Abstract:

As part of the Consortium for Advanced Research on Marine Mammal Health Assessment (CARMMHA) project, expert elicitation (EE) workshops were conducted to generate probabilistic distributions which could be used to update population models used in Deepwater Horizon (DWH) Natural Resources Damage Assessment (NRDA) integrating the latest and best data sources (DWH MMIQT, 2015; Schwacke, et al 2017). Between 2020-01-13 to 2020-01-16, we undertook two expert elicitation workshops to better understand the effect of DWH oiling on a range of Gulf of Mexico cetacean species at an individual and population level. This dataset contains a document summarizing the process and results of the expert elicitation study and a spreadsheet detailing the results.

Suggested Citation:

Booth, Cormac, Tiago Marques, Lori Schwacke and Len Thomas. 2020. Statistical distributions of model parameters generated by expert elicitation to inform cetacean population recovery trajectory models. Distributed by: GRIIDC, Harte Research Institute, Texas A&M University–Corpus Christi. doi:10.7266/n7-9hsz-sn16

Purpose:

Predicting population recovery requires assumptions about whether individuals with reduced survival and fecundity rates will recover and, if so, at what rate. For the population projections made in the DWH NRDA, this information was elicited informally from subject matter experts: a linear recovery trajectory was assumed, and the experts were asked to estimate the time to full recovery. In situations like this, where hard information is lacking, elicitation of parameter distributions from experts is an established approach and one that is growing in the application in ecology, for example, to examine the population consequence of acoustic disturbance on marine mammals. However, expert elicitation is not straightforward – there are many pitfalls and potential biases. We undertook a formal elicitation process about health effects on dolphins at different levels of oiling, to elicit recovery trajectories on survival and fecundity for populations of interest. We focussed initially on bottlenose dolphins, for which there is the most health and related information. We then asked how other species exposed to similar levels of oiling may be expected to differ – we expected much greater levels of uncertainty here. Outputs of this process will form some of the inputs to the population dynamics models.

Data Parameters and Units:

Parameters submitted are Species.group (common name, general group), Model.parameter, Distribution (beta or gamma), Distribution.parameter (controlling parameter for function), and Distribution.parameter.value. The following parameters were elicited: • The shape of density-dependent fecundity response in: – Bay, Sound and Estuary (BSE) and coastal bottlenose dolphins – Sperm whales • The survival of adult female sperm whales • The effect of oiling on the survival of different ‘offshore’ species groupings: – Pelagic dolphins (Shelf bottlenose dolphins, Atlantic spotted dolphins, Clymene dolphins, Fraser’s dolphins, Pantropical spotted dolphins, Spinner dolphins, Striped dolphins) – Bryde’s whale – Mesopelagic whales and dolphins (Risso’s dolphins, Oceanic bottlenose dolphins, Rough-toothed dolphins, Pygmy sperm whales, Short-finned pilot whales, False killer whales, Melon-headed whales, Pygmy killer whales) – Sperm whales & beaked whales • Proportion of population that recover in survival to baseline survival: – BSE bottlenose dolphins – All ‘offshore’ species

Methods:

Expert elicitation is a structured process by which expert knowledge regarding an uncertain quantity is translated into a probability distribution. This formal technique, first developed in the 1950s and 60s (Brown 1968; O’Hagan et al. 2006), is widely used in a range of scientific fields to combine the opinions of many experts in situations where there is a relative lack of data but an urgent need for conservation or management decisions (Runge et al. 2011; Martin et al. 2012). The formal process of expert elicitation, therefore, avoids many of the well-documented problems, heuristics, and biases that arise when quantitative judgements are made by experts or where expert knowledge is sought in an unstructured matter (Kynn 2008; Morgan 2014).

Provenance and Historical References:

Brown, B. B. (1968). Delphi process: a methodology used for the elicitation of opinions of experts (No. RAND-P-3925). Rand Corp Santa Monica CA. DWH MMIQT (2015). Models and Analyses for the Quantification of Injury to Gulf of Mexico Cetaceans from the Deepwater Horizon Oil Spill. Available at https://www.fws.gov/doiddata/dwh-ardocuments/876/DWH-AR0105866.pdf Kynn M. (2008). The ‘heuristics and biases’ bias in expert elicitation. Journal of the Royal Statistical Society: Series A (Statistics in Society) 171:239-264. doi:10.1111/j.1467-985X.2007.00499.x Morgan MG. (2014). Use (and abuse) of expert elicitation in support of decision making for public policy. Proceedings of the National Academy of Sciences 111:7176-7184. doi:10.1073/pnas.1319946111) Martin TG, Burgman MA, Fidler F, Kuhnert PM, Low-Choy S, McBride M, Mengersen K. (2012). Eliciting expert knowledge in conservation science. Conservation Biology 26:29-38. doi:10.1111/j.1523-1739.2011.01806.x O’Hagan A, Buck CE, Daneshkhah A, Eiser JR, Garthwaite PH, Jenkinson DJ, Oakley JE, Rakow T. (2006). Uncertain judgements: eliciting experts’ probabilities. John Wiley & Sons. doi:10.1002/0470033312 Runge MC, Converse SJ, Lyons JE. (2011). Which uncertainty? Using expert elicitation and expected value of information to design an adaptive program. Biological Conservation 144:1214-1223. doi:10.1016/j.biocon.2010.12.020 Schwacke et al. (2017) Quantifying injury to common bottlenose dolphins from the Deepwater Horizon oil spill using an age-, sex- and class-structured population model. Endangered Species Research. 33:265-279. doi:10.3354/esr00777 Schwacke, L., Thomas, L., Wells, R., McFee, W., Hohn, A., Mullin, K., Zolman, E., Quigley, B., Rowles, T. and Schwacke, J. (2017). Quantifying injury to common bottlenose dolphins from the Deepwater Horizon oil spill using an age-, sex- and class-structured population model. Endangered Species Research, 33, 265–279. doi:10.3354/esr00777

Individual Files: