Please contact Mark Sorel for questions about the code or data: [email protected]{.email}
Secondary contact: Sarah Converse ([email protected]{.email})
A primary challenge in conservation is the evaluation of management actions in terms of their ability to achieve desired outcomes. For species with complex life histories or life histories that play out across different habitats, this evaluation is particularly challenging. Imperiled migratory salmonids in North America face a wide variety of potentially threatening factors, most notably the “Four Hs”: habitat degradation, hydropower development, overharvest, and effects of hatcheries. Management actions for addressing these factors have often been evaluated in terms of proximal responses (e.g., area of habitat restored or number of additional juveniles produced), rather than long-term population-scale responses (e.g., changes in naturally-produced spawner abundance or quasi-extinction risk). To understand population-scale responses, quantitative models are needed to understand how effects on particular life stages translate to fundamental management objectives, such as an acceptably low extinction risk or large abundance. We modeled the effects of two types of management actions on achievement of recovery objectives for endangered Chinook salmon (Oncorhynchus tshawytscha) in the Wenatchee River Basin of Washington State. Specifically, we considered juvenile rearing habitat restoration at different spatial scales and hatchery supplementation actions. We evaluated these actions using simulations built on the structure and parameter estimates produced by an integrated population model that accounted for density-dependence operating in juvenile rearing habitat. Given substantial uncertainty regarding the effects of habitat restoration on survival of juvenile salmon, we conducted a sensitivity analysis to evaluate different magnitudes of effects. Simulations indicated that there were compounding benefits of restoring habitat in both natal streams, where juveniles are born, and in downstream areas where a portion of juveniles rear. Simulations further indicated that the proportion of juveniles rearing downstream, and the benefit of restoring downstream habitat, increased with population size. However, only the most optimistic predictions of effects of juvenile survival (i.e., near doubling of juvenile survival in restored habitat) in the most comprehensive habitat restoration strategy resulted in the population meeting recovery goals. We also found that the marginal benefit of producing more hatchery fish declined both as juvenile survival was assumed to improve through habitat restoration and as the number of hatchery fish produced increased, highlighting the effect of survival probabilities in the natural environment on the benefit of hatchery supplementation. Our results underscore the importance of considering the whole system when working to recover migratory salmonids, and in assessing resource allocation strategies in terms of fundamental management objectives. To recover spring-run Chinook salmon in the Wenatchee River, management actions that focus only on habitat restoration may not be adequate to meet recovery criteria.
Contains scripts to run all analyses.
IPM_4_3_hatch.R Constructs the integrated population model from an Rdata object data/dat_IPM_hatch.Rdata with data and a TMB model src/IPM_non_centered_hatchery_scenarios.cpp. Then posterior samples are drawn using tmbstan.
new_plots.R Conducts population projection simulations assuming different habitat and hatchery management strategies. The outputs from the simulations are saved in the results folder. The script also summarizes the results of the simulations in figures.
IPM_non_centered_hatchery_scenarios.cpp The integrated population model written in TMB.
Contains processed data and model inputs. For raw data and to see the steps for generating model inputs, see this repository.
dat_IPM_hatch.Rdata Data inputs to the integrated population model.
par_IPM_hatch.Rdata Initial parameter inputs to the model.
map_IPM_hatch.Rdata Parameters to be treated as random effects.
rand_par_IPM_hatch.Rdata Parameters to be fixed at initial values during fitting
broodstock_remova.xlsx Historical numbers of natural origin fish removed for hatchery broodstock.
proj_arrays_hatch_8_11_2022.Rdata Simulated future trajectories of environmental variables included as covariates in the population .
Contains raw and processed results
ipm_fit_non_centered_5_02_hatch_scen.Rdata The TMB integrated population model object. This is used to conducted the population projections using posterior samples of parameters, simulated environmental variables, and random draws of random effects of year.
list_of_draws.rda This file contains posterior samples from the integrated population model.
2_panel_map_elev_9242021.png Map used in the manuscript.
base_4.2.2
dplyr_1.1.0
forcats_1.0.0
ggplot2_3.4.1
graphics_4.2.2
grDevices_4.2.2
here_1.0.1
lubridate_1.9.1
methods_4.2.2
purrr_1.0.1
readxl_1.4.3
stats_4.2.2
stringr_1.5.0
tibble_3.1.8
tidyr_1.3.0
tidyverse_2.0.0
TMB_1.9.3
TMBhelper_1.4.0
tmbstan_1.0.7
utils_4.2.2
viridisLite_0.4.1
Sorel MH, RW Zabel, AR Murdoch, and SJ Converse. In prep. Management modeling of salmon habitat restoration and hatchery supplementation.
The new_plots.R file located in the 'scripts' folder is the main scripts for the analysis and can be used to generate the results in the paper. The IPM_4_3_hatch.R file in th 'scripts' folder draws posterior samples using tmbstan, but is very computationally and time intensive, so samples have been saved in the results and archived on Zenodo.