurements have been carried out in the surf zone of a wave flume. The purpose of the measurements is to determine the length scale of the turbulence generated by the wave breaking. The length scale of the turbulence is estimated on basis of the correlation between simultaneous measurements of the vertical turbulent fluctuations, taken at different levels above the bed.     

Analysis of Thomsen parameters for finely layered VTI media

Since the work of Postma and Backus, much has been learned about elastic constants in vertical transversely isotropic (VTI) media when the anisotropy is due to fine layering of isotropic elastic materials. Nevertheless, there has continued to be some uncertainty about the possible range of Thomsen's anisotropy parameters ε and δ for such media. We use both Monte Carlo studies and detailed analysis of Backus' equations for both two- and three-component layered media to establish the results presented. We show that ε lies in the range ?3/8 ε ½[〈v²_p〉〈v^?2_p〉?1], for finely layered media; smaller positive and all negative values of ε occur for media with large fluctuations in the Lamé parameter λ in the component layers. We show that δ can also be either positive or negative, and that for constant density media, sign (δ) = sign (〈v^?2_p〉 ? 〈v^?2_s〉〈v²_s/v²_p〉). Monte Carlo simulations show that among all theoretically possible random media, positive and negative δ are equally likely in finely layered media. (Of course, the δs associated with real earth materials may span some smaller subset of those that are theoretically possible, but answering this important question is beyond our present scope.) Layered media having large fluctuations in λ are those most likely to have positive δ. This is somewhat surprising since ε is often negative or a small positive number for such media, and we have the general constraint that ε ? δ > 0 for layered VTI media. Since Gassmann's results for fluid-saturated porous media show that the mechanical effects of fluids influence only the Lamé parameter λ, not the shear modulus μ, these results suggest that small positive δ occurring together with small positive ε (but somewhat larger than δ) may be indicative of changing fluid content in a layered earth.     

Automated GPS Data Analysis Service

Automatic analysis of geodetic-quality GPS data is available with the use of e-mail and ftp (file transfer program) as an interface to a computer at the Jet Propulsion Laboratory (JPL), where precise transmitter parameters – GPS ephemerides and clock errors – are computed regularly. The interface is such that e-mail from an external user causes the JPL computer to fetch the user's data. The computer than analyzes the data, and places the results in an area accessible to the user. An e-mail to the user gives information on the location of the analysis results, which the user can subsequently fetch. Operations on the JPL computer are entirely automatic, and require essentially no labor. ? 1999 John Wiley & Sons, Inc.     

Improving neural network performance on the classification of complex geographic datasets

Neural Networks are now established computational tools used for search minimisation and data classification. They offer some highly desirable features for landuse classification problems since they are able to take in a variety of data types, recorded on different statistical scales, and combine them. As such, neural networks should offer advantages of increased accuracy. However, a barrier to their general acceptance and use by all but `experts' is the difficulty of configuring the network initially.  This paper describes the architectural problems of applying neural networks to landcover classification exercises in geography and details some of the latest developments from an ongoing research project aimed at overcoming these problems. A comprehensive strategy for the configuration of neural networks is presented, whereby the network is automatically constructed by a process involving initial analysis of the training data. By careful study of the functioning of each part of the network it is possible to select the architecture and initial weights on the node connections so the constructed network is `right first time'. Further adaptations are described to control network behaviour, to optimise functioning from the perspective of landcover classification. The entire configuration process is encapsulated by a single application which may be treated by the user as a `black box', allowing the network to the applied in much the same way as a maximum likelihood classifier, with no further effort being required of the user.