Minimum Survival Rates
We calculated the minimum survival rates across four river reaches and assessed detection probabilities for each cross-channel array Actual detections at each bridge do not fully represent successful fish migration, as some tagged fish were not detected at the bridge but were later recorded by receivers downstream To estimate the minimum number of fish that successfully migrated to each bridge, we combined actual detections with those from lower reaches This approach also informed the detection probability for each cross-channel array These estimates are conservative, as some fish may have successfully navigated through a reach without any subsequent detection Reach-specific survival is defined as the percentage of fish detected at the beginning of a reach that were subsequently detected at the next bridge or reach.
Transit Times and Rates
The "transit time" of a fish refers to the duration between its last detection at one listening station and its first detection at another This interval can be converted into a movement rate measured in meters per second (m/s) by dividing the distance traveled by the elapsed time For subsequent analyses, distances between locations were sourced from the NMFS Tracking Consortium database or estimated using Google Earth maps.
We measured the acclimatization period by analyzing the time taken from Elkhorn Boating Launch to Benicia Bridge, avoiding movements in this section To assess overall transit times, we focused on individuals from each species detected at both Benicia Bridge and the Golden Gate Given the non-normal distribution of transit times for both species, we plotted the frequency distribution and calculated the median for each.
To determine reach-specific transit times, we employed two complementary methods by analyzing individuals detected at four gateway arrays: Benicia, Carquinez, Richmond, and Golden Gate We calculated the overall transit time for each species by measuring the interval between the last detection at Benicia Bridge and the first detection at Golden Gate Additionally, we assessed transit times for each river section: Benicia to Carquinez, Carquinez to Richmond, and Richmond to Golden Gate It is crucial to note that the total transit times for each reach may not sum to the overall transit time, as individual fish may linger at bridges for several minutes or move upstream to previous reaches.
The transit rate represents the overall speed of movement through a river section, calculated based on a consistent speed and the shortest path distance To determine these transit rates, we converted transit times using estimated shortest distances between bridges.
•Richmond to Golden Gate: 17.72 km
We conducted a repeated measures MANOVA test to identify significant differences in transit times and rates among various species and river sections These differences may be attributed to factors such as swimming speeds, resting phases, current velocity and direction, or the specific pathways taken through each river section.
We conducted an independent analysis of each river section by considering all individuals detected at both the entrance and exit arrays to enhance our sample size, as some fish were only detected at certain gateway arrays For data that did not follow a normal distribution, we presented medians and quartiles To assess differences in transit rates among species and river reaches, we employed non-parametric statistics, specifically the Kruskal-Wallis test Additionally, to address the bimodal distribution observed in transit rates, we categorized individuals of each species as 'fast' or 'slow' based on the cutoff rates identified in the frequency graphs.
We analyzed the correlation between individual fish weight and transit rates, focusing on the prominent bimodal distribution observed in the upper reach To explore the relationship between transit rates and river flow, we compared the transit rates of fish in the Benicia-Carquinez reach with the river discharge volume at Rio Vista, the nearest data point Additionally, we examined the timing of detections for both fast and slow-moving fish at Decker Island, the closest underwater receiver to Rio Vista, in relation to the river discharge at that location.
At the primary dredge placement site in San Pablo Bay, we analyzed the instantaneous transport rates between the SP Control and SP Array 1, located just upstream of the placement area.
The study utilized SP Array 1 and SP Array 2 to analyze fish movement through a river section by employing a non-parametric test to compare transit rates with the overall fish population Notably, Chinook salmon exhibited a significant tendency to undertake multiple trips in both upstream and downstream directions, complicating the determination of precise transit rates Consequently, individual positions were tracked over time, allowing for the estimation of transit rates for each movement Additionally, tide height data from the nearby Richmond Station was overlaid to examine the correlation between tidal states and the frequency of upstream and downstream movements In contrast, steelhead showed no detection at both gateway arrays, Carquinez and others.
Richmond) displayed this upstream/downstream movement, we performed a similar analysis on all fish which did display this behavior.
Residency and exposure times
Residency indicates the duration a fish spends at a specific listening station, but for salmonid smolts migrating downstream to the sea, the term 'exposure time' is more suitable This refers to the time these fish spend at key locations in the river system, including dredged areas like marinas, shipping channels, and dredge material placement sites such as those in Alcatraz and San Pablo Bay.
To accurately determine exposure time at a specific site with one receiver, individual detections must be converted into time intervals between each detection It is crucial to establish a cut-off time interval, which distinguishes subsequent detections as new visits rather than extensions of previous ones This cut-off is influenced by factors such as the tag's blanking interval, the receiver's detection efficiency—which can vary over time due to environmental changes—and the species' behavior A viable method for estimating this cut-off time interval is through log-survivorship analysis.
We employed the method established by Fagan & Young (1978) to analyze the intervals between detections, grouping them by minute For each species, we aggregated data from all receivers at the dredged sites, assuming consistent behavior across these locations Initially, we created a length frequency graph, followed by plotting the logarithm of the frequency of detections exceeding each time interval This distribution follows a negative exponential pattern, where the plotted intervals between detections form a curve, with the slope reflecting the likelihood of a detection occurring after the previous one.
The inflection point of a detection curve indicates a behavioral change, with intervals shorter than this point suggesting multiple detections during a single visit Missed detections can occur due to various factors, such as tag interference, signal decay, or external noise In contrast, intervals longer than the inflection point typically indicate a genuine absence from the site rather than missed detections This inflection point is often visually identifiable, as demonstrated by Klimley and Holloway (1984).
After establishing a cutoff interval for each species, we calculated the duration of visits at each site and assigned a mean blanking interval value for single detections—30 seconds for Chinook salmon and one minute for steelhead This report highlights the total exposure time at each site within the receiver's coverage area, which is approximately 17,620 m², assuming a detection distance of 75 meters Consequently, when evaluating exposure across an entire site, it is essential to adjust the time based on the specific area involved.
We analyzed the percentage of fish that crossed the Benicia Bridge into the study area and assessed their detection rates at various dredged sites, along with the potential exposure durations.
Due to data limitations, we conducted separate analyses for the two primary dredge material placement sites None of the fish from our study were detected at the Alcatraz placement receivers, and the Alcatraz Control 1 receiver was lost, preventing meaningful comparisons at this site Consequently, we estimated exposure time for each individual in the same manner as for the dredged sites.
At the San Pablo dredge material placement site, we analyzed the exposure time of fish in areas affected by dredged material placement compared to a control area We combined fish detections from SP Array 1 and 2, receivers A-D into the Disposal Site, and from receivers E-H into the Control Site Given that the sites are monitored by multiple receivers, we calculated exposure time by summing consecutive detections at each site, assuming any interval between detections occurred within the site's influence For single detections, we assigned a duration based on the average blanking interval for the respective fish tags (30 seconds for Chinook salmon and one minute for steelhead) A matched pairs test was conducted to assess whether salmonid smolts spent more time at either site.
Figure 13 Receivers used to determine exposure time of salmonids smolts to San Pablo Bay Disposal Site (SP
Array 1 & 2, receivers A-D) and adjacent control site (SP Array 1 & 2, receivers E-H).
Migratory pathways
Receivers deployed across various channels in the upper and lower bay were analyzed to assess the depth preferences of smolts during their outmigration, focusing on the Benicia array.
The study analyzed detections of various species across multiple arrays, including the Bridge, Carquinez Bridge, San Pablo Arrays (Control, 1, and 2), San Pablo/Richmond Bridge, and Golden Gate Arrays (west and east line) Data was visualized through graphs depicting species detections and channel depth profiles for each cross-section array Additionally, regression analyses were conducted to assess overall depths across all sites for each species, as well as individual site depths by species.
Comparative analyses of Chinook salmon and steelhead detections were conducted between shoals and dredged channels The data exhibited equal variances; however, it followed a non-normal distribution, prompting the application of a non-parametric Wilcoxon test for evaluation.
Interannual variations
We conducted a comparison of the origins of salmonid smolts utilized in previous years' data analyses with those from the current year To ensure consistency in transit times and rates with past analyses, we calculated the transit times and rates for Chinook salmon and steelhead between Richmond Bridge and the Golden Gate, as detailed in section 2.7.2.
Due to variations in receiver locations and experimental design, a detailed comparison of smolt exposure time at the SP and Alcatraz arrays was not feasible Instead, we analyzed the number of individuals detected and the mean exposure time at comparable sites, focusing on either identical receiver locations or closely proximate alternatives.
Green sturgeon analyses
In 2008/9, the California Fish Tracking Consortium database revealed the presence of green sturgeon in the Bay Area, particularly during the migration of salmonid smolts from March to June 2009 We assessed the proportion of green sturgeon detected at dredged and dredge placement sites and estimated their exposure time at these locations using the methodologies described in section 2.7.3.
Out of 500 Chinook salmon tagged and released at Elkhorn Landing on the Sacramento River, 309, representing 61.8%, successfully reached the study area's starting point at the bottom of Suisun Bay, near the Benicia Bridge.
Out of 500 tagged steelhead, 47.6% successfully returned to the same location, while 12.8% of tagged steelhead and 17.8% of Chinook salmon managed to migrate to the Golden Gate Detection efficiencies at various bridges ranged from 47% to 94%, although the Golden Gate's estimates were inflated due to reliance on the Point Reyes array for assessing fish survival to the ocean In total, 64 tagged steelhead and 89 Chinook salmon reached the Golden Gate array, with 52% of steelhead and 62% of Chinook salmon passing through the SF 10 dredge placement site in San Pablo Bay.
Table 3 Numbers of salmonid smolts detected at each bridge array and estimated detection efficiencies Figures for
Golden Gate in italics are estimates based on fish detected at Point Reyes.
The Elkhorn Boating Facility, located 204.92 km from the Golden Gate, serves as the release site for Chinook salmon and steelhead The river section from this site to the Benicia Bridge is designated as acclimatization habitat, where transit time and residency analyses commence upon detection of individuals at Benicia The overall transit time from the Benicia Bridge to the Golden Gate, covering a distance of 50.69 km, ranged from 1.14 to 11.6 days for Chinook salmon (median: 2.2 days) and from 1.0 to 17.7 days for steelhead (median: 1.9 days).
Figure 14 Total transit time (in days) for Chinook salmon (LFC) and steelhead (STH) from Benicia Bridge to the
Of the 500 individuals of each species released, 85 Chinook salmon (17 %) and 62 steelhead (12.4
At the Golden Gate, a total of 30 detections were recorded, with 20 (23.5%) from the former group and 10 (16.1%) from the latter Notably, only a subset of these detections was observed at all four bridges designated for separating river sections: Benicia and Carquinez.
Steelhead salmon exhibited faster transit rates than Chinook salmon in river reaches between Benicia and Carquinez, as well as between Richmond and the Golden Gate However, transit rates were comparable for both species in San Pablo Bay, from Carquinez to Richmond While there was no significant difference in transit rates between the species overall (MANOVA; p=0.0608), a significant difference was observed in transit rates between different river reaches (MANOVA; p=0.0009), with steelhead transiting more slowly through the Carquinez to Richmond reach.
The study found that the overall transit time for all fish species between the Benicia Bridge and the Golden Gate was consistent, with no significant differences observed (Welch’s test, p=0.722) Additionally, the transit rates did not vary by river stretch or species, as indicated by the Kruskal-Wallis test (p=0.67) [Fig 15].
Table 4 Transit time (h:mm:ss) and rate (ms -1 ) of Chinook salmon and steelhead which were detected through each section of the river system.
Species Tag ID Time Rate (ms -1 ) Time Rate (ms -1 ) Time Rate (ms -1 )
Figure 15 Transit time of Chinook salmon (LFC, grey) and steelhead (STH, white) in three sections of the
The Sacramento River system encompasses the segments from Benicia (BEN) to Carquinez (CAR), Carquinez (CAR) to Richmond (RIC), and Richmond (RIC) to the Golden Gate (GG) The data visualization features boxes representing the upper and lower quartiles, with a line indicating the median, while caps at each end of the box mark extreme values Outliers are denoted with a “+” symbol, and the sample size for each group is prominently displayed in bold above the boxes.
The transit rates between bridge arrays along the river exhibited a distinct bimodal distribution for both fish species observed between Benicia and Carquinez One group migrated slowly, averaging a modal speed of approximately 0.25 m/s, while another group displayed rapid migration, frequently exceeding speeds of 1 m/s Notably, there were very few fish migrating at intermediate speeds of 0.5 to 0.9 m/s, a pattern that was less pronounced in other river sections.
Figure 16 Transit rates (ms-1) for Chinook salmon (filled bars) and steelhead (white bars) from Benicia Bridge to
Figure 17 Transit rates (ms-1) for Chinook salmon (filled bars) and steelhead (white bars) from Carquinez Bridge to Richmond Bridge, 2009
Figure 18 Transit rates (ms-1) for Chinook salmon (filled bars) and steelhead (white bars) from Richmond Bridge to the Golden Gate, 2009
The size of the fish (represented as its wet weight, in grams) did not display any significant relationship with the transit speed, for either species (Figs 19 and 20).
Figure 19 Relationship between fish weight (g) and transit rate (ms -1 ) of Chinook salmon through river section fromBenicia to Carquinez.
Figure 20 Relationship between fish weight (g) and transit rate (ms -1 ) of steelhead through river section from Benicia to Carquinez.
Chinook salmon migrating through the Benicia-Carquinez river section exhibited two distinct groups based on their transit speeds: those moving faster than 0.6 ms^-1 and those at slower rates Most fish from both groups were observed migrating during peak flow conditions, with only three individuals recorded during slack tide, all of which were from the faster-moving group Additionally, there was no significant difference in river discharge volumes during the passage of either fast or slow salmon.
Steelhead typically migrate during peak ebb tides, but some slower individuals also move during slack or flood tides The downstream flow seems to influence their transit rate significantly.
The transit rate of fish along the Benicia-Carquinez river stretch was not influenced by their upstream movement and did not impact their eventual arrival at the Golden Gate.
Table 5 Proportion of fish which transited at ‘fast’ or ‘slow’ rates through Benicia-Carquinez river stretch and whether they displayed upstream-downstream movements and reached the Golden Gate
Figure 21 River discharge volume (from Rio Vista c-dec) at first detection of Chinook salmon through Decker
Figure 22 Correlation of transit rates (ms -1 ) of Chinook salmon through river section Benicia-Carquinez, with corresponding river discharge volume (from Rio Vista C-dec) at first detection at Decker Island
Figure 23 River discharge volume (from Rio Vista) at first detection of steelhead through Decker Island
Figure 24 Correlation of transit rates (ms -1 ) of Chinook salmon through river section Benicia-Carquinez, with corresponding river discharge volume (from Rio Vista C-dec) at first detection at Decker Island
A total of 36 Chinook salmon were detected at the Carquinez and Richmond bridges, with 22 fish migrating through the array once and 14 (39%) showing upstream movements at least once The instantaneous transit rates for these fish, as they moved between SP Control and SP Array 1, and from SP Array 1 to SP Array 2, were notably high, exceeding 1 ms-1, compared to an overall average rate of 0.44 ms-1 Significant differences were observed between the entire river stretch and the SP Control-SP Array 1 section (p < 0.001), as well as between the SP Array 1-SP Array 2 section, indicating that fish transit faster through the SP Control-SP Array 1 section (p < 0.001).
Table 6 Instantaneous transit rates (ms -1 ) of Chinook salmon through sections of the San Pablo Array compared with overall transit times from Carquinez to Richmond Bridge.
Chinook salmon #31142, measuring 191 mm in fork length and weighing 85.2 g, was released on February 27th and reached Carquinez Bridge by March 7th at 23:06 The fish was detected multiple times at the bridge throughout the night until 10:20 am the next day, starting its downstream movement during slack high tide It was later detected at SP Control, indicating a downstream transit rate of 0.53 m/s, primarily during the ebb tide Subsequent movements showed the fish traveling downstream through SP Arrays 1 and 2 at rates of 0.9 m/s and 0.5 m/s, respectively, before returning upstream to SP Control as the tide shifted On March 9th at 16:00, the fish repeated a similar pattern, ending further upstream at SP Buoy 10, where it remained for 12 hours before moving back to Carquinez on the flood tide at a transit rate of 1.28 m/s It then returned downstream to SP Control on the next ebb tide at 0.8 m/s and made another upstream journey back to Carquinez on the subsequent flood tide at 0.58 m/s.
The vessel traversed the full length of the bay, reaching Richmond Bridge at 19:21, with a transit speed of 0.5 ms⁻¹ It then made a return trip to SP Array 2 during the subsequent flood tide, after which it was no longer detected.
Figure 25 Movement of Chinook salmon #31142 from Carquinez to Richmond Bridge, related to tidal state.