2_Data.Rmd

<!-- ** Turn off numbering for inline tables ** -->

\captionsetup[table]{labelformat=empty,justification=raggedright,font=bf, singlelinecheck=false}

#Data

##Biological Parameters \label{bio-params}

\vspace{.5cm}
###Weight-Length

The weight-length relationship is based on the standard power function: $W = \alpha(L^\beta)$ where $W$ is individual weight (kg), $L$ is length (cm), and $\alpha$ and $\beta$ are coefficients used as constants.

To estimate this relationship, 12,778 samples with both weight and length measurements from the fishery independent surveys were analyzed. These included 6,354 samples from the NWFSC Combo survey, 5,085 from the Triennial survey, and 1,339 from the Hook and Line survey. All Hook and Line survey samples were from the Southern area, along with 910 samples from the other two surveys (Figure \ref{fig:weight-length}). 

A single weight-length relationship was chosen for females and males in both areas after examining various factors that may influence this relationships, including sex, area, year, and season. None of these factors had a strong influence in the overall results. Season was one of the bigger factors, with fish sampled later in the year showing a small increase in weight at a given length (2-6% depending on the other factors considered). However, season was confounded with area because most of the samples from the Southern area were collected from the Hook and Line survey which takes place later in the year (mid-September to mid-November) and the resolution of other data in the model do not support modeling the stock at a scale finer than a annual time step.

Males and females did not show strong differences in either area, and the estimated differences were in opposite directions for the two areas, suggesting that this might be a spurious relationship or confounded with differences timing of the sampling relative to spawning.

The estimated coefficients resulting from this analysis were $\alpha = 1.1843e-05$ and $\beta = 3.0672$.

\vspace{.5cm}
###Maturity And Fecundity

Maturity was estimated from histological analysis of 141 samples collected in 2016. These include 96 from the NWFSC Combo survey, 25 from mid-water catches in the NWFSC acoustic/trawl survey, 13 from the Hook and Line survey, and 7 from Oregon Department of Fish and Wildlife. The sample sizes were not adequate to estimate differences in maturity by area. Length at 50\% maturity was estimated at 42.49cm (Figure \ref{fig:maturity}) which was consistent with the range 37-45cm cited in the previous assessment [@Wallace2005].


\vspace{.5cm}
###Natural Mortality \label{nat-mort}

Hamel [-@Hamel2015] developed a method for combining meta-analytic approaches to relating the natural mortality rate M to other life-history parameters such as longevity, size, growth rate and reproductive effort, to provide a prior on M. In that same issue of ICESJMS, Then et al. [-@Then2015], provided an updated data set of estimates of M and related life history parameters across a large number of fish species, from which to develop an M estimator for fish species in general.
They concluded by recommending M estimates be based on maximum age alone, based on an updated Hoenig non-linear least squares estimator $M=4.899A_{max }^{-.916}$. 

The approach of basing M priors on maximum age alone was one that was already being used for west coast rockfish assessments. However, in fitting the alternative model forms relating M to Amax, Then et al. did not consistently apply their transformation. In particular, in real space, one would expect substantial heteroscedasticity in both the observation and process error associated with the observed relationship of M to Amax. Therefore, it would be reasonable to fit all models under a log transformation. This was not done.

Re-evaluating the data used in Then et al. [-@Then2015] by fitting the one-parameter Amax model under a log-log transformation (such that the slope is forced to be -1 in the transformed space (as in @Hamel2015), the point estimate for M is $M=5.4/Amax$

This is also the median of the prior. The prior is defined as a lognormal with mean $ln(5.4/Amax)$ and SE = 0.4384343. 

Initial natural mortality priors for these models were based on examination of the 99\% quantile of the observed ages from early in the time-series, before the full impact of fishing would have taken place. For the Northern model, these quantiles were approximately 35 years for females and 45 years for males, resulting in median M values of 0.15 and 0.12 for females and males. For the Southern model, the 99\% quantile of the early age observations were approximately 30 and 40 years for females and males, resulting in median M prior values of 0.18 and 0.135, respectively. In both models, M for males was represented as an offset from females.

\vspace{.5cm}
###Aging Precision And Bias

Age error matrices were developed for double-reads at the PFMC aging lab in Newport, OR and for double reads within the WDFW aging lab.  The Newport lab has done all of the Survey aging for the NWFSC, along with some commercial ages and  the 400 fish from the Small Study.  WDFW provided the bulk of recreational and commercial ages.  Between-lab differences in aging were minute, as were within-lab differences.  This result is supported by the primary age reader's assessment:  Yellowtail Rockfish are extremely easy to age (B. Kamikawa, pers. comm.).


##Biological Data and Indices

Data used in the Northern and Southern Yellowtail Rockfish assessments are summarized in Figures \ref{fig:data_plot.N} and \ref{fig:data_plot.S}.

Data sources for the two models are largely distinct.  Northern fisheries and surveys had very sparse data (if any) for the south and vice-versa.  Among the 12 data sources referenced below, only 2 data sources are common to both models.  These are the MRFSS/RecFIN recreational dockside survey, which focuses on California and Oregon, and the CalCOM California commercial dataset, which contributed data from the northern-most California counties (Eureka and Del Norte) to the Northern model.  The CalCOM data account for less than five percent of the commercial landings in the Northern model, and less than 1% of the biological samples. 

Commercial landings are not differentiated in either model.  For the Northern model, this is due to the very small portion (1.15 %) of the landings that are attributed to non-trawl gear.  For the Southern model, this is due to the paucity of data.

A description of each model's data sources follows.

<!-- ************FISHERY-DEPENDENT DATA************************************ -->


##Northern Model Data
\vspace{.5cm}

<!-- ***********Indices of abundance summary table************************* -->
```{r, results = "asis"}
    Data_sources = read.csv("./txt_files/02_Summary_N_Data_Sources.csv")
      colnames(Data_sources) = c("Source",
                                  "Landings",
                                  "Lengths",
                                  "Ages",
                                  "Indices",
                                  "Discard",
                                  "Type")

   # Index of abundance summary, create table   
     Data_sources.table = xtable(Data_sources, 
                                 caption = c("Summary of the data sources in the Northern model."),
                                 label = "tab:Data_sources",
                                 align = c("l","l", rep("c", 5),"r"),
                                 digits = 0) 

    # Print index summary table
      print(Data_sources.table, 
            include.rownames=FALSE, 
            caption.placement="top",
            sanitize.text.function = function(x){x})
```
<!-- ---------------------------------------------------------------------- -->

###Commercial Fishery Landings

**Washington and Oregon Landings**
The bulk of the commercial landings for Washington and Oregon came from the from the Pacific Fisheries Information Network (**PacFIN**) database.

**Washington Catch Information**   
The Washington Department of Fisheries and Wildlife (**WDFW**) provided historical Yellowtail catch for 1889–1980.  Landings for 1981-2016 came from the PacFIN database.  WDFW also provided catches for the period 1981 – 2016 to include the re-distribution of the unspeciated "URCK" landings in PacFIN; this information is currently not available from PacFIN.  

**Oregon Catch Information**   
The Oregon Department of Fisheries and Wildlife (**ODFW**) provided historical Yellowtail catch from 1892-1985.  ODFW also provided estimates of Yellowtail Rockfish in the in the un-speciated PacFIN "URCK" and "POP1" catch categories for recent years, and those estimates were combined with PacFIN landings for 1986-2016.

**Northern California Catch**   
The California Commercial Fishery Database (**CalCOM**) provided landings for the Northern model for the two counties north of  $40^\circ 10^\prime$ (Eureka and Del Norte) for 1969-2016.

**Hake Bycatch**   
The Alaska Fisheries Science Center (**AFSC**) provided data for Yellowtail bycatch in the hake fishery from 1976-2016.

###Sport Fishery Removals

**Washington Sport Catch**   
WDFW provided recreational catches for 1967 and 1975-2016.

**Oregon Sport Catch**   
ODFW provided recreational catch data for 1979-2016.

**MRFSS and RecFIN**
Data from Northern California came from the Marine Recreational Fisheries Statistical Survey (**MRFSS**) and from the Recreational Fisheries Information Network (**RecFIN**).  These are dockside surveys focused on California and Oregon.  MRFSS was conducted from 1980-1989 and 1993-2003, RecFIN from 2004 to the present.

###Estimated Discards

**Commercial Discards**   
The West Coast Groundfish Observing Program (**WCGOP**) is an onboard observer program that has extensively surveyed fishing practices since 2002, with nearly 100% observer coverage in the trawl sector in recent years.  WCGOP provided discard ratios for Yellowtail Rockfish from 2002 to 2015.

**Pikitch Study**  
The Pikitch study was conducted between 1985 and 1987 [@Pikitch1988].  The northern and southern boundaries of the study were $48^\circ 42^\prime$ N latitude and $42^\circ 60^\prime$ N. latitude respectively, which is primarily within the Columbia INPFC area [@Pikitch1988 ; @Rogers1992].  

Participation in the study was voluntary and included vessels using bottom, midwater, and shrimp trawl gears. Observers of normal fishing operations on commercial vessels collected the data, estimated the total weight of the catch by tow and recorded the weight of species retained and discarded in the sample.  
 
Pikitch study discards were aggregated due to small sample size and included in the data as representing a single year mid-way through the study.

###Abundance Indices \label{abundance-indices}

Two fishery-dependent abundance indices were developed for this analysis that were discovered in course of review to be based on incomplete information about how the commercial trawl and Hake fisheries were operated in the late 1980s through the late 1990s.  Representatives from WDFW and from the Council's Groundfish Advisory Panel raised numerous concerns about the Commercial Trawl Index and the Hake Bycatch Index, respectively, and they were ultimately removed from the model.

The commercial trawl index used the species composition of catch to infer the potential for Yellowtail Rockfish in each haul, however the way in which market categories were changing throughout the period of interest made the species composition of catch  led to concerns about the consistency of the resolution of catch reporting over time (Theresa Tsou, pers. comm.). The Hake fishery was explained to have had greater heterogeneity among the boats used in the fishery than had been assumed in developing the index (Dan Waldeck, pers. comm.).

Give the unknown impact of incomplete information used in developing these indices which could not be adequately addressed during the review, and that there were fishery-independent indices covering the period in question, the decision was made to withdraw these two indices.  They are described in Appendix B for completeness.

<!-- ************FISHERY-INDEPENDENT DATA*********************************** -->

###Fishery-Independent Data

**Alaska Fisheries Science Center (AFSC) Triennial Shelf Survey**    

Research surveys have been used since the 1970s to provide fishery-independent information about the abundance, distribution, and biological characteristics of `r spp`.  A coast-wide survey was conducted in 1977 [@Gunderson1980] by the Alaska Fisheries Science Center, and repeated every three years through 2001.  The final year of this survey, 2004, was conducted by the NWFSC according to the AFSC protocol. We refer to this as the **Triennial Survey**.

The survey design used equally-spaced transects from which searches for tows in a specific depth range were initiated. The depth range and latitudinal range was not consistent across years, but all years in the period 1980-2004 included the area from 40$^\circ$ 10'N north to the Canadian border and a depth range that included 55-366 meters, which spans the range where the vast majority of Yellowtail encountered in all trawl surveys. Therefore the index was based on this depth range.  The survey as conducted in 1977 had incomplete coverage and is not believe to be comparable to the later years, and is not used in the index.

An index of abundance was estimated based on the VAST delta-GLMM model as described for the NWFSCcombo Index above. In this case as well, Q-Q plots indicated slightly better performance of the lognormal over gamma models for positive tows (Figure \ref{fig:VAST_QQ}). The index shows a gradual decline from 1980 to 1992 followed by high variability in the final 4 points spanning 1995-2004. The distribution of estimated densities was more variable that in the NWFSCcombo survey, but the relatively higher densities in the northern part of the coast were similar (Figure \ref{fig:VAST_Dens}).

**Northwest Fisheries Science Center West Coast Groundfish Bottom Trawl Survey**

In 2003, the NWFSC took over an ongoing slope survey the AFSC had been conducting, and expanded it spatially to include the continental shelf. This survey, referred to in this document as the **NWFSCcombo Survey**, has been conducted annually since. It uses a random-grid design covering the coastal waters from a depth of 55 m to 1,280 m from late-May to early-October [@Bradburn2011 ; @Keller2017]. Four chartered industry vessels are used each year (with the exception of 2013 when the U.S. federal-government shutdown curtailed the survey). 

The data from the NWFSCcombo survey was analyzed using a spatio-temporal delta-model [@Thorson2015], implemented as an R package VAST [@Thorson2017] and publicly available online (https://github.com/James-Thorson/VAST).  Spatial and spatio-temporal variation is specifically included in both encounter probability and positive catch rates, a logit-link for encounter probability, and a log-link for positive catch rates.  Vessel-year effects were included for each unique combination of vessel and year in the database.

The patterns of estimated density for each year showed consistently higher biomass in the Northern part of the Northern area (Figure \ref{fig:VAST_Dens}). Both lognormal and gamma distributions were explored for the positive tows and produced similar results with the lognormal model showing slightly better patterns in Q-Q plot (Figure \ref{fig:VAST_QQ}). The index shows variability with an overall gradual increase from 2003 to 2013 with high estimates near the end of the time series in 2014 and 2016 (Figure \ref{fig:VAST_compare_NWFSCcombo}). A design-based index extrapolated from swept area densities without any geostatistical standardization shows a more dramatic increase from 2015 to 2016 (Figure \ref{fig:VAST_compare_NWFSCcombo})

Length and age compositions were also developed from this survey.


<!--************BIOLOGICAL DATA*********************************************-->
###Biological Samples

**Length And Age Compositions**   
Length composition data were compiled from PacFIN for Oregon and Washington for the Northern model and combined
with raw (unexpanded) length data from CalCOM for the two California counties north of 40$^\circ$ 10'N (Eureka and Del Norte counties).


Length compositions were provided from the following sources:

\vspace{.5cm}

<!-- ***********Indices of abundance summary table************************* -->
```{r, results = "asis"}
    Length_sources = read.csv("./txt_files/02_Summary_N_Length_sources.csv")
      colnames(Length_sources) = c("Source",
                                  "Type",
                                  "Lengths",
                                  "Tows",	
                                  "Years")

   # Lengths summary, create table   
     Length_sources.table = xtable(Length_sources, 
                                  caption = c("Summary of the time series of lengths used in the stock assessment."),
                                 label = "tab:Length_sources", digits = 0)    

    # Print index summary table
      print(Length_sources.table, 
            include.rownames=FALSE, 
            caption.placement="top",
            sanitize.text.function = function(x){x})
```
<!-- ---------------------------------------------------------------------- -->

The expanded table detailing the length data is Table \ref{tab:Northern_length}.  The names in this table are truncated so that the data can be compared side-by-side, but should be obvious: "C.Trawl" is the Commercial Trawl fishery.

Age structure data were available from the following sources:
\vspace{.5cm}

<!-- ***********Indices of abundance summary table************************* -->
```{r, results = "asis"}
    Age_sources = read.csv("./txt_files/02_Summary_N_Age_sources.csv")
      colnames(Age_sources) = c("Source",
                                  "Type",
                                  "Ages",
                                  "Tows",	
                                  "Years")

   # Index of abundance summary, create table   
     Age_sources.table = xtable(Age_sources, 
                                  caption = c("Summary of the
                                              time series of age data used in the stock
                                              assessment."),
                                 label = "tab:Age_sources",
                                 digits = 0)    

    # Print index summary table
      print(Age_sources.table, 
            include.rownames=FALSE, 
            caption.placement="top",
            sanitize.text.function = function(x){x})
```
<!-- ---------------------------------------------------------------------- -->

The expanded table detailing the ages can be found in Table \ref{tab:Northern_age}

<!-- ************FISHERY-DEPENDENT DATA************************************ -->
\FloatBarrier
\newpage
\clearpage
\FloatBarrier

##Southern Model Data

\vspace{.5cm}

<!-- ***********Data sources summary table************************* -->
```{r, results = "asis"}
    Data_sources = read.csv("./txt_files/02_Summary_S_Data_Sources.csv")
      colnames(Data_sources) = c("Source",
                                  "Landings",
                                  "Lengths",
                                  "Ages",
                                  "Indices",
                                  "Discard",
                                  "Type")

   # Index of abundance summary, create table   
     Data_sources.table = xtable(Data_sources, 
                                 caption = c("Summary of the data source in the Southern model."),
                                 label = "tab:Data_sources",
                                 align = c("l","l", rep("c", 5),"r"),
                                 digits = 0)    

    # Print index summary table
      print(Data_sources.table, 
            include.rownames=FALSE, 
            caption.placement="top",
            justify=c("l","c","c","c","c","c","c"),
            sanitize.text.function = function(x){x})
```
<!-- ---------------------------------------------------------------------- -->

###Commercial Fishery Landings

**California Commercial Landings**   
The California Commercial Fishery Database (**CalCOM**) provided landings in California south of 40$^\circ$ 10'N for 1969-2016. Because this fishery is known to have begun in the 1880s, we added catch as a linear ramp from 1889 (the earliest catch in the Northern model) to the 2016 value.

**Historical Data**
A reconstruction of the historical commercial fishery south of Cape Mendocino was provided by the Southwest Fisheries Science Center (**SWFSC**) for 1916-1968 [@Ralston2010].


###Sport Fishery Removals

**MRFSS Estimates and RecFIN**  
The California Department of Fish and Wildlife (**CDFW**) provided estimated Yellowtail removals for the Marine Recreational Fisheries Statistical Survey (**MRFSS**) from 1980-1989, 1993-2003.  The Recreational FIsheries Information Network, (**RecFIN**) provided landings for 2004-2016.

**Historical Data**
A reconstruction of the historical recreational fishery south of Cape Mendocino was provided by the Southwest Fisheries Science Center (**SWFSC**) for 1928-1980 [@Ralston2010].  Yellowtail Rockfish have been identified as a sigificant component of the catch since the earliest days of the fishery.  The catch at Monterey in 1935 was 7.9% Yellowtail Rockfish (with Bocaccio and Chillipepper Rocfish comprising 70.2%) [@FishBull1936], at a time of rapid expansion in the fishery [@Phillips1939].


**Small Research Study**
California Cooperative Groundfish Survey CPFV Sampling, 1978-1984.  Commercial port samplers with the California Cooperative Groundfish Survey sampled landings from CPFVs operating north of Point Conception in the late 1970s and early 1980s.  This data set represents the only source of sex-specific length information available for Yellowtail Rockfish in California. 

###Estimated Discards

No discard data were available for the Southern model.

###Abundance Indices

**MRFSS Index**  
From 1980-2003, the Marine Recreational Fisheries Statistics Survey (MRFSS) executed a dockside (angler intercept) sampling program in Washington, Oregon, and California. Data from this survey are available from the Recreational Fisheries Information Network (RecFIN). The Recreational Fishieries Information Network (RecFIN) serves as a repository for recreational fishery data for California, Oregon, and Washington (http://www.recfin.org). RecFIN is currently undergoing a transition to a relational database design. Catch estimates for years 1980-2003 were downloaded prior to the transition.

MRFSS-era recreational removals for California were estimated for two regions: north and south of Point Conception. No finer-scale estimates of landings are available for this period. Catches were downloaded in weight.  MRFSS sampling was temporarily suspended from 1990-1992, and we left the catch in these years as missing values rather than performing any interpolation.

MRFSS was replaced with the California Recreational Fisheries Survey (CRFS) beginning January 1, 2004. Among other improvements to MRFSS, CRFS provides higher sampling intensity, finer spatial resolution (6 districts vs. 2 regions), and onboard CPFV sampling. Estimates of catch from 2004-2016 were provided by RecFIN staff. We  and aggregated CRFS data to match the structure of the MRFSS data.

**California Onboard Surveys**  

*1987-1998*
This assessment uses two indices derived from onboard CPFV observer data and collected during different time periods of the fishery. The primary advantage of onboard observer data is that catch and effort data are based on individual fishing stops (or “drifts”), rather than aggregated at the trip level, and information about actual fishing locations is available, rather than port of landing or interview site. This location information, when combined with recent maps of rocky reef habitat, allows us to associate catch rates with reefs of known area and produce habitat area-weighted CPUE indices.

The CDFW (formerly CDFG) Central California Marine Sport Fish Project sampled the Northern and Central California CPFV fleet using onboard observers from 1987-1998. Observers recorded the total catch (kept and released fish) of a subset of anglers during each fishing drift. Catches from drifts occurring at a single CDFW fishing site were aggregated into a “fishing stop.” Each stop in the database is associated with the closest reef structure. Retained fish were measured at the end of the fishing day. Additional details about the survey design, data collected, spatial associations between fishing stops and reef habitat, and the structure of the relational database are described in [@Monk2016].

*1999-2016*
California onboard CPFV observer data, spanning the years 1999-2016 was provided by the SWFSC [@Monk2014]. Each observation included a unique trip and drift identifier, and a subset of anglers was observed at each drift. Drift-level information included catch of blue rockfish in numbers (kept and discarded) including zeros, number of observed anglers, time fished (in minutes), location where drift began (latitude and longitude), year, month, county, CRFS district, depth (in feet), distance from nearest reef habitat (in meters), and unique reef identified.

Indices from these datasets were provided by the SWFSC according to the methods described in [@Monk2016].

**Juvenile Pelagic Index**
The Fishery Ecology Division of the Southwest Fishery Science Center has conducted a standardized pelagic juvenile trawl survey during May-June every year since 1983 [@Williams2002]. The primary purpose of the survey is to estimate the abundance of pelagic juvenile rockfishes (Sebastes spp.) and to develop indices of year-class strength for use in groundfish stock assessments on the U. S. West Coast. The survey samples young-of-the-year rockfish when they are ~100 days old, an ontogenetic stage that occurs after year-class strength is established, but well before cohorts recruit to commercial and recreational fisheries [@Ralston2013],[@Sakuma2016]. 

The survey has encountered tremendous interannual variability in the abundance of the ten species that are routinely indexed, as well as high apparent synchrony in abundance among the ten most frequently encountered species [@Ralston2013].

The abundance index was developed using a delta-GLM within a hierarchical Bayesian framework using the R package \emph{rstanarm}, and used as an indicator of age-0 fish.  Further details of the analysis are available in Appendix C.


<!-- ************FISHERY-INDPENDENT DATA*********************************** -->

###Fishery-Independent Data

**Hook and Line Survey**  
The NWFSC Hook and Line survey provided data for an index in the Southern California Bight from 2004-2016.  The Yellowtail index of abundance is based on numbers of fish provided by the Northwest Fisheries Science Center’s Hook and Line survey in the Southern California Bight. This index used survey data from 2004-2016 and was created following the methods put forth in [@Harms2010], after those methods were updated to create models with greater parsimony.  In addition, the final index is averaged over all crew staff and sites. (Note that vessels are confounded with crew staff.)  Two vessels were employed for the survey in 2004-12 and three vessels in 2013-16.  Data from inside the Cowcod Conservation Area (CCA) was not used in this index.

The 2016 index value differs from previous years in that certain variables such as sea surface temperature and tide flow were not available for this analysis, due to an ongoing upgrade in data collection software.

Variables in the binomial model with logit link:

NumYTRK ~ Year + SiteName + CrewStaff + DropNum + HookNum + poly(WaveHt.m,3) + poly(SwellHt.m, 3) + poly(PctLiteR, 2) + poly(MoonPct, 3)

Where poly(…, X) identifies the Xth degree polynomials for continuous variables, and a colon (‘:’)  represents an interaction term. ‘PctLite’ is the percent of daylight that has passed at the time the drop occurs on a given day.

The posterior median index values and their associated posterior log-SD are from a converged, 2.5 million draw MCMC.
 

<!--************BIOLOGICAL DATA*********************************************-->
###Biological Samples
Length composition samples were available for the Southern model from 5 sources, and ages from 3.


Length compositions were provided from the following sources:

\vspace{.5cm}

<!-- ***********Length summary table************************* -->
```{r, results = "asis"}
    Length_sources = read.csv("./txt_files/02_Summary_S_Length_sources.csv")
      colnames(Length_sources) = c("Source",
                                  "Type",
                                  "Lengths",
                                  "Tows",	
                                  "Years")

   # Lengths summary, create table   
     Length_sources.table = xtable(Length_sources, 
                                  caption = c("Summary of the time series of lengths used in the stock assessment."),
                                 label = "tab:Length_sources_South", digits = 0)    

    # Print index summary table
      print(Length_sources.table, 
            include.rownames=FALSE, 
            caption.placement="top",
            sanitize.text.function = function(x){x})
```
<!-- ---------------------------------------------------------------------- -->

The expanded table with detailed lengths is Table \ref{tab:Southern_length}

Age structure data were available from the following sources:


\vspace{.5cm}

<!-- ***********Indices of abundance summary table************************* -->
```{r, results = "asis"}
    Age_sources = read.csv("./txt_files/02_Summary_S_Age_sources.csv")
      colnames(Age_sources) = c("Source",
                                  "Type",
                                  "Ages",
                                  "Years")

   # Index of abundance summary, create table   
     Age_sources.table = xtable(Age_sources, 
                                  caption = c("Summary of the
                                              time series of age data used in the stock
                                              assessment."),
                                 label = "tab:Age_sources_South",
                                 digits = 0)    

    # Print index summary table
      print(Age_sources.table, 
            include.rownames=FALSE, 
            caption.placement="top",
            sanitize.text.function = function(x){x})
```
<!-- ---------------------------------------------------------------------- -->

The expanded table with detailed age information is Table \ref{tab:Southern_Age}

\clearpage


###Environmental Or Ecosystem Data Included In The Assessment

No environmental or ecosystem data were included in either model.