High-precision TIMS U-series and AMS C-14 dating of a coral reef lagoon sediment core from southern South China Sea

Accurate dating of lagoon sediments has been a difficult problem, although lagoon profiles, usually with high deposition rates, have a great potential for high-resolution climate reconstruction. We report 26 high-precision TIMS U-series dates (on 25 coral branches) and five AMS C-14 dates (on foraminifera) for a 15.4-m long lagoon core from Yongshu Reef, Nansha area, southern South China Sea. All the dates are in the correct stratigraphical sequence, providing the best chronology so far reported for lagoon deposits. The results reveal a similar to 4000-a continuous depositional history, with sedimentation rates varying from 0.8 to 24.6 mm a(-1), with an average of 3.85 mm a(-1), which corresponds to an average net carbonate accumulation rate of similar to 2700 g CaCO3 m(-2) a(-1), significantly higher than the mean value (800 +/- 400 g CaCO3 m(-2) a(-1)) used for lagoons in general in previous studies of global carbonate budget. Episodes of accelerated depositions within the last 1000 years correlate well with strong storm events identified by U-series dates of storm-transported coral blocks in the area. However, in the longer term, the sedimentation rates during the past 1000 years were much higher than earlier on, probably due to more vigorous wave-reef interaction as a result of relative sea-level fall since 500 AD and expansion of reef flat area, supplying more sediments. The coral TIMS U-series ages and foraminifera AMS 14C dates reveal intriguing apparent radiocarbon reservoir ages (R) from 572 to 1052 years, which are much higher than global mean values of similar to 400 years. (c) 2006 Elsevier Ltd. All rights reserved.

Precision-recall operating characteristic (P-ROC) curves in imprecise environments

Traditionally, machine learning algorithms have been evaluated in applications where assumptions can be reliably made about class priors and/or misclassification costs. In this paper, we consider the case of imprecise environments, where little may be known about these factors and they may well vary significantly when the system is applied. Specifically, the use of precision-recall analysis is investigated and compared to the more well known performance measures such as error-rate and the receiver operating characteristic (ROC). We argue that while ROC analysis is invariant to variations in class priors, this invariance in fact hides an important factor of the evaluation in imprecise environments. Therefore, we develop a generalised precision-recall analysis methodology in which variation due to prior class probabilities is incorporated into a multi-way analysis of variance (ANOVA). The increased sensitivity and reliability of this approach is demonstrated in a remote sensing application.

Active node supporting context-aware vertical handover in pervasive computing environment with redundant positioning

A major requirement for pervasive systems is to integrate context-awareness to support heterogeneous networks and device technologies and at the same time support application adaptations to suit user activities. However, current infrastructures for pervasive systems are based on centralized architectures which are focused on context support for service adaptations in response to changes in the computing environment or user mobility. In this paper, we propose a hierarchical architecture based on active nodes, which maximizes the computational capabilities of various nodes within the pervasive computing environment, while efficiently gathering and evaluating context information from the user's working environment. The migratable active node architecture employs various decision making processes for evaluating a rich set of context information in order to dynamically allocate active nodes in the working environment, perform application adaptations and predict user mobility. The active node also utilizes the Redundant Positioning System to accurately manage user's mobility. This paper demonstrates the active node capabilities through context-aware vertical handover applications.