Potential geographic distribution and host utilization of Alectorobius puertoricensis (Ixodida: Argasidae) in the Caribbean Basin: a putative vector of African swine fever virus Authors: Taylor G Donaldson, Hsiao-Hsuan Wang, Charluz M Arocho Rosario, Pete D Teel, William E Grant, Gabriel L Hamer, Kathleen C O’Hara, Angela M James, Phillip E Kaufman Published: 2026 ABSTRACT Abstract Alectorobius puertoricensis (Fox 1947) (Ixodida: Argasidae) is a nidicolous soft tick distributed throughout the Caribbean Basin and has been experimentally shown to transmit African swine fever virus (ASFV). Given the reemergence of African swine fever (ASF) in the Caribbean in 2021, defining the potential distribution, host utilization, and environmental conditions of this tick species is essential for assessing its regional ASFV risk in the western hemisphere. This study compiled historical occurrence records and hosts associated with A. puertoricensis from published literature and diagnostic reports. These data were used to develop a species distribution model using Maxent to estimate the potential geographic distribution of A. puertoricensis across the Caribbean Basin. Environmental predictors were selected based on biological relevance and collinearity assessment within a defined area. Model performance indicated good discriminatory ability, with the potential geographic distribution concentrated in lowland tropical and subtropical regions characterized by warm temperatures and variable precipitation. Host records demonstrated a broad variety of vertebrate hosts, with rodents representing the most frequently documented hosts, consistent with a potential role in maintaining local tick populations. Several occurrence records placed A. puertoricensis in peridomestic and livestock-associated environments relevant to swine exposure. When interpreted in the context of regional animal movement and connectivity, areas where environmental suitability for A. puertoricensis overlaps with ports, coastal communities, and population centers may represent elevated concern for ASF exposure should the virus be introduced. These findings provide a framework for incorporating soft tick ecology into ASF preparedness and surveillance efforts in the Caribbean. Introduction Alectorobius puertoricensis (Ixodida: Argasidae) was first described in 1947 from specimens collected from brown rats (Rattus norvegicus) in Santurce, San Juan, Puerto Rico (Fox 1947). A collection had been made in Panama prior to its description in 1931, but specimens were misidentified as Ornithodoros talaje (Fairchild et al. 1966). A redescription of Ornithodoros (Alectorobius) puertoricensis by Endris et al. (1989) provided a comprehensive list of known hosts and geographic distribution of tick collections known at that time. The taxonomic status of Argasidae has been complex and has led to ongoing debates (Mans et al. 2019). Filippova (1961) had assigned it to the subgenus Alectorobius, which was originally proposed by Pocock (1907). Recent genetic analyses support the elevation of Alectorobius to a full genus (Mans et al. 2021). This tick exhibits a nidicolous life history, dwelling in the nest/burrow of its hosts. Larvae of Alectorobius are long-duration feeders of 3 to 10 days (Davis 1955, Clifford et al. 1964, Endris et al. 1991b, Robayo-Sánchez et al. 2022). This long feeding period for larvae may serve as an important mode of dispersal, whereas nymphs and adults are rapid feeders ∼20 to 60 min (Endris et al. 1991b). Under optimal conditions, nymphs and adults have been known to live up to 8 years without feeding (Davis 1955). Temperature and humidity affect oviposition, survivability, and duration of development of each life stage. Endris et al. (1991b) proposed that A. puertoricensis might not reproduce continuously throughout the year but would exhibit 1 or 2 generations that correspond with the rainy season, when humidity levels would be higher (>60%). While the list of potential hosts of A. puertoricensis has been expanded to include mammals, birds, reptiles, and amphibians, and records of its distribution suggest its presence throughout the Caribbean basin, no comprehensive assessments have been made of its potential distribution. A. puertoricensis is a vector of importance to both human and animal health, transmitting the relapsing fever spirochete Borrelia puertoricensis (Bermúdez et al. 2021, Vázquez-Guerrero et al. 2025) and parasitizing humans and poultry (Bermúdez et al. 2017, López et al. 2021). African swine fever (ASF) is a major concern that currently threatens the Caribbean region and whose spread to the United States mainland would devastate the US swine industry. Golnar et al. (2019) conducted a literature review identifying the potential vectors and hosts of African swine fever virus (ASFV) transmission in the United States and ranked A. puertoricensis as a high-risk vector. ASF is a viral disease caused by a double-stranded DNA virus (Asfarviridae; Asfivirus), which causes high morbidity and mortality (>90%) in both domestic (Sus domesticus) and feral swine (Sus scrofa). Animals infected with highly virulent strains may present as sudden deaths, with moderate strains having more generalized signs, including high fever, appetite loss, vomiting, diarrhea, lethargy, and weakness, accompanied by visible signs, including cutaneous red patches and hemorrhagic skin lesions, which are most frequently observed on the ears, distal limbs, and snout (Oura 2024). There are over 24 different genotypes of ASFV (Tulman et al. 2009). Transmission of ASFV to swine can be via (i) horizontal transmission from swine to swine through direct contact, (ii) indirect transmission via ingestion of contaminated food and water or exposure to contaminated fomites (e.g. boots, trucks beds, equipment), and (iii) tick-borne transmission via Ornithodoros and Alectorobius species. Hess et al. (1987) reported that A. puertoricensis that had fed on infected pigs were able to transmit the virus to susceptible pigs after 20 days. Tick transmission to susceptible pigs was documented up to 239 days after feeding on infected pigs; beyond this time frame, transmission remains unknown. Additionally, A. puertoricensis was shown to exhibit transovarial transmission of ASFV (Hess et al. 1987). A detailed study of vector potential was conducted by Endris et al. (1991a), which confirmed the findings of Hess et al. (1987). It was also noted that mortality among ASFV-infected nymphs of A. puertoricensis was 43.2% compared to 23.2% among uninfected nymphs (Endris et al. 1992). These transmission studies were conducted with an early strain of ASFV (Dominican Republic strain—DR-II). To date, transmission studies evaluating A. puertoricensis with current ASFV genotypes have not been conducted. Thus, any deleterious viral effects these genotypes have on A. puertoricensis remain unknown. The current ASFV strain in the Dominican Republic and Haiti belongs to Genotype II (Ramirez-Medina et al. 2022). The Caribbean Basin remained free of ASF until 1971, when ASFV was introduced into Cuba. Subsequent outbreaks in the Dominican Republic and Haiti throughout the 1970s prompted the creation of local commissions and international eradication programs (Alexander 1992, Simeón-Negrín and Frías-Lepoureau 2002). In 1980, the Dominican Republic declared successful eradication of ASF and Haiti later followed suit in 1984 (Alexander 1992). The aim of this paper is to (i) project the potential geographic distribution of A. puertoricensis throughout the Caribbean Basin, (ii) characterize the specific environmental conditions, biomes, and ecoregions associated with known occurrence records, and (iii) compile a list of known host species. This information is important for assessing potential threats of the introduction, spread, and establishment of ASFV and for providing a foundation for targeted and effective ASF response and control plans. Materials and Methods Occurrence and Hosts Records The majority of the occurrence records for A. puertoricensis (n = 60) were obtained from 23 literature sources (Table 1) (Fox 1947, Kohls et al. 1965, Jones et al. 1972, Capriles and Gaud 1977, Butler et al. 1985, Endris et al. 1989, Paternina et al. 2009, Bermúdez et al. 2010, 2013, 2017, 2021, López et al. 2021, Quintero et al. 2012, Basu and Charles 2017, Londoño et al. 2017, Guzmán-Cornejo et al. 2019, Faccini-Martínez et al. 2022, Robayo-Sánchez et al. 2022, Rodriguez et al. 2022, Ballados-González et al. 2023, Alvarez-Londoño et al. 2024, Romero et al. 2024, Vázquez-Guerrero et al. 2025). Three additional occurrences were obtained from the National Veterinary Services Laboratories, Ames, IA. Occurrences were georeferenced using the GEOLocate Web Application (Rios and Bart 2010) (http://www.geo-locate.org), a commonly used platform for georeferencing natural history collection locality data. Locality records were spatially thinned at a radius of 5 km to reduce autocorrelation (Boria et al. 2014), which resulted in the removal of 4 locality points. Two other points were removed: 1 had no corresponding environmental data, and 1 was a point from the Isla Clarion, an island in the Pacific where a collection was made from a burrowing owl (Athene cunicularia) burrow. We felt this second point was disjunct from the main population of A. puertoricensis. The resulting 57 occurrence points were used for all subsequent analyses. These occurrence records included the hosts from which A. puertoricensis were collected as well as those collections made from host nest/burrows. Collectively, these records permitted assessment of host utilization. When given, developmental stages were included in our dataset. Table 1.Classification of A. puertoricensis occurrence records based on collection site and descriptionCollection siteCollection description/entryNumber of collection recordsReferences Hosts Artibeus lituratus (Great fruit-eating bat)1 Jones et al. 1972 Canis familiaris (Dog)5National Veterinary Services Laboratories; Paternina et al. 2009, Faccini-Martínez et al., 2022, Romero et al. 2024 Conepatus semistriatus (Striped hog-nosed skunk)1 Jones et al. 1972 Dasyprocta fuliginosa (Black agouti)1 Jones et al. 1972 Didelphis virginiana (Virginia opossum)2 Ballados-González et al. 2023, Vázquez-Guerrero et al. 2025 Domestic Fowl(chickens, ducks, turkeys) 4 López et al. 2021, Faccini-Martínez et al. 2022 Iguana sp.1 Jones et al. 1972 Lontra annectens (Northern neotropical river otter)1 Alvarez-Londoño et al. 2024. Marmosa robinsoni (Robinson’s mouse opossum)2 Jones et al. 1972 Marmosa robinsoni (Robinson’s mouse opossum) and Sylvilagus floridanus (Eastern Cottontail)1 Jones et al. 1972 Monodelphis brevicudata (Northern red-sided opossum)2 Jones et al. 1972 Proechimys guyannensis trinitatus (Cayenne spiny rat)2 Kohls et al. 1965, Jones et al. 1972 Proechimys semispinosus (Tome’s spiny rat)6 Jones et al. 1972 Rattus rattus (Black rat) and Mus musculus (House mouse)1 Londoño et al. 2017, Faccini-Martínez et al. 2022 Rattus rattus (Black rat) and Rattus norvegicus (Brown rat)2 Fox 1947 Rattus sp. (Rat)3 Quintero et al. 2012, Faccini-Martínez et al. 2022 Rhinella marina (Cane toad)1 Bermúdez et al. 2013; Faccini-Martínez et al. 2022 Tamandua tetradactyla (Southern tamandua (anteater))1 Jones et al. 1972 Tamandua mexicana (Northern tamandua (anteater))1 Faccini-Martínez et al. 2022, Rodriguez et al. 2022 Urva auropunctatus (Small Indian mongoose)2National Veterinary Services Laboratories Zygodontomys brevicauda (Short-tailed cane mouse)1 Jones et al. 1972 Burrow/Nest “from straw and earth in a fowl run”1 Kohls et al. 1965 “mongoose nest at base of tree”1 Butler et al. 1985 “rat and burrowing owl nest at swine market”1 Butler et al. 1985 “rat and owl nests in dry gulley”1 Butler et al. 1985 “rat burrow”1 Butler et al. 1985 “rat nest—dry habitat”1 Butler et al. 1985 “rat/mongoose nest”2 Butler et al. 1985 “rat/owl nest”4 Butler et al. 1985 House/dwelling Homo sapiens (Human)3 Bermúdez et al. 2017, Faccini-Martínez et al. 2022 Unknown Unknown2 Bermúdez et al. 2010, Basu and Charles 2017, Faccini-Martínez et al. 2022Number of collection records and references also are noted. Classification of A. puertoricensis occurrence records based on collection site and description Domestic Fowl Domestic Fowl (chickens, ducks, turkeys) (chickens, ducks, turkeys) Number of collection records and references also are noted. Defining Accessible Area Species distribution models (SDMs) often include biotic and abiotic factors but rarely include factors related to dispersal/movement capabilities of the species. Thus, projections of such SDMs do not consider the crucial role of the “accessible area,” which can be defined as the geographical area that has been available to (reachable by) a species through dispersal over selected periods of time (Barve et al. 2011). Identification of the accessible area helps guide the selection of environmental variables and focuses the determination of correlations among model variables (see next section) on data from geographical areas the species can reach. Additionally, a large study area extent can lead to inflation of the area under the curve (AUC) score (Amaro et al. 2023). The accessible area for A. puertoricensis was defined using a kernel method that is often used in home range analyses to estimate the geographical area used by a species or its utilization distribution (Worton 1989). Estimation of the accessible area for A. puertoricensis was conducted in R v.4.5.2 (R Core Team 2025) using the “adehabitatHR” package (Calenge 2024). For this estimation, we evaluated a series of inclusion levels (95% to 99%) and selected the 1 that allowed the inclusion of all occurrence points (Fig. 1). A shapefile of the boundary of the estimated accessible area was created and readied for clipping environmental data. Locations of A. puertoricensis occurrence data used in SDMs (yellow dots). From these occurrences, the accessible area of A. puertoricensis (dashed orange line) was defined using a kernel utilization estimation method. Raster values (red shading) are percentage isopleths ranging from 0 to 100. Lower values show higher density (dark red), while higher values show lower density (pale red/peach). Environmental Data A total of 20 environmental variables (19 bioclimatic variables and elevation) were obtained from the WorldClim v.2.1 dataset (http://www.worldclim.org; Fick and Hijmans 2017) at a spatial resolution of 2.5 arc-minutes (∼4.5 × 4.5 km2). These files were clipped in ArcGIS Pro v.3.6.0 (Esri) to produce 2 datasets: (i) the study area focused on the Caribbean Basin clipped to a set extent (top: 29.06, bottom: −1.1, left: −100.26, right: −55.57 [decimal degrees]), which would serve as the projection base environmental input files for Maxent v.3.4.4; and (ii) the accessible area, which was produced from the kernel analysis and used for variable selection analysis. To determine correlations among the accessible area variables, a correlation matrix was calculated in R using the “raster” package (Hijmans 2025). As recommended by Booth (2022), environmental variables were checked for discontinuities prior to elimination. These discontinuities are artifacts within the raster files that show a sudden and sharp change in the data values that do not reflect a real-world environmental gradient (Booth 2022). BIO18 and BIO19 showed discontinuities and thus were removed. Elimination of the remaining correlated variables was based on Pearson’s correlation values ≥0.80 or ≤ −0.80 (Supplementary Fig. S1). Selection for elimination between each pair of correlated variables was based upon their biological relevance to A. puertoricensis and followed the approach of Sage et al. (2017), who selected minimum and maximum monthly variables over those that were quarterly and annual. This approach encompassed the extreme conditions that could affect the survivability and distribution of A. puertoricensis. Out of the 20 variables, only 7 were kept (BIO2, 4, 5, 6, 13, 14, and 15), and summary statistics of these uncorrelated environmental variables were calculated (Supplementary Table S1). To further understand the environments from which A. puertoricensis were collected, raster cell values of the environmental data were extracted for each occurrence point and summarized across all 20 environmental variables. To assess the biomes and ecoregions from which A. puertoricensis were collected, a shapefile of ecoregions was downloaded from the RESOLVE Ecoregions database (www.ecoregions.appspot.com; Dinerstein et al. 2017), converted to a raster, and cell values for each occurrence point were extracted. Species Distribution Models SDMs were developed using Maxent (Phillips et al. 2006), which has been used to develop SDMs for various species of Ornithodoros (Donaldson et al. 2016, Sage et al. 2017, Moradi-Asl and Jafari 2020, Wu et al. 2022, Botero-Cañola et al. 2024). Six models were run, each with a different value for the regularization multiplier (0.1, 0.5, 1, 2, 5, and 10). Adjustment of the regularization multiplier prevents overfitting and controls model complexity (Cao et al. 2013, Radosavljevic and Anderson 2014). Additional model settings included: setting algorithm parameters to auto, a maximum of 10,000 background points, model output formatted as Cloglog (complementary log–log), a subsampling of 10 model runs with a random test percentage of 20%, and a maximum iteration of 5,000. To evaluate the accuracy of each model, the AUC score and the uncertainty/variability value (SD), as well as the mean True Skill Statistic were calculated. The percent contribution and permutation importance of each variable, jackknife results, and response curves were also evaluated. Resulting SDMs were processed and visualized in ArcGIS Pro, with the geographic area restricted to the Caribbean Basin. Results Occurrences and Host Utilization A total of 57 occurrences of A. puertoricensis were spread across the Caribbean Basin within the following administrative areas: Colombia (n = 13 occurrences), the Dominican Republic (n = 7), El Salvador (n = 1), Guadeloupe (n = 1), Haiti (n = 4), Jamaica (n = 1), Mexico (n = 2), Panama (n = 5), Puerto Rico (n = 4), Trinidad (n = 1), the US Virgin Islands (n = 1), and Venezuela (n = 17) (Fig. 1). These occurrences were classified into 4 categories based on their site of collection (Host, Burrow/Nest, House/Dwelling, and Unknown) (Table 1). Most collections represented the removal of A. puertoricensis attached to hosts (n = 40; 70.18%), followed by the removal from host burrow/nest (n = 12; 21.05%), human house/dwelling (n = 3; 5.26%), and unknown (n = 2; 3.51%). A total of 23 host species for A. puertoricensis were documented, representing 4 animal classes (Amphibia, Reptilia, Aves, and Mammalia) (Table 1). The most common hosts were rodents (6 species) and opossums (3 species) (Table 1). Other hosts included 1 amphibian (cane toad, Rhinella marina) and 1 reptile (iguana, Iguana sp.). Domesticated animals such as dogs (Canis familiaris) and fowl (chickens, ducks, and turkeys) also were hosts. Host records did not specify which species of fowl, and these records were not included in our host species totals. Likewise, records for mongooses did not specify species, but we assume they refer to the small Indian mongoose (Urva auropunctatus; formerly Herpestes auropunctatus), which was originally introduced to the Caribbean Basin to control rat populations in sugarcane fields (Horst et al. 2001). Human host records were classified in the house/dwelling category, since individuals recording bites were frequently collected from these areas. One bat species (Greater fruit-eating bat, Artibeus lituratus) was reported as a host by Jones et al. (1972), but it was noted that this host record should be treated as questionable. Other host records included black agouti (Dasyprocta fuliginosa), northern neotropical river otter (Lontra annectens), northern and southern tamandua (anteaters, Tamandua mexicana, and Tamandua tetradactyla), and eastern cottontail (Sylvilagus floridanus). Rodents represented all but 2 host collections in the Burrow/Nest category. The 2 non-rodent collections included 1 from a mongoose nest and another from “straw and earth in a fowl run.” Many records classified in the Burrow/Nest category used the general term “Rat Nest,” without specifying the species, but likely referring to the black rat (Rattus rattus) or the brown rat (R. norvegicus). A few of the rodent Burrow/Nest records came from burrows co-occupied by burrowing owls (A. cunicularia) or small Indian mongooses (U. auropunctatus). No A. puertoricensis collections were made from burrowing owls, but given their use of burrows, they should be considered a suspected host. Butler et al. (1985) summarized collections of A. puertoricensis from nests used by burrowing owls and rats in Haiti and by rats and/or mongooses in the Dominican Republic, all of which were in the vicinity of swine at the time of collection or within the previous few months. Summary of Habitat Conditions at Occurrence Points A summary of the environmental data correlated with A. puertoricensis occurrences (Table 2) revealed that this species is adapted to warm tropical and subtropical climates and that its distribution is limited by cold temperatures. Collections of A. puertoricensis occurred in regions where the annual mean temperature (BIO1) ranged between 20.21 and 28.17 °C (x̄ = 26.32 °C) and where the minimum temperature of the coldest month (BIO6) did not fall below 13.27 °C (x̄ = 20.1 °C). High temperatures may be tolerated, with maximum monthly temperatures of the warmest month (BIO5) reaching 35.72 °C (x̄ = 32.38 °C). Occurrence points occurred in regions encompassing a wide range of precipitation, with annual precipitation (BIO12) as low as 82 mm and as high as 2772 mm (x̄ = 1433.28 mm). Occurrence points occurred at elevations from 4 to 1,228 m above sea level, although occurrences were more common in lowland areas (x̄ = 164.26 m). Table 2.Summary of environmental data correlated with A. puertoricensis occurrence recordsEnvironmental variableValue rangeMean (±SD)Median BIO1—Annual Mean Temperature 20.21 to 28.17 °C26.32 (1.51) °C26.58 °C BIO2—Annual Mean Diurnal Range 5.7 to 12.73 °C9.46 (1.75) °C9.49 °C BIO3—Isothermality 66.85% to 92.56%77.78 (6.05)%76.53% BIO4—Temperature Seasonality 21.49% to 225.16%85.83 (48.08)%69.91% BIO5—Max Temperature of Warmest Month 25.99 to 35.72 °C32.38 (1.98) °C32.74 °C BIO6—Min Temperature of Coldest Month 13.27 to 23.11 °C20.10 (2.28) °C20.71 °C BIO7—Annual Temperature Range 7.34 to 19.03 °C12.28 (2.71) °C11.71 °C BIO8—Mean Temperature of Wettest Quarter 20.38 to 28.76 °C26.46 (1.5) °C26.75 °C BIO9—Mean Temperature of Driest Quarter 19.57 to 28.32 °C25.89 (1.88) °C26.53 °C BIO10—Mean Temperature of Warmest Quarter 20.67 to 29.32 °C27.28 (1.56) °C27.68 °C BIO11—Mean Temperature of Coldest Quarter 19.43 to 27.62 °C25.22 (1.73) °C25.39 °C BIO12—Annual Precipitation 82 to 2,772 mm1,433.28 (582.71) mm1,305 mm BIO13—Precipitation of Wettest Month 82 to 452 mm214.81 (92.74) mm189.5 mm BIO14—Precipitation Seasonality (CV) 3 to 79 mm30.74 (19.81) mm22.5 mm BIO15—Precipitation Seasonality 28.13% to 97.87 %54.14 (14.66) %54.48 % BIO16—Precipitation of Wettest Quarter 221 to 1,300 mm568.64 (241.83) mm492.5 mm BIO17—Precipitation of Driest Quarter 16 to 298 mm117.26 (68.52) mm97 mm BIO18—Precipitation of Warmest Quarter 104 to 675 mm329.31 (135.25) mm317.5 mm BIO19—Precipitation of Coldest Quarter 42 to 1,300 mm327.9 (277.25) mm210.5 mm ELEV—Elevation 4 to 1228 m164.26 m (231.28) m80 mVariable abbreviations and definitions as described in WorldClim v.2.1 dataset (http://www.worldclim.org; Fick and Hijmans 2017). Summary of environmental data correlated with A. puertoricensis occurrence records Variable abbreviations and definitions as described in WorldClim v.2.1 dataset (http://www.worldclim.org; Fick and Hijmans 2017). Occurrences of A. puertoricensis were recorded across 6 biomes (Table 3), with most collections being made from forested biomes (n = 43; 75.44%). These forested biomes consisted of 3 distinct categories: (i) Tropical and Subtropical Moist Broadleaf Forests, which comprised most of the collections (n = 31; 54.39%), (ii) Tropical and Subtropical Dry Broadleaf Forests (n = 11; 19.3%), and (iii) Tropical and Subtropical Coniferous Forests (n = 1; 1.75%). Drier biomes in which collections were made included Tropical and Subtropical Grasslands, Savannas, and Forest (n = 2; 3.51%) and Deserts and Xeric Shrublands (n = 8; 14.04%). A few collections were also made along coastal areas occurring in the Mangroves biome (n = 4; 7.02%). Table 3.Classification of A. puertoricensis occurrences by biome and ecoregion based on the RESOLVE Ecoregions shapefile (www.ecoregions.appspot.com; Dinerstein et al. 2017) (1) Biome: desert and Xeric Shrublands: 8 occurrences (14.04%) Ecoregions: Caribbean Shrublands = 1 (1.75%) Guajira-Barranquilla Xeric Scrub = 1 (1.75%) La Costa Xeric Shrublands = 1 (1.75%) Paraguaná Xeric Scrub = 5 (8.77%) (2) Biome: tropical and subtropical dry broadleaf forests: 11 occurrences (19.3%) Ecoregions: Hispaniolan Dry Forests = 6 (10.53%) Lara-Falcón Dry Forests = 2 (3.51%) Maracaibo Dry Forests = 2 (3.51%) Veracruz Dry Forests = 1 (1.75%) (3) Tropical and subtropical moist broadleaf forests: 31 occurrences (54.39%) Ecoregions: Cordillera La Costa Montane Forests = 3 (5.26%) Hispaniolan Moist Forests = 2 (3.51%) Isthmian-Atlantic Moist Forests = 4 (7.02%) Jamaican Moist Forests = 1 (1.75%) Leeward Island Moist Forests = 1 (1.75%) Magdalena-Urabá Moist Forests = 11 (19.3%) Northwest Andean Montane Forests = 1 (1.75%) Puerto Rican Moist Forests = 4 (7.02%) Rio Negro Campinarana = 1 (1.75%) Trinidad and Tobago Moist Forests = 1 (1.75%) Venezuelan Andes Montane Forests = 1 (1.75%) Yucatán Moist Forests = 1 (1.75%) (4) Tropical and subtropical grasslands, savannas, and shrublands: 2 occurrences (3.51%) Ecoregions: Llanos = 2 (3.51%) (5) Tropical and subtropical coniferous forests: 1 occurrence (1.75%) Ecoregions: Central American Pine-Oak Forests = 1 (1.75%) (6) Mangroves: 4 occurrences (7.02%) Ecoregions: Bahamian-Antillean Mangroves = 3 (5.26%) Southern Mesoamerican Pacific Mangroves = 1 (1.75%) Classification of A. puertoricensis occurrences by biome and ecoregion based on the RESOLVE Ecoregions shapefile (www.ecoregions.appspot.com; Dinerstein et al. 2017) (1) Biome: desert and Xeric Shrublands: 8 occurrences (14.04%) (1) Biome: desert and Xeric Shrublands: 8 occurrences (14.04%) Ecoregions: Ecoregions: Caribbean Shrublands = 1 (1.75%) Caribbean Shrublands = 1 (1.75%) Guajira-Barranquilla Xeric Scrub = 1 (1.75%) Guajira-Barranquilla Xeric Scrub = 1 (1.75%) La Costa Xeric Shrublands = 1 (1.75%) La Costa Xeric Shrublands = 1 (1.75%) Paraguaná Xeric Scrub = 5 (8.77%) Paraguaná Xeric Scrub = 5 (8.77%) (2) Biome: tropical and subtropical dry broadleaf forests: 11 occurrences (19.3%) (2) Biome: tropical and subtropical dry broadleaf forests: 11 occurrences (19.3%) Ecoregions: Ecoregions: Hispaniolan Dry Forests = 6 (10.53%) Hispaniolan Dry Forests = 6 (10.53%) Lara-Falcón Dry Forests = 2 (3.51%) Lara-Falcón Dry Forests = 2 (3.51%) Maracaibo Dry Forests = 2 (3.51%) Maracaibo Dry Forests = 2 (3.51%) Veracruz Dry Forests = 1 (1.75%) Veracruz Dry Forests = 1 (1.75%) (3) Tropical and subtropical moist broadleaf forests: 31 occurrences (54.39%) (3) Tropical and subtropical moist broadleaf forests: 31 occurrences (54.39%) Ecoregions: Ecoregions: Cordillera La Costa Montane Forests = 3 (5.26%) Cordillera La Costa Montane Forests = 3 (5.26%) Hispaniolan Moist Forests = 2 (3.51%) Hispaniolan Moist Forests = 2 (3.51%) Isthmian-Atlantic Moist Forests = 4 (7.02%) Isthmian-Atlantic Moist Forests = 4 (7.02%) Jamaican Moist Forests = 1 (1.75%) Jamaican Moist Forests = 1 (1.75%) Leeward Island Moist Forests = 1 (1.75%) Leeward Island Moist Forests = 1 (1.75%) Magdalena-Urabá Moist Forests = 11 (19.3%) Magdalena-Urabá Moist Forests = 11 (19.3%) Northwest Andean Montane Forests = 1 (1.75%) Northwest Andean Montane Forests = 1 (1.75%) Puerto Rican Moist Forests = 4 (7.02%) Puerto Rican Moist Forests = 4 (7.02%) Rio Negro Campinarana = 1 (1.75%) Rio Negro Campinarana = 1 (1.75%) Trinidad and Tobago Moist Forests = 1 (1.75%) Trinidad and Tobago Moist Forests = 1 (1.75%) Venezuelan Andes Montane Forests = 1 (1.75%) Venezuelan Andes Montane Forests = 1 (1.75%) Yucatán Moist Forests = 1 (1.75%) Yucatán Moist Forests = 1 (1.75%) (4) Tropical and subtropical grasslands, savannas, and shrublands: 2 occurrences (3.51%) (4) Tropical and subtropical grasslands, savannas, and shrublands: 2 occurrences (3.51%) Ecoregions: Ecoregions: Llanos = 2 (3.51%) Llanos = 2 (3.51%) (5) Tropical and subtropical coniferous forests: 1 occurrence (1.75%) (5) Tropical and subtropical coniferous forests: 1 occurrence (1.75%) Ecoregions: Ecoregions: Central American Pine-Oak Forests = 1 (1.75%) Central American Pine-Oak Forests = 1 (1.75%) (6) Mangroves: 4 occurrences (7.02%) (6) Mangroves: 4 occurrences (7.02%) Ecoregions: Ecoregions: Bahamian-Antillean Mangroves = 3 (5.26%) Bahamian-Antillean Mangroves = 3 (5.26%) Southern Mesoamerican Pacific Mangroves = 1 (1.75%) Southern Mesoamerican Pacific Mangroves = 1 (1.75%) At a finer scale, occurrences of A. puertoricensis were recorded across 24 ecoregions (Table 3), with most collections from the Magalena-Urbá Moist Forest (n = 11; 19.3%), followed by the Hispaniolan Dry Forest (n = 6; 10.53%) and the Paraguaná Xeric Scrub (n = 5; 8.77%). None of the remaining ecoregions had more than 1 or 2 collections recorded. Maxent Analyses Performance metrics for the 6 SDMs with different adjustments to the regularization multiplier value are summarized in Table 4. Based on these performance metrics, we selected the model with the regularization value of 1. This model had an AUC score of 0.872, which indicates considerable discriminative ability, a relatively low model uncertainty (SD = 0.055), and a TSS score of 0.706, which indicates fair model performance in distinguishing between presence and pseudoabsences. Table 4.Summary of model performance metrics for the 6 SDMs for A. puertoricensis with different adjustments to the regularization multiplier valueRegularization value0.10.5 1 a 2510 Mean AUC 0.8570.840.8720.8740.8140.852 Mean SD 0.0540.0510.0550.0690.0430.038 Mean TSS 0.6410.6590.7060.7240.6190.667 PC PI PC PI PC PI PC PI PC PI PC PI BIO2 7.26.94.55.83.25.10.30.70000 BIO4 12.14.310.9127.798.48.710.54.87.211.1 BIO5 3.96.511.70.72.10.40.50.10.100 BIO6 24.13228.225.628.919.531.225.232.922.827.121.5 BIO13 35.22.739.76.642.6741.612.136.521.544.234.5 BIO14 10.633.49.134.710.945.111.635.511.730.217.622.9 BIO15 6.814.26.613.66.112.16.517.48.420.63.910aIndicates the best model.For abbreviations and definitions of environmental variables (BIO2 to BIO15), see Table 2.AUC, area under the curve; TSS, true skill statistic; PC, variable percent contribution; PI, variable permutation importance. Summary of model performance metrics for the 6 SDMs for A. puertoricensis with different adjustments to the regularization multiplier value Indicates the best model. For abbreviations and definitions of environmental variables (BIO2 to BIO15), see Table 2. AUC, area under the curve; TSS, true skill statistic; PC, variable percent contribution; PI, variable permutation importance. The predicted distribution of suitable habitat (%) for A. puertoricensis throughout the Caribbean Basin is shown in Fig. 2. The southernmost Florida Keys were predicted to be of moderate suitability (∼62%). Moving southeast, the islands from the Bahamas to the Turks and Caicos Islands showed a gradual shift from low to high suitability, with suitability increasing toward Cuba and Hispaniola (Haiti and the Dominican Republic). Jamaica’s northern, northwestern, and southern coasts also contained areas of high suitability, whereas the interior areas were of low suitability. Maxent model prediction (regularization value = 1) of suitable habitat for A.puertoricensis across the Caribbean Basin. Hotter colors (red, orange, and yellow) indicate higher habitat suitability while cooler colors (blues) indicate lower suitability. In Cuba, the model identified moderate suitability along the eastern region, whereas the northern and central regions had low suitability. On the island of Hispaniola, areas of low elevation, the coast, and semi-arid regions showed high suitability. Specifically, in the Dominican Republic, low suitability occurred in the interior, whereas the northwest, the east, and regions along the Haitian border were highly suitable. In Puerto Rico, the island where A. puertoricensis was originally described, areas of high suitability occurred along the northern and southern coasts. The central part of the island is mountainous with high rainfall and cooler conditions compared to the coast, which is relatively flat, hotter, and drier. The Caribbean islands east of Puerto Rico from Vieques to Guadeloupe showed areas of very high suitability. This includes the British and US Virgin Islands, both with predicted suitability greater than 95%. Farther south, suitability for the Windward Islands decreased from Guadeloupe, but gradually increased with increasing proximity to the South American mainland. East of these chain islands lies Barbados, with suitable habitat for A. puertoricensis ranging from ∼40% to 80%. On the North, Central, and South America mainland, the model identified areas of high suitability occurring in the Yucatan Peninsula of Mexico, Honduras, Panama, Colombia, and Venezuela. Although 1 occurrence record from El Salvador was included in the model, predictions for suitability were low throughout the country (<10%). The island of Trinidad showed higher suitability along the coastal regions. The 3 most important predictor variables in the model included: Precipitation of the Wettest Month (BIO13, contributing 42.6%), Minimum Temperature of the Coldest Month (BIO6, contributing 28.9%), and Precipitation of the Driest Month (BIO14, contributing 10.9%). These 3 variables accounted for 82.4% of the model’s explanatory power. The remaining 17.6% was provided by: Temperature Seasonality (BIO4; 7.7%), Precipitation Seasonality (BIO15; 6.1%), Mean Diurnal Range (BIO2; 3.2%), and Maximum Temperature of the Warmest Month (BIO5; 0.6%). Average jackknife results (Supplementary Fig. S2) showed that Precipitation of the Wettest Month (BIO13) had the highest gain, indicating that this variable contributed the most useful information by itself. Omission of Precipitation of the Driest Month (BIO14) decreased the model gain, indicating this variable contributed information that is not present in all the other environmental variables included in the model. The response curves (Fig. 3) of the 7 bioclimatic variables in the model indicated that occurrence of A. puertoricensis was associated with the Precipitation of the Wettest Month (BIO13, less than approximately 225 mm), Minimum Temperature of the Coldest Month (BIO6, greater than approximately 20 °C), Precipitation of the Driest Month (BIO14, less than approximately 40 mm), Temperature Seasonality (BIO4, less than approximately 120%), Precipitation Seasonality (BIO15, less than approximately 55%), and Mean Diurnal Range (BIO2, greater than approximately 9 °C). Response curves of the 7 bioclimatic variables for the best Maxent model (regularization value = 1) for A. puertoricensis. These response curves show how changes in the predicted probability of occurrence changes with respect to each environmental variable while keeping all other environmental variables at their average sample value. Red lines represent the mean (n = 10), and blue lines represent the standard deviation. For abbreviations and definitions of variables, see Table 2. Discussion This study provides an integrated assessment of the potential geographic distribution, environmental suitability, and host utilization of A. puertoricensis across the Caribbean Basin. By combining historical occurrence data, host records, and species distribution modeling, the results identify regions where ecological conditions may support persistent populations of this nidicolous soft tick. These findings are relevant given the re-emergence of ASFV in the Caribbean and the demonstrated vector competence of A. puertoricensis under experimental conditions with 1 strain of ASFV (Hess et al. 1987, Endris et al. 1991a, 1992). The SDM predicted various areas of suitability for A. puertoricensis through lowland tropical areas and subtropical regions characterized by warm temperatures and variable precipitation. The influence of temperature and precipitation-related variables (Minimum Temperature of the Coldest Month [BIO6], Maximum Precipitation of the Wettest Month [BIO13], and Minimum Precipitation of the Driest Month [BIO14]), is consistent with the known physiological limitations of argasid ticks under cold and dry conditions (Endris et al. 1991b). However, the nidicolous ecology of A. puertoricensis suggests that local microhabitats may partially buffer populations from broader environmental variability (Davis 1955, Endris et al. 1991b, Mans et al. 2019). As demonstrated for nidicolous Ornithodoros species, ecological niche models may identify areas of suitable habitat beyond those in which occurrence has been recorded, supporting the interpretation that limited sampling can contribute to apparent patchiness in documented distributions (Sage et al. 2017). Even though BIO6, BIO13, and BIO14 showed clear trends of A. puertoricensis preferred environmental conditions, Mean Diurnal Range (BIO2) and Temperature Seasonality (BIO4) showed a bimodal pattern (Fig. 3). The response curve of BIO2 indicated that A. puertoricensis prefers environments where the average difference between the monthly maximum and minimum temperatures over the year is less than 7 °C or more than 9 °C. This is consistent with A. puertoricensis being found in humid tropical regions, which have a lower and narrower mean diurnal range versus that of arid regions, where the mean diurnal range is higher and broader. Utilization of host burrows and cavities allows A. puertoricensis to survive these varied conditions. The response curve of BIO4 indicated that the probability of occurrence is higher when temperature variation over a year is low, and as temperature variability increases, the probability of A. puertoricensis occurrences decreases, even though there is a peak between 50% and 200%. This indicates that A. puertoricensis prefers a more stable, year-round temperature threshold. A common issue modelers face when building SDMs is data availability for both locality information and environmental data layers while ensuring an overlapping time frame, where both datasets have the same temporal scales. In this study, locality information for A. puertoricensis spanned from 1946 to 2023, whereas the WorldClim v.2.1 data covered 1970 to 2000. Temporal mismatches could create biases in the model and influence model interpretations, especially as climatic conditions change. Matching temporal scales leads to a reduction in the number of locality points and model performance (Bracken et al. 2022). If temporal matching for the A. puertoricensis SDM were to occur in this study, only 17 localities would be available and would not cover the whole Caribbean Basin. Additionally, a study by Bracken et al. (2022) found that model performance is less affected when using all available data rather than matching timespan and relying on persistent datasets (elevation, seasonal precipitation) rather than impersistent datasets (land-cover, land-use, vegetation potential, vegetation type). Since our model used only bioclimatic variables, which are relatively stable across decades, the impact of the temporal mismatch is likely minimized. In addition, we ensured that both minimum and maximum environmental datasets of temperature and precipitation were included in the model to understand the extremes that influence habitat suitability and which could shed light on the extremes that may be experienced under increased climatic changes. The results of this study highlight the geographical overlaps between areas that were predicted to be suitable for A. puertoricensis and where outbreaks of ASF have occurred across Cuba, Haiti, and the Dominican Republic. In Cuba, ASF cases were documented during the 1980s outbreak in the provinces of Holguín, Guantánamo, Granma, and Santiago de Cuba (Simeón-Negrín and Frías-Lepoureau 2002). Predicted mean habitat suitability of these provinces for A. puertoricensis ranged from 21.3% to 51.88% (Fig. 4; Supplementary Table S2). In Haiti, ASF outbreaks have occurred in 7 of the 10 administrative departments (Nord-Ouest, Artibonite, Nord, Nord-Est, Ouest, Grande-Anse, and Sud) (Ter Beek 2022; World Animal Health Information System, World Organisation for Animal Health, www.wahis.woah.org). Most of the departments had a mean habitat suitability for A. puertoricensis ranging from 21.3% to 70.7% but the Grande-Anse and Sud departments, along the Tiburon Peninsula, had lower suitability (9.82% to 21.29%) (Fig. 4; see Supplementary Table S2 for individual province habitat suitability values). A total of 22 provinces in the Dominican Republic were noted by Schambow et al. (2025) as having reports of ASF in 2022. Habitat suitability of these provinces ranged from 21.30% to 98.94% (see Supplementary Table S2 for individual province habitat suitability values). Experimental studies confirm the capacity of A. puertoricensis to acquire, maintain, and transmit ASFV. These evaluations were conducted using historical virus strains, and the vector competence of this species for the contemporary genotype II ASFV remains uncharacterized. Though the epidemiological role of A. puertoricensis was not identified in these past outbreaks, there remains a persistent threat. A spillover of ASF from infected swine to A. puertoricensis could make future outbreaks significantly more difficult to eradicate. Map of predicted mean habitat suitability for A. puertoricensis for each of the provinces of Cuba and the Dominican Republic and the departments of Haiti. See Supplementary Table S2 for specific suitability values. Host utilization records indicate that A. puertoricensis parasitizes a wide range of vertebrate hosts, with rodents representing the most frequently documented group. Tome’s spiny rat (Proechimys semispinosus) accounted for the highest number of collection records (Fox 1947, Jones et al. 1972, Endris et al. 1989). Commensal rodents may serve as potential maintenance hosts because they occupy burrows and peridomestic habitats that favor long-term tick persistence by providing repeated feeding opportunities (Butler et al. 1985). Several occurrence descriptions further place A. puertoricensis in environments directly relevant to swine exposure, including swine markets and animal holding areas, where livestock presence attracts rodents that establish burrows nearby in response to food availability (Butler et al. 1985). These settings create opportunities for ticks to move among wildlife hosts and domestic animals. In many Caribbean islands, free-roaming pigs are increasingly observed in rural, peri urban, and currently urban areas, including Vietnamese pot-bellied pigs (Sus scrofa domesticus), which were initially acquired through the unregulated pet trade and later released or abandoned (Coto 2020). Once free roaming, these animals may persist in peridomestic environments where they overlap spatially with commensal rodent populations and domestic animals, creating conditions that increase opportunities for contact among domestic swine, humans, wildlife hosts, and tick vectors (Coto 2020). Such peridomestic interfaces increase opportunities for contact among ticks, wildlife, and livestock, potentially facilitating the local circulation of A. puertoricensis and its associated pathogens (Butler et al. 1985). Recognizing the risk posed by highly dense populations of peri-urban swine in Puerto Rico, the US Department of Agriculture recently has implemented aggressive population reduction strategies across the island (K. O’Hara, personal communication May 2026). This effort may serve as a model for ASF risk mitigation in other countries in the region where swine-tick interfaces exist. The potential concerns of other soft tick species from the Caribbean Basin becoming involved in the spread and maintenance of ASF through transmission of ASFV are not well understood. One species of concern is Alectorobius talaje (formerly O. talaje), which has a wide distribution that extends from Texas down into Central America (Kohls et al. 1965). This distribution is shared with other species, which could be potential vectors: Ornithodoros turicata, Ornithodoros coriaceus, and A. puertoricensis. Furthermore, hosts of A. talaje include domestic swine along with rodents and humans (Valadez 1923, Cooley and Kohls 1944, Guglielmone et al. 2006). Currently, no transmission studies for A. talaje have been conducted. The following argasid species, also present in the Caribbean, would likely be of low concern with respect to ASFV due to host range or habitat. Four species of Argas have been noted within the region. Argas miniatus, Argas persicus, and Argas radiatus have been widely collected throughout the region (Hooker 1909, Vigueras 1934, Osorno-Mesa 1940, Cooley and Kohls 1944, Fairchild et al. 1966, Kohls et al. 1970, Capriles and Gaud 1977, Keirans 1984, Maes et al. 1989, Muñoz-Leal et al. 2018, Alvarez et al. 2020). Argas dulus has only been documented in the Dominican Republic from the nest of the passerine bird Dulus dominicus (palm chat) (Keirans et al. 1971). A transmission study found that A. persicus failed to transmit ASFV (Luo et al. 2026), and no trials have been conducted with A. miniatus, A. radiatus, nor A. dulus. As all 4 Argas species are strongly associated with birds, they are unlikely to play a significant role in the spread and maintenance of ASFV. Other bird-associated argasid ticks from this region include Alectorobius capensis recorded from Trinidad (Aitken et al. 1968; Jonkers et al. 1973) and Alectorobius denmarki from Cuba (Danielová et al. 1982). Currently, no ASFV transmission studies have been conducted on these species. Antricola spp. also occur and are associated with bats. Unlike other tick species where all 3 stages blood-feed, only the larval stage of Antricola is hematophagous; the nymphs and adults are free-living and feed on bat guano (De la Cruz 1973, Ribeiro et al. 2012). A total of 11 species of Antricola are found in Cuba (Antricola silvai, Antricola granasi, Antricola habanensis, Antricola cernyi, Antricola occidentalis, Antricola martelorum, Antricola naomiae, Antricola armasi, Antricola centralis, Antricola siboneyi, Antricola marginatus) (Alvarez et al. 2020). Outside of genus Antricola, Alectorobius and Subparmatus species also parasitize bats (Alectorobius azteci, Alectorobius brodyi, Alectorobius dusbabeki, Alectorobius kelleyi, Alectorobius natalinus, Alectorobius tadaridae, Subparmatus viguerasi). Lastly Otobius megnini is found throughout the United States, Mexico, Central and South America, and the Caribbean (Van Volkenberg 1929, Cooley and Kohls 1944, Thompson 1950, Guzmán-Cornejo et al. 2019) with hosts including livestock, but they are not known vectors of any known human or animal pathogens. Interpreting the SDM Outputs in the Animal Movement Context Caribbean basin areas predicted by our SDM to be suitable for A. puertoricensis habitat overlap with areas of concentrated human movement, including ports, coastal communities, and population centers. Network analyses of live pig and pork product trade among Caribbean countries (Arcega Castillo et al. 2026) also reveal spatial connectivity with SDM-predicted suitable habitat for A. puertoricensis. The probability of ASFV exposure events may be higher than in equally suitable habitats with limited host connectivity. If ASFV were introduced into areas supporting established A. puertoricensis populations, experimental evidence indicates the tick can maintain infection for extended periods and transmit ASFV to susceptible pigs. This experimental evidence is consistent with evidence of transovarial ASFV maintenance under laboratory conditions (Hess et al. 1987, Endris et al. 1991a, 1992). However, these studies were conducted with an earlier ASFV strain, and comparable evaluations using the contemporary Genotype II strains circulating in Hispaniola have not been conducted, which represents a key knowledge gap (Tulman et al. 2009, Ramirez-Medina et al. 2022). The SDM of A. puertoricensis has projected areas of habitat suitability in the Caribbean basin and has suggested ecological factors that support these projections based upon prior tick collections. The SDM of this ASFV vector may serve future pathway analyses for risk assessment, preparedness, and response and mitigation planning in the Caribbean basin and surrounding countries. Concerns for expansion of ASF beyond Hispaniola in the Caribbean basin and beyond the distribution of A. puertoricensis could involve a number of different pathways. Generally, the illegal movement of infected swine products is considered the highest risk route for introduction to the United States. ASFV introductions are likely less dependent on conventional commercial live swine exchanges and more dependent on infected products or fomites carried through human movement or migration pathways, including passenger baggage, mail shipments, and informal maritime traffic. Regional spread through the movement of infected ASFV vectors is considered less likely due to the need not just to introduce an infected vector but for that vector to also find and transmit the virus to a susceptible host. Rather, ASFV vectors play a role in maintaining local transmission, thereby extending the duration of outbreaks and resulting in ongoing pressure and sustained risk for spread in the region. Customs Border Protection and enforcement agencies describe efforts to prevent the entry of prohibited animal products via maritime routes and other traveler associated pathways, emphasizing enforcement at ports and points of entry as a core prevention strategy (Mason 2024). Regionally, the Caribbean Agricultural Health and Food Safety Agency (CARICOM) and associated animal health networks have advised the intensification of protective measures following ASF detection in the region, including the strengthening of border surveillance and the coordination of preventive actions (CARICOM 2021). Supplementary Material