Juvenile Production Estimates (JPE) Methods through the years

"The JPE Subteam’s recommendation uses "Method 2" as described in O’Farrell et al. (2018) to calculate the JPE for BY 2025 and includes data through December 31, 2025 which results in the following estimates for BY 2025 natural-origin and hatchery-origin JPE:

JPE_Natural = 1,057,452
JPE_Hatchery = 130,096
JPE_BCJumpstart = 5,186

…

JPE Subteam Recommended Method for BY 2025
The JPE Subteam recommends the previously described inputs and the following equations be used for estimating the BY 2025 natural-origin (Equation 1) and hatchery-origin (Equations 2 and 3) JPE:

Equation 1:
JPE_Natural = JPI_Fry x Survival_Fry-to-Smolt x Survival_Smolt
= 6,007,515 × 0.5217 × 0.3374 = 1,057,452

Equation 2:
JPE_Hatchery = N_Hatchery x Survival_{HatcherySmolt}
= 470,000 × 0.2768 = 130,096

Equation 3:
JPE_BCJumpstart = N_BCJumpstart x Survival_BCJumpstart
= 217,000 × 0.0239 = 5,186"

"The JPE Subteam used "Method 2" as described in O’Farrell et al. (2018) to calculate the JPE for BY 2024 and includes data through December 31, 2024, which results in the following estimates for BY 2024 natural-origin and hatchery-origin JPE:

JPE_Natural = 98,893
JPE_Hatchery = 135,342
JPE_BCJumpstart = 2,868

…

JPE Subteam Estimates of the JPEs for BY 2024
The JPE Subteam used the previously described inputs and the following equations for estimating the BY 2024 natural-origin (Equation 2) and hatchery-origin (Equations 3 and 4) JPEs:

Equation 2:
JPE_Natural = JPI_Fry x Survival_Fry-to-Smolt x Survival_Smolt
= 587,167 × 0.5149 × 0.3271 = 98,893

Equation 3:
JPE_Hatchery = N_Hatchery x Survival_{HatcherySmolt}
= 515,000 × 0.2628 = 135,342

Equation 4:
JPE_BCJumpstart = N_BCJumpstart x Survival_BCJumpstart
= 120,000 × 0.0239 = 2,868"

"Estimated egg-to-fry survival for BY 2023

The approach described above allows us to back-calculate egg-to-fry survival based on estimates of the number of successful female spawners (N_Spawners), average female fecundity (AF), and JPI, as described under "Fry Production" and in Equation 1. This calculation can be a useful metric to compare with average or expected survival to identify mortality occurring during egg incubation and fry emergence. Using this equation, estimated BY 2023 egg-to-fry survival for winter-run Chinook Salmon is 0.2494.

Equation 1:
Survival_Egg-to-Fry = JPI_Fry / (N_Spawners x AF)

Although uncertainty exists in all three variables used to estimate egg-to-fry survival, uncertainty in female spawners and fecundity is not quantified during JPE development. Uncertainty in the JPI is quantified by 90 percent confidence intervals around the estimate of fry equivalents; based on these confidence intervals, egg-to-fry survival is estimated to be between 0.1515 and 0.3319.

Winter-run Chinook Salmon in 2023 spawned during one of the wettest water years on record, providing ample cold water for spawning and egg incubation below Keswick Reservoir. The NMFS model (Martin et al. 2017) estimates mean annual temperature dependent mortality of winter-run Chinook Salmon eggs at 0 percent (25–75% confidence interval of 0–1%), based on measured water temperatures and mapped winter-run Chinook Salmon spawning locations in the Sacramento River in 2023 (SWFSC 2023).

Thiamine deficiency complex syndrome contributed to early life stage mortality in 2023, thought to be the result of shifts in marine forage fish species and a dominance of northern anchovy in Chinook Salmon diets off the coast of California. An unpublished model estimated thiamine-dependent fry mortality at 51 percent (95% credible interval of 43-58%). This estimate is based on thiamine concentrations in egg samples from 59 females spawned at LSNFH in 2023, which showed 88 percent of females with thiamine levels low enough (<5.9 nmol/g) that some fry mortality would be expected and 56 percent with critically low levels (<2.7 nmol/g) (SWFSC, pers. comm; NOAA Fisheries 2024). These estimates of mortality related to temperature and thiamine deficiency and assuming background survival of 32 percent predict an estimated egg-to-fry survival of 16 percent. The Winter-run PWT notes that the uncertainty for this prediction is not yet quantified.

Winter-Run PWT Recommended Method for BY 2023

The Winter-Run PWT recommends the previously described inputs and the following equations be used for estimating the BY 2023 natural-origin (Equation 2) and hatchery-origin (Equations 3 and 4) JPE:

Equation 2:
JPE_Natural = JPI_Fry x Survival_Fry-to-Smolt x Survival_Smolt
= 1,458,089 x 0.5066 x 0.3180 = 234,896

Equation 3:
JPE_Hatchery = N_Hatchery x Survival_{HatcherySmolt}
= 760,934 x 0.2544 = 193,582

Equation 4:
JPE_BCJumpstart = N_BCJumpstart x Survival_BCJumpstart
= 140,530 x 0.0239 = 3,359

It is the opinion of the Winter-Run PWT that this method represents the best available science for estimating a JPE given currently available data. The JPE and confidence intervals account for detection probabilities and quantify uncertainty associated with estimates of JPI_Fry, fry-to-smolt, and smolt survival rates, which are used to develop the 95 percent confidence intervals for the JPE forecast."

"Thiamine deficiency complex syndrome contributed to early life stage mortality in 2023, thought to be the result of shifts in marine forage fish species and a dominance of northern anchovy in Chinook Salmon diets off the coast of California. An unpublished model estimated thiamine-dependent fry mortality at 51 percent (95% credible interval of 43-58%). This estimate is based on thiamine concentrations in egg samples from 59 females spawned at LSNFH in 2023, which showed 88 percent of females with thiamine levels low enough (<5.9 nmol/g) that some fry mortality would be expected and 56 percent with critically low levels (<2.7 nmol/g) (SWFSC, pers. comm; NOAA Fisheries 2024). These estimates of mortality related to temperature and thiamine deficiency and assuming background survival of 32 percent predict an estimated egg-to-fry survival of 16 percent. The Winter-run PWT notes that the uncertainty for this prediction is not yet quantified."

2023-01-20

"Discussion on low estimated egg-to-fry survival for BY 2022

The approach described above allows us to back-calculate egg-to-fry survival based on estimates of the number of successful female spawners (N_Spawners), average female fecundity (AF), and JPI, as described under "Fry Production" and in Equation 1. This calculation can be a useful metric to compare to average or expected survival in order to identify mortality occurring during egg incubation and fry emergence. Using this equation, estimated BY 2022 egg-to-fry survival for winter-run Chinook Salmon is 0.0217.

Equation 1:
Survival_Egg-to-Fry = JPI_Fry / (N_Spawners x AF)

Although uncertainty exists in all three variables used to estimate egg-to-fry survival, uncertainty in female spawners and fecundity is not quantified during JPE development. Uncertainty in the JPI is quantified by 90 percent confidence intervals around the estimate of fry equivalent; based on these confidence intervals, egg-to-fry survival is estimated to be between 0.0135 and 0.0299.

Winter-run Chinook Salmon in 2022 spawned during one of the warmest and driest years on record. Average air temperatures in the Sacramento Drainage from July through September were the third highest in the last century (NOAA 2022), and streamflow was approximately one third of average flow since 1981. The NMFS model (Martin et al. 2017), estimates mean annual temperature dependent mortality of winter-run Chinook Salmon eggs at 17 percent (25–75% confidence interval of 13–29%), based on measured water temperatures and mapped winter-run Chinook Salmon spawning locations in the Sacramento River in 2022 (SWFSC 2022).

Additional early life stage mortality was likely due to thiamine deficiency complex syndrome, thought to be the result of shifts in marine forage fish species off the coast of California. Thiamine concentrations in egg samples from 28 females spawned at LSNFH in 2022 showed 93 percent of females with thiamine levels low enough (<4.9 nmol/g) that some fry mortality would be expected and 71% with critically low levels (<2.9 nmol/g) (SWFSC, pers. comm.).

These estimates of mortality related to temperature and thiamine deficiency do not completely account for the low egg-to-fry survival estimates calculated using Equation 1. There are likely other sources of mortality upstream of Red Bluff that are not accounted for in the models. Because the USFWS did not observe any abnormal fry behavior at the RBDD RSTs (B. Poytress, USFWS, pers. comm.) and pathology results of fish collected at the traps did not show abnormally high levels of pathogens (S. Foott, USFWS, pers. comm.), suggesting that any acute It is important to note that because the method used to calculate the JPE uses the JPI approach, any uncertainty due to the cause of mortality does not affect the JPE.

Winter-Run PWT Recommended Method for BY 2022

The Winter-Run PWT recommends the previously described inputs and the following equations be used for estimating the BY 2022 natural-origin (Equation 2) and hatchery-origin (Equations 3 and 4) JPE:

Equation 2:
JPE_Natural = JPI_Fry x Survival_Fry-to-Smolt x Survival_Smolt
= 311,058 x 0.4946 x 0.3245 = 49,924

Equation 3:
JPE_Hatchery = N_Hatchery x Survival_{HatcherySmolt}
= 741,000 x 0.2577 = 190,956

Equation 4:
JPE_BCJumpstart = N_BCJumpstart x Survival_BCJumpstart
= 193,000 x 0.0206 = 3,976

It is the opinion of the Winter-Run PWT that this method represents the best available science for estimating a JPE given currently available data. The JPE and confidence intervals account for detection probabilities and quantify uncertainty associated with estimates of JPI_Fry, fry-to-smolt, and smolt survival rates, which are used to develop the 95 percent confidence intervals for the JPE forecast."

"7/ Estimate of fry-to-smolt survival rate based on O'Farrell et al. (2018), updated using data from BY 1998-2017."

2022-01-20

"Discussion on low estimated egg-to-fry survival for BY 2021

The approach described above allows us to back-calculate egg-to fry survival based on estimates of the number of successful female spawners (N_Spawners), average female fecundity (AF), and JPI, as described under "Fry Production" and in Equation 1. This calculation can be a useful metric to compare to average or expected survival in order to identify mortality occurring during egg incubation and fry emergence. Using this equation, estimated BY 2021 egg-to-fry survival for winter-run Chinook Salmon is 0.0256. The two primary factors contributing to low egg-to-fry survival in BY 2021 are thought to be temperature dependent mortality and thiamine deficiency complex.

Equation 1:
Survival_Egg-to-Fry = JPI_Fry / (N_Spawners x AF)

Winter-run Chinook Salmon in 2021 spawned during one of the warmest and driest years on record, and Sacramento River water temperatures during the majority of the incubation period exceeded limits for successful egg incubation. Using the Martin et al. (2017) model, NMFS estimated mean annual temperature dependent mortality of winter-run Chinook Salmon eggs at 75 percent (25–75% confidence interval of 64–81%), based on measured water temperatures and mapped winter-run Chinook Salmon spawning locations in the Sacramento River in 2021 (SWFSC, 2021).

Additional early life stage mortality was likely due to thiamine deficiency complex syndrome, thought to be the result of shifts in marine forage fish species off the coast of California. Thiamine concentrations in egg samples from 30 females spawned at LSNFH in 2021 showed 83 percent of females with thiamine low enough where some fry mortality would be expected (T. Lipscomb, USFWS, pers. comm.). Any thiamine deficiency impacts manifested in egg viability or early fry stages will lead to a reduced JPI compared to what would have been observed absent thiamine deficiency impacts. USFWS had only one observation of abnormal fry behavior at the RBDD rotary screw traps (B. Poytress, USFWS, pers. comm.), suggesting that mortality caused by thiamine deficiency occurred primarily upstream of RBDD, though there may be latent impacts to young-of-year winter-run Chinook Salmon downstream of RBDD that cannot be estimated based on information available this year. The assumption that most mortality would occur prior to outmigration is consistent with observations at Central Valley hatcheries, where mortality and behavioral abnormalities associated with thiamine deficiency in hatchery-origin juveniles were documented soon after hatch. Survival studies of untreated fish would be necessary to understand lower survival due to latent effects of thiamine deficiency.

Uncertainty exists within all three of the variables used to calculate an estimate of egg-to-fry survival. Female spawners and fecundity estimates are not used in the JPE calculation, and their uncertainty is not quantified during JPE development. Uncertainty in the JPI is quantified, and it is a factor considered in the JPE calculation and in the back-calculation of egg-to-fry survival. For 2021, the nonoperation of the juvenile traps at RBDD for two days during a substantial storm event in October, and potential underestimate of passage during that event for the JPI, may contribute to the relatively low estimate of egg-to-fry survival. Standard methods used to interpolate juvenile fish passage data for unsampled days (see Voss and Poytress 2020) during the October flow event likely resulted in a slight negative bias to the juvenile passage estimates for those days. However, the impact of those two interpolated days on the total JPI calculated for the entire BY 2021 outmigration season is likely captured within the uncertainty (confidence intervals) of the 2021 JPI. The current range of uncertainty around the preliminary point estimate JPI would result in an egg-to-fry survival estimate of between 0.0166 and 0.0353.

It is unknown how much each of these factors may be contributing to the low estimated egg-to-fry survival for BY 2021, but there are ongoing efforts to better understand the contribution of each and any interactions between them. It is important to note that because the method used to calculate the JPE uses the JPI approach, any uncertainty about the mortality does not affect the JPE. Uncertainty in the JPE as a result of uncertainty in the JPI is captured in the 95 percent confidence intervals shown in Table 1.

Winter-Run PWT Recommended Method for BY 2021

The Winter-Run PWT recommends the previously described inputs and the following equations be used for estimating the BY 2021 natural-origin (Equation 2) and hatchery-origin (Equations 3 and 4) JPE:

Equation 2:
JPE_Natural = JPI_Fry x Survival_Fry-to-Smolt x Survival_Smolt
= 798,183 x 0.4429 x 0.3537 = 125,038

Equation 3:
JPE_Hatchery = N_Hatchery x Survival_{HatcherySmolt}
= 537,771 x 0.2818 = 151,544

Equation 4:
JPE_BCJumpstart = N_BCJumpstart x Survival_BCJumpstart
= 139,000 x 0.0526 = 7,311

It is the opinion of the Winter-Run PWT that this method represents the best available science for estimating a JPE given currently available data. It accounts for detection probabilities and quantifies uncertainty associated with estimates of JPI_Fry and smolt survival rates, which are used to develop the 95 percent confidence intervals for the JPE forecast."

"7/ Estimate of fry-to-smolt survival rate based on O'Farrell et al. (2018), updated using data from BY 1998-2016."

"11/ Estimated Battle Creek Jumpstart release as of December 15, 2022 (100% tagged and marked)."

2021-01-25

"JPE Recommendations

The Winter-Run PWT identified several factors in calculating the JPE that we advise be continued or updated for BY 2020. We considered one method for forecasting natural-origin JPE—The "Method 2" approach used last year for the BY 2019 JPE and described in O'Farrell et al. (2018). The data inputs for the calculations include estimates of the following parameters for calculating JPE for natural-origin BY 2020 winter-run Chinook (JPE_Natural) (Figure 1):

1. Number of winter-run fry-equivalents passing Red Bluff Diversion Dam(RBDD)(JPI_Frysub>)
2. Survival rate of natural origin fry to smolts (Survival_Fry-to-Smolt)
3. Survival rate of smolts from RBDD to Delta entry (defined as Sacramento at the I-80/I-50 Bridge) (Survival_Smolt)

Hatchery Release JPE Recommendations

Additionally, we used the number of winter-run hatchery smolts expected to be released from Livingston Stone National Fish Hatchery (LSNFH) in February 2021 (N_Hatchery) and their predicted survival rate (Survival_{HatcherySmolt}) to estimate a JPE of hatchery-origin winter-run juveniles in the Delta (JPE_Hatchery) (Figure 1). We present the data inputs used in the calculations in Table 1 and describe each in the sections below.

For the first time, we also include an estimate of hatchery-origin winter-run smolts released in Battle Creek as part of the "Jumpstart" reintroduction (N_BCJumpstart), their survival (Survival_BCJumpstart), and a forecast of the number entering the Delta (JPE_BCJumpstart). Although there was natural spawning in Battle Creek in 2020, we do not differentiate naturally produced juveniles from Battle Creek from Sacramento River juveniles, and they are included in the JPI_Fry.

Winter-Run JPE Methods for 2020-2021

The Winter-Run PWT focused on a single method for forecasting the JPE for BY 2020. This method was recommended in O'Farrell et al. (2018) and was the chosen method for BY 2019. It is the opinion of the Winter-Run PWT that this method represents the best available science given currently available data.

Juvenile Production Index -For the BY 2020 JPE, the Winter-Run PWT continues to recommend using the Juvenile Production Index (JPI_Fry or JPI), which is based on an estimate of fry-equivalents at RBDD. The JPI has been used in the calculation since 2014 and better represents the response of fish to annual environmental conditions during spawning, egg incubation, and outmigration, as compared to the long-term average egg-to-fry survival rate used in the JPE prior to 2014."

"Winter-Run PWT Recommended Method for BY 2020

The Winter-Run PWT recommends that Method 2, using the JPI estimate from method Te42, be used for estimating the BY 2020 natural-origin (Equation 1) and hatchery-origin (Equations 2 and

3) JPE:

Equation 1:
JPE_Natural =JPE_Fry x Survival_Fry-to-Smolt x Survival_Smolt
= 2,232,811 x 0.4475 x 0.3304 = 330,130

Equation 2:
JPE_Hatchery = N_Hatchery x Survival_{HatcherySmolt}
= 310,953 x 0.3148 = 97,888

Equation 2:
JPE_BCJumpstart = N_BCJumpstart x Survival_BCJumpstart
= 237,148 x 0.1570 = 37,232

It is the opinion of the Winter-Run PWT that this method represents the best available science given currently available data. Method 2 accounts for detection probabilities and quantifies uncertainty associated with estimates of JPIFry and smolt survival rates, which are used to develop the 95 percent confidence intervals for the JPE forecast."

"7/ Estimate of fry-to-smolt survival rate based on O'Farrell et al. (2018) applied to catch at RBDD."

"11/ Estimated Battle Creek Jumpstart release as of January 11, 2021 (100% tagged and marked)"

2020-02-03

"The process for developing the JPE was similar to what was done for BY 2018. A technical team from the Interagency Ecological Program (IEP), the Winter-run Project Work Team (WRPWT), met at the end of 2019 and provided recommendations to NMFS (Enclosure 2) on January 23, 2020. The method used to calculate the 2019 JPE is derived from the USFWS' estimated number of juveniles passing RBDD. This estimate is known as the Juvenile Production Index, or JPI, and is based on fry-equivalents at RBDD.

Until this year, a fry-to-smolt survival rate of 0.59, based on fall-run Chinook salmon, has been used since 1993 as a surrogate for winter-run Chinook salmon fry-to-smolt survival. This value is based on previous studies by Hallock (undated), and confirmed through a literature review in 1995 (B. Poytress, USFWS, personal communication). Without this factor, survival from fry to smolts is assumed to be 100 percent, which is unrealistic. The WRPWT has expressed reservations about the accuracy of the 0.59 term, and thus has recommended an alternative approach. The WRPWT reviewed a fry-to-smolt survival rate forecasting method developed by O'Farrell et al. (2018), which uses more recent winter-run Chinook salmon survival data and is updated with new survival data annually. The fry-to-smolt survival rate of 0.4651 was based on peer-reviewed methodologies and more recent winter-run Chinook salmon data, and therefore, improves the calculation of the JPE as compared to values used previously.

Winter Run JPE Methodologies Considered for 2019-2020

The Winter-Run PWT considered two similar methodologies for estimating the JPE for 2019-2020. Method 1 is the method that has been used in recent years, and we recommend updating several terms if this method is used for the 2019-2020 estimate. Method 2 has similar structure, but it forecasts the smolt survival rate differently than Method 1 and accounts for observation error when making the JPE forecast (see O'Farrell et al. 2018). After review of both methods, the Winter-Run PWT recommends using Method 2 to forecast the JPE for 2019-2020. Both methods and their inputs are described below."

"Winter Run PWT Recommended Method for 2019-2020

The Winter Run PWT recommends that Method 2 be used for estimating the BY 2019

natural-origin (Equation 1) and hatchery-origin (Equation 2) JPE:

Equation 1:
JPE_Natural =JPE_Fry x Survival_Fry-to-Smolt x Survival_Smolt
= 4,762,142 x 0.4651 x 0.3860 = 854,941

Equation 2:
JPE_Hatchery = N_Hatchery x Survival_{HatcherySmolt}
= 256,381 x 0.3687 = 94,528

It is the opinion of the Winter Run PWT that Method 2 represents the best available science given currently available data. Method 2 accounts for detection probabilities and quantifies uncertainty associated with estimates of JPIFry and smolt survival rates, which are used to develop the 95 percent confidence intervals for the JPE forecast (Table 1). Because Method 2 does not capture process error, or the variation in true survival rates from year to year, these confidence intervals likely underestimate the uncertainty in the JPE forecast.

We acknowledge that Method 2 still has considerable uncertainty, and that confidence intervals may not have utility to water managers under the current management setting. However, there is uncertainty with any forecast method for JPE, and we believe there is value in quantifying and reporting that uncertainty. This is an advantage of Method 2 over Method 2 and methods recommended in previous years."

"5/ Back calculated estimated survival between eggs laid in-river and fry production estimates at RBDD based on numbers of fry equivalents (JPI) using the 0.4651 fry-to-smolt survival rate estimate based on method described in O'Farrell et al. (2018).

6/ Preliminary number of fry equivalents estimated on December 31, 2019 plus 3.1% interpolation to account for remainder of estimated passage for the 2019 brood year at RBDD; using 0.4651 fry-to-smolt survival rate estimate (Bill Poytress, USFWS, pers. comm.).

7/ Estimate of fry-to-smolt survival rate based on O'Farrell et al. (2018).

8/ Variance-weighted mean survival rate of acoustically tagged winter-run from 2013 to 2019 between RBDD and I-80/Tower Bridge in Sacramento. Survival is estimated from the Salt Creek receiver site, located 3 miles downstream of RBDD, to estimate survival from RBDD for acoustic tag studies. See Smolt Survival section for an explanation of the different values displayed here.

9/ Estimated LSNFH production release as of December 31, 2019 (100% tagged and adipose clipped). Up to 1,500 juveniles may be retained as broodstock or to conduct fish health assessments.

10/ Variance-weighted mean survival rate of acoustically tagged winter-run from 2013 to 2019 between release location and I-80/Tower Bridge in Sacramento."

2019-02-13

"The process for developing the JPE was the similar to what was done for BY 2017. A technical team from the Interagency Ecological Program (IEP), the Winter-run Project Work Team (WRPWT), met at the beginning of 2019 and provided recommendations to NMFS (Enclosure 2). The method used to calculate the 2018 JPE is derived from the number of juveniles passing RBDD, as estimated by USFWS. This estimate is known as the Juvenile Production Index, or JPI, and it is based on fry-equivalents at RBDD."

"The Winter Run PWT recommends that the following methods be used for estimating the BY 2018 natural-origin (Equation 1) and hatchery-origin (Equation 2) JPE:

Equation 1: JPE (natural origin) = 1,460,216 (JPI) x 0.59 (fry-to-smolt survival) x 0.5028 (smolt survival of acoustic tag hatchery winter run from RBDD to the Delta) = 433,176

Equation 2: JPE (hatchery origin) = 223,817 (LSNFH release) x 0.3853 (smolt survival of acoustic tag hatchery winter-run from release location to Delta) = 86,237

While we acknowledge that there will still be uncertainty in the JPE estimate even if these recommendations are incorporated, we believe it to be the best information currently available from which to derive a JPE."

"6/ Preliminary number of fry-equivalents estimated on January 14, 2019 plus 2.1 % interpolation at RBDD using traditional (59%) fry to smelt survival estimates (Bill Poytress, USFWS, personal communication"

"8/ Average weighted survival of acoustically tagged winter-run in 2013, 2014, 2015 (2 values), 2016 (2 values), 2017, and 2018 between RBDD and I80 Tower Bridge in Sacramento. Survival is estimated from the Salt Creek receiver site, located 3 miles downstream of RBDD, to estimate survival from RBDD for acoustic tag studies (M. O'Farrell, NMFS, personal communication)"

2018-01-29

"The process for developing the JPE was … similar to what was done in 2016. A technical team from the Interagency Ecological Program (IEP), the Winter-run Project Work Team (WRPWT), met throughout the year and provided recommendations in a letter to NMFS (Enclosure 2). The method used to calculate the 2017 JPE is derived from the number of juveniles passing RBDD, as estimated by USFWS. This estimate is known as the Juvenile Production Index, or JPI, and it is based on fry-equivalents at RBDD. This is the same method that was used in 2016, however for BY 2017, the USFWS provided alternative JPI estimates to account for errors associated with using LAD criteria to determine Chinook salmon race/run (see Enclosure 2).

In 2017, several juvenile spring-run-sized Chinook salmon were sampled at RBDD, although very few spring-run Chinook salmon adults were detected in spawning surveys upstream of RBDD. These spring-run were identified using the river LAD criteria, but the timing and numbers suggested these fish sampled at RBDD were likely late-emerging winter-run. Tissue samples were genetically analyzed, which determined that many of the LAD spring-run were genetic winter-run. Alternative JPIs have been calculated incorporating the additional winter-run. These genetically-identified run results from the RBDD rotary screw traps have been independently supported by water temperature data, carcass surveys, and redd observations from 2017 field surveys. The JPE calculation methods that use this corrected number are denoted with an "a" both in the text of this letter and within Enclosure 2.

The JPE for BY 2017 incorporates the recommendations of the Independent Review Panel (IRP, Delta Science Program 2014) and advice from the WRPWT (Enclosure 2). The WRPWT identified four factors in calculating the JPE, similar to last year, that they would advise continuing or updating for BY 2017:

1. Estimated number of fry passing the RBDD
2. Survival rate of natural-origin fry to smolts
3. Survival rate of natural-origin smolts from RBDD to Delta entry
4. Estimated survival rate of hatchery-origin winter-nm to be released from LSNFH in January or February of 2018

A fry-to-smolt survival rate of 0.59, based on fall-run Chinook salmon, has been used as a surrogate for winter-nm fry-to-smolt survival since 1993. This value is based on previous studies by Hallock (undated), and confirmed through a literature review in 1995 (B. Poytress, USFWS, personal communication). Without this factor, survival from fry to smolts is assumed to be 100 percent, which is unrealistic. The WRPWT has expressed reservations about the accuracy of the 0.59 term, and thus have interest in considering alternative approaches.

As part of the "Species in the Spotlight" initiative. the NMFS Southwest Fisheries Science Center (SWFSC) was awarded funding in 2016 to develop alternative approaches for forecasting the natural-origin winter-run JPE. The alternative forecasting methods follow the same basic JPE model adopted for use in recent years (Method 1). Appendix A in Enclosure 2 describes two alternative methods (Methods 2 and 3) for forecasting the natural-origin JPE, both of which include a new approach for forecasting the fry-to-smolt survival rate. The fry-to-smolt survival rate forecast described in Appendix A of Enclosure 2 is based on fitting a zero-intercept linear model to estimates of hatchery-origin and natural-origin juvenile survival rates from the same brood years. This new methodology results in a lower fry-to-smolt survival rate for both Method 2 and Method 3, as shown in Table 1 and Appendix A in Enclosure 2.

Based on these latest recommendations, NMFS examined three methods for calculating the JPE for BY 2017, as follows:

1. Method I (JPI) uses the estimated juvenile passage at RBDD based on rotary screw trap monitoring. The JPI is a better estimate of fry survival to RBDD than the long-term average egg-to- fry survival rate used prior to 2014 in the JPE.
2. Method 2 uses the same general form as Method 1, but replaces the fry-to-smolt survival rate of 0.59 with 0.4725. It also accounts for observation error in the JPI, the fry-to-smolt survival rate, and the smolt survival rate (Appendix A in Enclosure 2).
3. Method 3 has the same general form as Methods 1 and 2. However, Method 3 assumes that the JPI error distribution is Lognormal. The approach used to forecast the fry-to-smolt survival rate is the same as described for Method 2. The forecast of smolt survival to the Delta is based on application of a hierarchical Bayesian model fitted to acoustic tag data (Appendix A in Enclosure 2).

As mentioned earlier, Methods la, 2a and 3a incorporate the additional estimated number of genetic winter-run due to misclassification of spring-run using the river LAD criteria. Each of the methods used to calculate the JPE begins with the same adult escapement estimate from CDFW, which uses data from carcass surveys to estimate the number of adult female spawners.

Estimates of egg-to-fry survival rate (SI) are based on the JPI estimate at RBDD. The JPI method is considered a more accurate estimate of S 1 because it is an annual estimate which better represents the response of fish to the environmental conditions at the time of spawning (see recommendations from the WRPWT in Enclosure 2). Differences in SI among the various methods considered here are due to the alternative JPI calculations used for Methods 2 and 3 (Appendix A in Enclosure 2).

The survival of juvenile winter-run to the Delta (S2) is based on assumed environmental conditions (e.g., temperature, flows, and turbidity) in the Sacramento River. However, actual environmental conditions, which may occur after the JPE is calculated, may be different than those assumed in the calculation of the JPE. This year, based on recommendations from the WRPWT, smolt survival (S2) to the Delta was calculated based on a weighted average of acoustically-tagged hatchery winter-run releases in 2013, 2014, 2015, 2016, and 2017 (Enclosure 2) from RBDD to the Tower Bridge (at Sacramento). NMFS considers the Tower Bridge as the point of Delta entry.

Method Selection and JPE Calculation

After reviewing the recommendations from the WRPWT technical team (Enclosure 2), NMFS chose Method 1 a (JPI method) to calculate the BY 2017 natural-origin winter-run JPE. There is significant interest in continuing to explore the use of alternative methods in the future (Methods 2 and 3), however complete documentation of these forecast methods is currently in preparation and has not been subject to peer review. These alternative methods will continue to be considered in subsequent years following adequate review and vetting. There was unanimous agreement within the WRPWT on the inclusion of the additional winter-run genetically identified from those initially identified as spring-run using the river LAD. Using the JPI, and based upon the WRPWT recommendation to use Method 1 a, NMFS estimates a JPE of 201,409 natural-origin juvenile winter-run entering the Delta during WY 2018 (Table 1, Enclosure 2). Winter-run juveniles are expected to emigrate into the Delta from November 2017 through April 2018, based upon CDFW historical monitoring data at Knights Landing rotary screw traps."

"4/ Preliminary (subject to change) average # eggs/female from weighted estimate based on two year and three [year] old fish spawned at Livingston Stone Hatchery and from the carcass survey"

"5/ Preliminary number of re-equivalents estimated on December 16, 2017 plus 4.1% interpolation at RBDD – JPI – Bill Poytress, (USFWS), personal communication"

"7/ Average weighted survival of acoustically tagged winter-run in 2013, 2014, 2015 (2 values in 2015) and 2016 (2 values in 2016) and one value in 2017 between RBDD and I80 Tower Bridge in Sacramento - A. Ammann, NMFS, personal communication. Survival is estimated from the Salt Creek receiver site, located 3 miles downstream of RBDD, to estimate survival from RBDD for acoustic tag studies (See Appendix A, for further description of how the weighted average acoustic tag survival was used for methods 2, 2a, 3 and 3a). This is the derivation of the numbers of fish arriving to Salt Ck (RBDD) and Sacramento (Tower, Sacl80/50) from the acoustic: tagged LSNFH winter run pre-smolts. The number of fish arriving to Salt Ck and Sacl80/50 are derived by looking at receivers at this location and all locations downstream. It uses fish directly detected at that site plus any fish that were missed at that site but detected at a site downstream. This was done for all years 2013-2017. The 2017 year was the only year that had significant flows over Fremont Weir. Fish that went over the weir and into the Yolo Bypass would bypass the Sacramento I80/50 Bridge and other nearby receiver locations. If these fish were detected at locations downstream of Liberty Island they were considered to have survived to the Delta and thus added to the number of the fish counted at Sacramento I80/50 Bridge.

8/ LSNFH estimated production release as of 1/8/17 (100% tagged and adipose clipped). 500 to 1000 of these fish may be held back for the captive broodstock program. Final number should be updated when the number of those held back are finalized.

9/ Weighted average of acoustically tagged winter-run survival in 2013, 2014 and 2015 and 2016 (2 values in 2015 and 2016) and 2017 between release location and I80/I50 Bridge in Sacramento. The estimates were derived in a similar manner as those from Salt Creek to Sacramento (Footnote 7)."

2017-01-28

"The process for developing the JPE was the similar to what was done in 2015. A technical team from the Interagency Ecological Program, the Winter-run Project Work Team (WRPWT), met throughout the year and provided recommendations in a letter to NMFS (enclosure 2). The method used to calculate the 2016 JPE is derived from the number of juveniles passing Red Bluff Diversion Dam estimated by the U.S. Fish and Wildlife Service (USFWS) and is the same method as was used in 2015, except for a change in calculation of fecundity. The fecundity rate was lower in this year's calculation as a result of a larger than average return of 2-year old, smaller, females (see USFWS Memo within enclosure 2)."

"2/ In-river females derived from carcass survey. Males derived using sex ratio at Keswick trap."

"4/ fecundity calculated based on weighted average of 2-yr old + 3-yr old females (USFWS Memo)"

"6/ Average egg-fry survival based on 15 years (2002-2016)"

"9/ LSNFH final count after tagging, Niemela Pers. comm. 1/12/17 (100% ad-clipped & CWT)."

"10/ Hatchery survival based on weighted average of acoustic tags (2013-2016) Caldwell Park to Sacramento."

2016-01-28

"Methods used for 2015 JPE calculation (Table 1):

1. NMFS (JPE) spreadsheet method
2. Cramer Fish Sciences (CFS) model: a systems dynamics computer simulation model that includes Monte Carlo stochastic simulations (CFS 2010)
3. Juvenile Production Index (JPI) method: real-time passage estimates at Red Bluff Diversion Dam (RBDD)

After reviewing the recommendations from the WRPWT technical team (enclosure 2), NMFS chose the JPI method to calculate the winter-run JPE from broodyear 2015, since it more closely represented the actual hydrologic conditions experienced by winter-run egg and fry in 2015 and because the loss of juveniles due to high temperatures is not accurately represented in either the NMFS or CFS models. Using the JPI method, and based upon the best available information,"

2015-01-16

"This year's JPE reflects a number of significant changes as a result of: (I) the Independent Review Panel (IRP) review of the JPE in November 2014 (DSP 2014), (2) the Winter-Run Project Work Team (WRPWT) technical review of the survival terms used to calculate the JPE (enclosure 2), and (3) internal discussions with the NMFS-Southwest Fisheries Science Center. Both the IRP and the WRPWT found that the JPE methodology did not adequately represent mortality due to upstream water temperatures and, therefore, overestimates the number of juveniles entering the Delta. This year, based on recommendations from the IRP, we are providing several different methods of calculating the JPE.

JPE methods considered:

1. NMFS spreadsheet model: same as 2013 methodology, however, with different survival terms
2. Cramer Fish Sciences (CFS) model: a systems dynamics computer simulation model that includes Monte Carlo stochastic simulations (CFS 2010)
3. Juvenile Production Index (JPI) method: based on real-time passage estimates at Red Bluff Diversion Dam (RBDD) rotary screw traps, which includes fry-to-smolt survival

Changes this year to the JPE included the survival rate from egg-to-fry (S1). Historically, the S1 survival rate has been based on an average over the last 15 years (i.e., 0.27). This rate was calculated from the adult escapement and passage estimates at RBDD. However, due to the extreme hydrologic conditions in 2014, mortality was very high at this early life-stage and S1 survival was calculated at 0.05 (or 95% mortality) based on the JPI (enclosure 2). Since neither

the NMFS nor CFS models accurately represent the extreme conditions and associated loss of juveniles due to high temperatures that occurred in 2014, NMFS decided that using the JPI method was a better fit to the actual hydrologic conditions that winter-run experienced.

In addition, each of the three methods, above, was calculated using three different survival estimates to the Delta (S2) based on the latest winter-run acoustic tag information, for a total of nine separate JPE estimates (Table 1). The method and the survival rate to the Delta (S2) were chosen before seeing the results of the JPE calculations in Table 1. The survival rates were obtained from 2013 and 2014 releases made in February at Caldwell Park near Redding. These survival data represent an improvement over previously-used late-fall run Chinook surrogate releases, which behave differently than winter-run as they migrate downstream."

"6/ S1 from RBDD passage estimate (JPI) of fry equivalents from Bill Poytress, USFWS 12/23/14. Includes temperature related mortality by back calculating survival to this point."

"8/ In-river survival (S2) estimated from 2014 winter-run acoustic tag data (S=0.42). Source: Arnold Ammann.

9/ LSNFH estimated release as of 12/03/14 (100% tagged & clipped). Tripled production due to drought year.

10/ Survival based on average of 2013 and 2014 acoustic tag studies (Ammann & Hassrick, pers. comm. 2014"

2014-02-21

"The methodology (i.e., Cormack-Jolly-Seber Model) used by CDFW to calculate winter-run escapement in 2013 was the same as in 2012. This method allows the calculation of confidence intervals. The 90 percent confidence interval for total estimate (6,075) is from 5,275 to 6,677 fish.

This year the Interagency Ecological Program's winter-run Project Work Team (WRPWT) conducted a technical review of the survival terms used to calculate the JPE based on the most recent acoustic tag studies in the Sacramento River. The WRPWT review found that the current JPE overestimates the number of juveniles entering the Delta on average by 400 percent (Table 1) based on four years of genetic studies at Chipps Island (Pyper et al. 2013). The positive identification of juvenile winter-run captured at Chipps Island allowed for comparisons between abundance estimates using actual observed data from 2008-2011 instead of modeled data. …

In addition, 2013 was the first year that in-river survival was directly estimated for hatchery released juvenile winter-run (using acoustic JSAT tags). This allowed for the comparison of survival rates from the 2013 data to those previously used in the calculation of the JPE. Direct estimates of survival are considered the best method of estimating natural juvenile winter-run survival and abundance between reaches on the Sacramento River. Previously, NMFS used survival rates indirectly by comparing the differences in ocean recovery rates of hatchery CWT coded wire tagged (CWT) late-fall run releases made at Battle Creek and in the Delta.

For the 2013 broodyear, NMFS has revised the survival terms in the JPE calculator based on the latest acoustic tag studies and abundance estimates at Chipps Island (Table 1). Smolt survival in from Red Bluff to the Delta was changed based on: (1) 2013 survival of acoustically-tagged hatchery winter-run (hydrologic conditions at release were similar to water year 2014); and (2) the average survival of 4 years (2007-2010) of acoustically-tagged late fall-run hatchery releases (excluding 2011 releases because it was a wet year with high survival). In 2013, the habitat and environmental conditions were considered to be similar to the dry year survival estimates from the acoustic tag data. Survival from Red Bluff to the Delta was derived by equally weighting the survival of both the winter-run and the average of the late fall-run releases (see enclosure 3). Various survival terms in the JPE calculator may be revised for the 2014 broodyear based on additional data from the 2014 release of acoustically-tagged winter-run."

"7/ Weighted average (50/50) of winter-run (2013) and late-fall run (2007-2010) acoustic tag data"

"9/ Hatchery survival estimated from 2013 acoustic tag study (Hassrick, unpublished)"

2013-01-31

"The methodology used to calculate the annual winter-run escapement was modified this year to the use of the Cormack-Jolly-Seber model and incorporate an estimate of uncertainty (see Enclosure 1). CDFW provided 90 percent confidence intervals around its escapement estimate upon which the JPE is calculated (i.e., 2,451 to 2,896). NMFS has accepted this change in methods as it improves the accuracy of the escapement estimate and provides an indication of the error associated with the estimate."

"1/ Total in-river escapement (90% CI) using modified Cormack-Jolly-Seber (CJS) model, includes 803 hatchery fish from CDFW letter 1/14/13, but does not include 93 removed for LSNFH.
2/ The number of females is derived from carcass surveys by applying the CJS model and then the number of males is derived using the sex ratio at the Keswick trapping station."

2012-01-26

"The methodology used to calculate the JPE was updated this year to reflect new information on juvenile survival. The Interagency Ecological Program winter-run Project Work Team (PWT) reviewed the last five years of data and pertinent new studies, along with annual updates made to various factors (e.g., fecundity rate, egg loss due to temperature, etc.) used in the calculation. Revisions to be made in the JPE calculations for water year 2012 incorporate the best available data from studies conducted since the last review in 2005. These changes include in-river survival estimates based on the latest acoustic tagging studies (see PWT notes). NMFS has accepted the PWT recommended revisions to in-river survival and included these changes in the calculation of the JPE for water year 2012. However, NMFS has not incorporated the PWT recommendation of a new tern1 for through-Delta survival (i.e., 35.9 percent survival from Freeport to Chipps Island). NMFS agrees that this additional term is important to determine the freshwater survival of juvenile winter-run to Chipps Island, but has determined that this additional survival estimate, if included, would change the definition of JPE as used in CVP/SWP Opinion. The definition of JPE as used in the CVP/SWP Opinion is the number of juvenile winter-run that enter the Delta. This line of reasoning is consistent with the notes from the December 7, 2011, PWT meeting, which stated that "The original need.for a relatively low take level was to account for indirect mortality of winter-run in the Delta. Now that indirect mortality in the Delta is incorporated into the JPE calculation, with the term for estimated survival through the Delta, the need/or the low take level was questioned."

Details of this year's calculation are described in the NMFS JPE Estimator Program (Enclosure 2) and its footnotes. As a method of checking the confidence of this year's estimate, NMFS compared the JPE to the U.S. Fish and Wildlife Service (USFWS) juvenile production index (JPI). The JPI is an independently-developed estimate based on real-time catch data of juvenile winter-run passage at Red Bluff Diversion Dam (RBDD). The USFWS estimated brood year total as of January 17, 2012, is 855,163 juvenile winter-run have passed RBDD (Enclosure 3). The JPE method estimated 512,192juvenile winter-run would reach RBDD, which is close to the lower 90 percent confidence interval of 611,409 in the JPI. In addition, NMFS hired a consultant in 2010, Cramer Fish Science Inc., to develop a model for the JPE that would provide some measure of the error associated with the calculation. This model was completed in 2011 and used for the first time last year. NMFS used the Cramer Fish Science model in checking the accuracy of the JPE Estimator Program in this year's estimate. The end point for juvenile winter-run survival in the Cramer model is Chipps Island. As stated above, NMFS defines the JPE as the number of winter-run entering the Delta, but used the JPE-based number of juveniles at Chipps Island to compare the results with the Cramer model. The NMFS JPE-based estimate of winter-run to Chipps Island (65,586) was within the modeled results for Dry and Critical year hydrology shown below (i.e., comparable to this year's hydrologic conditions)."

"2/ Sex ratio from 2011 Keswick trap instead of carcass surveys to account for bias in males. Number of females estimated is from DFG letter 12/6/11."

2011-01-12

"The methodology used to calculate the escapement was the same as that used last year with annual updates made to the various factors (i.e., fecundity, unspawned females, and survival rates)."

2010-02-24

"The methodology used to calculate the escapement was the same as that used last year with annual updates made to the various factors (i.e., fecundity, unspawned females, and survival rates)."

"2/ Sex ratio was based on data collected in 2009 carcass survey (USFWS pre-release report 1/25/10)"

2009-01-12

"The methodology used to calculate the escapement was the same as that used last year with annual updates made to the various factors (i.e., fecundity, unspawned females, and survival rates)."

"2/ Used sex ratio from Keswick trap (53%) instead of carcass surveys (78%) to account for bias in males""

2007-02-15

"The methodology used to calculate the escapement was the same as that used last year with annual updates made to the various factors (i.e., fecundity, unspawned females, and survival rates)."

2006-02-10

"The methodology used to calculate the escapement was the same as that used last year with annual updates made to the various factors (i.e., fecundity, unspawned females, and survival rates)."

"3/ Pre-spawn mortality (# adult females x 2.2% unspawned) Alice Low, DFG, per. Comm. WRSPWT 2005"

"7/ Average of ocean recoveries of paired Late-Fall Chinook from Coleman National Fish Hatchery released at Battle Cr. And Delta locations from 1994-2001 (USFWS, unpubl. Data, Chippstable 2005)"

"8/ Livingston Stone National Fish Hatchery conservation program release, all tagged and fin-clipped. Survival calculated based on 2003 winter-run tag recoveries at Chipps Is (n=34)"

2004-02-08

"The methodology used to calculate the escapement was the same as that used last year, except that it was updated with the most recent in-river survival rates. Details of the change made to this year's calculation are described in the JPE Estimator (Enclosure 2) and the survival estimates (Enclosure 3)."

2004-01-21

"The adult escapement estimate for 2003-2004 has been derived from the upper Sacramento River carcass surveys conducted by the California Department of Fish and Game (CDFG). Only one change occurred this year in the methodology used to calculate the number of females. Details of the change made to this year's calculation are described in the enclosed spreadsheet (Enclosure 1) and copy of CDFG letter dated January 10, 2004 (Enclosure 2). NOAA Fisheries will provide the CALFED Science Program with all pertinent data used in the calculation of the 2003-2004 estimate upon request, for their review and comment. This year's Sacramento River winter-run Chinook salmon JPE was developed by NOAA Fisheries after discussions with the Interagency Ecological Program Winter-run [Salmon] Project Work Team (WRSPWT)."

2003-02-19

2002-02-22

"After several meetings and discussion with CDFG and FWS, the methodology used to calculate the JPE has been changed from methods used in previous years. This change provides a more precise estimate and is based on the best information currently available. These changes are based, in part, upon a recommendation from CDFG (letter to Rodney McInnis dated January 16, 2002). The adult escapement estimate for 2001-2002 is derived from the upper Sacramento River carcass surveys, instead of the previously used Red Bluff Diversion Dam (RBDD) counts. The rationale for use of the carcass survey this year instead of the RBDD count is discussed in Attachment 1 of CDFG's letter. In addition, survival estimates for juvenile winter-run chinook salmon were revised based on more recent information from FWS (2001 and 2002). Details of the changes made to this year's calculation are described in a two page summary (Attachment 2). NMFS plans to provide the CALFED Science Program with all pertinent data used in the calculation of the 2001-2002 estimate, for their review and comment."

2001-01-29

"Two changes to this year's method of calculating the estimate were made at the discretion of NMFS. These changes involved using the grilse rate from Keswick instead of Red Bluff because of the higher number of winter-run adults observed there (102 vs 24 respectively), and increasing the fecundity based on the larger size of adults this year combined with actual egg counts at Livingston Stone Hatchery. It was felt that these changes would increase the accuracy of the juvenile production estimate."

"An additional 154,000 winter-run chinook salmon juveniles propagated at the Livingston Stone National Fish Hatchery were released in the upper Sacramento River on January 28, 1999. All hatchery -produced winter-run chinook salmon juveniles are marked with an adipose fin clip and coded-wire-tag (CWT). The NMFS will work with the Data Assessment Team (DAT) to monitor losses of CWT winter-run chinook salmon at the Delta fish salvage facilities."

"This year's juvenile winter-run chinook salmon production estimate was developed in coordination with the Interagency Ecological Program's Central Valley Salmon Project Work Team (CVSPWT). … On-going research by the U.S. Fish and Wildlife Service (FWS) and the California Department of Fish and Game (DFG) has provided additional site-specific information related to several of the parameters used to estimate juvenile chinook production. … The group concluded that use of the modified parameters did not significantly change this season's estimate of the number of juvenile winter-run salmon arriving in the Sacramento-San Joaquin Delta. Based on this sensitivity analysis and the preliminary nature of the results of FWS and DFG field investigations, the winter-run team recommended NMFS continue to use the traditional juvenile production estimation method during 1997."

"This total includes the contribution to winter-run chinook production from fish relocated from Battle Creek..."

"All hatchery-produced winter-run chinook salmon juveniles are marked with a coded-wire-tag (CWT). The initial identification of wild (unmarked) juvenile winter-run chinook salmon shall be based on the Delta size model developed by FWS in cooperation with the winter-run technical team. As additional information becomes available through genetic analysis of tissue samples and other fisheries monitoring programs in the Central Valley, this estimate of incidental take based on the Delta size model will be re-evaluated by the winter-run chinook technical team, the salmon genetics team, and the CALFED Operations Group Data Assessment Team."