How Accurate Are Population Predictions? Wind Farms and Egyptian Vultures as a Case Study

Abstract

It is clear that scientists' predictions must be rigorous and based on scientific evidence, but, even more, it is crucial to review scientific predictions after a reasonable time. However, predictions of published PVAs have rarely been contrasted with real populations' trends over time. This is worrisome because this is the only way to keep learning and improve our ability to make more accurate predictions. In addition, conservation efforts can shift the initial predictions for the viability of threatened populations; thus, the evaluation of initial predictions becomes required over time. This is the case of the Egyptian vulture in Spain, where trajectories of real populations over the years differ from large-scale predictions. Its extinction in the Iberian Peninsula-due to mortality in wind farms, among other causes-was predicted by 2020, according to published viability analyses; yet, 14 years after this publication, not only did it not happen, but its national (and European) population remains stable and is even slightly increasing (+2.6%). These differences between predicted and observed trajectories of populations show the limitations of the simulations as a conservation tool and offer the opportunity to evaluate the used PVAs and the shortcomings that affected the assessment of the real trajectory of the species. With only four years of data available to simulate and generate 100-year predictions, along with the distribution and variance of mortality rates by collision in wind farms and an overestimation the number of pairs in risk areas, a clear relationship was assumed between predicted risk according to distances and the actual recorded mortality at wind farms, even though it is known that these are not closely related.

Keywords: Neophron percnopterus; PVA; bird mortality; observed versus predicted population trajectories; prediction of population trajectories; viability of population; wind farm.

Figure 1

Predicted (according to [10]) versus observed trajectories for the Southern population of Egyptian vultures. As we can see, trajectories differ dramatically. In fact, ref. [10] predicted complete extinction of the species from the Southern population (Andalusia) by 2018. Breeding pairs and years are presented in the axis.

Figure 2

Predicted (according [10]) versus observed trajectories for the Iberian Peninsula population of Egyptian vultures. As we can see, trajectories differ dramatically. In fact, ref. [10] predicted complete extinction of the species from the Iberian Peninsula by 2021. On the contrary, this population experienced an increase of 2.6%. Breeding pairs and years are presented in the axis.

Figure 3

Wind farm annual mortality of Egyptian vultures in Andalusian population during 2004–2022. This mortality follows a Poisson distribution (lambda = 0.57, Kolmogorov–Smirnov d = 0.56; X² = 0.535; p = 0.464) and not a normal one (dotted line; Kolmogorov–Smirnov d = 0.35; X² = 105.419; p < 0.001).

Figure 4

Predicted (according to new simulations) versus observed trajectories for Andalusian population of Egyptian vultures. As we can see, using the new values for productivity and mortality, both trajectories do not differ dramatically.