Integration

Marine Bird Diversity

The upper-left figure depicts interpolated marine bird diversity throughout the study region. The top 20% of predicted diversity is bounded by a thin black line. Because bird diverstiy was significantly correlated with survey effort, we have also provided a mask (cross hatched area) indicating where residual estimates provided evidence that diversity was lower than expected given the amount of effort spent there (rediduals were among the lowest third). Interpret with caution in this area, as the expression of high diversity under the mask may actually be an artifact of high sampling effort.

The upper-right figure depicts interpolated bird density. Again, the top 20% of this estimate is bounded by a thin black line. No statistical relationship was found between density and effort, therefore no residual mask is provided for this model. Figures a and b have both been clipped using the standard error estimate for the interpolated surfaces (see data inventory). This was done to avoid unsupported extrapolation into poorly sampled areas.

The lower figure depicts the overlap between the top 20th percentile for diversity and density.

INTEGRATION BIRD DIVERSITY STANDARD ERROR

This figure depicts the standard error of interpolated bird diversity throughout the study region. This standard error map represents the interpolation (kriging) error and was used to clip the bird diversity interpolations so that only those areas that fell in the region of lowest 20% standard error (depicted in dark blue) are used for subsequent display and analysis. This technique provides a non-arbitrary method of restricting the analysis extent to those areas that are relatively well sampled.

INTEGRATION BIRD DIVERSITY RESIDUAL

This figure depicts the interpolated residual bird diversity throughout the study region. The residuals represent the variation in bird diversity that is not explained by the linear relationship between diversity and the natural log of survey effort. Residual maps are used to correct for differences in the distribution of survey effort and are necessary because of the correlation between diversity and effort.

INTEGRATION BIRD DENSITY STANDARD ERROR

This figure depicts the standard error of interpolated bird density throughout the study region. This standard error map represents the interpolation (kriging) error and was used to clip the bird density interpolations so that only those areas that fell in the region of lowest 20% standard error (depicted in dark blue) are used for subsequent display and analysis. This technique provides a non-arbitrary method of restricting the analysis extent to those areas that are relatively well sampled.