Mapping Stream Temperature and Salmon Health of the Sacramento River
CVTEMP forecasts how meteorology, hydrology, and Shasta Dam operations shape river temperature and translates it into survival estimates for endangered winter-run Chinook salmon eggs. Built by the UC Santa Cruz Fisheries Collaborative Program and NOAA's Southwest Fisheries Science Center.
Use the menu on the left to explore the site— from meteorology down to salmon. Each part has tabs for an overview, plots of observations and predictions, as well as options to download data and images.
A summary of conditions is also shown below — quick, at-a-glance cards highlighting the most recent model run's meteorology, watershed, reservoir, river, and salmon conditions, each linking to the full analysis and the latest model report.
The range reflects the spread across the historical redd-distribution year assumed — i.e., which past year's spawn timing and location is used to place the eggs in the model. It is not a forecast confidence interval.
An automated report is generated following each model run, documenting simulated reservoir conditions, river temperatures, and Winter-run Chinook salmon egg-survival forecasts.
Operational forecasting tools for Sacramento River salmon.
Developed at UC Santa Cruz with NOAA's Southwest Fisheries Science Center, CVTEMP forecasts water temperature across the Sacramento River system to inform Chinook salmon management. It links reservoir, river, and meteorological models to deliver information on thermal habitat, redd mortality risk, and TCD gate operations — helping protect cold-water refugia.
UC Santa Cruz
Lead Institution
NOAA SWFSC
Partner Agency
Operational
Daily Updates
How It Works
A plain-language walkthrough of the full forecasting workflow — from the weather forecast and reservoir models through to the river-temperature and salmon egg-survival results.
Four integrated models for the Sacramento River system.
Meteorology
Weather drives water temperature. CVTEMP ingests NOAA's GFS forecast (updated daily, out to 16 days) and compares it against the NARR historical archive to put current conditions in a long-term context.
Three main rivers — the Pit, McCloud, and upper Sacramento — carry snowmelt and runoff into Shasta each spring. CVTEMP tracks inflow volume and temperature to show how much cold water is being stored and how that compares to seasonal expectations.
Shasta Reservoir is simulated with CE-QUAL-W2, a physics-based model that tracks temperature from the surface to the bottom. The model can simulate which TCD gate to open — blending warm and cold water to hit a target release temperature.
Once water leaves Keswick Dam it continues to warm as it flows south. RAFT — the River Assessment for Forecasting Temperature — models this hour by hour from Keswick to Red Bluff (47 miles), accounting for solar heating, air temperature, and tributary inflows.
RAFT temperatures feed directly into the Temperature-Dependent Mortality (TDM) model, which estimates what fraction of Winter-run Chinook salmon eggs will survive the summer. Results can help inform decisions about TCD gate settings and reservoir operations.
River and reservoir temperature are in part driven by weather. Air temperature, sunlight, cloud cover, humidity, and wind all determine how much heat the water gains or loses on any given day. CVTEMP pulls meteorological data from two sources — a short-range weather forecast for the next 16 days and a historical weather archive spanning 35 years (1990–2024) — to drive all of the water temperature models and provide context for how current conditions compare to the past.
The GFS is NOAA's primary global weather model, updated four times a day and producing forecasts out to 16 days. CVTEMP extracts GFS output at locations across the Shasta Dam watershed and Sacramento River corridor.
Variables used: air temperature, humidity, cloud cover, solar radiation, wind speed
Drives the reservoir model (CE-QUAL-W2) and the river model (RAFT)
Updated daily as new GFS model runs become available
NARR is a historical reconstruction of past weather that blends observed station data, weather balloon soundings, and satellite data into a consistent long-term archive. It serves as a benchmark for comparing current forecasts against historical norms.
Covers 1990–2024 — 35 years of continuous data
Shown as the green shaded band in the meteorology charts
Helps answer: is this year's forecast warmer or cooler than average?
device_thermostatWhy Meteorology Matters
Weather drives water temperature
A hot, sunny forecast means the river and reservoir will warm faster, putting more pressure on Shasta Dam's cold-water pool and leaving less margin for salmon protection.
Air temperature — directly heats or cools the water surface
Solar radiation — the largest heat input to the reservoir in summer
Wind & humidity — drive evaporative cooling at the water surface
Cloud cover — reduces solar input and moderates nighttime cooling
What This Tab Shows
16-day Forecast
Sub-daily meteorological conditions over the next two weeks — air temperature, humidity, cloud cover, solar radiation, and wind speed.
Timeseries (full simulation)
Daily-average meteorological variables for the full model period, from the same GFS and NARR data that drive the water temperature simulations. Useful for understanding the seasonal weather context behind any given forecast run.
Meteorology — Outputs
What you're looking at: Sub-daily weather conditions over approximately the next two weeks — use the sidebar to switch between air temperature, humidity, cloud cover, solar radiation, and wind speed. The dashed vertical line marks today, separating recent observed conditions (left) from the forecast (right). Blue line and shaded band = GFS forecast averaged across the full Sacramento River corridor (Shasta Dam to Red Bluff); shaded region is the 5th–95th percentile spread across those locations. Navy dashed line = GFS forecast at Keswick Dam specifically.
Variables
Air Temperature
Near-surface air temp (°F)
Relative Humidity
Surface relative humidity (%)
Cloud Cover
Total cloud coverage (%)
Solar Radiation
Downward solar radiation (W/ft²)
Wind Speed
Wind speed (miles/hour)
Air TemperatureNear-surface air temp (°F)
Variable: Air Temperature |
Units: °F |
Source: GFS / NARR
Meteorology — Timeseries (full simulation)
Variables
Air Temperature
Near-surface air temp (°F)
Relative Humidity
Surface relative humidity (%)
Cloud Cover
Total cloud coverage (%)
Solar Radiation
Downward solar radiation (W/ft²)
Wind Speed
Wind speed (miles/hour)
What you're looking at: Daily-average weather for the full model run. Blue line = GFS watershed mean (Shasta Dam to Red Bluff). Navy dashed line = GFS forecast. Green band = NARR historical range (1990–2024). The dashed grey vertical line marks today — hindcast to the left, forecast to the right.
Discovering runs…
Air TemperatureNear-surface air temp (°F)
Variable: Air Temperature |
Units: °F |
Source: W2 met forcing (GFS + NARR analog) |
Resolution: Daily mean
Watershed — About
Shasta Reservoir is fed by three primary rivers draining the southern Cascade Range and Klamath Mountains — the Pit, McCloud, and upper Sacramento Rivers. How much water flows in, and how warm or cold that water is, directly determines the temperature structure inside the reservoir and how much cold water is available for release into the Sacramento River below. During the hindcast period, CVTEMP drives the reservoir model using inflow volumes and temperatures derived from observed data from CDEC (California Data Exchange Center). For the forecast period, inflow volumes are based on the B120 seasonal forecast or NOAA's Hydrologic Ensemble Forecast Service (HEFS) from the California Nevada River Forecast Center, while inflow temperatures follow a historical climatology of each river — the methods differ because flow can be forecast from hydrology but temperature this far ahead cannot.
terrainSacramento Watershed Tributaries
Three branches feeding Shasta Reservoir
Each tributary brings water of a different volume and temperature. Cold tributaries cool the reservoir; warm, low-flow years leave managers with less cold water to work with come summer.
Pit River — largest tributary by volume; originates on the Modoc Plateau
McCloud River — cold, spring-fed tributary flowing through volcanic terrain
Sacramento River — main stem above Shasta, carrying Cascade and Klamath snowmelt
Each branch provides daily inflow volume (Flow) and inflow temperature (Temperature) to the reservoir model.
waterInflow Boundary Conditions
How CVTEMP sets inflow boundary conditions
During the hindcast period, both inflow volume and inflow temperature come straight from observed CDEC gauge data: tributary volumes are anchored so their total matches observed Shasta inflow, and each river's temperature is taken directly from its gauge. These observation-based inflows are the solid lines in the Reservoir Inputs plots.
For the forecast period, inflow volume and inflow temperature are predicted in two different ways. Forecast volumes come from the B120 seasonal forecast (CDWR's runoff estimate) or NOAA's Hydrologic Ensemble Forecast Service (HEFS), which blend weather outlooks with watershed hydrology to project a range of likely future flows.
Forecast inflow temperature, however, does not come from B120 or HEFS. Because river temperature this far ahead cannot be predicted from weather, CVTEMP instead uses a historical climatology — the typical range of each river's temperature for that time of year, built from many past years of observations. Each scenario picks how warm or cool the incoming water is assumed to run (a normal, warmer, or cooler year), giving a realistic seasonal temperature pattern for every tributary through the forecast.
Forecast volume and temperature are shown as dashed lines; dotted lines show actual CDEC gauge readings for comparison against the forecast
compare_arrowsB120 Seasonal Forecast Comparison
How is this year tracking?
The B120 forecast is the official seasonal outlook issued by the California Department of Water Resources, estimating total runoff into Shasta Reservoir for the year. Because B120 is what drives the model's forecast inflows, comparing actual CDEC gauge readings against B120 shows where the model inputs may diverge from reality — important context for interpreting reservoir temperature forecasts.
Tracks cumulative inflow volume (acre-feet) vs. B120 expectation
Also compares inflow temperature against the B120 seasonal forecast
Helps managers assess how well the B120 assumptions held through the season
What This Tab Shows
Reservoir Inputs
Daily inflow volume and inflow temperature for each of the three tributary branches — Pit River, McCloud River, and upper Sacramento River. Solid lines are the hindcast inflows, whose total is anchored to observed CDEC Shasta inflow and split across tributaries by fixed proportions; dashed lines are the forecast-period inflow scenarios.
Analysis
Season-to-date comparison of observed tributary inflows against the B120 seasonal forecast — tracking cumulative volume and flow-weighted temperature differences across all three branches.
What you're looking at: The volume of water (flow, in cubic feet per second) and its temperature entering Shasta Reservoir from the three main tributaries — Pit, McCloud, and upper Sacramento Rivers. The dashed vertical line separates what has already happened (hindcast, left) from what is projected (forecast, right). Solid hindcast lines are anchored to observed CDEC inflow, with each river's share set by fixed split proportions; dashed forecast lines come from the B120 / HEFS seasonal outlook. More flow = more cold water filling the reservoir, which gives operators more to work with later in the summer. Warmer inflow temperatures reduce the cold-water pool and make protecting salmon downstream harder.
What you're looking at: How actual tributary inflow volumes and temperature into Shasta Reservoir compare to what was forecasted. For inflow volume, B120 provides a forecasted total inflow volume (in Thousand Acre-Feet, TAF) at each major tributary.
Cumulative inflow difference (top chart): Each day, the observed inflow from the Pit River, McCloud River, and upper Sacramento River is summed across all three tributaries and converted to TAF. That daily observed total is subtracted from the corresponding B120 daily forecasted inflow and the difference is accumulated from the start of the hindcast period. A running positive value means more water has arrived than B120 forecasted; a negative value means less. Blue shading = above expected; red shading = below expected.
Inflow temperature difference (bottom chart): For each day, the flow-weighted mean inflow temperature is computed across all three tributaries — each tributary's temperature is weighted by its share of total inflow volume observed on that day, so a larger tributary contributes proportionally more to the mean. The difference between observed and forecasted temperature is then accumulated over time (in °F·days). A positive running total means inflows have been warmer than expected on a flow-weighted basis; negative means cooler.
Why it matters: When inflows are lower and warmer than forecasted, the reservoir typically enters summer with less total cold water and a warmer initial reservoir profile.
Initializing…
Inflow Volume — Observed vs. B120 Forecasted
Variable: Cumulative Inflow Difference (Observed − B120 Expected) |
Units: Thousand Acre-Feet (TAF) |
Branches: Pit River + McCloud River + Sacramento River |
Blue = above expected ·
Red = below expected
Inflow Temperature — Observed vs. Forecasted
Variable: Cumulative Flow-Weighted Mean Temperature Difference (Observed − B120 Expected) |
Units: °F·days |
Branches: Pit River + McCloud River + Sacramento River (weighted by observed Flow) |
Red = warmer than expected ·
Blue = cooler than expected
Network Overview
Monitoring Stations
sensors
47
Total Stations
check_circle
44
Active
warning
3
Flagged / Offline
Station Directory
ID
Station Name
District
Type
Status
AND
Anderson
SAC
COOP
Active
BND
Bend
SAC
COOP
Active
BSF
Balls Ferry
SAC
ASOS
Active
CCR
Clear Creek
SAC
COOP
Active
SAC
Sacramento Executive
SAC
ASOS
Active
JLF
Jellys Ferry
SAC
COOP
Active
RDB
Red Bluff Municipal
SAC
ASOS
Active
STK
Stockton Metro
DLT
ASOS
Active
FRS
Fresno Yosemite Intl
FRS
ASOS
Active
TLR
Tulare
TLR
COOP
Flagged
KRN
Kern County
KRN
COOP
Active
BAK
Bakersfield Meadows
BAK
ASOS
Active
MDL
Modesto City
SJV
ASOS
Offline
VIS
Visalia Municipal
TLR
COOP
Active
River — About
Once water flows out of Keswick Dam, its temperature continues to change as it travels down the Sacramento River. Air temperature, sunlight, and tributary inflows all alter this temperature change. CVTEMP's River Assessment for Forecasting Temperature (RAFT) model tracks these dynamics hour by hour from Keswick Dam to Red Bluff — a stretch of about 47 river miles — covering the critical window when Winter-run Chinook salmon eggs are incubating in the riverbed.
waterHow RAFT Works
River Assessment for Forecasting Temperature
RAFT routes water downstream from Keswick Dam, continuously calculating how much heat is gained or lost through the water surface and from tributary inflows as the water travels south.
Driven by Keswick Dam outflow temperature and discharge
Uses GFS meteorological forcing (air temperature, solar radiation, humidity, wind) along the river corridor
Runs at sub-daily (hourly) resolution to capture diurnal temperature swings
Covers April through October — the incubation window for Winter-run eggs
placeMonitoring Sites
CDEC gages — Sacramento River corridor
Four California Department of Water Resources (CDEC) temperature gages provide real-time observed readings that are compared directly against RAFT model predictions.
Sacramento (SAC) — ~5 miles below Keswick
Clear Creek (CCR) — ~10 miles below Keswick
Balls Ferry (BSF) — ~13 miles below Keswick
Red Bluff (RDB) — ~47 miles below Keswick
thermostatTemperature Thresholds
Salmon protection targets
Regulatory temperature limits define the maximum acceptable river temperature at key locations to protect incubating Winter-run Chinook eggs during summer and fall.
53.5°F at Clear Creek (CCR) — primary compliance target
56°F at Balls Ferry (BSF) — downstream compliance point
The Analysis tab counts the number of days each threshold is exceeded across the season.
What This Tab Shows
Temperature Landscape
A heatmap of river temperature across every river mile and every day of the season — from Keswick Dam to Red Bluff. Quickly reveals where and when the river runs warm or cool.
Timeseries (2-week)
Sub-daily temperature at each monitoring site over the two-week forecast window. Model predictions side-by-side with CDEC observed readings.
Timeseries (full season)
Daily-average temperature at each site for the full simulation period, with hindcast and forecast periods clearly marked.
Analysis
Count of days above the 53.5°F (CCR) and 56°F (BSF) thresholds, based on observed CDEC data through the current date.
Temperature Landscape
Site Lines
What you're looking at: A heatmap showing predicted water temperature at every point along the Sacramento River from Keswick Dam (0 miles) to Red Bluff (47 miles) across the full season. Each column of color is one day; each row is a different river location. Blue = cold; orange/red = warm. The dashed vertical line separates hindcast (left, matched to observed conditions) from forecast (right). Horizontal dashed lines mark real monitoring stations. Warm colors spreading downstream early in summer mean less cold water is reaching salmon spawning areas.
Discovering runs…
°F
River — Temperature Timeseries (2 weeks)
Predicted and observed CDEC hourly temperature · forecast window
Sites
What you're looking at: Predicted and observed river temperature at four monitoring stations, zoomed into the current 2-week forecast window. Solid lines are the RAFT model prediction; dots are real-time readings from CDEC water temperature gages. Close agreement between lines and dots means the model is performing well. The grey dashed line is the temperature management target — days when the predicted line rises above it signal potential stress for incubating salmon eggs.
Select a run and scenario above
Solid lines: RAFT predicted (daily avg) |
Dashed lines: Observed CDEC |
Window: forecast start −5 to +15 days
River — Temperature Timeseries (full simulation)
Daily-average predicted and observed temperature · full simulation period
Sites
What you're looking at: Daily-average river temperature at four monitoring stations over the full April–October season. The dashed vertical line separates what already happened (hindcast, left) from what is projected (forecast, right). The grey dashed line is the temperature management target for Clear Creek (CCR). Days where model lines exceed the target indicate periods when incubating salmon eggs may be at risk. Compare scenarios using the dropdown above to see how different management choices affect river temperature.
What you're looking at: Daily observed river temperatures at two CDEC monitoring stations compared against regulatory salmon protection thresholds, from May 15 through the hindcast end date. Each bar represents one day; the dashed horizontal line marks the compliance threshold.
Monitoring stations and thresholds:Clear Creek (CCR) is located approximately 10 river miles below Keswick Dam and has a compliance threshold of 53.5°F. Balls Ferry (BSF) is approximately 13 miles below Keswick and has a threshold of 56°F. These thresholds are established in the National Marine Fisheries Service (NMFS) Biological Opinion and are designed to protect incubating Winter-run Chinook salmon eggs during their summer incubation period. CCR is the primary compliance point; BSF serves as a downstream check.
How values are calculated: Sub-daily (15-minute) CDEC water temperature readings at each station are averaged to produce a daily mean temperature. Each bar's height is the daily mean, colored red if it equals or exceeds the threshold and blue if below. The exceedance count in the summary cards is the number of days on or above the threshold out of all days with available data since May 15. This is observed conditions only — no model output is used here.
Why it matters: Winter-run Chinook salmon spawn in late spring and early summer, burying their eggs in the river gravel where they incubate through the summer. Sustained temperatures above these thresholds increase egg mortality. Operators use Shasta Dam's cold-water pool and the Temperature Control Device (TCD) to manage releases and keep downstream temperatures within these limits. The cumulative count of exceedance days is a key indicator of thermal stress on the current year's egg cohort.
Loading temperature data…
Clear Creek Gage (CCR) — daily mean water temperature. Red = exceeds 53.5°F; blue = below threshold.
Balls Ferry (BSF) — daily mean water temperature. Red = exceeds 56°F; blue = below threshold.
72-Hour Outlook
Temperature Forecast
Today
74°
High / 52° Low
wb_sunny
Clear skies. Valley winds 5–10 mph. Low relative humidity expected through afternoon.
Tomorrow
71°
High / 49° Low
wb_cloudy
Increasing cloud cover by afternoon. Marine layer possible in northern sections overnight.
Day 3
68°
High / 47° Low
grain
Slight chance of light precipitation in foothills. Valley floor stays dry. Cooling trend continues.
WFO Forecast vs Model Guidance
District
WFO High
Model High
Diff
WFO Low
Model Low
Sacramento Valley
78°
76°
+2°
52°
51°
San Joaquin Valley
82°
80°
+2°
54°
53°
Fresno District
84°
83°
+1°
55°
54°
Tulare Basin
80°
79°
+1°
51°
50°
Kern County
86°
85°
+1°
58°
57°
Delta / Stockton
75°
74°
+1°
50°
49°
Salmon — About
The Sacramento River below Keswick Dam is one of the last remaining spawning habitats for Sacramento River Winter-run Chinook salmon, a federally endangered species. Unlike other Chinook runs, Winter-run salmon spawn from late spring into summer, so their eggs incubate through the hottest months of the year. Keeping the spawning habitat section of the river cool enough to protect those eggs is the central goal of Shasta Dam cold-water management. CVTEMP's Temperature-Dependent Mortality (TDM) model uses RAFT river temperature predictions to estimate how many eggs are expected to survive in any given forecast scenario.
set_mealWinter-run Chinook Salmon
Federally Endangered — Sacramento River
Adults enter the Sacramento River in winter and spring, holding in cold, deep pools near Shasta Dam
Spawning occurs April through August in the mainstem river below Keswick Dam
Females dig nests — called redds — in the gravel riverbed and bury their eggs
Eggs incubate in the gravel for several months, exposed to whatever river temperatures occur above
Without cold releases from Shasta Dam, summer river temperatures would be lethal to eggs
calendar_todayRedd Years & Spawning Cohorts
Not all nests face the same temperatures
Because spawning spans several months, different nests are started at different times. A redd year represents one spawning cohort — the group of eggs deposited during a particular part of the season. Eggs buried in late April will experience different temperatures than eggs buried in late July.
CVTEMP runs TDM for multiple redd years simultaneously
Each cohort has a unique temperature exposure window
Results reveal which part of the spawning season is most at risk under a given forecast
scienceTemperature-Dependent Mortality (TDM)
Egg survival model
Takes RAFT predicted river temperatures as input for each scenario
Applies temperature–dependent egg mortality relationships for Winter-run Chinook based on the stage-independent (Martin et al. 2017) and stage-dependent (Anderson et al. 2022) formulations
Estimates cumulative percent egg mortality across the entire incubation period
Can be run across multiple redd-year distributions and other scenario conditions to represent forecast uncertainty
What This Tab Shows
TDM Table
A summary of predicted egg mortality (%) for each model run, scenario, and meteorological year — broken out by redd year cohort so managers can see which part of the spawning season carries the highest risk.
TDM Landscape
A spatial heatmap showing mortality risk across river miles and time — revealing where and when along the river thermal stress is highest for a given redd year cohort.
Analysis
Aggregated mortality across all redd years and meteorological scenarios, with uncertainty bands and trend comparisons across recent model runs.
Aggregate Mortality by Run & Scenario
Predicted egg mortality by run, scenario, and TDM model · averaged across redd years
What you're looking at: A row-by-row summary of predicted Winter-run Chinook egg mortality for every forecast scenario and model run. The Mean Mortality column is the estimated percentage of eggs expected to die from warm river temperatures, averaged across all redd years — lower is better. The TDM Model column reports that value under each temperature–dependent mortality formulation: Stage Independent (Martin et al. 2017) and Stage Dependent (Anderson et al. 2022, which weights warm exposure by the egg's developmental stage), so each scenario appears once per model. The remaining columns describe the run: Run Date is when the forecast was produced, Scenario and Start Date identify the management run, and Met Year is the historical weather year used to drive it. Redd Year Range gives the minimum–maximum mortality across spawning cohorts — a redd year is one cohort, eggs deposited during a particular part of the April–August spawning season — so a wide range means risk depends strongly on when eggs were laid. Compare rows to see how hydrology and management choices change expected egg survival.
What you're looking at: A heatmap showing predicted egg mortality at every location along the Sacramento River over the spawning season for one selected redd year — a group of eggs deposited at the same time. The X-axis is time; the Y-axis is river miles below Keswick Dam. Blue = low mortality (eggs surviving); red = high mortality (eggs dying from heat). Hot spots reveal where and when temperature conditions were most dangerous for eggs. The heatmap itself is the same for every redd year — it is the scenario's thermal-risk field — while the white dots mark where and when each cohort's eggs were actually deposited — a larger dot means more redds at that time and location (see the size key below the heatmap). Select All Redd Years in the Redd Year dropdown to overlay the redds from every spawning cohort on that shared field.
Discovering scenarios…
0%100%Stage Independent TDM
Redds (egg nests)
1
10
30
Each white circle marks redds deposited at one time & river location; circle area is proportional to the number of redds there (number of nests labeled above).
Mortality by Redd-Timing Distribution
What you're looking at: Mean predicted egg mortality for the currently selected run and scenario, broken out by redd year. The key point: every point on this chart uses the same modeled river temperatures — the thermal field shown in the heatmap above — and varies only the redd distribution, i.e. the timing and location of egg deposition observed in each historical year. So the value at, say, 2015 is this run's temperatures applied to 2015's spawn-timing pattern — not the mortality that actually occurred in 2015. The chart therefore isolates how sensitive egg survival is to spawn timing, holding this season's thermal conditions fixed. Line color matches the selected TDM model (Stage Independent or Stage Dependent); when the Stage Independent model is shown, the shaded band spans the 10th–90th Monte Carlo percentiles (MC10–MC90), reflecting uncertainty in the model's temperature-tolerance parameters.
TDM Forecast Evolution
Mean Stage Independent TDM (%) · One Point per Scenario · Tracked Across Run Dates
What you're looking at: How the salmon egg-mortality forecast has evolved across successive model runs. The horizontal axis is the model run date and the vertical axis is predicted mean temperature-dependent mortality (%), so lower points are better for salmon. Each dot is one scenario's forecast for a single meteorological year, colored by management scenario (see legend); filled vs. open dots distinguish the forecast start date, and connecting lines trace each scenario across run dates (dashed = later start date). Faint diamonds show the average mortality for each redd year (spawning cohort) under that scenario. As the season progresses and conditions become clearer, the forecast tends to settle; a downward trend over time means conditions are improving for salmon.
Discovering runs…
Each dot is one scenario × meteorological-year forecast of mean temperature-dependent egg mortality. Diamonds show per-redd-year averages (across met years); lines connect each scenario's met-year mean across successive run dates.
Stage Dependent TDM Evolution
Mean Stage Dependent TDM (%) · One Point per Scenario · Tracked Across Run Dates
What you're looking at: The same forecast-evolution view as above, but for Stage Dependent TDM — the Anderson Type-O temperature-dependent mortality formulation — instead of the Stage Independent TDM shown above. Axes, colors, and markers are identical: the horizontal axis is the model run date and the vertical axis is predicted mean Stage Dependent TDM egg mortality (%), so lower points are better for salmon. Comparing this plot against the one above shows how sensitive the egg-survival outlook is to the choice of mortality model.
Each dot is one scenario × meteorological-year forecast of mean Stage Dependent TDM egg mortality. Diamonds show per-redd-year averages (across met years); lines connect each scenario's met-year mean across successive run dates.
Reservoir — About
Shasta Reservoir is the largest reservoir in California by volume, storing water from the Sacramento, Pit, and McCloud Rivers high in the southern Cascade Range and Klamath Mountains. The temperature of water released from Shasta Dam into the river below is one of the most powerful tools available to water managers for protecting cold-water fisheries. CVTEMP uses the CE-QUAL-W2 hydrodynamic model to simulate how temperature evolves through the reservoir — and what temperature water will be when it exits the dam.
layersTemperature Stratification
How deep reservoirs store cold water
Sunlight warms the upper layers of Shasta Reservoir while cold winter runoff sinks to the bottom, creating a stable layered structure — warm on top, cold below — called thermal stratification.
Surface water can exceed 70–75°F in summer
Deep water near the dam face can stay below 50°F year-round
The model tracks this temperature profile at every depth throughout the season
tuneTemperature Control Device (TCD)
Shasta Dam's depth-selective intake
Shasta Dam is fitted with a Temperature Control Device (TCD) — a structure on the dam face with four intake gates at different elevations. Operators open specific gates to draw water from the depth that hits a target release temperature, blending warm surface and cold deep water as needed.
Side gate — 750 ft (deepest, coldest)
Lower gate — 817 ft
Middle gate — 922 ft
Upper gate — 1,022 ft (warmest)
Gate flows are shown as overlay bands on the reservoir heatmap in the Landscape tab.
model_trainingCE-QUAL-W2 Reservoir Model
Two-dimensional hydrodynamics
Simulates water temperature in two dimensions — along the length of the reservoir and vertically with depth
Driven by watershed inflows and atmospheric forcing (air temperature, solar radiation, humidity, wind)
Captures seasonal stratification, mixing, and reservoir turnover
Outputs the full temperature profile at the dam face — directly determining what the TCD can deliver to the river below
What This Tab Shows
Landscape (Heatmap)
Temperature at every depth in Shasta Reservoir over the simulation time period. Cold blues at depth, warm reds near the surface. TCD gate flows shown as overlay bands.
Vertical Profile
How temperature changes with depth at a given point in the season. Model predictions compared against USBR observed profile measurements.
Timeseries
Daily outflow temperature from Shasta and Keswick Dams, compared against CDEC observed readings. Hindcast and forecast periods shown together.
Shasta — Observed vs. Predicted Vertical Profile
What you're looking at: Temperature at different depths inside Shasta Reservoir over time. Each colored line is the CE-QUAL-W2 model prediction at one elevation; colored dots are real measurements taken periodically by the Bureau of Reclamation. Close agreement between lines and dots means the model is accurately capturing the reservoir's temperature structure. The spread between warm water at the surface and cold water at the bottom is called thermal stratification — it's what makes selective cold-water releases possible.
What you're looking at: The temperature of water released from Shasta Dam and Keswick Dam (the afterbay just downstream) over the season. CE-QUAL-W2 model predictions and observed readings from CDEC gages are shown in addition to the target temperatures for management objective to protect salmon eggs (typically at a reference location of either CCR or BSF) and the Shasta release temperature to achieve the river target. When the Keswick line approaches the target, it may indicate operators will need to open lower TCD gates to pull colder water from deeper in the reservoir.
Select a run and scenario above
Variable: Outflow Temperature |
Units: °F |
Source: CE-QUAL-W2 two_21.opt, column T
Shasta and Keswick - Outflow Volume
What you're looking at: The volume of water released from Shasta Dam and the downstream Keswick Dam over the season, in cubic feet per second (cfs). The dashed vertical line separates hindcast (left, matched to observed operations) from forecast (right). Release volume is the other half of temperature management: higher flows push more cold water downstream to salmon, but also draw down the reservoir and its cold-water pool faster.
What you're looking at: A heatmap showing simulated temperature inside Shasta Reservoir at every depth over the season. The X-axis is time; the Y-axis is elevation above sea level — higher = closer to the water surface, lower = deeper in the reservoir. Blue = cold; warm colors = hot. The dark overlaid bands show when each Temperature Control Device (TCD) gate was open. The pool of cold blue water near the bottom is the cold-water resource operators draw on to keep the river cool for salmon.
Discovering runs…
Temperature (°F)
Gate flow:
lowhigh
Reservoir Release Analysis — Hindcast Period
What you're looking at: Two hindcast-only diagnostics — how much water was actually released at Keswick relative to the scenario's plan, and how accurately the model reproduced Shasta release temperature — covering only the period before the forecast start date.
Cumulative release volume difference (top chart): Each day, the observed Keswick release (CDEC daily mean, cfs) and the scenario's planned Keswick schedule are each converted to Thousand Acre-Feet (TAF). The daily difference (observed minus planned) is accumulated from April 1 forward into a running cumulative total. Blue = actual releases are running ahead of the planned schedule; red = below plan. A persistent negative trend means operations released less water than the scenario's plan called for. (Older runs that predate the planned-release export fall back to a model-Keswick vs prescribed-Shasta-outflow comparison; the chart caption notes which is shown.)
Cumulative release temperature bias (bottom chart): Each day, the W2_SHA model temperature at Shasta Dam (converted to °F) is subtracted from the CDEC-observed temperature at Shasta Dam (SHD gage) for the same day. These daily differences are accumulated into a running cumulative sum in °F·days. Red = model is running warmer than observed; blue = model is cooler than observed. A large positive cumulative bias means the model has been overpredicting release temperature; a large negative bias means it has been underpredicting. This diagnostic is primarily a model accuracy check — close tracking near zero indicates good calibration of the reservoir thermal structure.
Why it matters: Release volume and temperature together determine how much cold water enters the river below Keswick. If cumulative releases are below plan and temperatures are warmer than expected, less cold water than planned is reaching salmon habitat downstream.
Discovering runs…
Release Volume — Observed vs. Planned
Cumulative Keswick release difference (TAF) — observed release minus planned scenario schedule. Blue = released more than planned; red = released less than planned.
Release Temperature — Model vs. Observed
Cumulative Shasta release temperature difference (°F·days) — model minus SHD observed. Red = model warmer than observed; blue = model cooler than observed.
References
About This Page
This page provides references for the scientific models, data sources, and methods used throughout the CVTEMP platform. CVTEMP integrates several independently developed tools — a reservoir hydrodynamics model, a river temperature model, and a salmon egg survival model — each grounded in peer-reviewed research. The publications listed below describe the underlying science behind those components.
The observational data driving the models come from publicly accessible federal and state monitoring networks. Links to those data sources and to related decision-support tools for the Sacramento River are provided below.
Publications
Anderson, J. J., Beer, W. N., Israel, J. A., & Greene, S. (2022). Targeting river operations to the critical thermal window of fish incubation: Model and case study on Sacramento River winter-run Chinook Salmon. River Research and Applications, 38(5), 895–905. https://doi.org/10.1002/rra.3965
Cole, T. M., & Wells, S. A. (2006). CE-QUAL-W2: A two-dimensional, laterally averaged, hydrodynamic and water quality model, version 3.5 (Instruction Report EL-06-1). U.S. Army Engineer Research and Development Center.
Daniels, M. E., Sridharan, V. K., John, S. N., & Danner, E. M. (2018). Calibration and validation of linked water temperature models for the Shasta Reservoir and the Sacramento River from 2000 to 2015 (NOAA Technical Memorandum NMFS-SWFSC-597). U.S. Department of Commerce. https://doi.org/10.7289/V5/TM-NMFS-SWFSC-597
Martin, B. T., Pike, A., John, S. N., Hamda, N., Roberts, J., Lindley, S. T., & Danner, E. M. (2017). Phenomenological vs. biophysical models of thermal stress in aquatic eggs. Ecology Letters, 20(1), 50–59. https://doi.org/10.1111/ele.12705
National Centers for Environmental Prediction (NCEP). (2004). Global Forecast System (GFS) atmospheric model. NOAA National Centers for Environmental Prediction. https://repository.library.noaa.gov/view/noaa/11406
Pike, A., Danner, E., Boughton, D., Melton, F., Nemani, R., Rajagopalan, B., & Lindley, S. (2013). Forecasting river temperatures in real time using a stochastic dynamics approach. Water Resources Research, 49(9), 5168–5182. https://doi.org/10.1002/wrcr.20389
CVTEMP brings together four scientific models and a number of monitoring networks, each with its own vocabulary. This page defines the model names, technical terms, units, monitoring sites, and agencies referenced throughout the platform. Terms are grouped by topic in roughly the order they appear in the workflow — from weather, to the reservoir, to the river, to salmon.
Models
CVTEMP
The integrated forecasting system itself — four linked models (weather → reservoir → river → salmon) for the Sacramento River below Shasta Dam.
CE-QUAL-W2 (W2)
A physics-based, two-dimensional hydrodynamic and water-quality model used to simulate Shasta Reservoir temperature from the surface to the bottom, including which TCD gate to open.
RAFT
River Assessment for Forecasting Temperature. Models river temperature hour by hour from Keswick Dam to Red Bluff (47 miles), accounting for solar heating, air temperature, and tributary inflows.
TDM
Temperature-Dependent Mortality model. Estimates the fraction of Winter-run Chinook salmon eggs that will survive the summer, driven by RAFT river temperatures.
Weather & Meteorology
GFS
NOAA's Global Forecast System — the primary global weather model, updated four times a day with forecasts out to 16 days. CVTEMP extracts GFS output across the watershed and river corridor.
NARR
North American Regional Reanalysis — a historical weather reconstruction blending station, weather-balloon, and satellite data. Serves as the benchmark for comparing current forecasts against historical norms.
NOAA
National Oceanic and Atmospheric Administration — source of the GFS forecast and the NARR historical archive.
Reservoir & Dam Operations
Shasta Dam & Reservoir
The major storage reservoir on the upper Sacramento River, simulated by CE-QUAL-W2. Holds the cold-water pool used for downstream temperature management.
Keswick Dam
The re-regulating dam just below Shasta. Its release is the upstream boundary where RAFT begins routing water down the river.
TCD
Temperature Control Device — the gated intake structure on Shasta Dam that blends warm surface water and cold deep water to hit a target release temperature.
Gate / outlet
An individual TCD intake at a fixed elevation (side, lower, middle, upper). Gate selection controls how cold the released water is.
Cold-water pool
The volume of cold water stored deep in Shasta Reservoir — the limited resource that constrains how much salmon protection is possible later in the summer.
Hydrology & Flow
CDEC
The California Data Exchange Center, operated by DWR — the source of real-time observed streamflow and water-temperature gage data.
B120
DWR Bulletin 120 water-supply and seasonal runoff forecast, used to drive reservoir inflow scenarios.
CNRFC
California Nevada River Forecast Center — a NOAA office that produces the streamflow forecasts entering the reservoir.
Hindcast
The modeled past period, reconstructed from observed conditions. On charts it is shown to the left of the dashed divider, typically as solid lines.
Forecast
The modeled future period, driven by the weather forecast and operational scenarios. Shown to the right of the divider, typically as dashed lines.
TAF
Thousand acre-feet — the volume unit used for reservoir storage and cumulative inflow.
cfs
Cubic feet per second — the flow-rate unit used for river and inflow discharge.
Salmon & Biology
Winter-run Chinook
An endangered run of Chinook salmon that spawns below Keswick in summer.
Redd
A salmon nest excavated in river gravel, where eggs are deposited and incubate. "Redd year" groups eggs by the year they were laid.
Cohort
A group of eggs or fish spawned in the same period, tracked together as they move through the mortality model.
Stage Independent TDM
The temperature-dependent mortality formulation of Martin et al. (2017). It applies a single, constant thermal-tolerance threshold across the whole incubation period — warm exposure is treated the same regardless of how far the egg has developed.
Stage Dependent TDM
The temperature-dependent mortality formulation of Anderson et al. (2022). It weights warm exposure by the egg's developmental stage, concentrating sensitivity within a critical thermal window of incubation.
MC10 / MC50 / MC90
Monte Carlo percentiles of the TDM estimate. The model is run many times while resampling uncertain field-derived parameters; MC10, MC50, and MC90 are the 10th, 50th (median), and 90th percentiles of the resulting egg-mortality values — together they express the uncertainty band around the forecast.
Gage Sites (CDEC IDs)
Four-letter station identifiers used by the California Data Exchange Center.
