Skip to main content
. 2014 Jul 22;281(1787):20132935. doi: 10.1098/rspb.2013.2935

Figure 5.

Figure 5.

Predicted median false-positive (a,c) and false-negative (b) rates based on simulations that assumed the estimated magnitude of autocorrelated process noise and observation error for each of the 627 time series of population abundance from the global population dynamics database; note the different scales. Data are plotted over a range of observation windows (number of years of data) and thresholds for classification as being at risk of extinction (x-axis, fractional decline in abundance over observation-window years). The predicted classification errors are based on simulations assuming either density-dependent (a) or density-independent (b,c) dynamics.