BackgroundGlobally, coral bleaching has been responsible for a significant decline in both coral cover and diversity over the past two decades. During the summer of 2010–11, anomalous large-scale ocean warming induced unprecedented levels of coral bleaching accompanied by substantial storminess across more than 12° of latitude and 1200 kilometers of coastline in Western Australia (WA).Methodology/Principal FindingsExtreme La-Niña conditions caused extensive warming of waters and drove considerable storminess and cyclonic activity across WA from October 2010 to May 2011. Satellite-derived sea surface temperature measurements recorded anomalies of up to 5°C above long-term averages. Benthic surveys quantified the extent of bleaching at 10 locations across four regions from tropical to temperate waters. Bleaching was recorded in all locations across regions and ranged between 17% (±5.5) in the temperate Perth region, to 95% (±3.5) in the Exmouth Gulf of the tropical Ningaloo region. Coincident with high levels of bleaching, three cyclones passed in close proximity to study locations around the time of peak temperatures. Follow-up surveys revealed spatial heterogeneity in coral cover change with four of ten locations recording significant loss of coral cover. Relative decreases ranged between 22%–83.9% of total coral cover, with the greatest losses in the Exmouth Gulf.Conclusions/SignificanceThe anomalous thermal stress of 2010–11 induced mass bleaching of corals along central and southern WA coral reefs. Significant coral bleaching was observed at multiple locations across the tropical-temperate divide spanning more than 1200 km of coastline. Resultant spatially patchy loss of coral cover under widespread and high levels of bleaching and cyclonic activity, suggests a degree of resilience for WA coral communities. However, the spatial extent of bleaching casts some doubt over hypotheses suggesting that future impacts to coral reefs under forecast warming regimes may in part be mitigated by southern thermal refugia.

We introduce TextWorld, a sandbox learning environment for the training and evaluation of RL agents on text-based games. TextWorld is a Python library that handles interactive playthrough of text games, as well as backend functions like state tracking and reward assignment. It comes with a curated list of games whose features and challenges we have analyzed. More significantly, it enables users to handcraft or automatically generate new games. Its generative mechanisms give precise control over the difficulty, scope, and language of constructed games, and can be used to relax challenges inherent to commercial text games like partial observability and sparse rewards. By generating sets of varied but similar games, TextWorld can also be used to study generalization and transfer learning. We cast text-based games in the Reinforcement Learning formalism, use our framework to develop a set of benchmark games, and evaluate several baseline agents on this set and the curated list.

No abstract

No abstract

Vertebrate animals localize sounds by comparing differences in the acoustic signal between the two ears and, accordingly, ear structures such as the otoliths of fishes are expected to develop symmetrically. Sound recently emerged as a leading candidate cue for reef fish larvae navigating from open waters back to the reef. Clearly, the integrity of the auditory organ has a direct bearing on what and how fish larvae hear. Yet, the link between otolith symmetry and effective navigation has never been investigated in fishes. We tested whether otolith asymmetry influenced the ability of returning larvae to detect and successfully recruit to favourable reef habitats. Our results suggest that larvae with asymmetrical otoliths not only encountered greater difficulties in detecting suitable settlement habitats, but may also suffer significantly higher rates of mortality. Further, we found that otolith asymmetries arising early in the embryonic stage were not corrected by any compensational growth mechanism during the larval stage. Because these errors persist and phenotypic selection penalizes asymmetrical individuals, asymmetry is likely to play an important role in shaping wild fish populations.

Monitoring changes in coral cover and composition through space and time can provide insights to reef health and assist the focus of management and conservation efforts. We used a meta-analytical approach to assess coral cover data across latitudes 10–35°S along the west Australian coast, including 25 years of data from the Ningaloo region. Current estimates of coral cover ranged between 3 and 44% in coral habitats. Coral communities in the northern regions were dominated by corals from the families Acroporidae and Poritidae, which became less common at higher latitudes. At Ningaloo Reef coral cover has remained relatively stable through time (∼28%), although north-eastern and southern areas have experienced significant declines in overall cover. These declines are likely related to periodic disturbances such as cyclones and thermal anomalies, which were particularly noticeable around 1998/1999 and 2010/2011. Linear mixed effects models (LME) suggest latitude explains 10% of the deviance in coral cover through time at Ningaloo. Acroporidae has decreased in abundance relative to other common families at Ningaloo in the south, which might be related to persistence of more thermally and mechanically tolerant families. We identify regions where quantitative time-series data on coral cover and composition are lacking, particularly in north-western Australia. Standardising routine monitoring methods used by management and research agencies at these, and other locations, would allow a more robust assessment of coral condition and a better basis for conservation of coral reefs.

Anthropologists have become increasingly aware of the importance of population as a factor in a systematic view of human biological and cultural development. This awareness has generated an interest in the field of demography, and consequently, techniques once utilized almost exclusively by demographers are now frequently utilized for anthropological studies. Anthropological-paleodemographic inquiry traditionally starts with the excavation of a skeletal population sample. The sample is aged and sexed, and the data are put into a descriptive analytic model–the life table. The life table, through a process of inference, is taken to represent the life processes of a local biological population and often forms the basis for further inference on the relationships between populational and cultural processes (Green and others 1974; Howell-Lee 1971).Critics have questioned the assumption of life table construction and cited various sources of error in data collection to argue against the use of life tables as a source of inference concerning the biological population. In this paper we will attempt to address some of these sources of error. Specifically, we will discuss the effects of enumeration errors, population growth, and small population size on life table values. To assess the impact of these errors on life table values of “anthropological” populations, we make use of computer simulation. We conclude that, once the implications of these factors are understood, the life table can provide a useful model for paleodemographic research.

No abstract