Model FAQs
DSM FAQ
This document contains Frequently Asked Questions on the CVPIA DSMs. It has been modified only to include relevant sections to the Reorienting to Recovery version of the CVPIA DSMs, and links have been updated to point to R2R DSM inputs and modeling repositories. The original FAQ document can be found here. Additional FAQs regarding model code and repository structure are located on the DSM documentation sites.
Contents
What is the model in a nutshell?
How many DSMs are there and how do they differ?
What are the outputs from the model?
What is the geographic scope of the model?
What are the source data for temperature inputs to the model?
What are the source data for flow inputs to the model?
What are the source data for habitat inputs to the model?
FAQs
What is the model in a nutshell?
The Decision Support Models (DSMs) are stochastic stage-based models. Stochastic models are models where randomness is introduced into the modeling process; for the DSMs, this occurs by drawing random values from probability distributions throughout the model. The models can be run in a stochastic or a deterministic mode.
The DSM operates on a monthly time step and simulates 20 years (see model outputs).
The conceptual diagram below gives an overview of how fish move throughout the system (from Peterson and Duarte 2020).
In each year the following actions occur:
Adults return from the ocean to the watershed and ocean and in river harvest is applied. Adults experience 2 types of survival before spawning: en-route survival and prespawn survival.
The en-route survival submodel represents prespawn morality as it is applied to adults returning to their natal tributaries and the pre-spawn survival submodel represents mortality while adults are at the spawning grounds.
En route survival is a function of migratory temperatures, whether the bypasses are overtopped (this represents fish loss due to stranding), and the adult harvest rate.
Pre-spawn mortality is a function of temperature, specifically the number of degree days that a fish experiences before spawning.
The number of juveniles produced is calculated based on the number of spawners, fecundity (varies with age of fish), and an egg-to-fry survival sub model. Egg-to-fry survival is a function of the temperature, the probability of the nest being scoured, and the proportion of natural fish spawning.
In each month the following actions occur:
Juveniles rear in-channel or on the floodplain, or migrate downstream depending on habitat availability and size of the juvenile. Tributary habitat capacity to support juvenile rearing is determined based on the total habitat in a tributary and a size-dependent territory requirement.
Growth is applied each month and differs with habitat type and temperature: seasonally inundated (floodplain, including the Yolo and Sutter bypasses) and perennially inundated (in-channel). Juveniles rearing on floodplains grow at a faster rate than juveniles who rear in-channel.
Rearing survival is applied and is a tributary-specific function based on water temperature, water diversions, weeks of floodplain inundation (when applicable), and predator prevalence.
The remaining juveniles not assigned to rear in a tributary will leave the watershed and migrate downstream. As fish migrate downstream a migratory survival rate is applied.
When habitat capacity is non-limiting, fish outmigrate when they reach the "very large" size class or at the end of the rearing season. The exception is for the springRunDSM where fish that are still small or medium size in their natal tributaries at the end of the outmigration window will remain in the natal tributaries as yearlings until the next year’s outmigration window.
After a juvenile makes it out to the ocean, ocean survival is applied, and they are assigned to return to their natal tributary one to three years later.
A manuscript describing the model can be found here.
Back to Contents.
How many DSMs are there and how do they differ?
There are four DSMs, one representing each run of Chinook salmon (fall-run, late-fall-run, winter-run, and spring-run). The late-fall-run DSM is still considered in “beta” mode. The DSMs differ with respect to timing of life history events, inputs, yearling dynamics, and juvenile movement rulesets.
Timing:
Late-fall-run:
Adults return (Oct-Dec)
Spawning (Oct-Dec)
Incubation (Nov-Dec)
Juvenile Rearing and outmigration (April - Nov)
Fall-run:
Adults return (Oct-Dec)
Spawning (Oct-Dec)
Incubation (Nov-Dec)
Juvenile Rearing and outmigration (Jan-Aug)
Winter-run:
Adults return (Jan-April)
Spawning (May-July)
Incubation (May-Aug)
Juvenile Rearing and outmigration (Sep-May)
Spring-run:
Adults return (Mar-Aug)
Spawning (Sep-Oct)
Incubation (Sep-Nov)
Juvenile Rearing and outmigration (Nov-Aug)
Inputs:
DSM habitat inputs vary depending on run, see FAQ on habitat inputs below for more information. All models use the same flow and temperature inputs.
Yearling Dynamics:
The springRunDSM allows for yearling dynamics where juveniles can spend a whole year in their natal tributary before outmigration. This only occurs for the spring-run model.
Back to Contents.
What are the outputs from the model?
The model generates outputs for each tributary and year of the 20-year model simulation:
the number of spawners,
the proportion of natural-origin spawners,
cohort replacement rate,
harvested adults
Back to Contents.
What is the geographic scope of the model?
The model includes the mainstem Sacramento River and San Joaquin River and their major tributaries, the Sutter and Yolo Bypasses, and the North and South Delta.
The model includes the following model segments:
Upper Sacramento River (Keswick to Red Bluff)
Antelope Creek
Battle Creek
Bear Creek
Big Chico Creek
Butte Creek
Clear Creek
Cottonwood Creek
Cow Creek
Deer Creek
Elder Creek
Mill Creek
Paynes Creek
Stony Creek
Thomes Creek
Upper-mid Sacramento River (Red Bluff to Wilkins Slough)
Sutter Bypass
Bear River
Feather River
Yuba River
Lower-mid Sacramento River (Wilkins Slough to American)
Yolo Bypass
American River
Lower Sacramento River (American to Freeport)
Calaveras River
Cosumnes River
Mokelumne River
Merced River
Stanislaus River
Tuolumne River
San Joaquin River
To see Chinook salmon spawning and rearing habitat ranges for the model segments look at the DSM habitat map. A map of the entire system is shown below (from Peterson and Duarte 2020).
Back to Contents.
What are the source data for temperature inputs to the model?
Stream temperature for a model segment is represented with one value that varies for each month and year of the simulation. Temperature inputs are sourced from a variety of models; more detailed information can be found here. Temperature is used in the model to calculate pre-spawn survival, egg-to-fry survival, rearing survival, and migratory survival.
Documentation of the temperature inputs can be found here DSMtemperature.
Back to Contents.
What are the source data for flow inputs to the model?
Streamflow for a model segment is represented with one value that varies for each month and year of the simulation. Monthly flow inputs are CALSIM II output, the model used to simulate California State Water Project (SWP) / Central Valley Project (CVP) operations. The CALSIM II node used to represent each model section is documented here. Flow is used to model fish migration and to calculate available habitat using stream-specific flow-area relationships.
Full documentation of the flow inputs can be found here DSMflow.
Back to Contents.
What are the source data for habitat inputs to the model?
Habitat for a model segment is represented with one value for instream and one value for floodplain that varies for each month and year of the simulation (except for the mainstem Sacramento River, which is split into four sections).
Where available:
Spawning and rearing habitat inputs are sourced from flow to suitable area relationships generated using results from Instream Flow Incremental Methodology (IFIM) studies.
Floodplain habitat inputs are sourced from flow to suitable area relationships generated from results of floodplain hydraulic modeling studies.
For model segments without IFIM studies or hydraulic modeling, suitable instream and floodplain areas were estimated by scaling proximal geomorphically-similar model segments. Habitat is used in the model to limit the number of fish that can spawn or rear in a tributary.
Documentation of the habitat inputs can be found here DSMhabitat.
Back to Contents.
Last updated