DSLCM

Delta Smelt Life Cycle Model (DSLCM)

Criterion	Explanation
General Description	The Delta Smelt Life Cycle Model is a Bayesian state-space model designed to assess and predict the effects of management actions on the delta smelt population, with the long-term goals of population recovery and sustainability.
Model Domain	Sacramento-San Joaquin Delta. Western limit is mid San Pablo Bay.
Developer	US Fish and Wildlife Service
Hardware computing requirements	Not specified
Code language	R, JAGS
Original application	Assess and predict the effects of management actions on the delta smelt population, with the long-term goals of population recovery and sustainability.
Public/proprietary and cost	Model was publicly funded but source code is not publicly available. No licensing costs reported.
Physically or empirically based	Empirical
Mathematical methods used	Recruitment modeled with Beverton-Holt model. Markov chain Monte Carlo (MCMC) used to generate samples of states and parameters from the conditional distribution. Stochastic Search Variable Selection (SSVS) used to search the space of candidate models for the best model.
Input data requirements	SWP and CVP salvage data used for estimating entrainment into water export facilities. Twenty millimeter survey, Fall Midwater Trawl, and Spring Kodaiak Trawl data used for estimating abundance of different life stages. DSM2 PTM used for modeling entrainment. Food, X2, outflow, and predators as covariates.
Outputs	Primary outputs are vital rates (juvenile recruitment, juvenile survival, and subadult survival), population growth, and change in population over time.
Pre-processing and post-processing tools	R scripts for pre-processing data are described in materials downloaded from DSLCM webpage but are not available from that webpage.
Representation of uncertainty	Stochasticity and uncertainty is explicit component of Bayesian state-space modeling.
Prevalence	There are numerous conceptual, statistical, and life cycle models for Delta Smelt. Community support has seemingly not coalesced around any one model, including the DSLCM.
Ease of use for public entities	The model uses advanced statistical techniques that would require training to understand.
Ease of obtaining information and availability of technical support	No user group or commercial help desk
Source code availability	Source code is not publicly available
Status of model development	The model framework is mature. Future work not explicitly described in available documentation.
Challenges for integration	No major barriers to integration