Fig 1.
The foraging range for (A) 11 juvenile and (B) 55 adult shy albatrosses from the Albatross Island population.
White dots show the locations returned from satellite transmitter (PTT) tracking devices and the colour scale shows percentage utilization calculated using the kernel density method. A 1 degree grid is imposed; black outlines indicate grid cells that have been excluded from the calculation because they substantially cover land. A red dot marks the location of Albatross Island.
Fig 2.
Annual totals for fishing effort in millions of hooks or thousands of trawl operations for pelagic longline, demersal longline, and trawl fishing within the area modelled.
Table 1.
Fisheries included in the model, showing the super-fleets they have been grouped into and their total effort within the bird foraging area.
Table 2.
Biological data available from long-term monitoring studies on Albatross Island (taken from [14]).
Table 3.
Probability of first breeding at given age (unpublished data) expressed as a percentage.
Fig 3.
Decadal averages for (A) rainfall and (B) number of days over 23°C at Albatross Island during the chick rearing season for the historical period (red), and hind- and forecasts from climate models using the B1 (black lines) or A2 (grey lines) scenarios.
Six variants are available for each of the B1 and A2 scenarios.
Fig 4.
The estimated relationships between a multiplier on chick survival and (A) rainfall, (B) number of days with a maximum temperature over 23°C, and (C) SSHA.
The histograms reflect frequency distribution of rainfall and temperature from the B1 (black bars) and A2 (grey bars) climate models. Vertical lines show the range of the historical measurements. Note that SSHA forecasts are not available, but an observation for 2011 (the model ends with 2011) was used. The functional forms shown use different values for a power parameter b which is chosen to be 0.5, 1 or 2.
Table 4.
Negative log likelihoods (-lnL), estimated adult natural mortality rates (M) and productivity (the breeding success rate for the pristine stock) with standard errors (SE) for the models that use each of the three alternative values for b, the power parameter for the relationship that relates environmental variables to chick mortality.
Fig 5.
Observed (circles) and expected (lines) annual time series of (A) the number of breeding pairs in the Albatross island population; (B) apparent juvenile survival rate to 5 years old (note that apparent survival for recent years is low because young birds have not all had time to mature); (C) adult survival rate (the grey line marks the maximum value possible: 100%); and (D) annual breeding success (the grey line shows the expected breeding success for the model that ignores environmental factors).
Fig 6.
Annual observed minus expected breeding success plotted against the standardized environmental covariate values: (A) rainfall (R), (B) number of days over 23°C (T), and (C) SSHA (S) for the model that allows environmental factors to influence chick mortality (‘RTS’); and (D,E,F) the model that ignores environmental factors (‘xxx’).
The values for 1999 and 2006 are shown as open circles because these were not used in the model, having been years of notable pox infestation.
Table 5.
Negative log-likelihood values (-lnL) for each of the models considered, the improvement over the null model (xxx) is shown (2*Diff) along with the significance (p) of the inclusion of n additional parameters calculated using the likelihood ratio test.
Fig 7.
For every year and all 12 future climate scenarios (6 models each using a B1 or an A2 scenario), the estimated breeding success from the three models that used b values of 0.5, 1 or 2, are plotted against those from the b = 1 model.
Results are shown for future projections that (A) use future rainfall but keep future temperature at its historical mean, (B) keep rainfall at its historical mean but use future temperature, (C) use both future rainfall and temperature.
Fig 8.
Projected number of shy albatross breeding pairs on Albatross Island in year 2100 if rainfall (“Rain”) and temperature (“TT”) are drawn from a normal distribution with historical mean and variance (“Both zero”) or if either or both are taken from the six B1 and A2 climate projections.
Standard box and whisker plots are shown with the actual projected population sizes overlaid (grey circles). Horizontal black lines show the median population size in the climate neutral scenario. Results are shown for the models that use (A) b = 0.5; (B) b = 1; and (C) b = 2.
Fig 9.
Projected number of shy albatross breeding pairs on Albatross Island in year 2100 based on rainfall and temperature time series from the six (A) B1 or (B) A2 climate projections.
Incidental capture by the trawl fishery is assumed to remain the same (i.e. no mitigation: “0% mit”) or to be stopped entirely (100% mitigation: “100% mit”) and values in between. Alternatively, future SSHA values are assumed to be between 100% (“100% SSHA”) and 300% (“300% SSHA”) greater than their historic mean. Box and whisker plots are shown with the projected population sizes from which they are constructed, overlaid (grey circles). These are shown because box and whisker plots based on such a small number of data points can be misleading. Horizontal black lines show the median population size in the climate neutral scenario.