M.S.

Extensive development has destroyed and fragmented wildlife habitat in the Lower Florida Keys. The Lower Keys marsh rabbit (LKMR; Sylvilagus palustris hefneri) and the silver rice rat (SRR; Oryzomys argentatus) are listed by the United States Fish and Wildlife Service (USFWS) and the Florida Fish and Wildlife Conservation Commission (FFWCC) as endangered species. Both species depend on coastal prairies, freshwater marshes, and intertidal salt-marsh zones. The objective of this study was to meet specific, species-level recovery goals and to add reliable information that may modify or support current recovery plans. Specifically, I (1) evaluated the use of LKMR reintroduction to suitable habitat, (2) examined characteristics of habitat used by LKMR, and (3) surveyed the Lower Florida Keys for SRRs, documenting current range and examining survey results for the past decade. I reintroduced 7 rabbits (3 males, 4 females) to suitable habitat on Water Key, and monitored their survival and release-site fidelity. All reintroduced rabbits survived and some reproduced, suggesting these translocation techniques are a viable tool for recovery. On Boca Chica Key, I radio-collared 13 LKMRs and compared vegetation characteristics between core-use and avoided areas within home ranges. Binary logistic regression associated rabbit use with high vegetation heights (7–8 dm), low canopy coverage (<=10%), high bunchgrass densities (2.5–3.8/sq m), and forb presence (>5%), supporting the hypothesis that LKMRs may be detrimentally impacted by hardwood encroachment into salt-marsh habitats. For LKMR recovery, I recommend management to resist hardwood encroachment, together with active predator control. I surveyed 36 locations on 18 islands for SRRs, capturing rats on 12 islands, including 2 on which SRRs had not previously been found. Comparisons of my data with historic data suggest SRRs either have increased in abundance over the past decade or that previous trapping efforts were not effective. Abundance of SRRs does not appear to be significantly different from that of populations of rice rats on the mainland. The USFWS and FFWCC should consider revising the conservation status of the SRR; however, it still should be regarded as a unique evolutionary unit with a very limited potential range.

Swamp rabbit and eastern cottontail populations have declined substantially in Illinois within the last half-century. Habitat loss and fragmentation, particularly of bottomland hardwood (BLH) forest, have left swamp rabbit populations patchily distributed along major rivers in the southern portion of the state. In addition, the decline of early-successional upland habitats due to changes in farming techniques have led to as much as 90% declines in cottontail populations in Illinois. Managers need information on basic vital rates and habitat use to conserve both species in southern Illinois; however, many questions regarding demography and behavior have not been investigated. My research examined the importance of several factors that may influence survival and habitat use among swamp rabbits and eastern cottontails. My specific research objectives were to 1) estimate seasonal and annual survival rates and identify primary mortality agents, 2) examine the influence of intrinsic factors and habitat variables on annual and seasonal 50% core area (CA) and 95% home range (HR) sizes, 3) assess differences in space use and habitat use by season, and 4) evaluate differences in space and habitat use between species for swamp rabbits and cottontails in BLH forests in southern Illinois. During December-March 2009-2013, I live-trapped and radiocollared adult swamp rabbits (>1.9 kg) and cottontails (>1.0 kg) at 7 BLH sites along the Cache River and Cypress Creek within the Cypress Creek National Wildlife Refuge in southern Illinois. I monitored rabbits for survival every 24-48 hr and estimated radiolocations by triangulation more than twice weekly during morning (0500-0900 hr), daytime (0900-1700 hr), and evening (1700-2400 hr) time periods on a rotating schedule. I delineated annual and seasonal 50% CA and 95% HR isopleths using kernel density estimators, comparing winter-spring (W/S; 21 Dec-19 Jun) and summer-fall (S/F; 20 Jun-21 Dec) seasons. During May-August 2012, I sampled microhabitat in 0.02-ha circular plots (3.5 plots/ha) randomly placed throughout home ranges of rabbits at each site. In ArcGIS, I classified macrohabitat patches into 1 of 4 cover types: agriculture, early-successional BLH (EBLH), mature BLH (MBLH), or upland. I also measured the distance from all radiolocations to the nearest patch of each cover type and the nearest river or creek. I modeled the influence of species, sex, and season on annual survival using parametric survival regression models. Then, I used the best model (based on AICc) to examine the influence of habitat covariates on survival. I also modeled the influence of habitat on annual and seasonal CA and HR sizes using generalized linear and generalized linear mixed effects models with AIC model selection. Finally, I estimated conspecific and heterospecific CA and HR overlap for swamp rabbits and cottontails, and examined differences between species in space and habitat use. I documented causes of mortality and estimated survival for 129 swamp rabbits and 75 cottontails during the period of study. Predation (71%) was the primary mortality agent for both species, followed by weather (9%) and hunter harvest (6%). Models with survival rates differing by species and season received the most support; swamp rabbits had higher estimated annual survival (0.37  0.05) than did cottontails (0.20  0.05), and survival for both species was lowest during the W/S season (βW/S=-2.28  0.46). None of the habitat covariates that I measured apparently influenced survival. Core areas and HRs were estimated for 60 swamp rabbits (SR; 34 M, 26 F) and 21 cottontails (CT; 10 M, 11 F) during the W/S season; of these, 57 swamp rabbits and 11 cottontails lived long enough to estimate S/F seasonal and annual home ranges. The average annual CA and HR for swamp rabbits were 2.49  1.42 ha and 11.60  5.81 ha, respectively. Cottontails had an average annual CA and HR size of 2.48  1.26 ha and 13.54  7.24 ha, respectively. Core areas and HRs for both species during the were larger during W/S than S/F (CA: βW/S=0.59  0.11, wi=1.0; HR: βW/S=0.53  0.11, wi=1.0). Seasonal CA sizes increased with decreasing proportions of woody ground cover within CAs (SR: βShrubs=-2.75  0.50, wi=1.0; CT: βShrubs=-2.30  0.74, wi=0.91). Few macrohabitat variables influenced space use for either species. The coefficient of variation in patch size within 1 km of study sites was positively associated with space use for both swamp rabbits (CA; β=0.01  0.004, wi=0.81; HR: β=0.01  0.003, wi=0.95) and cottontails (CA; β=0.02  0.01, wi=0.29; HR: β=0.02  0.01, wi=0.23); I did not detect significant differences between species in pairwise conspecific overlap within CAs or pairwise conspecific volume of intersection (VI) over the entire utilization distribution. Median conspecific CA overlap was higher during W/S than S/F, with mean CA overlap proportions of 0.20 ± 0.21 (range: 0−0.87) and 0.10 ± 0.18 (range: 0−0.73) respectively. Median conspecific VIs also differed between seasons, with mean pairwise VIs of 0.32 ± 0.19 (range: 0−0.77) during winter-spring and 0.19 ± 0.18 (range: 0−0.73) during S/F. Heterospecific CA overlap was 48% and 46% lower than conspecific overlap during W/S and S/F, respectively. Differences in habitat use between species were apparent. Swamp rabbits had HRs and CAs in areas with higher basal area compared to cottontails during both seasons (all U≥483, Z≥2.83, p≤0.005). Multivariate tests for both CAs and HRs indicated a significant effect of species on habitat use (CA: T2=0.58, F3, 71=13.70, p<0.001; HR: T2=0.24, F3, 71=5.80, p=0.001). During W/S, swamp rabbits had CAs and HRs composed of significantly higher proportions of EBLH (CA: F1, 73=16.46, p<0.001; HR: F1, 73=8.55, p=0.005) and MBLH (CA: F1, 73=17.99, p<0.001; HR: F1, 73=7.78, p<0.007). Swamp rabbits were located significantly closer to a permanent watercourse (F1, 79=24.18, p<0.001) than cottontails. Indeed, 95% of all swamp rabbit radiolocations were ≤332.0 m away from a permanent watercourse (mean=169.0  100.0 m; range=1.0−571.0 m; Figure 18), whereas 95% of cottontail radiolocations were ≤536.0 m away from a permanent watercourse (mean=289.0  142.0 m; range=1.7−670.0 m; Figure 18). Swamp rabbits also were significantly closer to MBLH patches (F1, 79=9.05, p=0.003) and farther from agriculture (F1, 79=12.36, p=0.001) than cottontails. My study represents the most complete record to date on survival and habitat use by swamp rabbits. Rabbit survival was positively associated with basal area so management actions that provide woody cover for concealment and thermoregulation may benefit both species. Although cottontails in my study used early-successional BLH, patterns of space and habitat use described here demonstrate that cottontails remained on the periphery of bottomlands. My study confirms the utility of early-successional BLH to both species; however, stands that are located too far from a permanent water sources are unlikely to be used by swamp rabbits, and may be less suitable for other BLH specialists as well. Allowing grasslands and crop fields to succeed into old fields containing bushes, vines, and other woody species will benefit both cottontails and swamp rabbits. Within BLH forests, canopy gaps can be created to promote tree regeneration and woody ground vegetation such as vines and shrubs. Finally, upland early-successional habitats that border bottomland forests are especially important as refugia for swamp rabbits during flooding.

Reforestation of bottomland hardwood (BLH) forests has occurred within the Lower Mississippi Alluvial Valley (LMAV), USA, to support a wide range of ecosystem services, but especially wildlife habitat enhancement. As ecosystem restoration efforts proceed in BLH ecosystems, managers and policymakers are seeking criteria to evaluate wildlife habitat enhancement goals. Specialist wildlife that evolved within forest ecosystems can be sensitive to the composition, structure, and function of an ecosystem in relation to the system's natural or historical range of variation and thereby serve as indicators of habitat quality. The swamp rabbit (Sylvilagus aquaticus) is a specialist of BLH forests throughout the LMAV and therefore may be an appropriate indicator species for this ecosystem. In Chapter 1, I reviewed peer-reviewed literature to evaluate the utility of swamp rabbits as an indicator species according to three commonly-used criteria: habitat factors defining swamp rabbit relationships to BLH forests, the importance of swamp rabbit habitat to other wildlife, and the efficiency of swamp rabbit monitoring. I concluded that the swamp rabbit is a suitable indicator of wildlife habitat quality in BLH ecosystems in the LMAV because they evolved and remain endemic to the ecosystem, use habitat that integrates desirable characteristics that positively influence wildlife biodiversity, and are easy to monitor routinely.

Genetic analyses of two endangered species of mammals in the Lower Keys of Florida (Lower Keys marsh rabbit, LKMR, Sylvilagus palustris hefneri; silver rice rat, SRR, Oryzomys palustris natator) were performed to evaluate the genetic structure of their populations. Mitochondrial sequence data (control region; 763 base pairs (bp), LKMR; 788 bp, SRR) were used to explore patterns of genetic variation within and among island populations in both species. Analysis of the SRR also included 8 polymorphic nuclear microsatellite loci (9 to 16 alleles). Phylogenetic analyses of mitochondrial sequence data for both species revealed two main lineages corresponding to eastern and western localities, with high levels of genetic structuring (LKMR FST = 0.982, SRR ΦST = 0.916). The two species differed in the level of sequence divergence between eastern and western populations (LKMR, 19 bp; SRR 4 bp). In addition to an overall similar pattern of genetic subdivision, populations of both species possessed low levels of mtDNA variation (haplotypic diversity in the LKMR = 66.1%, SRR = 58.6%). Microsatellite analyses of the SRR revealed subdivision between eastern and western regions. Although less pronounced than the structure observed in mtDNA, the overall pattern was still apparent. Additional examination of divergence between mainland and Lower Keys rice rats revealed a genetic division that indicated a lack of recent gene exchange between the regions (i.e. no shared haplotypes, the presence of private alleles, and distinctive separation in numerous analyses). Although this degree of division does not warrant species designation, the levels and patterns of divergence, both morphological and genetic, do suggest genetic isolation of mainland and island forms. This fact, along with restricted gene flow between the Lower Keys and the Everglades, suggests that the SRR is on an evolutionary trajectory separate from its mainland counterparts and validates its identification as a separate subspecies, Oryzomys palustris natator. Finally, the genetic division between eastern and western populations of the SRR and LKMR suggests that populations of both species in these two regions of the Lower Keys should be treated as separate management units, especially when considering the enhancement of populations via translocations.

Subspecific taxonomic designations solely based on morphological characters can often lead to erroneous assumptions about the evolutionary history of populations. This study sought to investigate evolutionary questions and conservation implications associated with morphological subspecific designations of island populations. To this end, I focused my attention on the Lower Keys of Florida, a unique chain of islands with well-known geologic history and rich in endemic, endangered subspecies. I employed genetic analyses to evaluate historical variation and contemporary restriction of gene flow between the endangered Lower Keys marsh rabbit (Sylvilagus palustris hefneri) and its sister mainland taxa. A Bayesian phylogeny using 1063 base pairs of the mitochondrial cytochrome b gene did not recover reciprocal monophyly of the three named subspecies, and a 95% statistical parsimony haplotype network showed haplotypes being shared among subspecies. Furthermore, clustering analyses using 10 microsatellite loci identified a break within the Lower Keys, separating the western Lower Keys from the island of Big Pine Key. Surprisingly, Big Pine Key grouped with mainland populations and exhibits higher genetic diversity than the western Lower Keys islands. These unexpected findings suggest either a stepping-stone colonization pattern or recent gene flow between the mainland and Big Pine Key via natural dispersal or undocumented man-mediated transfers. Although these results suggest that subspecies designations within S. palustris are unwarranted, this study supports the designation western Lower Keys population as a discrete unit of conservation with regard to both DPS and ESU criteria. The importance of using several lines of evidence to uncover the evolutionary history of populations and implications for the conservation of island populations are discussed.

The Lower Keys marsh rabbit (Sylvilagus palustris hefneri, LKMR), 1 of 3 subspecies of Sylvilagus palustris, is endemic to the Lower Florida Keys. The LKMR is listed as an endangered species due to predation by feral and free roaming domestic cats (Felis catus) and raccoons (Procyon lotor), road mortality, effects of storm surges, sea level rise, the small declining metapopulation size, and possible habitat loss from hardwood encroachment. The purpose of this study was to determine the current LKMR density on lands managed by the United States Navy, Naval Air Station Key West and evaluate how vegetation structure and patch elevation effect LKMR population density. I conducted fecal pellet counts to determine LKMR density, collected vegetation data using percent composition of ground cover, Robel range pole, and point-centered quarter methods, and obtained data on patch area and elevation. I used simple linear regression to assess the relationship between LKMR density and 9 measured vegetation characteristics, patch area, and patch elevation to determine which variables have an influence on LKMR density and the relationship between them. In my examination of the simple regression models, 6 out of the 11 variables appeared to influence LKMR population density. The average per patch percent composition of nonliving material and grasses, maximum height of vegetation at the range pole, distance to nearest woody vegetation, patch elevation, and visual obstruction readings (VOR) individually accounted for 26.4%, 30.4% , 18.1%, 8.5%, 6.8%, and 1.4% of the variability in LKMR density, respectively. According to the regression models, LKMR density increased in patches with greater amounts of grasses and with greater distance to woody vegetation. Habitat management is vital to the recovery of the LKMR and needs to focus on providing greater amounts of grasses and reducing the amount of woody vegetation encroachment to enhance LKMR population density.

Listed as federally-endangered in 1990, the Lower Keys marsh rabbit (LKMR, Sylvilagus palustris hefneri) exists as a metapopulation in patches of wetland habitat in Florida’s Lower Keys. This study sought to address 2 priority actions identified by the LKMR Recovery Team: (1) monitoring of populations and (2) reintroduction. Monitoring the distribution and status of LKMR populations is critical for targeting future management actions. Informal transects for rabbit fecal pellets were used to survey habitat patches documented in1988–1995 surveys and to identify additional patches of occupied and potential habitat. Next, a buffer was created around patches to help managers account for uncertainty in rabbit movements and to identify groups of patches that might function as local populations. Surveys included 228 patches of occupied and potential habitat, 102 of which were occupied by rabbits. Patches were arranged in 56 occupied and 88 potential populations. Surveys revealed new patches of both occupied and potential habitat. Considering only areas included in 1988–1995 surveys, however, revealed a net decrease in the number of occupied patches. Many of the recently extirpated populations, which tended to occupy the periphery of larger islands or small neighboring islands, were unlikely to be recolonized without human intervention. Reintroduction provides a means of artificially recolonizing potential habitat. Two pilot reintroductions were conducted to evaluate this conservation strategy for the species. The second reintroduction was postponed, but the first effort met all criteria for short-term success, including survival comparable to a control group, fidelity to release sites, and evidence of reproduction. There are a limited number of potential source populations for translocations. Future efforts should consider using in-situ captive breeding to prevent potential long-term impacts to these populations. Few potential release sites exhibited suitable habitat quality and landscape context. Thus, for reintroduction to be more widely-applied for this species, it must be part of a comprehensive management plan involving land acquisition, control of secondary impacts from development, and habitat restoration and enhancement.

Listed as federally-endangered in 1990, the Lower Keys marsh rabbit (LKMR, Sylvilagus palustris hefneri) exists as a metapopulation in patches of wetland habitat in Florida’s Lower Keys. This study sought to address 2 priority actions identified by the LKMR Recovery Team: (1) monitoring of populations and (2) reintroduction. Monitoring the distribution and status of LKMR populations is critical for targeting future management actions. Informal transects for rabbit fecal pellets were used to survey habitat patches documented in1988–1995 surveys and to identify additional patches of occupied and potential habitat. Next, a buffer was created around patches to help managers account for uncertainty in rabbit movements and to identify groups of patches that might function as local populations. Surveys included 228 patches of occupied and potential habitat, 102 of which were occupied by rabbits. Patches were arranged in 56 occupied and 88 potential populations. Surveys revealed new patches of both occupied and potential habitat. Considering only areas included in 1988–1995 surveys, however, revealed a net decrease in the number of occupied patches. Many of the recently extirpated populations, which tended to occupy the periphery of larger islands or small neighboring islands, were unlikely to be recolonized without human intervention. Reintroduction provides a means of artificially recolonizing potential habitat. Two pilot reintroductions were conducted to evaluate this conservation strategy for the species. The second reintroduction was postponed, but the first effort met all criteria for short-term success, including survival comparable to a control group, fidelity to release sites, and evidence of reproduction. There are a limited number of potential source populations for translocations. Future efforts should consider using in-situ captive breeding to prevent potential long-term impacts to these populations. Few potential release sites exhibited suitable habitat quality and landscape context. Thus, for reintroduction to be more widely-applied for this species, it must be part of a comprehensive management plan involving land acquisition, control of secondary impacts from development, and habitat restoration and enhancement.

Rapid development and urbanization of the Lower Florida Keys in the last 30 years has fragmented the habitat of the Lower Keys marsh rabbit (Sylvilagus palustris hefneri) and threatened it with extinction. Current threats exist at multiple spatiotemporal scales and include threats due to development, invasive species, and global climate change. On Boca Chica Key, the Lower Keys marsh rabbit (LKMR) exists as a metapopulation on Naval Air Station-Key West (NASKW). I conducted a population viability analysis to determine the metapopulation's risk of extinction under multiple management scenarios by developing a spatially-explicit, stage-structured, stochastic matrix model using the programs RAMAS Metapop and ArcGIS. These management scenarios include clearance of airfield vegetation, habitat conversion, and control of feral cats as an invasive species. Model results provided the Navy with relative risk estimates under these different scenarios. Airfield clearance with habitat conversion increased extinction risk, but when coupled with feral cat control, risk was decreased. Because of the potential of sea-level rise due to human-induced global climate change, and its projected impact on the biodiversity of the Florida Keys, I estimated the impacts of rising sea levels on LKMR across its geographic distribution under scenarios of no, low (0.3m), medium (0.6m), and high (0.9m) sea-level rise. I also investigated impacts due to 2 treatments (allowing vegetation to migrate upslope and not allowing migration), and 2 land-use planning decisions (protection and abandonment of humandominated areas). Not surprisingly, under both treatments and both land-use planning decisions, I found a general trend of decreasing total potential LKMR habitat with increasing sea-level rise. Not allowing migration and protecting human-dominated areas both tended to decrease potential LKMR habitat as compared with allowing migration and abandoning human-dominated areas. In conclusion, conservation strategies at multiple scales need to be implemented in order to reduce threats to LKMR, such as development, invasive species, and global climate change.