Modelling the probability of detecting mass mortality events

Abstract

While reports of mass mortality events (MMEs) are increasing in the literature, comparing the incidence of MMEs through time, among locations, or taxa is problematic without accounting for detection probabilities. MMEs involving small, cryptic species can be difficult to detect even during the event, and degradation and scavenging of carcasses can make the window for detection very short. As such, the number or occurrence rate of MMEs may often be severely underestimated, especially with infrequent observations. We develop a simple modeling framework to quantify the probability of detecting an MME as a function of the observation frequency relative to the rate at which MMEs become undetectable. This framework facilitates the design of surveillance programs and may be extended to correct estimates of the incidence of MMEs from actual surveillance data for more appropriate analyses of trends through time and among taxa.

Not All Mass Mortality Events are Equal

“…However, generating artificial MMEs has its own limitations, such as introducing carrion resources from animals that did not interact with the ecosystem or community prior to its death, and lacks a disturbance event that traditionally supersedes an MME (such as severe weather seen in (16,17)). With current research attempting to predict the detection of mortality events (as in (68)), there are opportunities to begin research on MMEs as they occur. Utilization of models, such as the one outlined in this paper and others predicting ecological outcomes of these events (See (2)) can provide pathways for hypothesis-driven research following less predictable (except in some instances, such as human induced mortalities) natural mortality events.…”