Determining Extinction for Small Cryptic Species: The Morro Bay Kangaroo Rat

Abstract

One third of missing mammal species thought to be extinct have been rediscovered. Determining extinction correctly, without misinterpreting negative evidence, is difficult and takes significant effort, especially for small, cryptic species. The Morro Bay kangaroo rat (MBKR), Dipodomys heermanni morroensis, is a small nocturnal rodent suspected of being extinct. This is because it has not been seen since 1986 despite three range-wide surveys conducted between 1995 and 2012, a recent scent-detecting dog survey, and over a dozen localized surveys. Causes of decline have been reported and the primary causes are thought to be development-driven habitat loss and ecological succession. Given this, we suspect that if the MBKR is extant, then it occurs in the periphery of its historical range. We summarize a survey of the Morro Bay sandspit, an area not previously considered part of MBKR’s range but that has the potential to be occupied. Inferences from the subspecies’ closest relative, Dipodomys heermanni arenae, were used to inform surveys and detection probability estimates for MBKR. Visual surveys of the sandspit in areas with the greatest probability of displaying signs yielded few occurrences of possible signs. Camera traps were deployed in winter and summer at locations with possible signs, but, despite occupancy model detection probabilities of 0.88 in winter and 0.97 in summer, there were no detections of MBKR. Given detection probability estimates inferred from Dipodomys heermanni arenae, the conditional occupancy estimate that MBKR are present on the sandspit but were missed by all cameras on all nights of surveying is extremely low (5 × 10−6). We conclude that the MBKR is not present on the Morro Bay sandspit, at least not in the habitat where its presence was most likely to be detected. Future surveys for this small, cryptic species will need to adapt to a combination of low expected occupancy and high expected detection probability.