Trapping Giant Gartersnakes

All aspects of the giant gartersnake monitoring effort involved using trap transects composed of floating galvanized minnow traps (Casazza and others, 2000) for capture. Beginning in 2012, traps were modified to contain one-way valves constructed from cable ties placed in the small opening of the funnels. Beginning in 2013, traps also were modified to include two pieces of hardware cloth attached to each end of the funnel using zip ties (Halstead and others, 2013a). These modifications helped to direct snakes moving along the edge of a habitat into the trap and keep the snake within the trap, thus increasing capture probability.

Giant gartersnake occurrence and demography were monitored at 60 sites in 2018 and 2019, 40 sites in 2020 and 2023, 45 sites in 2021, 41 sites in 2022, and 44 sites in 2024. Between 2020 and 2022, we dealt with substantial challenges caused by the COVID-19 pandemic. In 2022, we faced additional obstacles, such as heat waves and drought that caused low water levels and high water temperatures. In 2023, there were unprecedented rainstorms and flooding that resulted in inaccessible roads and delays in setting certain traplines. Like the previous 3 years, in 2024, we were able to continue the 21-day sampling periods as opposed to the 14-day period in 2020. We selected which sites to sample using a Generalized Random Tessellation stratified approach to ensure a random, spatially dispersed sample (table 1; fig. 2). Random selection of sites allowed inference for TNBC reserves as a whole. To ensure adequate representation of each reserve, sampled sites consisted of individual wetland units (defined as being contained within water control structures) and canals adjacent to rice, and the selection of sampled sites was stratified by reserve area (23 sites in the North Basin reserve, 12 sites in the Central Basin reserve, and 9 sites in the Fisherman’s Lake reserve). Most sites in the Fisherman’s Lake reserve were wetland sites (7 of 9), whereas sites were nearly evenly divided between the two types in the Central Basin reserve (5 of 12 were wetlands) and in the North Basin reserve (13 of 23 sites were wetlands). At each site, one transect of 50 traps was deployed, with traps spaced approximately 10–20 meters (m) apart. In 2024, trap transects were deployed for approximately 11–21 days each, from as early as possible in the active season, beginning May 13, 2024, until September 13, 2024. Sampling did not take place on Memorial Day, Juneteenth, Independence Day, or Labor Day. Trap transects were positioned along the banks, at the edges of emergent vegetation in wetlands, or along the edges of canals because giant gartersnakes forage along habitat edges. Habitat edges also act as natural drift fences that direct snake movement to traps. Traps were checked daily while deployed.

For sites that were sampled as occupancy traplines before 2018, transects generally were placed in the same location in 2024; this approach maintains the same extent of sampling to provide inference about giant gartersnake occurrence in the same areas. These original occupancy traplines continued to be trapped in similar numbers after 2018; however, at the five demography sites on the Betts-Kismat-Silva (BKS), Lucich North, Lucich South, Natomas Farms, and Sills tracts, sampling after 2018 differed from the areas sampled from 2011 to 2017. Previously, three transects were sampled at each demography site each year, but we modified our study to allow for better inference of both occupancy and abundance through the Natomas basin. In 2018 and 2019 at the BKS tract, all three old demography traplines were sampled, and beginning in 2020 and continuing until 2024, only two old demography traplines were sampled (BKS 2 and 3). At both Lucich North and Lucich South, only one of the three old demography traplines (Lucich North 4 and Lucich South 3, respectively) has been sampled during the past 7 years. At the Natomas Farms tract, one old demography trapline (Natomas Farms 1) was sampled in 2018, 2019, 2020, and 2023, three old demography traplines (Natomas Farms 1, 2, and 3) were sampled in 2021, and two old demography traplines (Natomas Farms 1 and 2) were sampled in 2022 and 2024. At the Sills tract, two old demography traplines (Sills 2 and 3) were sampled in 2018 and 2019, whereas no old demography traplines were sampled from 2020 to 2023, and one old demography trapline (Sills 3) was sampled in 2024. These changes made to our study’s sampling methods allowed us to survey a broader area at most of the old demography sites (except in 2020 when the study was truncated due to the COVID-19 pandemic); however, they also meant that some areas that were sampled from 2011 to 2017 were not covered by trap arrays from 2018 to 2024. As noted above, these changes present ramifications for modeling giant gartersnake demography and interpreting changes in abundance estimates, as described below in the “Analytical Methods” section.

Environmental conditions relevant to giant gartersnake behavior were monitored daily at each trap transect, including water temperatures, air temperatures, and fluctuations in water level. To obtain a measure of the relative abundance and diversity of potential local aquatic prey, the contents of every fifth trap were recorded, and then, all contents were removed. All other traps were monitored, but prey items, such as fish, tadpoles, and small frogs, were left in the traps so that they became naturally baited over time. In some instances, large fluctuations in water level (draining of wetlands, canals, or ditches) necessitated opening traps temporarily or relocating transects to a suitable nearby location within the selected site.

Universal Transverse Mercator (UTM) coordinates of all trap locations were recorded, and vegetation and habitat surveys were undertaken at points within and adjacent to each trap transect. Percentage cover of habitat types (water, submerged aquatic vegetation, floating aquatic vegetation, emergent vegetation, terrestrial vegetation, rock, or bare ground) and vegetative composition (species or higher taxonomic category) was estimated within a 1-m radius of every fifth trap. Associated with each habitat and vegetation survey along the transect, a point was selected toward the terrestrial-side (odd-numbered traps) or the aquatic-side (even-numbered traps) of the transect at a random perpendicular distance of 2–5 m, and percentage cover of habitats and vegetative composition was estimated within a 1-m radius of this point to better characterize habitat surrounding the traps.

Each captured giant gartersnake was measured, sexed, and uniquely marked. Scale measurements in Rossman and others (1996) were used to verify the species of each captured gartersnake. Individual snout-vent lengths (SVL) and tail-vent lengths (TVL) were measured to the nearest millimeter (mm), and each individual was weighed to the nearest gram (g). The sex of each individual was determined by probing the cloaca to detect the presence or absence of hemipenes, indicative of male snakes. After examination, each individual that showed no sign of previous capture was given a unique brand on its ventral scutes (Winne and others, 2006) and, if large enough (greater than 35 g), implanted with a passive integrated transponder (PIT) tag. PIT tags were implanted using syringe injector needles swabbed with alcohol before each use, and the injection site on the snake was swabbed with alcohol before tag insertion. The tag was injected subcutaneously, approximately one-third of the SVL anterior to the cloaca. After insertion of the tag, cyanoacrylate glue was applied to the insertion site to seal the dermis and prevent tag loss. Most individuals were processed in the field within a few minutes of their capture. If snakes were held for more than a few minutes, they were kept in the shade in cooled and insulated containers to prevent overheating until they could be examined and released. Each individual was released at its location of capture immediately after processing.

Starting in 2024, as many as 10 gravid females were collected from the BKS Wetland complex as part of a captive-rearing project in coordination with the Larry Ellison Conservation Center for Wildlife Care (hereafter, Ellison Center). Females were brought into captivity to the U.S. Geological Survey (USGS) Dixon Field Station and held temporarily until the females, or neonates produced by the females, were transferred to the Ellison Center. All females were returned to their sites of capture after giving birth. Any neonates produced by the females will be reared at the Ellison Center for 1–2 years before they are transferred back to the Dixon Field Station and then released back into the BKS Wetland Complex.

Analytical Methods

Most analyses were done in a Bayesian analytical framework. In Bayesian analyses, the probabilities are interpreted slightly differently from traditional frequentist statistical analyses. The posterior probability is the probability of a random event or uncertain proposition given the data at hand and is how most people intuitively think about probability (for example, how people interpret a weather forecast). Bayesian analyses also require specification of a prior probability distribution, which allows for the inclusion of information obtained through other sources into the analysis. The prior can be an informative prior (in other words, a distribution based on previously collected data or a hypothesis about the probability distribution of interest) or an uninformative prior (meaning, a probability distribution that will have negligible effect on the outcome of the analysis). Unless otherwise noted, for all parameter estimates, posterior distributions are summarized as medians with equal-tailed 95-percent credible intervals in parentheses.

Distribution of Giant Gartersnake on Reserve Lands

Giant gartersnakes were detected at 20 of 44 sites sampled in 2024 (fig. 2). Of the 40 sites surveyed in both 2023 and 2024, giant gartersnakes were detected at four sites in 2023 where no giant gartersnakes were detected in 2024. There were five detections of giant gartersnakes in 2024 at sites where they were not detected in 2023. Of the four sites that were surveyed in 2024 but not surveyed in 2023, giant gartersnakes were detected at two sites. Of all the sites monitored in 2024, 21 (20–24) were estimated to be occupied (fig. 3).

The effects of habitat variables on the probability of occurrence were not supported in 2024 (table 2). The effect of water temperature and date of sampling on daily detection probability was not supported (table 3; fig. 4). Daily detection probabilities for giant gartersnake in 2024 at an occupied site on a day with average conditions (for example, average water temperature) were 0.27 (0.13–0.44). Over 21 days of trapping, this corresponded to a cumulative probability of detecting giant gartersnake at least once, given they occurred at a site in 2024, of 0.99 (0.94 to greater than 0.99).

The probability of occurrence of giant gartersnakes varied by reserve (in other words, North Basin, Central Basin, and Fisherman’s Lake). The probability of occurrence in wetlands in the North Basin reserve was 0.38 (0.18–0.69), the probability of occurrence in wetlands in the Central Basin reserve was 0.957 (0.74 to greater than 0.99), and the probability of occurrence in wetlands in the Fisherman’s Lake reserve was 0.05 (0.01–0.4).

The dynamic occupancy model including all sites in the Natomas basin indicated evidence for a slight decrease in the probability that sites on TNBC reserves were occupied by giant gartersnakes from 2011 to 2013, followed by a period of stability from 2013 to 2018, an increase from 2018 to 2019, a decrease from 2019 to 2021, and an increase in 2022 that remained stable through 2024 (fig. 5). The number of occupied sampled sites followed a similar pattern, with a decrease from 2019 to 2021 and an increase in 2022 that was stable in 2023 and increased again in 2024 (fig. 6). Both occupancy parameters were estimated with much greater precision in 2018 and 2019 compared to previous years when the number of sites increased to 60 as part of the revised sampling design. Conversely, precision of these parameters was much lower between 2020 and 2024 compared to 2018 and 2019 because of the decreased number of sites to 40 (2020), 45 (2021), 41 (2022), 40 (2023), and 44 (2024) due to the COVID-19 pandemic and stochastic weather patterns. The annual probability that occupied sites would become unoccupied (site extirpation) was generally low and stable between 2011 and 2019. Although it was higher from 2020 to 2023 compared to previous years, site-extirpation probability consistently declined after 2020, and in 2024, site extirpation was back down to levels seen previously (fig. 5). The annual probability that unoccupied sites were colonized was generally low and stable from 2011 to 2024 (fig. 5). The mean intrinsic growth rate of occupancy from 2011 through 2024 was −0.032 (−0.059 to −0.003; fig. 7), indicating a small but substantial decline in occupancy over time. From 2011 to 2024, the posterior probability that occupancy declined was 0.982 on average. Occupancy growth was most negative between 2019 and 2020, potentially because of the reduced number of sampled sites, and although it has rebounded slightly since the 2020–21 period, the trend over time remains negative (fig. 7). There was no support for effects of habitat on occupancy dynamics (table 4).

The dynamic occupancy model including only northern and central marsh sites managed by TNBC indicated evidence of a general decrease in the probability that managed marsh sites on northern and central TNBC reserves were occupied by giant gartersnakes from 2011 to 2018, followed by a general increase until 2022, and a plateau from then until 2024 (fig. 8). The number of occupied TNBC-managed marsh sites in the northern and central basin followed a similar pattern, with an overall decrease from 2011 to 2018 and an overall increase in 2019 that continued through 2024 (fig. 9). The annual probability that occupied managed marsh sites in the northern and central basin would become unoccupied (site extirpation) was generally low and stable with little annual variation at around 0.25 each year (fig. 8). The annual probability that unoccupied TNBC-managed marsh sites in the northern and central basin were colonized also was generally low and stable, though more variable than extirpation probabilities. Colonization probability was particularly low in 2015 and 2018, and higher than average in 2020, 2022, and 2024 (fig. 8). The mean intrinsic growth rate of occupancy for Natomas Basin Conservancy northern and central managed marsh sites from 2011 through 2024 was −0.021 (−0.053 to 0.011; fig. 10), indicating a 2-percent annual decline in occupancy over time. The posterior probability that occupancy of these sites declined from 2011 to 2024 was 0.878 on average. There was little evidence (Bayes factor=2.79 when compared to the null model) that terrestrial vegetation had a positive effect on occupancy and occupancy dynamics at managed marsh sites in the northern and central Natomas basin (table 5). The model including terrestrial vegetation was the top model selected, although the posterior probability for this model was small (table 5).

Demography

In addition to analyzing the occupancy and overall distribution of giant gartersnakes at the Natomas Basin Conservancy reserves, several metrics of giant gartersnake demography were also analyzed. The following sections detail giant gartersnake abundance estimated from closed population models, density estimates from spatially explicit capture-recapture models, and the sex ratio, size distribution, and survival estimates of giant gartersnakes in Natomas Basin Conservancy reserves.

Estimates of Abundance Using Closed Population Models

Capture probability averaged over all sites was 0.003 (0.001–0.007). The standard deviation of capture probability among sites was similar to the standard deviation of capture probability among days, and both were less than the variation among individuals (table 6). There was evidence for a positive ephemeral behavioral response to capture, and for a positive effect of water temperature on capture probability (fig. 11; table 6).

Table 6.    

Posterior distributions for capture parameters of the closed abundance model of giant gartersnakes in the Natomas basin, 2024.

[CI, confidence level; %, percent; p₀, daily capture probability; α_temp, logit-scale coefficient describing the effect of water temperature on capture probability; α_SVL, logit-scale coefficient describing the effect of snake snout-vent-length (SVL) on capture probability; α_sex, logit-scale coefficient describing the effect of snake sex on capture probability; α_behav, logit-scale coefficient describing the effect of snake behavior in response to capture on capture probability, describes whether a snake is more likely to be captured again if they were captured the day before; σ_variable, standard deviation describing random effects of site, individual, or day on capture probability, respectively]

Table 6.    Posterior distributions for capture parameters of the closed abundance model of giant gartersnakes in the Natomas basin, 2024.

Parameter	Median (95% CI)
p0	0.003 (0.001–0.007)
αtemp	0.287 (0.120–0.456)
αSVL	0.094 (−0.345–0.580)
αsex	−0.197 (−0.986–0.651)
αbehav	1.651 (1.081–2.191)
σsite	0.461 (0.108–1.120)
σind	1.082 (0.674–1.615)
σday	0.617 (0.412–0.845)

Seven instances of giant gartersnakes moving between traplines were detected in 2024. Two snakes were captured in multiple traplines within the Sills tract, three snakes were detected moving between a trapline in the BKS tract to a trapline in the Frazer South tract, and two snakes moved between traplines within the BKS tract. Traplines in close enough proximity that giant gartersnakes did or would be expected to move between traplines were grouped together and treated as a single site for this analysis (for example, all traplines in the Lucich North wetland complex were grouped together; and one trapline in the Frazer South tract was grouped with traplines in the BKS tract), resulting in eight demographic clusters. Estimates of abundance at each of the seven demographic clusters where giant gartersnakes were captured are summarized in table 7.

Table 7.    

Summary of giant gartersnake captures and abundance estimates, 2024.

[BKS, Betts-Kismat-Silva; N, estimated abundance]

Table 7.    Summary of giant gartersnake captures and abundance estimates, 2024.

Site	Individuals	Captures	N	Trap days	Traplines
Bennett North	6	12	34 (17–73)	4,198	4
Bennett South	1	1	33 (17–69)	1,048	1
Bianchi West	9	10	38 (22–76)	1,047	1
BKS	70	129	222 (136–454)	6,340	6
Lucich North	16	19	52 (31–105)	9,644	10
Lucich South	2	4	32 (15–70)	3,100	3
Sills	27	41	87 (53–175)	4,988	5
Total	131	211	497 (306–1,019)	30,365	30

At the six traplines within the BKS cluster (including one trapline from Frazer South), 70 individuals were captured 129 times over 6,340 trap days in 2024 (table 7). Nine of the individuals that were captured in 2023 were recaptured in 2024. Three snakes captured at BKS in 2022 were recaptured in 2024, three snakes captured in 2021 were recaptured in 2024, and no snakes captured before 2021 were recaptured at BKS in 2024. The estimated abundance in sampled areas at BKS in 2024 was 222 (136–454) individuals (fig. 12; table 7). The number of individuals captured and estimated abundance in 2024 was higher than in 2023 but not as high as in 2022 or 2020 (table 1.2).

At five traplines within the Sills cluster (including one trapline in the Tufts tract), 27 individuals were captured 41 times over 4,988 trap days in 2024 (table 7). Of the 27 individuals captured at Sills in 2024, 7 were recaptured in 2023. Two snakes captured at Sills in 2022 were recaptured in 2024, one snake captured at Sills in 2021 was recaptured in 2024, and one snake captured at Sills in 2020 was recaptured in 2024. The estimated abundance at Sills in 2024 was 87 (53–175; fig. 13; table 7). The number of individuals captured and the estimated abundance at Sills in 2024 was comparable to 2023 and 2021 but lower than in 2022 and 2020 (table 1.2).

At 10 traplines in the Lucich North cluster (including two traplines in the Frazer North tract and one in the Nestor tract), 16 individuals were captured a total of 19 times over 9,644 trap days in 2024 (table 7). Of the 19 individuals captured at Lucich North in 2023, two were recaptured in 2024. No snakes captured at Lucich North in 2024 were captured previously before 2023. The estimated abundance at Lucich North in 2024 was 52 (31–105) individuals (fig. 14; table 7). The estimated abundance in 2024 was higher than it had been in the previous four years, with the exception of 2020 (table 1.2).

At three traplines within the Lucich South demographic cluster, two individuals were captured four times over 3,100 trap days in 2024 (table 7). The two individuals captured at Lucich South were new to our study and had not been captured in previous years. The estimated abundance at Lucich South in 2024 was 32 (15–70) individuals (fig. 15; table 7). The number of captures and estimated abundance in 2024 was similar to the previous 4 years (table 1.2).

In the Fisherman’s Lake reserve area, no giant gartersnakes were captured at nine traplines in the Cummings, Natomas Farms, and Rosa tracts over 7,945 trap days in 2024. Likewise, in 2023, 2022, 2021, and 2020, no giant gartersnakes were captured at the three tracts (Cummings, Natomas Farms, and Rosa; table 1.2). In 2019, one giant gartersnake was captured at two traplines, whereas in 2018, no giant gartersnakes were captured at either trapline in the Fisherman’s Lake reserve area.

The other traplines at which snakes were captured were in the Bennett North, Bennett South, and Bianchi West tracts. At Bennett North, six individuals were captured a total of 12 times, and the median estimated abundance was 34 (17–73) snakes (table 7). At Bennett South, one individual was captured once, and the median estimated abundance was 33 (17–69) snakes (table 7). At Bianchi West, nine individuals were captured a total of ten times, and the median estimated abundance was 38 (22–76) snakes (table 7).

Size Distribution and Sex Ratio

The overall sex ratio of captured snakes in the Natomas basin was slightly female-biased. The sex ratio was 0.84 (0.59–1.15) males per female for all sites in the basin combined. Basin-wide mean SVL was 597 mm (575–619 mm), and basin-wide mean mass was 125.05 g (110.49–139.91 g). Mean female SVL (640 mm [607–673 mm]) was 97 mm (59–134 mm) greater than mean male SVL (543 mm [524–563 mm]) and mean female mass (128.53 g [108.27–152.27 g]) was 96.94 g (59.09–134.49 g) greater than mean male mass 75.13 g (67.37–83.84 g; fig. 16).

Survival Estimates from 2018 to 2024

Average annual recapture probabilities, given 21 days of sampling at a site, were 0.16 (0.11–0.27). There was no support for a difference in apparent survival rate between female and male giant gartersnake. Apparent survival varied among years and sites. At BKS, apparent survival was highest from 2021 to 2022 and lowest from 2023 to 2024 (fig. 17). At Lucich North, apparent survival was similar over each of the year-long intervals (fig. 18). At Lucich South, apparent survival was also similar for each interval (fig. 19). At the Sills tract, apparent survival estimates were highest for the intervals from 2022 to 2023 and 2023 to 2024 (fig. 20). The probability a snake was available on site for capture (γ) was slightly higher for female than male giant gartersnake (table 9). As with the closed CMR model, there was a positive relationship between capture probability and water temperature (table 9).

Table 9.    

Posterior summaries for parameters from the robust-design capture-mark-recapture model, 2018–24.

[CI, confidence level; %, percent; p, daily recapture probability; β_wt, coefficient for the effect of water temperature on capture probability; σ_site, standard deviation describing the random effect of site on capture probability; σ_year, standard deviation describing the random effect of year on capture probability; φ_female, apparent survival of females; φ_male, apparent survival of males; σ_φ, standard deviation describing variation in apparent survival among sites and years; γ_female, availability on site of females; γ_male, availability on site of males; σ_γ, standard deviation describing variation in availability among sites and years]

Table 9.    Posterior summaries for parameters from the robust-design capture-mark-recapture model, 2018–24.

Parameter	Median (95% CI)
Recapture
p	0.008 (0.006–0.015)
βwt	0.29 (0.20–0.39)
σsite	0.48 (0.11–1.33)
σyear	0.44 (0.26–0.73)
Survival
φfemale	0.46 (0.35–0.59)
φmale	0.47 (0.35–0.61)
σφ	0.72 (0.31–1.39)
Availability
γfemale	0.63 (0.46–0.83)
γmale	0.45 (0.17–0.91)
σγ	0.69 (0.06–1.85)

Habitat Connectivity

An assessment of habitat connectivity is incomplete without addressing the different means by which animal populations are connected. Connectivity generally occurs via the dispersal of individuals across the landscape. Little is known about reptile dispersal, but radio-telemetry studies suggest that most giant gartersnakes have small home ranges with an average core area of 3–20 ha using fixed kernel methods (Valcarcel, 2011) and 2.2 ha in another study using 95-percent adaptive local convex hulls (Nguyen and others, 2024), although individuals can move several kilometers (km) through appropriate habitat if necessary (Reyes and others, 2017; Nguyen and others, 2024). Two distinct forms of connectivity must also be considered. Demographic connectivity refers to the movement of individuals among (sub)populations to the extent that migration plays a role in population dynamics, potentially rescuing local populations from extirpation through migration into them from a source population (Mills, 2007). Genetic connectivity is the dispersal of enough individuals among populations to prevent genetic differentiation among them. A one-migrant-per-generation rule is often considered an adequate amount of connectivity to avoid the negative effects of inbreeding (Mills, 2007). In general, demographic connectivity requires the exchange of far more individuals than genetic connectivity. Both forms of connectivity are addressed in the following discussion.

Although portions of TNBC’s reserve system are well-connected, some notable exceptions exist. In particular, although surface water connects the Fisherman’s Lake reserve with other reserve areas, the northernmost suitable Fisherman’s Lake reserve tract (Natomas Farms), is approximately 15 km (by canal) south of the nearest suitable Central Basin reserve tract known to be occupied by giant gartersnake (Elsie). Giant gartersnakes have small home ranges and typically move relatively short distances (Valcarcel, 2011; Reyes and others, 2017; Nguyen and others, 2024) but nonetheless can move up to 5 km over multiple days (U.S. Geological Survey, unpub. data, 2019). Given the stretches of marginal habitat in canals that connect tracts, the surrounding land uses inhospitable to giant gartersnakes, potential fragmentation caused by Interstate 5 (I–5), and the distance between tracts of the Central Basin reserve and the Fisherman’s Lake reserve, it is unlikely that the Fisherman’s Lake reserve is currently demographically connected to the other reserves. Connectivity between the Fisherman’s Lake reserve and other habitats north of I–5 may have improved with the completion of the Giant Gartersnake Drainage Canal, constructed as mitigation for the Natomas Levee Improvement project that connects the North Drainage Canal just south of the Sacramento/Sutter County line with the West Drainage Canal just north of I–5. Most of this new canal appears to be suitable habitat for giant gartersnake and connects to the Fisherman’s Lake reserve through the West Drainage Canal, which is also suitable habitat. Within the Fisherman’s Lake reserve, three of the suitable tracts (Natomas Farms, Anne Rudin Preserve, and Cummings) are connected by approximately 3.5 km of canal habitats that compose Fisherman’s Lake, and by the intervening Sacramento Area Flood Control Agency (SAFCA) wetlands. The eastern boundary of the fourth suitable tract, Souza, is adjacent to the northernmost wetlands of the Natomas Farms and SAFCA tracts. The creation of the SAFCA wetlands provides much greater continuity of habitat within the Fisherman’s Lake area than was previously present. Movement data from radio-tagged snakes translocated to the SAFCA wetlands between 2019 and 2021 showed individual snakes moved between the Natomas Farms, SAFCA, and Cummings wetlands, but did not show signs of migration out of the Fisherman’s Lake reserve (Nguyen and others, 2024).

In contrast to the tracts of the Fisherman’s Lake reserve, the tracts of the Central Basin reserve are near those of the North Basin reserve, and these two areas are linked by a dense network of canals. The eastern edge of Ruby Ranch tract in the North Basin reserve is only approximately 3 km (by canal) from the Sills and Tufts tracts of the Central Basin reserve. Within the Central Basin reserve, tracts are nearly contiguous, with the exception of a 0.8-km gap between Bianchi West and Frazer South tracts. The intervening tract consists of rice agriculture and a canal with marginal habitat, so demographic connectivity among these tracts is likely and genetic connectivity is nearly certain. Perhaps a greater barrier to connectivity among Central Basin tracts is State Route 99 (SR 99), which lies between the Bianchi West and Sills tracts. Although this highway is a formidable barrier, it is possible for giant gartersnakes to cross it. A female giant gartersnake initially marked in 2010 at Bianchi West (east of SR 99) was captured at Sills (west of SR 99) three times in 2011. This individual almost certainly crossed through the 132-m-long single box culvert under SR 99, providing strong evidence for genetic connectivity across SR 99 in the Natomas basin (Halstead and others, 2013b). No such movements have been detected since, however. Given that the Sills tract and BKS tract contain the two most abundant populations of giant gartersnakes in the Central Basin reserve, connectivity across SR 99 could increase the probability of persistence of giant gartersnakes in this region as a whole.

Like the tracts of the Central Basin reserve, the tracts of the North Basin reserve are well-connected. No major highways fragment the North Basin tracts, and the only discontinuity between tracts containing suitable habitat is 1 km between the Lucich North and Nestor tracts. This gap occurs along the North Drain; however, we captured a snake in the Nestor tract in 2018 that was originally marked in the Lucich North tract in 2012, demonstrating at least some connectivity between these two areas. It is highly likely that all tracts in the North Basin reserve are genetically connected, and nearly all tracts are demographically connected with at least one other tract as well.

Overall, it is very likely that populations in all Central Basin and North Basin reserve tracts are genetically connected and that these tracts are also demographically connected to at least one other tract. These conditions help to promote genetic diversity, limit the effects of genetic drift and inbreeding depression, and may rescue small populations on some reserves by the migration of individual giant gartersnakes from neighboring reserves. In the future, maintaining this connectivity and its benefits to giant gartersnakes would include the continued availability of suitable habitat in canals that link wetland reserves. In contrast to the North and Central Basin reserves, connectivity between the Fisherman’s Lake reserve and the other reserves is far more tenuous. Although snakes in Natomas Farms and Cummings are almost certainly genetically connected and possibly demographically connected, the very small population in this area and isolation of these reserves from demographic rescue and genetic input from other, more abundant giant gartersnake populations to the north leaves them at risk for founder effects, inbreeding depression, and fixation of deleterious alleles through genetic drift, and it renders them very sensitive to both demographic and environmental stochasticity (random variation in birth/death rates or climatic conditions). When examining the posterior probability of occurrence at traplines within the Fisherman’s Lake area, we found it was mostly likely that none of the sampled traplines were occupied (posterior probability of 0.89). The establishment of these reserves and the additional marsh habitat created by SAFCA can provide the conditions that will allow this population to recover, but detailed demographic study of this population would ascertain whether more intensive management strategies (such as augmentation of the population with genetically distinct individuals to increase genetic diversity [Madsen and others, 1996, 2004]) would be beneficial in the Fisherman’s Lake area. The radio-telemetry study that began in 2018 is an important first step to determine the potential effectiveness of translocation of individuals from more abundant and presumably more genetically diverse populations, as a means to increase the abundance of snakes and the genetic diversity in sparse populations in the Fisherman’s Lake reserve. Individuals from the Central Basin reserve were translocated to the SAFCA wetlands in the Fisherman’s Lake reserve in 2019 and 2020 and tracked using radio-telemetry through 2021. Survival was much lower for these translocated individuals (8 percent of adult snakes survived >801 days) when compared to resident snakes (39 percent of adult snakes survived >1,154 days); however, juvenile snakes raised in captivity and released into the SAFCA wetlands had a relatively high survival rate of 60 percent during the 4-month period in which they were monitored (Nguyen and others, 2023).

These results indicate that captive rearing is a potential method for supplementing populations within the Fisherman’s Lake reserve. Further research would clarify the potential benefits and risks associated with this approach, including how supplemented individuals could influence the genetic makeup of the local population.

As seen in previous years, the habitat dynamics of canals with respect to giant gartersnakes can both improve or degrade from year to year based on annual fluctuations in water availability and growth of emergent vegetation. Monitoring these changes over time will be important to determine if any long-term trends exist and whether those trends are positive or negative for giant gartersnake persistence.

Distribution of Giant Gartersnake on Reserve Lands

The occupancy analysis for 2024 indicated that giant gartersnakes are expected to occur in approximately 47 percent of wetland and rice units on reserve lands, with occupancy highest in the Central Basin reserve. The probability of occurrence of giant gartersnakes in canals adjacent to rice fields did not differ from that of wetland sites, after accounting for variation in occupancy rates among the three reserve areas, in 2024. This, however, does not minimize the importance of rice agriculture as a supplemental wetland habitat for this conservation-reliant species (Halstead and others, 2019). It also should be noted that the Central Basin has historically had the highest proportion of sites occupied and the highest proportion of sites that are considered rice agriculture; these patterns remained evident in 2024. Because the Central Basin is dominated by rice and the Fisherman’s Lake reserve is dominated by wetlands, it is difficult to fully disentangle the effects of rice from geographic variation in probability of occurrence.

The lack of a strong effect of emergent vegetation on occurrence was notable in 2024. However, this likely does not necessarily mean that emergent vegetation is not important for giant gartersnakes; rather, given the model selection results, there could be additional benefits of rice agriculture, not measured in the current analysis, that affect giant gartersnake occurrence. Emergent vegetation is likely an important habitat component for giant gartersnakes, providing cover from predators and higher prey concentrations. Radio-telemetry study of giant gartersnake movement and habitat selection has shown that giant gartersnake preferentially select tules over other vegetation types (Halstead and others, 2016). Because tule marsh is historical habitat for giant gartersnakes, management for emergent vegetation, particularly tules, within a mosaic of other wetland vegetation is important.

The probability of occupancy was greatest in the Central Basin reserve, moderate in the North Basin reserve, and lowest in the Fisherman’s Lake reserve. Both the North Basin reserve and Central Basin reserve have a mix of rice and wetland habitat, whereas the Fisherman’s Lake reserve is primarily composed of recently created freshwater marsh.

Based on the dynamic occupancy model of all Natomas basin sites, the proportion of occupied wetland units on reserve lands decreased an average of 3 percent annually from 2011 through 2024. Relative to previous years, the probability that occupied sites became unoccupied (site extirpation) declined from 2023 to 2024 but was higher than in 2019. The probability that unoccupied sites became occupied (site colonization) in 2024 was higher than in 2023 and was similar to the average of all years of the study. The average 3-percent decline in occurrence was not steady, with some increases and decreases in the number of sites estimated to be occupied by giant gartersnake. The decline in the number of occupied sites was especially notable from 2011 to 2013 and from 2019 to 2021. Although apparently small in magnitude, if the long-term average decline in giant gartersnake occupancy continues, it would result in only 17 (6–53) occupied sites by 2050, assuming 60 occupied sites in 2011.

The dynamic occupancy model containing only TNBC-managed marsh sites in the northern and central basin had similar results to the model run on all sites but did show a smaller rate of decline in occupancy. The proportion of occupied TNBC-managed marshes within the northern and central basin decreased an average of 2 percent annually from 2011 through 2024. The probability that occupied sites within the northern and central managed marshes were extirpated has remained relatively low and stable from 2011 through 2024, and in comparison with the basin as a whole, site extirpation in the northern and central managed marshes is more stable than extirpation across all sites. This is encouraging given that most of the giant gartersnake population in the Natomas basin occurs in the northern and central regions. The probability that unoccupied sites became occupied (site colonization) was higher in 2024 than 2023; colonization of sites in the Northern and Central basins is more stochastic than colonization in the Natomas basin as a whole but seems to have a slight upward trend overall since 2018. The average 2-percent decline in occurrence fluctuated depending on the specific year but showed an overall downward trend in the number of North and Central managed marsh sites occupied from 2011 to 2018 and an overall increase in occupied North and Central managed marsh sites from 2019 through 2024, although still below the number originally occupied in 2011.

One potential mechanism leading to a decrease in the proportion of sites occupied across the Natomas basin is the extreme drought conditions from 2012 to 2015 and from 2021 to 2022. According to the California Department of Water Resources, California experienced the second driest water year (October 2020–September 2021) in 2021. Although water remained on TNBC reserves during the drought, it is unknown to what extent the source of water (surface water versus groundwater) affects giant gartersnake occupancy or demography, and precipitation may affect the productivity of lower trophic levels including giant gartersnake prey. Thus far, occupancy does not appear to have completely rebounded to earlier levels (for example, 2011), but 2022 showed a clear increase after 2 years of decline in 2020 and 2021 and appears to continue to increase in 2024. This was comparable to 2019, which showed a clear increase after 4 years of stability from 2015 to 2018. The rebound in occupancy in 2019 follows 3 out of 4 years of normal to above-average rainfall (2016, 2017, and 2019). Three years of favorable rainfall in a 4-year period might not be long enough for giant gartersnakes to recolonize every site from which they were extirpated during the drought, but 2019 showed some positive signs of recolonization.

Demography

In addition to monitoring where giant gartersnakes were distributed on Natomas Basin Conservancy reserves, another main objective was to provide information on the giant gartersnake populations within these sites. The following sections provide more context on the results of the demography analysis, including the abundance, density, survival, size distribution, and sex ratio of giant gartersnake populations in the Natomas basin.

Informing Effectiveness

Our study can inform the effectiveness of the NBHCP for conserving giant gartersnakes. The NBHCP is assessed based upon the acquisition of reserve lands; changes in the abundance or, preferably, demographic rates of giant gartersnakes; and land management activities to increase the distribution and health of giant gartersnakes in the Natomas basin (NBHCP; City of Sacramento and others, 2003).

The primary issue affecting giant gartersnakes throughout their range is habitat, and the Natomas basin is no different in this regard. Marshes that approximate natural tule marshes provide the best habitat for giant gartersnakes, promoting both higher densities and greater body condition than other habitats (Wylie and others, 2010). For example, a recent, long-term study of giant gartersnakes throughout the Sacramento Valley found that survival was positively related to the percentage cover of emergent vegetation at a site, though we note the maximum percentage cover of emergent vegetation in that study was 80 percent (Rose and others, 2018b); it is unlikely that further increases in emergent vegetation cover would increase giant gartersnake survival. Although giant gartersnakes have persisted in the rice agricultural landscape of the Sacramento Valley, the limited duration of rice fields as appropriate habitat (mid-May through August), the restriction of giant gartersnakes to structurally simple linear canals during the other 4 months of the active season, and the rice fallowing because of drought or late spring rains in recent years likely reduce the carrying capacity of agricultural habitats relative to natural or created marshes managed for giant gartersnakes. Nevertheless, rice agricultural habitats are likely the only agricultural habitats in which giant gartersnakes can persist (Halstead and others, 2010), and they provide connectivity between other patches of suitable habitat. Also, the survival rate of radio-tracked giant gartersnakes positively correlates to the area of rice agriculture surrounding their home range (Halstead and others, 2019). The Frazer South, Bianchi West, and Nestor tracts had the highest, second highest, and third highest density estimates, respectively, of giant gartersnakes in 2024, illustrating the importance of rice agricultural habitats in the Natomas basin, particularly in the central basin reserve. The change in density estimates from 2018 to 2023 (when the BKS wetlands had the highest density estimates) to 2024 (when density was highest in the Frazer South tract) also shows the importance of the surrounding rice as surrogate habitat for wetlands. As mentioned earlier, maintenance work on two wetlands in the BKS wetland complex occurred in 2024, resulting in the temporary draining of two wetlands and lowering water levels in two others. The subsequent increase in density in the Frazer South tract suggests that snakes from BKS likely moved to the Frazer South canal in response to the maintenance activities. Giant gartersnake density in the Lucich North cluster of sites, which is dominated by created marsh, was lower than many rice sites and remained lower than the giant gartersnake density at BKS despite wetland maintenance there. In 2022, there were low water levels at this location, with extensive areas of mudflat or shallow and hot water between emergent vegetation and deeper water. These conditions were likely caused by drought, which in turn resulted in only 18 percent of the normal contracted water allotment being provided to the Natomas basin in 2022. Although water levels in the Lucich North cluster improved dramatically in 2023 and 2024 with increased water allotments, there may have been a lag effect from 2022 that contributed to lower giant gartersnake densities at this cluster. Additional years of monitoring could be used to see if giant gartersnake densities, abundance, and survival at the BKS wetlands and the Lucich North Wetlands rebound after the water level changes at these sites in 2024 and 2022, respectively. Continued communication about management activities at these sites and others throughout the basin have been valuable in interpreting these results. TNBC has created managed marsh habitats and provides for the continuation of rice agriculture in the Natomas basin. It is beneficial to giant gartersnakes to maximize water levels in created marsh habitats, as in 2023 (all wetlands) and 2024 (all wetlands except those under construction and immediately adjacent in BKS).

Habitat for giant gartersnakes is only effective insofar as adequate water is supplied. The persistence of water adjacent to cover on the landscape throughout the active season is important for giant gartersnakes, and increased water availability is related to higher rates of survival for adult female giant gartersnakes (Reyes and others, 2017; Halstead and others, 2019). Drying of marshes, fallowing of rice fields in areas with high densities of snakes for more than a year, cultivation of alternative crops (especially if accompanied by lack of water in canals), and low or fluctuating water levels in marshes reduce the availability and quality of habitat for giant gartersnakes. Reducing the amount of rice grown in an area has the potential to negatively affect the survival of adult giant gartersnakes (Halstead and others, 2019). Our study found that managed marsh habitats within TNBC’s lands provided persistent aquatic habitat throughout the year and generally provided aquatic foraging and escape habitat adjacent to and below basking sites.

Another important component of giant gartersnake habitat is refuge from predators and environmental extremes. Mammal burrows and lodges and crayfish burrows offer important refuge for giant gartersnakes; maintaining these structures in association with marshes and canals likely benefits giant gartersnakes unless burrows threaten the integrity of the berms and levees, or they present a major hazard to humans or livestock. Muskrats (Ondatra zibethicus), California ground squirrels (Otospermophilus beecheyi), North American beavers (Castor canadensis), and crayfish likely improve habitat quality for giant gartersnakes by providing refugia in the form of burrows; muskrats and beavers further enhance habitat suitability by constructing piles of vegetation in and near water and reducing the density of cattails (thereby promoting the emergent vegetation/open water interface) through their foraging activity. Similar to muskrat lodges, tule thatch that accumulates naturally in mature tule marshes (like BKS) may also serve as important refuge from predators and temperature extremes and likely benefits giant gartersnakes. Giant gartersnakes actively select tule over other microhabitats available in their environment (Halstead and others, 2016).

Overall, giant gartersnake populations in the central basin reserves appear healthy. Fewer individuals were captured in the north basin reserves, and no giant gartersnakes were captured in the Fisherman’s Lake reserves. Our initial results indicate that conversion of additional habitats to created marshes and continuing to maintain high water levels in the marshes would benefit giant gartersnakes in the long term. Maintenance of rice agriculture also would help achieve connectivity, prey production, high adult survival in canals adjacent to rice fields, and other conservation goals. Minimizing ground disturbance, ensuring aquatic habitat is available in the spring for foraging upon emergence from winter refuges (Halstead and others, 2019), and maintaining stable and high water levels throughout the active season enhances habitat quality for giant gartersnakes. Lowering water levels in the early fall also may help to concentrate prey before hibernation, though further investigation would help ascertain the effectiveness of this practice.

Summary of Management Options that Likely Benefit Giant Gartersnakes

Based on the results of our study in the Natomas basin, the results of previous giant gartersnake studies, and existing literature on the species, management practices that are likely to benefit giant gartersnake populations involve maximizing the cover-open water interface, the availability of water from April through September, and maintaining hibernation habitat. Management practices would benefit from minimizing disturbance and certain types of vegetation. The following specific management options support these principles: