Fig 1.
Overview of the tītī population model.
The 15 age classes are represented by circles and intermediate calculations (chicks not harvested, and number of adults) by rectangles. The multipliers used to determine the number of individuals progressing to the next age-class are shown in red. The notation is as in the text, except that the dependence on year is suppressed for simplicity. Likewise, age classes 4 to 12 are suppressed for ease of presentation.
Table 1.
Overview of how the data inform the tītī population model.
Table 2.
Description of alternatives to the best model (S-1F0) that were fitted in the sensitivity analysis.
Table 3.
ΔW values and WAIC model weights for models with different relationships between adult survival and SOI, and fecundity and SOI.
Fig 2.
Fit of the demographic models to harvest data.
Observed and predicted catch per unit effort (CPUE) data from harvesting of tītī chicks in New Zealand during the period 1979–1998. In both plots the solid line shows the model-averaged posterior mean for CPUE. In (a) the dashed lines show the 95% model-averaged credible interval; in (b) the dots are the observed CPUE values and the dashed lines show the 95% model-averaged 95% prediction interval. Model averaging was performed using WAIC model weights.
Fig 3.
Southern Oscillation Index versus adult survival and fecundity of tītī.
Estimated relationship, based on the best model (S–1F0), between SOI and both adult survival rate and fecundity rate, for tītī in New Zealand during the period 1976–2005, plotted against the observed values of and
respectively (dots), together with 95% credible intervals (dashed).
Fig 4.
Effects of climate, harvesting, bycatch, and weka-depredation on tītī population dynamics.
Model-averaged posterior mean of population size of tītī in the New Zealand region for the period 1976–2005 (black), together with the model-averaged posterior mean of the predicted population size in the absence of harvest (blue), absence of bycatch (orange), absence of weka (green), and absence of a relationship between SOI and adult survival (red). Model averaging was performed using WAIC model weights.
Fig 5.
Effects of including process error in the best-fitting model on the relationship between SOI and both adult survival and fecundity.
Estimated relationships (solid) between a) adult survival from the best model (S–1F0) and ; b) adult survival from the modified best model and
; c) fecundity from the best model (S–1F0) and
; d) fecundity the modified best model and
. The dashed lines are 95% credible intervals.
Fig 6.
Effects of including process error in the best-fitting model on the estimation of CPUE and the effects of climate, harvesting, bycatch, and weka-depredation.
Posterior means (solid) and 95% credible intervals (dashed) for a) CPUE from the best model (S–1F0) and b) CPUE from the modified best model, plus posterior means of population size (black), predicted population size in the absence of harvest (blue), the absence of bycatch (orange), the absence of weka (green), and the absence of a relationship between SOI and adult survival (red) c) from the best model (S–1F0) and d) from the modified best model.
Fig 7.
Sensitivity analysis for the best-fitting demographic model.
Comparison of estimates, between models with different input parameter values, of a) the coefficient of the relationship between fecundity and SOI, b) the coefficient of the relationship between adult survival and SOI, c) the predicted population size in 2005 relative to the predicted population size in 2005 in the absence of harvesting, d) the predicted population size in 2005 relative to the predicted population size in 2005 in the absence of fisheries bycatch, e) the predicted population size in 2005 relative to the predicted population size in 2005 in the absence of weka depredation, and f) the predicted population size in 2005 relative to the predicted population size in 2005 in the absence of harvesting, bycatch and weka depredation. Dots are posterior means and error bars show the central 95% credible intervals.