Introduction

In 1988, the Suisun Marsh Salinity Control Gates (SMSCG) were installed in Montezuma Slough to meet water quality criteria in Suisun Marsh. The gates are located approximately 2.5 miles upstream from the confluence of the Sacramento and San Joaquin Rivers (Fig. 1). The structure consists of a section of removable flashboards, three tidally operated radial gates, and a boat lock (Fig. 2).

Fig. 1
figure 1

Location of Montezuma Slough in relation to the Sacramento/San Joaquin Delta

Full size image
Fig. 2
figure 2

Schematic of the Suisun Marsh Salinity Control Gates

Full size image

By raising the radial gates during ebb flow and lowering the gates during flood flow, the structure tidally pumps water from the Sacramento River through Montezuma Slough. This action moves less saline water downstream through the slough, making it available to the interior of Suisun Marsh. Salinity standards were set in Suisun Marsh to preserve the area as a brackish tidal marsh and to provide optimum source water for waterfowl food plant production (DWR 1997).

As part of the permit to build the gates, a monitoring program was developed to study the impacts of the gates and structure on migratory fish. Of special concern were federally listed winter-run Chinook salmon, that may be present in Montezuma Slough during regular operation of the SMSCG, October through May.

Biotelemetry studies began in 1993 at the SMSCG. Biotelemetry has been commonly used for studies of fish migration, orientation, mechanisms, and movement patterns at obstructions since 1956 (Stasko and Pincock 1977). Studies on gastrically implanted radio tags in adult Chinook salmon (Matter and Sandford 2003), concluded that implantation of a radio tag in adult spring-summer Chinook salmon at Bonneville Dam did not appear to have an effect on migration rate, and the assumption that transmitter attachment does not affect fish behavior is implicit in biotelemetry studies. Biotelemetry studies in 1993 and 1994 at the SMSCG found that the structure hindered salmon movement through Montezuma Slough up to 12 h a day when the radial gates were lowered and the flashboards installed (Edwards et al. 1996; Tillman et al. 1996). Brett (1995) noted that upstream migrating Pacific salmon have a fixed store of energy reserves and rely on ocean feeding to build up the stores. Winter-run salmon spawning is delayed for some time after river entry, so not only do they lose feeding time in the sea, but they must also maintain their ion balance, without feeding, in the osmotically rigorous freshwater environment for several months before spawning (Healy 1991).

Mitigation options were considered to facilitate salmon passage at the SMSCG including modifications to the structure. Proposed options included the removal of the top flashboard, modifications to the flashboards, and the installation of a fish ladder (DWR 1997). Biotelemetry studies conducted in 1998 and 1999 focused on the option of modification of the existing flashboards in the form of two 0.9 m by 20.1 m slot openings to allow passage of migrating salmon during full operation of the SMSCG. Studies done on the Columbia River involving the modification of passage orifices in a Weir to include overflow slots were shown to resolve passage problem with American shad at John Day Dam without impeding the passage of salmonids (Monk et al. 1989). Results from the 1998 and 1999 SMSCG study found that the modified flashboards had the opposite effect on salmon migration and recorded the lowest passage rate and longest delay times for both years (Vincik 2002).

The feasibility of using the open boat lock at the SMSCG for salmon passage was first tested during the 2001 SMSCG adult salmon passage study. The 2001–2004 adult salmon passage study monitored the passage rates and passage times of migrating adult Chinook salmon (Oncorhynchus tshawytscha) past the SMSCG in Montezuma Slough. The objective for the 2001–2004 studies was to determine if holding open the existing boat lock during periods of adult Chinook salmon migration would result in: 1) a significant difference in passage rates of salmon through the SMSCG; 2) a significant difference in passage times of salmon through the SMSCG; 3) a significant difference in the passage rates and times between each experimental operational configuration of the SMSCG.

Methods

The boat lock is located near the eastern bank of Montezuma Slough, having dimensions of 21.3 m long by 6.1 m wide with an average depth of 3.7 m. Modeling using the DWR Delta Simulation Model (Suisun Marsh Version) showed that the existing boat locks could be held open with only a minimum reduction in the effectiveness of the SMSCG (Chris Enright, DWR, pers. comm.). During the 2001 to 2004 studies, 3 operational configurations (phases) of the gates were studied. Configurations over all years included: Phase I, full open with boat lock closed (flashboards removed, gates up); Phase II, full operation with the boat lock open (flashboards installed, gates tidally operated); Phase III, full operation with the boat lock closed (flashboards installed, gates tidally operated) (Fig. 3). Tagged salmon were monitored through all four years using these three configurations (phases) with rates of passage and passage times though the facilities calculated for each year. The 2001–2004 studies were conducted from the end of September to the first week in November and the order of the gate configurations changed each year to eliminate the influence of migration run timing (Table 1).

Fig. 3
figure 3

Three Experimental Operational Phases of the Suisun Marsh Salinity Control Gates, 2001–2004

Full size image
Table 1 Experimental operating phases and phase timing of the Suisun Marsh Salinity Control Gates, Sep–Nov 2001–2004
Full size table

For the 2001–2004 studies, abundant fall-run Chinook salmon were used as a surrogate for endangered winter-run salmon. Adult salmon were captured using a drifted, large mesh gill net downstream from the SMSCG. Once a salmon was detected in the net, it was removed and placed into a 567.8 L holding tank filled with ambient river water. Prior to tagging, each fish was measured, visually sexed, and evaluated for tagging suitability based on overall health and swimming behavior (gills pink, upright position). Fish that were injured or exhibiting extensive stress from handling were not tagged. Transmitters were directed into the fish using a pill dispenser device, commonly used in livestock, via the mouth and down its esophagus, and finally into the gastric cavity. After tagging, fish were released upstream of the capture site approximately 2.4 km downstream from the SMSCG near the Grizzly Island boat ramp (Fig. 4).

Fig. 4
figure 4

Adult salmon capture and release area, 2001–2004

Full size image

During the 2001–2002 studies, Sonotronics® telemetry tags and recording equipment were used. Each tag was coded with a specific pulse interval and frequency to distinguish it from other tags, ranging from 65 to 81 kHz and a minimum battery life of 14 months. Tags were detected using Sonotronics® USR-90 automatic scanning receivers and DH2 hydrophones hooked up to portable computers (HP200LX and HP100LX) and powered by a 12 volt car battery. While this equipment had served our needs in the past, attrition and new technologies warranted the consideration of changing to equipment that is more reliable, easier to maintain, recorded data in a more usable format and is much more compact.

For the 2003 and 2004 studies, Vemco® telemetry tags and equipment were used. Each tag had a unique code with a randomized 20–60 second pulse interval with a frequency of 69 kHz and a minimum battery life of 12 months. Tags were detected using Vemco® VR2 receivers with an internal flash memory and a 15 month battery life.

Fixed monitoring stations with receivers were placed upstream, downstream and on the SMSCG (Fig. 5). For all years of the study, an additional Floy® t-bar tag was placed behind the dorsal fin. A return address was imprinted on each external tag to identify internally tagged fish that may be recaptured during the study and elicit possible angler response. Up to sixty-six tagged salmon were released at the beginning of each two week phase and monitored for passage rate and time through the SMSCG. During the 2004 study, additional Vemco® receivers were placed upstream and downstream from Montezuma Slough to detect tagged fish migration further into the Sacramento and San Joaquin Rivers (Fig. 6).

Fig. 5
figure 5

Hydrophone locations at the Suisun Marsh Salinity Control Gates, 2001–2004

Full size image
Fig. 6
figure 6

Additional hydrophone locations beyond the Suisun Marsh Salinity Control Gates, 2004

Full size image

Passage was deemed to be successful if a tag was detected at the downstream fixed monitoring stations and then detected later at the upstream monitoring stations. In addition to the stationary monitoring sites set up in Montezuma Slough, mobile monitoring was conducted for a minimum of 3 days during each phase after the required numbers of adult salmon were tagged. Each complete sweep of the tracking area took approximately 6 h. Fish were assumed dead if a tag was detected for more than 3 days at one specific location. Salmon detected by boat monitoring on the downstream side of the control gates and shortly thereafter on the upstream side were assumed to have passed through the gates. Tagged fish that were not detected by any of the onshore monitoring stations or in the boat monitoring area during the respective study phase were assumed to have left the study area without passing through the gates.

As in previous studies at the SMSCG, tagged salmon were assigned to a behavior mode defining their movement patterns:

  • Mode 1 – Passed the SMSCG and continued to move upstream.

  • Mode 2 – Passed the SMSCG then moved back downstream.

  • Mode 3 – Did not pass the SMSCG, moved back downstream.

  • Mode 4 – No records, tagging mortality or regurgitated tag

A 2 × 2 chi-square comparison (P < 0.05) was used to determine whether any differences in percent passage rates between gate configurations are significantly greater than could be attributed to random variation. Passage times were analyzed using ANOVA (P < 0.05) to detect significant differences across SMSCG operations.

Results

2001

Of the 198 adult salmon tagged during the 2001 study, 118 (60 %) passed through the SMSCG. The other 40 % of the tagged salmon did not pass through the gates and returned downstream after tagging. The highest rate of passage (77 %) occurred during Phase II (full operation, boat lock open), and the lowest rate of passage (61 %) occurred during Phase III (full operation, boat lock closed). Phase I (full open) had a 63 % passage rate (Fig. 7). Mean passage times for tagged salmon ranged from 15 to 47 h. The longest mean passage time (47 h) occurred during Phase III; the shortest mean passage time (15 h) occurred during Phase I. Phase II had a mean passage time of 26 h (Fig. 8). During Phase II, of the 44 tagged salmon to pass the gates, 14 (32 %) were recorded using the boat lock for passage. The average time spent in the boat lock was 13 min. Tagged fish ranged between 600 and 1070 mm fork length (FL). There was no significant difference in passage rates (χ² = P > 0.05) There was an overall significant difference in passage times between Phases (ANOVA, all P < 0.05) (Table 2).

Fig. 7
figure 7

Passage rates for the 2001–2004 SMSCG studies comparing three operational phases; gates up, flashboards removed boat lock closed (Phase I); gates operational, flashboards installed, boat lock open (Phase II); gates operational, flashboards installed, boat lock closed (Phase III)

Full size image
Fig. 8
figure 8

Mean passage times for the 2001–2004 SMSCG studies comparing three operational phases; gates up, flashboards removed, boat lock open (Phase I); gates operational, flashboards installed, boat lock open (Phase II); gates operational, flashboards installed, boat lock closed (Phase III)

Full size image
Table 2 Chi-square and probability for passage rates and times during the 2001–2004 SMSCG Study
Full size table

2002

Of the 198 adult salmon tagged during the 2002 study, 84 (42 %) passed through the SMSCG. The other 58 % of the tagged salmon did not pass through the gates and returned downstream after tagging. The highest rate of passage (58 %) occurred during Phase I, and the lowest rate of passage (36 %) occurred during Phase II. Phase III had a 47 % passage rate (Fig. 7). Mean passage times for tagged salmon ranged from 20 to 78 h. The longest mean passage time (78 h) occurred during Phase II; the shortest mean passage time (20 h) occurred during Phase I. Phase III had a mean passage time of 36 h (Fig. 8). During the Phase II configuration, of the 21 tagged salmon to pass the gates, eight (38 %) were recorded moving through the boat lock. The average time spent in the boat lock was 7 min. Tagged fish ranged between 600 and 1040 mm FL. There was a significant difference in passage rates (χ² = P < 0.05) between Phase I and Phase II. There was no significant difference in passage rates between Phase I and III and Phase II and III. There was a significant difference (ANOVA P < 0.05) in passage times between Phase I and Phase II, and between Phase II and Phase III. There was no significant difference between Phase I and III (Table 2).

2003

Of the 163 adult salmon tagged during the 2003 study, 103 (63 %) passed through the SMSCG. The other 37 % of the tagged salmon did not pass through the gates and returned downstream after tagging. The highest rate of passage (71 %) occurred during Phase I, and the lowest rate of passage (54 %) occurred during Phase III. Phase II had a 63 % passage rate (Fig. 7). Mean passage times for tagged salmon ranged from 36 to 52 h. The longest mean passage time (52 h) occurred during Phase II; the shortest mean passage time (36 h) occurred during Phase I. Phase III had a mean passage time of 39 h (Fig. 8). During Phase II, of the 28 tagged salmon to pass the gates, eight (29 %) used the boat lock for passage. Tagged fish ranged in size from 600 and 1120 mm and were evenly distributed between male and female. There was a significant difference in passage rate between Phase II and Phase III (χ² = P < 0.05). There was no significant difference in passage rate between Phase I and Phase II, or between Phase I and Phase III. There was no significant difference between the passage times of the three Phases (ANOVA, P > 0.05) (Table 2)

2004

Of the 197 adult salmon tagged during the 2004 study, 97 (49 %) passed through the SMSCG. The other 51 % of the tagged salmon did not pass through the gates and returned downstream after tagging. The highest rate of passage (58 %) occurred during Phase I, and the lowest rate of passage (35 %) occurred during Phase III. Phase II had a 55 % passage rate (Fig. 7). Mean passage times for tagged salmon ranged from 28 to 61 h. The longest mean passage time (61 h) occurred during Phase III; the shortest mean passage time (28 h) occurred during Phase II. Phase I had a mean passage time of 37 h (Fig. 8). During Phase II, of the 36 tagged salmon to pass the gates, twelve (33 %) were recorded moving through the boat lock. The average time spent in the boat lock was 14 min. Tagged fish ranged between 600 and 1150 mm FL. Phase I had the best passage rate and was significantly different (χ² = P < 0.05) from Phase III, but not significantly different from Phase II. There was a significant difference in passage rates between Phase II and Phase III. There was no significant difference in passage times between all three phases (ANOVA, P > 0.05) (Table 2).

2004 migration behavior

At the upstream locations, 97 tagged salmon were recorded at the Collinsville site leaving Montezuma Slough. Fifty-eight were recorded further upstream at Rio Vista and 49 at Hood, the uppermost Sacramento River site. Of the tagged salmon that were recorded further upstream at Hood, 30 (61 %) had passed through the SMSCG, and 19 (39 %) were recorded as having gone back around the SMSCG. Two salmon were recorded at Mossdale Landing on the San Joaquin River, one having passed through the SMSCG and the other having gone around. Several tagged salmon were recorded at the downstream sites as well, with 85 tagged salmon recorded at the Grizzly Bay site. The site located furthest downstream at the Carquinez Strait had three records of tagged fish and may have had more except the unit stopped working on October 8th due to a leak in the battery compartment.

Multi-year comparison

Chi-square comparisons of passage rates between Phase II over all years of the study show a significant difference between 2002 and 2001 and 2002 and 2003, but were not significant between 2002 and 2004 (Table 3). There is also a significant difference between the 2001 and 2004 results. Mean passage times varied greatly between study years with Phase I having the lowest delay time (27.8 h) compared with Phase II (48.3 h) and Phase III (38.4 h) (Table 4).

Table 3 Chi-square comparison of passage rates for 2001–2004 Phase II (boat lock open configuration) at the SMSCG
Full size table
Table 4 Salmon mean passage times during operational phases, 2001–2004
Full size table

Fish passage behaviors

The highest percentage of tagged salmon passing the gates (58 %) occurred during the 2003 study and the lowest percentage (29 %) during the 2002 study (Table 5). This behavior is hard to quantify since during the 2004 study, of the tagged salmon that did not pass the gates (n = 100), eight were recorded upstream of the gates and recovered at hatcheries (3), caught by anglers (4), and recovered during a carcass survey (1). Mean time between tagging and recovery was 30.6 d (SD 9.2).

Table 5 Behavior modes of tagged adult salmon 2001 through 2004
Full size table

Discussion

Access to the boat lock appears to have improved fish passage in some years, and was presumably related to lower energy costs for fish, whereas access to the boat lock was related to longer passage time in only one year. Phase I had a significantly higher passage rate than Phase III in 2004 only. Phase II had a significantly higher passage rate than Phase III in 2003 and 2004. Phase II had a significantly shorter passage time than Phase III in 2001. Phase II had a significantly shorter passage time than Phase III in 2001, but a longer passage time in 2002.

Pacific salmon must rely solely on energy reserves to carry out their upriver migration, to complete their gonad development, and to spawn. Energy saving behaviors, such as the selection of low-velocity or reverse-current paths may thus be important migration tactics for these upriver migrants. This behavior may be particularly crucial at sites of difficult passage, where energetic demands may exceed the migrant’s metabolic scope (Hinch and Bratty 2000). Studies by Hallock et al. (1982) and the US Fish and Wildlife Service (1988) hypothesize that a delay of mature salmonids could result in fecundity declines, increased prespawning mortality, and spawning in less desirable habitat. Montezuma Slough provides an alternative, low velocity pathway for migrating adult salmon.

Previous studies at the SMSCG showed that the gates delayed salmon migration up to 12 h a day when the gates were lowered and the flashboards installed. Studies done on adult sockeye salmon (Oncorhynchus nerka) found that unsuccessful migrants had longer residency in the approach to a fishway than did successful migrants (Hinch and Bratty 2000). Russell et al. (1998) found that the upstream migration of Atlantic salmon (Salmo salar) seemed to be delayed by the presence of a barrage with many fish holding station close to the barraged structure.

While environmental factors unrelated to the gate operations may have affected fish passage rates and times, I do not think this was the case. Following the 1998–1999 studies, in 2000, a Technical Team was put together to determine if there were any environmental factors that could have influenced salmon passage that might have had an influence on future studies. The team looked at ten environmental factors including; tidal stage, flow direction, types of tide (spring, neap), phase of the moon, day vs. night, salinity within Montezuma Slough, duration of gate openings, water temperature, dissolved oxygen, magnitude of flow, and Delta outflow. The Team concluded that these environmental factors were not a significant influence on salmon passage and that the modified flashboards hampered salmon passage and that to improve passage, the focus must shift back to the structure itself (DWR 2001).

The behavioral modes assigned to tagged adult salmon during each study have shown that fish may pass the gates, but subsequently move back downstream again. This “milling” behavior has been noted in previous studies at the SMSCG (Edwards et al. 1996; Tillman et al. 1996) as well as salmon studies in other estuarine conditions similar to Montezuma Slough and the Suisun Marsh area (Groot et al. 1975; Stasko et al. 1976). There can be a great deal of variability between movement patterns of adult salmon. In an ultrasonic tracking study on Fraser River sockeye salmon (Stasko, et al. 1976), movement patterns of individual fish were unpredictable. Some fish moved actively northward; some drifted most of the time; still others traveled southward away from their presumed destination. This milling behavior in adult Chinook salmon was also observed in Montezuma Slough where tagged salmon were recorded heading back downstream after tagging and some passing through the gates more than once. Behavior modes assigned to tagged salmon over all years of the study confirm that a percentage of salmon move back downstream after tagging. This trait has been noted in previous studies on externally tagged salmon (Hallock et al. 1970; Gray and Haynes 1979) where tagged salmon dropped back downstream immediately after release. During the 2004 study, several tagged salmon exited the slough downstream towards Grizzly Bay, then migrated back upstream via the Sacramento River and were recorded at Rio Vista and Hood while other tagged salmon that passed the gates headed downstream to Grizzly Bay before turning around and migrating upstream towards the Sacramento and San Joaquin Rivers. Because salmon have been known to delay migration in estuaries up to one month (Hallock et al. 1970; Olson and Quinn 1993) it is not unusual to find this behavior in SMSCG tagged fish.

The open boat lock can provide passage during flow or ebb tidal cycles allowing migrating salmon to swim into an attraction flow or move during reverse flow conditions in Montezuma Slough. Brown and Geist (2006) found that on average, swim speed was significantly higher when tagged adult Chinook salmon were moving through tailraces in Bonneville Dam. Flow velocities through the boat lock were rarely greater than 2.4 m/s, and DWR and DFG biologists agreed that neither the instantaneous or average velocities appear to hinder adult salmon passage (DWR, DFG 2002). The overall trend across study years confirmed that the open boat lock presents an alternate passage route to migrating adult salmon when available.

As a result of the 2001–2004 passage studies at the SMSCG, a proposal to improve fish passage was recommended by the SMSCG Steering Committee (DWR 2005). The National Oceanic and Atmospheric Administration (NOAA) Biological Opinion on the Long-Term Central Valley Project and State Water Project Operations Criteria and Plan (OCAP) in October 2004 included terms and conditions for Suisun Marsh Operations: “DWR shall operate the SMSCG to minimize delay and blockage of adult Sacramento River winter-run Chinook salmon, Central Valley spring run Chinook salmon and Central Valley steelhead migration upstream.” And, “Beginning no later than November 15, 2004, hold the boat lock “open” at all times when the flashboards are installed at the SMSCG.” This includes feasible measures to remove and reinstall the flashboards when the SMSCG are not required to meet Suisun Marsh water quality standards.

While the SMSCG Steering Committee and the NOAA OCAP Biological Opinion have already recommended keeping the boat lock open for fish passage to aid in salmon passage through the SMSCG, there are still questions that need to be addressed following the current study. In 2002 and 2003, the boat lock open phase (Phase II) recorded the longest passage times through the gates compared with 2001 and 2004 and in 2002, Phase II recorded the lowest percentage of passage compared with other years. Further studies at the SMSCG using biotelemetry that more fully examine relationships between passage success and migration delays, dry vs. wet years, straying and temperature could further help answer questions posed through all years of the study. Researchers on the Columbia River have studied hydrosystem effects for decades, but the complex effects of altered environment and hydrosystem operations on adult behavior and survival are still not fully understood (Keefer et al. 2004). The Sacramento/San Joaquin River hydrosystem has been dramatically altered regarding the migration environment for anadromous salmonids and new proposals for water diversions are currently being planned. Results and lessons learned from studies at the SMSCG along with further multi-year studies using biotelemetry can aid in the design of future water diversions and help in the adaptive management of salmonids in anadromous waters.