Summary:
Roads are a source of mortality for federally threatened eastern massasauga rattlesnakes (EMR; Sistrurus catenatus) in Michigan. However, a clear understanding of space use and habitat selection of EMR near roads is lacking. Identifying when, where, and why EMR use spaces near roads could inform decisions that reduce the risk of road-related mortalities. Objectives of this research were to survey for EMR at sites along state trunkline roadsides, and report on occupancy probability, demonstrate how weather influences EMR visibility during visual encounter surveys, and explore environmental factors resulting in high EMR space use near roads. During 54 formal (i.e., following a previously published survey protocol) EMR surveys at 35 sites, five EMR were detected (at two sites) along state trunklines. A site refers to an area within the modeled habitat range of EMR, <300m from a roadside and consisting of vegetation cover types (e.g., wet prairie, emergent marsh, early successional forest edges, and upland grassland lightly interspersed with woody vegetation) suitable for EMR use during the overwintering and active seasons. Individual EMR surveys had low site-level detection probability (mean = 0.16 (SE = 0.11)), but repeated surveys increased detection probability to near 1.00 at some sites. In total, 24 EMR were radio-tagged at the two occupied sites and collected locations 1-5 times per week. During relocations, observers recorded percent of the body visible and if the EMR was in a burrow (i.e., unavailable for detection). A Bayesian beta regression was used to model EMR body exposure and burrow use during the active season (April-October; 2020-2022) to discern when EMR are most visible. At the time of relocation, it was found that estimated body exposure (x = 42%, SE = 3%, range = 0-100%) was positively correlated with ambient temperature, increasing from ~25% exposure at ~15° C (59° F) to ~55% at ~40° C (104° F). Also, the probability of burrow use was quadratically related to Julian day broadly representing the transition from inactive to active seasons. Probability of EMR burrow use was low (<0.10) from June to August. Male EMR in this study were less likely (x = ~0.02%, range = ~2 - 4%) to use burrows during the active season than females (x = ~8%, range = ~2 - 42%), likely because the majority of tagged EMR in this study were gravid females with more demanding thermoregulatory needs. Brownian Bridge kernel home ranges were calculated for four (3 – female, 1 – male) telemetered EMR with ≥20 relocations and combined their estimated home ranges into a single spatial layer representing intensity of snake use. A Bayesian beta regression was used to model intensity of snake use based on environmental covariates for each site. At Site 1, probability of use was greatest within 70m of the road and correlated negatively with canopy cover (use decreased from 45% at 0% canopy cover to 20% at 90% canopy cover). At Site 2, distance to the road had a slightly positive impact on EMR probability of use, where probability of use increased from by 5% (42% to 47%) as distance from the road approached 100m. However, as distance from the road approached 300 m, use declined sharply by 23% (50% to 27%). No ground-layer vegetation characteristics were significant to EMR probability of use but note that the vegetation plots occurred in areas used by telemetered EMR (i.e., mostly grass and forb dominated), so variation among plots was minimal. Temperature loggers were systematically located above and below ground near roads at sites with telemetered EMR. Above ground temperatures were highly variable with no consistent pattern throughout the active season. However, below ground loggers showed that temperatures were, on average, 1.27° C (34.3° F) warmer closer to roads during the active season, potentially explaining the affinity for the gravid females in this study to use roadsides. Consistent with other studies on factors affecting EMR detection probability, the results confirm that ambient temperature is one of the most important factors affecting EMR, and visual encounter surveys should be conducted during optimal temperature ranges (~30-40° C (86- 104° F) in this study). Additionally, space use of EMR occupying habitat within 300m of roadsides had core use areas close (e.g., within 20 m) to roads. This is likely due to suitable habitat conditions provided by vegetation height management (e.g., mowing), and the physical properties of the roadbed (i.e., warmer below ground temperatures). Gravid female EMR (n=5) and their birthing sites were located within 10 m of state trunkline, an area typically mowed on a regular basis during the whole summer season. Based on these observations, it would be advantageous to avoid mowing in late July and early August during the EMR birthing and brooding period, and from late May to Early June and late August to September when EMR are moving between active season and over-winter habitats
Publications:
