Predicted growth in plastic waste exceeds efforts to mitigate plastic pollution.

Plastic pollution is a planetary threat, affecting nearly every marine and freshwater ecosystem globally. In response, multilevel mitigation strategies are being adopted but with a lack of quantitative assessment of how such strategies reduce plastic emissions. We assessed the impact of three broad management strategies, plastic waste reduction, waste management, and environmental recovery, at different levels of effort to estimate plastic emissions to 2030 for 173 countries. We estimate that 19 to 23 million metric tons, or 11%, of plastic waste generated globally in 2016 entered aquatic ecosystems. Considering the ambitious commitments currently set by governments, annual emissions may reach up to 53 million metric tons per year by 2030. To reduce emissions to a level well below this prediction, extraordinary efforts to transform the global plastics economy are needed.

No-U-turn sampling for fast Bayesian inference in ADMB and TMB: Introducing the adnuts and tmbstan R packages.

Statistical inference is a widely-used, powerful tool for learning about natural processes in diverse fields. The statistical software platforms AD Model Builder (ADMB) and Template Model Builder (TMB) are particularly popular in the ecological literature, where they are typically used to perform frequentist inference of complex models. However, both lack capabilities for flexible and efficient Markov chain Monte Carlo (MCMC) integration. Recently, the no-U-turn sampler (NUTS) MCMC algorithm has gained popularity for Bayesian inference through the software Stan because it is efficient for high dimensional, complex hierarchical models. Here, we introduce the R packages adnuts and tmbstan, which provide NUTS sampling in parallel and interactive diagnostics with ShinyStan. The ADMB source code was modified to provide NUTS, while TMB models are linked directly into Stan. We describe the packages, provide case studies demonstrating their use, and contrast performance against Stan. For TMB models, we show how to test the accuracy of the Laplace approximation using NUTS. For complex models, the performance of ADMB and TMB was typically within +/- 50% the speed of Stan. In one TMB case study we found inaccuracies in the Laplace approximation, potentially leading to biased inference. adnuts provides a new method for estimating hierarchical ADMB models which previously were infeasible. TMB users can fit the same model in both frequentist and Bayesian paradigms, including using NUTS to test the validity of the Laplace approximation of the marginal likelihood for arbitrary subsets of parameters. These software developments extend the available statistical methods of the ADMB and TMB user base with no additional effort by the user.

ss3sim: an R package for fisheries Stock Assessment simulation with stock synthesis.

Simulation testing is an important approach to evaluating fishery stock assessment methods. In the last decade, the fisheries stock assessment modeling framework Stock Synthesis (SS3) has become widely used around the world. However, there lacks a generalized and scriptable framework for SS3 simulation testing. Here, we introduce ss3sim, an R package that facilitates reproducible, flexible, and rapid end-to-end simulation testing with SS3. ss3sim requires an existing SS3 model configuration along with plain-text control files describing alternative population dynamics, fishery properties, sampling scenarios, and assessment approaches. ss3sim then generates an underlying 'truth' from a specified operating model, samples from that truth, modifies and runs an estimation model, and synthesizes the results. The simulations can be run in parallel, reducing runtime, and the source code is free to be modified under an open-source MIT license. ss3sim is designed to explore structural differences between the underlying truth and assumptions of an estimation model, or between multiple estimation model configurations. For example, ss3sim can be used to answer questions about model misspecification, retrospective patterns, and the relative importance of different types of fisheries data. We demonstrate the software with an example, discuss how ss3sim complements other simulation software, and outline specific research questions that ss3sim could address.

Estimating historical eastern North Pacific blue whale catches using spatial calling patterns.

Blue whales (Balaenoptera musculus) were exploited extensively around the world and remain endangered. In the North Pacific their population structure is unclear and current status unknown, with the exception of a well-studied eastern North Pacific (ENP) population. Despite existing abundance estimates for the ENP population, it is difficult to estimate pre-exploitation abundance levels and gauge their recovery because historical catches of the ENP population are difficult to separate from catches of other populations in the North Pacific. We collated previously unreported Soviet catches and combined these with known catches to form the most current estimates of North Pacific blue whale catches. We split these conflated catches using recorded acoustic calls from throughout the North Pacific, the knowledge that the ENP population produces a different call than blue whales in the western North Pacific (WNP). The catches were split by estimating spatiotemporal occurrence of blue whales with generalized additive models fitted to acoustic call patterns, which predict the probability a catch belonged to the ENP population based on the proportion of calls of each population recorded by latitude, longitude, and month. When applied to the conflated historical catches, which totaled 9,773, we estimate that ENP blue whale catches totaled 3,411 (95% range 2,593 to 4,114) from 1905-1971, and amounted to 35% (95% range 27% to 42%) of all catches in the North Pacific. Thus most catches in the North Pacific were for WNP blue whales, totaling 6,362 (95% range 5,659 to 7,180). The uncertainty in the acoustic data influence the results substantially more than uncertainty in catch locations and dates, but the results are fairly insensitive to the ecological assumptions made in the analysis. The results of this study provide information for future studies investigating the recovery of these populations and the impact of continuing and future sources of anthropogenic mortality.