Africa Wild

Evaluating uncertainty in estimates of large rhinoceros populations
Sam M Ferreira1* and Danie Pienaar11Scientific Services, SANParks, Skukuza, 1350, South Africa*corresponding author: sam.ferreira@sanparks.org

Abstract
Estimates of the numbers of living rhinoceroses inform management interventions. Several techniques assist authorities in obtaining estimates. For large populations, authorities use sample-based methods. Estimates for the number of rhinos living in Kruger National Park (Kruger) make use of sample-based block surveys. Critics of this approach allege that the authorities place sample blocks mostly in areas with high rhino numbers and that this, together with correction for various biases, inflates estimates. The critics also claim that the percentage confidence intervals (PCIs) associated with estimates are too large and propose total area counts as an alternative. We assess these criticisms by comparing the results of a sample-based survey with those of a near concurrent total count. We found that sample surveys appeared to focus on areas with higher rhino densities, but rhino movements in and out of the survey area confounded results. Moreover, total counts do not produce reliable estimates when surveyors fail to account for biases inherent to all sampling procedures. Bias corrections used by sample surveys most likely underestimate the number of rhinos that surveyors miss, contrary to the allegations of critics that sample-based techniques inflate population estimates. Even so, estimates that transparently report uncertainties detected a significant decline in white rhinos from 8,968 (95% CI: 8,394–9,564) in 2013 to 4,116 (95% CI: 2,994–5,726) in 2018. The trends in the black rhino population also indicate a decline from 627 (95% CI: 588–666) in 2009 to 291 (95% CI: 151–441) in 2018. Conducting block-based sample surveys for large populations that correct for biases provides useful information for decision makers. Given that South Africa, specifically Kruger makes substantial contributions to continental rhino numbers, reporting to international bodies such as CITES should transparently include the uncertainties associated with population estimates.