Draft Summary of Data Used in JPE

Overview

The spring run juvenile production estimate (SR JPE) contains a suite of submodels that rely on different datasets. This document is currently focused on Chinook monitoring data that will be used to fit the stock recruit model and within season models. The within season model relies on juvenile data and the stock recruit relies on both juvenile and adult datasets.

Additional sections will be added to highlight survival, genetics, and covariate data.

Modeling Data Selection

Gathering and identifying spring run Chinook monitoring data to use in the spring run juvenile production estimate (SR JPE) was an iterative, multistep process.

Filtering applied to juvenile RST data - necessary for both within season and stock recruit model data.

  1. We acquired, processed and described monitoring that we presented in workshops to define a ruleset for evaluating years to exclude based on missing data (see years to include vignette for an in depth overview of this process). This filtered juvenile dataset is used for both the within season and stock recruit models.

Additional filters required for the stock recruit:

  1. We selected data where there was both adult and juvenile data.
  2. We selected a single adult dataset to use for each year and stream.

The diagram below visualizes this process and shows the total number of datasets we retained along the way.

Distinct annual datasets are identified by looking at distinct year, stream, and monitoring types

Monitoring Data Overview

Extensive spring run Chinook monitoring is conducted throughout the Central Valley. Monitoring programs have been collecting data since the 1990s and monitoring has increased through time to expand monitoring types and data collection coverage across the seven tributaries and mainstem Sacramento River (identified as the geographic scope for the SR JPE).

The plot below shows all monitoring coverage of rotary screw trap, adult upstream passage, adult holding, adult redd, and adult carcass survey data.

Plot does not include Sacramento River mainstem monitoring. This data may also be utilized in the SR JPE but will not be included in the stock recruit models. For some monitoring types data is collected annually or monthly communicated through gaps in the colored lines.

Excluding Incomplete Years

Although the SR JPE aims to utilize all available data, there are some years where data is sparse and is excluded from analysis. This decision process is documented in the Years to Include Analysis.

Juvenile datasets filtered to exclude incomplete years.

Gray boxes indicate that monitoring occurred but data is not being utilized. Data may be excluded because the data was considered incomplete for a given monitoring type and year.

Adult dataset filtered to exclude incomplete years

Gray boxes indicate that monitoring occurred but data is not being utilized. Data may be excluded because the data was considered incomplete for a given monitoring type and year.

Summary Table of Years to Use for Within Season

The summary table below describes which years remain after filtering out incomplete juvenile data. This table accounts for years excluded based on data quality and completeness considerations.

Stream Years to Include in Model Total Number of Years
battle creek 1999 - 2000, 2002 - 2006, 2008 - 2014, 2016 - 2025 24
butte creek 1995 - 1996, 1999 - 2004, 2007 - 2008, 2012 - 2018, 2020 - 2025 23
clear creek 1998 - 2025 28
deer creek 1995 - 1996, 2000 - 2003, 2005, 2007, 2009 - 2010, 2023 - 2025 13
feather river 1997 - 2020, 2022 - 2025 28
mill creek 1995 - 1996, 2000 - 2003, 2005 - 2008, 2010, 2022 - 2025 15
sacramento river 1995 - 2001, 2003 - 2025 30
yuba river 1999, 2001 - 2002, 2004 - 2009, 2022 - 2025 13

This table shows us that there are a total of 174 stream years to use for developing the SR JPE stock recruit model.

Filtering data to years that contain both juvinile and adult monitoring

After excluding incomplete years, the stock recruit data requires additional filtering. In order to utilize data for the stock recruit models, we need both juvenile data inputs (RST data) and adult data inputs (adults) for each year. The plot below shows the time windows where there is overlapping juvenile and adult monitoring on each tributary. While this limits the amount of data, it still provides a few large windows for most monitoring programs.

Gray boxes indicate that monitoring occurred but data is not being utilized for the stock recruit. Data may be excluded because there is not adult and juvenile annual overlap or because the data was considered incomplete for a given monitoring type and year.

Summary Table of Years to Use for Stock Recruit

The summary table below describes which years have overlapping RST and adult data. This table accounts for years excluded based on data quality and completeness considerations.

Stream Years to Include in Model Total Number of Years
battle creek 2005 - 2006, 2009 - 2014, 2016, 2020 - ongoing 11
butte creek 2001 - 2004, 2007 - 2008, 2013 - 2014, 2016 - 2018, 2020 - ongoing 14
clear creek 2003 - 2010, 2012 - 2015, 2021 - ongoing 13
deer creek 1992, 1995 - 1996, 2000 - 2003, 2005, 2007, 2009 - 2010, 2022 - ongoing 11
feather river 2010 - 2012, 2018 - 2020, 2022 - ongoing 7
mill creek 2000 - 2003, 2005 - 2008, 2010 9
yuba river 2004 - 2008 5

This table shows us that there are a total of 70 stream years to use for developing the SR JPE stock recruit model.

Selecting Adult Data Type to Use

There are 35 stream year combinations where there are multiple adult data types to pick from. These occur for: “butte creek”, “battle creek”, “clear creek”, and “feather river.”

The process of determining the appropriate adult data to use in stock recruit modeling for the SR JPE is currently ongoing. This information will be updated as the Modeling Team evaluates the tradeoffs associated with various data types in more detail.

We used the following logic to select data to use in these cases:

  • if stream = "butte creek" we used carcass estimate data (carcass estimate methods include mark recapture)
  • if stream = "battle creek" we used…TODO
  • if stream = "clear creek" we used…TODO
  • if stream = "feather river" we used…TODO
Year Stream Sampling Type Selected Adult Data Type
1999 clear creek rst , redd , upstream_estimate TBD
2001 butte creek rst , carcass_estimate, holding carcass estimate
2001 clear creek rst , redd , upstream_estimate TBD
2002 battle creek rst , redd , upstream_estimate TBD
2002 butte creek rst , carcass_estimate, holding carcass estimate
2002 clear creek rst , redd , upstream_estimate TBD
2003 battle creek rst , redd , upstream_estimate TBD
2003 butte creek rst , carcass_estimate, holding carcass estimate
2003 clear creek rst , redd , upstream_estimate TBD
2004 battle creek rst , redd , upstream_estimate TBD

… with 51 more rows

Proposed Data to Use in SR JPE

The table below compiles all the filtering above to provide a summary of the data utilized for the stock recruit models for each stream and year.

Year Stream Juvenile Data Type Adult Data Type
1995 deer creek rst holding
1996 deer creek rst holding
1996 mill creek rst redd
1998 clear creek rst upstream_estimate
1999 battle creek rst upstream_estimate
1999 clear creek rst TBD
2000 battle creek rst upstream_estimate
2000 clear creek rst upstream_estimate
2000 deer creek rst holding
2000 mill creek rst redd

… with 101 more rows