Skip to content

Onboarding

Jump to bottom
Grant Adams edited this page May 7, 2026 · 11 revisions

Background

CEATTLE is a multi-species statistical catch-at-age model that fits to standard stock assessment data to fit single-species and multi-species models. The multi-species models estimate time, age, and sex varying predation mortality conditioned on diet based predator-prey preference (suitability) and energetic consumption demand (ration). CEATTLE was first developed by Holsman et al (2016) in ADMB as an expansion of Jurado-Molina et al (2005) to incorporate temperature-dependent bioenergetics based consumption for pollock, cod, and arrowtooth flounder in the Bering Sea. CEATTLE utilizes MSVPA type 2 functional form for predation mortality (see Magnusson 1995 for derivation) that is estimated iteratively.

Adams et al (2022) converted the model to TMB and generalized the model to fit any number of species, one/two sex, surveys, and fisheries with a variety of parameterizations for its application to groundfish in the Gulf of Alaska. Wassermann et al (2024) expanded the TMB model for its application for understanding cannibalism in Pacific hake.

Software

Rceattle is the TMB based R package used to run CEATTLE models. The “dev-name-change” branch is the most up to date and should be a starting point for further development.

Inputs: Rceattle reads an excel file with data inputs and observation model specification (selectivity, catchability, etc) into R as a list of data.frames and vectors. Rceattle uses the data list to fit models using R and TMB using the "fit_mod” function which has switches to specify the population model (recruitment, mortality, etc). The data file can also be written to excel then altered and read back in to R using “read_data” and “write_data”. The excel files have a meta-data sheet with specifics and I’ve slowly put together a documentation vignette. However the helper functions using “?” will be the most up to date.

Model code: The basis of Rceattle is the TMB cpp file that contains the model equations and likelihood functions. The cpp file is designed to fit any number of species, surveys, and fisheries and is able to fit one and two sex models in single-species or multi-species mode. Rceattle automatically automatically builds a parameter list given switches in the data and then relies heavily on the map function to turn on and off parameters for the specified model parameterizations. For example, estimation of time-varying selectivity deviates will be turned off and fixed to the default value of 0 if time varying selectivity is not selected in the “fleet_control” excel sheet.

The “fit_mod” function takes an Rceattle data object and fits the model and outputs an “Rceattle” model object. There are a number of wrapper functions to do MSE, simulation, diagnostics, plots, retrospectives, etc that take the “Rceattle” model object as an input.

Getting started

For getting started there is a vignette on GitHub to walk through examples and a few different example models and applications in the examples folder.

Development

There are multiple branches for development. My approach has been that updates from the various dev branches are pushed to master once a project with the added developments has been completed.

Any feature branches should be branched from the “dev” branch and then merged back into dev. For adding spatial components, the sex dimension could be converted to “group” and a transition matrix used.

Development is ongoing and any updates will be pushed to dev. If anything doesn’t work please do a pull request, post an issue, or just let Grant know!

Deprecated branches:

These only allow single-sex models.

  • ceattle_classic: The original Holsman et al. (2016) model implemented in TMB/R. Uses ceattle_v01_02.cpp
  • ceattle_classic_reorganized: The original Holsman et al. (2016) model implemented in TMB/R but allowing more flexible data inputs via excel files. Also includes multiple additional suitability parameterizations to be estimated. Uses ceattle_v01_04.cpp

Pacific hake implementation

Pacific hake is one of the most valuable fisheries on the US West Coast and is one of the most abundant predators. The original interest in a CEATTLE model for hake was to explore their role in the California Current Ecosystem food web, with a specific interest in cannibalism. The code for the hake implementation is on GitHub and the diet data are stored in the California Current Trophic Database. The majority of the code was developed by Sophia Wassermann, so get in touch with her if you have questions. Further development of the model is of interest to the Joint Technical Committee of the Pacific Hake/Whiting Treaty, both as part of the ‘ecosystem considerations’ portion of the assessment and to inform age- and time-varying estimates of natural mortality.

Other implementations

There are a few toy implementations of single-species models in Rceattle in Rceattle-models (note, some may need to be updated with recent package updates).

The GOA arrowtooth assessment is found here

The 2022 GOA 4-spp multi-species paper implementation can be found here, but has since been updated annually here.

Kirstin’s Rceattle branch for the EBS 3-spp multi-species model can be found here.

A set of MSE runs for the GOA and EBS 3-spp multi-species models are here.

Key references

Adams, G. D. et al. (2022) ‘An ensemble approach to understand predation mortality for groundfish in the Gulf of Alaska’, Fisheries Research, 251(October 2021), p. 106303. doi: 10.1016/j.fishres.2022.106303.

Collie, J. S. and Gislason, H. (2001) ‘Biological reference points for fish stocks in a multispecies context’, Canadian Journal of Fisheries and Aquatic Sciences, 58(11), pp. 2167–2176. doi: 10.1139/cjfas-58-11-2167.

Collie, J. S. et al. (2016) ‘Ecosystem models for fisheries management: Finding the sweet spot’, Fish and Fisheries, 17(1), pp. 101–125. doi: 10.1111/faf.12093.

Curti, K. L. et al. (2013) ‘Evaluating the performance of a multispecies statistical catch-at-age model’, Canadian Journal of …, 484(January), pp. 470–484. doi: 10.1139/cjfas-2012-0229.

Gislason, H. (1999) ‘Single and multispecies reference points for Baltic fish stocks’, ICES Journal of Marine Science, 56(5), pp. 571–583. doi: 10.1006/jmsc.1999.0492.

Hollowed, A. (2000) ‘Are multispecies models an improvement on single-species models for measuring fishing impacts on marine ecosystems?’, ICES Journal of Marine Science, 57(3), pp. 707–719. doi: 10.1006/jmsc.2000.0734.

Holsman, K. K. et al. (2016) ‘A comparison of fisheries biological reference points estimated from temperature-specific multi-species and single-species climate-enhanced stock assessment models’, Deep Sea Research Part II: Topical Studies in Oceanography, 134, pp. 360–378. doi: 10.1016/j.dsr2.2015.08.001.

Jurado-Molina, J., Livingston, P. A. and Ianelli, J. N. (2005) ‘Incorporating predation interactions in a statistical catch-at-age model for a predator-prey system in the eastern Bering Sea’, Canadian Journal of Fisheries and Aquatic Sciences, 62(8), pp. 1865–1873. doi: 10.1139/f05-110.

Kaplan, I. et al. (2019) ‘A multi-model approach to understanding the role of Pacific sardine in the California Current food web’, Marine Ecology Progress Series, 617(8), pp. 307–321. doi:

Kinzey, D. and Punt, A. E. (2009) ‘Multispecies and Single-Spieces Models of Fish Population Dynamics: Comparing Parameter Estimates’, Natural Resource Modelling, 22(1), pp. 67–104.

Koehn, L.E. et al. (2016) ‘Developing a high taxonomic resolution food web model to assess the functional role of forage fish in the California Current Ecosystem’, Ecological Modelling, 355, pp. 87–100.

Magnusson, K. (1995) ‘An overview of the multispecies VPA—theory and applications’, Reviews in Fish Biology and Fisheries, 212, pp. 195–212. doi: 10.1007/BF00179756.

Malick, M. et al. (2020) ‘Relationships between temperature and Pacific hake distribution vary cross latitude and life-history stage’, Marine Ecology Progress Series, 639, pp. 185–197.

Plagányi, É. E. et al. (2014) ‘Multispecies fisheries management and conservation: Tactical applications using models of intermediate complexity’, Fish and Fisheries, 15(1), pp. 1–22. doi: 10.1111/j.1467-2979.2012.00488.x.

Phillips, E.M. et al. (2023) ‘The influence of temperature on Pacific hake co-occurrence with euphausiids in the California Current Ecosystem’, Fisheries Oceanography, 23, pp. 267–279.

Tsehaye, I. et al. (2014) ‘A multispecies statistical age-structured model to assess predator–prey balance: application to an intensively managed Lake Michigan pelagic fish community’, Canadian Journal of Fisheries and Aquatic Sciences, 71(4), pp. 627–644. doi: 10.1139/cjfas-2013-0313.

Wassermann, S. N. et al. (2024) ‘Even low levels of cannibalism can bias population estimates for Pacific hake’, ICES Journal of Marine Science. Edited by M. Hidalgo, 3001(0), pp. 2994–3001. doi: 10.1093/icesjms/fsae064.

Trijoulet, V., Fay, G. and Miller, T. J. (2020) ‘Performance of a state-space multispecies model: What are the consequences of ignoring predation and process errors in stock assessments?’, Journal of Applied Ecology, 57(1), pp. 121–135. doi: 10.1111/1365-2664.13515.

Van Kirk, K. F., Quinn, T. J. and Collie, J. S. (2010) ‘A multispecies age-structured assessment model for the Gulf of Alaska’, Canadian Journal of Fisheries and Aquatic Sciences, 67(7), pp. 1135–1148. doi: 10.1139/F10-053.

Wells, B.K. et al. (2024) ‘Trophoscapes of predatory fish reveal biogeographic structuring of spatial dietary overlap and inform fisheries bycatch patterns’, Marine Ecology Progress Series, 741:47–70.

Clone this wiki locally