Halstead, B.J. Thompson, M.E. Amarello, Melissa Smith, J.J. Wylie, G.D. Routman, E.J. Casazza, M.L. 2017 Tabular Digital Data (CSV) Denver, CO U.S. Geological Survey data release https://doi.org/10.5066/F73F4NJK Brian J. Halstead Michelle E. Thompson Melissa Amarello Jeffrey J. Smith Glenn D. Wylie Eric J. Routman Michael L. Casazza 2017 Publication (journal) Journal of Wildlife Mangement The Wildlife Society These data are multi-state capture histories of 273 individual San Francisco gartersnakes collected at a site before and after a portion of the site was burned. Data collection began in 2008 and continued until 2013, and the prescribed fire was applied in the fall of 2010. Data were obtained to assess the effects of prescribed fire on San Francisco gartersnakes to inform conservation and management decisions. 2008 2013 ground condition Complete None planned USGS Thesaurus population dynamics controlled fires capturing (animals) endangered species None survival capture-mark-recapture USGS Metadata Identifier USGS:59a06fc7e4b038630d03062f Getty Thesaurus of Geographic Names San Mateo County, CA none The authors of these data require that users direct any questions pertaining to appropriate use or assistance with understanding limitations and interpretation of the data to the individuals/organization listed in the Point of Contact section. Brian J Halstead U.S. Geological Survey, Western Ecological Research Center mailing and physical

800 Business Park Drive

Suite D

Dixon CA 95620 United States 530-669-5076 bhalstead@usgs.gov U.S. Geological Survey, U.S. Fish and Wildlife Service, Peninsula Open Space Trust, National Science Foundation Graduate Research Fellowship Program USGS Biocomplexity Thesaurus Reptiles Integrated Taxonomic Information System (ITIS) 2017 ONLINE_REFERENCE Washington, D.C. Integrated Taxonomic Information System (ITIS) http://itis.gov expert advice;;identification keys;; All species in these data were robust and comprehensively identified by coloration, scale counts, and general appearance. All reptiles were identified to species or subspecies. The only subspecies included in these data is the San Francisco gartersnake (Thamnophis sirtalis tetratenia) Kingdom Animalia Subkingdom Bilateria Infrakingdom Deuterostomia Phylum Chordata Subphylum Vertebrata Infraphylum Gnathostomata Superclass Tetrapoda Class Reptilia Order Squamata Suborder Serpentes Infraorder Alethinophidia Family Colubridae Subfamily Natricinae Genus Thamnophis Species Thamnophis sirtalis Subspecies Thamnophis sirtalis tetrataenia San Francisco Garter Snake San Francisco Gartersnake Custom-written R code to run a multi-state Jolly-Seber model in JAGS. Model code is included as supplemental material in the journal publication. Brian J Halstead U.S. Geological Survey, Western Ecological Research Center mailing and physical

800 Business Park Drive

Suite D

Dixon CA 95620 United States 530-669-5076 bhalstead@usgs.gov Brian Halstead Unknown R code Data were cross-referenced with field notebooks. Capture histories for individual snakes were cross-referenced with morphological information to ensure that marks were read correctly. Data set is considered complete for the information presented, as described in the abstract. Users are advised to read the rest of the metadata record carefully for additional details. No formal positional accuracy tests were conducted No formal positional accuracy tests were conducted A database was created of data collected in the field. 2013 Data were extracted into a spreadsheet and a pivot table created to generate daily capture histories for each year, with "0" indicating a day that the individual wasn't captured and "1" indicating a day that an individual was captured. 2014 Capture histories were modified so that captures that occurred outside of burned areas were coded as "2" instead of "1." 2014 Field In 2007, we constructed and installed 24 drift fence and funnel trap arrays at random locations, stratified by upland habitat (grassland, shrubland, or forest) and proximity to aquatic habitats that serve as foraging sites, at the site (Fig. 1). Traps were located 55 ± 61 m (range = 1‒234 m) from aquatic habitat features. Eleven trap arrays were placed in the area burned in 2005 and 2010 (93 ± 38 m [35‒150 m] from 2010 burn boundary), and 13 trap arrays were placed in unburned areas (309 ± 304 m [12‒809 m] from 2010 burn boundary). To prevent desiccation and thermal stress of trapped individuals, we placed moistened sponges in the traps, placed shade covers on top of them, and checked traps twice daily while they were open. Trapping occurred in spring of each year from 2008 through 2013, with effort varying among years. Trapping was supplemented with cover object searches and visual surveys as time allowed, and we treated each day a survey was conducted as a sample. In total, we surveyed the site 40 days (40 days of trapping for 960 array-days) in 2008, 35 days (29 days of trapping for 666 array-days) in 2009, 59 days (56 days of trapping for 1320 array-days) in 2010, 21 days (21 days of trapping for 504 array-days) in 2011, 62 days (60 days of trapping for 1440 array-days) in 2012, and 47 days (46 days of trapping for 1104 array-days) in 2013. In 2009 and 2010, some arrays were opened late because of flooding. In 2008–2010, we processed all captured individuals in daily batches; thereafter, we processed them individually in the field immediately after removal from the trap. In the former case, we transported captured individuals in cotton sacks to a temporary field station where they were maintained overnight in climate-controlled chambers to prevent thermal stress until they could be processed the following morning. We measured (snout-vent length [SVL], tail length [TVL], and mass), determined the sex of, uniquely marked (passive integrated transponder [PIT] tag or unique brand [Winne et al., 2006]), and photographed each individual. We released all individuals at their location of capture as soon as possible following processing. All individuals were handled in accordance with the University of California, Davis, Animal Care and Use Protocol 9699 and as stipulated in U.S. Fish and Wildlife Service Recovery Permits TE-170403 (2007 and 2008), TE-020548-5, and California Fish and Game MOU (SC-009313, SC-009315). See Halstead et al. (2011) for additional details about field methods. We used a Bayesian implementation of multi-state Jolly-Seber models to examine the effects of prescribed fire on the apparent survival (ϕ, the probability of surviving and remaining on the study site), entry probability (γ, the probability an individual enters the population through birth or immigration), abundance (N), and population growth (λ) of San Francisco gartersnakes. We used the multi-state implementation of the Jolly-Seber model (Kéry and Schaub 2011), modified to include two observable states (in burned area or outside burned area, hereafter referred to as location) and to accommodate the robust design (Pollock 1982), with each day of sampling explicitly accounted for in the analysis (File S1, available online in Supporting Information). Our full model consisted of effects of prescribed fire and random annual heterogeneity on ϕ (ϕlocation × fire + j, where location indicates the burned or unburned area, fire indicates post-burn [2011 ‒ 2013], and j indicates year), and prescribed fire and random daily heterogeneity on daily capture probability (plocation × fire + t, where t indicates day). We also allowed γ to vary by year and location (γlocation × j), but we treated year as fixed, rather than random, effects. We modeled prescribed fire effects on demographic parameters as persistent through the end of the study (three years post-burn). We also allowed transition probabilities (ψ, movement from the burned area to the unburned area and vice versa) to vary before and after the fire (ψlocation × fire). We specified that fire affected p and ϕ at each location separately, with the model selection constraint that if fire affected one location, its effect was estimated for the other (i.e., the model was constrained to have fire affect both or neither of the areas). The net effect of fire on p and ϕ was then estimated as the coefficient for the effect of fire in the unburned area minus the coefficient of the effect of fire in the burned area. All parameters, except annual heterogeneity in ϕ and daily heterogeneity in p, were estimated separately for each location. We parameterized the state transition process as: Not yet recruited Alive in burned area Alive in unburned area Dead Not yet recruited 1 – (γB + γU) γB γU 0 Alive in burned area 0 (1 – ψBU)ϕB ψBUϕU 1 – ((1 – ψBU)ϕB + ψBUϕU) Alive in unburned area 0 ψUBϕB (1 – ψUB)ϕU 1 – (ψUBϕB + (1 – ψUB)ϕU) Dead 0 0 0 1 , and the observation process, conditional on location, as: Observed in burned area Observed in unburned area Not observed Not yet recruited 0 0 1 Alive in burned area pB 0 1 – pB Alive in unburned area 0 pU 1 – pU Dead 0 0 1 . Note that this parameterization assumes that individuals transition between locations first, then survive. We think that this is a reasonable assumption because our sampling occurs at the beginning of the active season each year, and likely captures individuals in the location at which they completed the previous active season prior to redistribution. We fitted the model using data augmentation by adding 2000 all-zero capture histories for pseudo-individuals to estimate abundance as a derived parameter (Royle et al. 2007, Kery and Schaub 2011, Tenan et al. 2013). Pseudo-individuals are unobserved individuals, and abundance estimation proceeds by estimating how many of these pseudo-individuals were part of the population, but unobserved. This number of pseudo-individuals was deemed adequate because the posterior distribution of the sum of the superpopulation for both locations was well below 2000 (Royle and Dorazio 2008). We calculated the following derived parameters for burned and unburned areas separately: annual abundance, Nj; the number of snakes present during the entire study, Nsuper; annual recruitment (the number of individuals born in year j − 1 and surviving until year j, plus immigrants between year j − 1 and year j), Bj, from which we calculated annual per capita recruitment rates as Bj / Nj−1; and the population growth rate, λj, calculated as Nj+1 / Nj. We selected among models to evaluate the support for an effect of fire using indicator variables on model coefficients including random year or day effects (Kuo and Mallick 1998, Royle and Dorazio 2008). All parameters were given vague priors (logit-scale intercepts and coefficients = normal[mean = 0, SD = 10]; logit-scale SD of temporal random effects = uniform[min = 0, max = 10]; state-specific annual entry probability = uniform[0, 0.5]; and indicator variables = Bernoulli[0.5]). We ran the model on 5 chains of 10,000 iterations each after a burn-in of 10,000, and thinned the output by a factor of five to base posterior inference on 10,000 samples. We called JAGS (JAGS Version 3.4.0, http://sourceforge.net/projects/mcmc-jags/files/, accessed 21 October 2014) from R 3.1.0 (R Version 3.1.0, http://cran.us.r-project.org/, accessed 21 October 2014) using the package ‘rjags’ (https://cran.r-project.org/web/packages/rjags/index.html, accessed 21 October 2014) to obtain samples from the posterior distribution of each parameter. We examined history plots and the Gelman-Rubin statistic (Gelman and Rubin 1992) for each parameter for evidence of lack of convergence, and found no evidence for lack of convergence. For all parameters, we represent posterior distributions with the median and 95% symmetrical credible interval (i.e., the 0.025 and 0.975 quantiles of the distribution). Please refer to the journal atricle for a complete list of references cited. Brian J. Halstead Michelle E. Thompson Melissa Amarello Jeffrey J. Smith Glenn D. Wylie Eric J. Routman Michael L. Casazza 2017 Publication (journal) Location data are withheld because of sensitive nature of data (endangered species). Tsir_Fire_MS_CH_data.csv Multi-state capture histories of San Francisco gartersnakes Field-collected data year Calendar year in which the data were collected Producer defined 2008 2013 year id Unique identifier for an individual snake Producer defined 1 273 unique code for observed individual d1-d62 Indicators for capture and location of capture of individual snakes on days 1 - 62 of sampling within each year Producer defined 0 not captured Producer-defined 1 captured in the burned area Producer-defined 2 captured in the unburned area Producer-defined U.S. Geological Survey, ScienceBase mailing address

Denver Federal Center

Building 810

Mail Stop 302

Denver CO 80225 United States 1-888-275-8747 sciencebase@usgs.gov Unless otherwise stated, all data, metadata and related materials are considered to satisfy the quality standards relative to the purpose for which the data were collected. Although these data and associated metadata have been reviewed for accuracy and completeness and approved for release by the U.S. Geological Survey (USGS), no warranty expressed or implied is made regarding the display or utility of the data on any other system or for general or scientific purposes, nor shall the act of distribution constitute any such warranty. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government. This .csv file contains data that can be used transformed into individual x day x year arrays and analyzed using the model in supplemental material associated with the journal article. 20200830 Brian J Halstead U.S. Geological Survey, Western Ecological Research Center Research Wildlife Biologist mailing and physical

800 Business Park Drive

Suite D

Dixon CA 95620 United States 530-669-5076 bhalstead@usgs.gov FGDC Biological Data Profile of the Content Standard for Digital Geospatial Metadata FGDC-STD-001.1-1999