Impacts of abiotic and biotic factors on terrestrial leeches in Indonesian Borneo

Haemadipsid leeches are ubiquitous inhabitants of tropical and sub-tropical forests in the Indo-Pacific region. They increasingly used as indicator taxa for biomonitoring, yet very little is known about their basic ecology. For example, to date no study has assessed occurrence distribution haemadipsid across naturally occurring gradients within intact habitats. We analyzed a long-term data set (2012–2020) on closely related tiger (Haemadipsa picta) brown spp.) leech species investigate if how abiotic biotic factors influence gradient forest types at an undisturbed rainforest site Indonesian Borneo. compared series negative binomial mixed models found that, factors, soil moisture had largest positive effect encounter rates both species. Among type differential effects counts two species: while were greater low elevation types, high types. Additionally, we that presence one other Finally, our results show narrower distribution, being restricted lower with higher water retention, suggesting could be more sensitive levels. Abstract available online material. Pacet adalah penghuni hutan tropis dan sub-tropis, yang hidup di kawasan Indo-Pasifik. Mereka seringkali digunakan sebagai taksa indikator untuk namun sangat sedikit diketahui tentang ekologi dasar mereka. Sebagai contoh, sampai saat ini tidak ada penelitian menilai keberadaan distribusi pacet seluruh gradien alami dalam habitat masih utuh. Kami telah menganalisis dari kumpulan jangka panjang spesies harimau coklat memiliki kekerabatan dekat, menyelidiki apakah bagaimana faktor abiotik biotik dapat memengaruhi kemunculannya tipe pada suatu lokasi hujan terganggu Kalimantan, Indonesia. membandingkan serangkaian model campuran negatif menemukan bahwa, abiotik, kelembaban tanah efek positif terbesar tingkat pertemuan kedua tersebut. Di antara faktor-faktor biotik, pengaruh berbeda terhadap jumlah spesies: ditemukan lebih besar dengan elevasi rendah, sedangkan tinggi. Selain itu, kami bahwa satu lainnya. Pada akhirnya, hasil menunjukkan sempit, terbatas rendah retensi air Hasil lain juga bisa sensitif rendah. Studying wildlife dense often difficult, researchers benefit from having broad toolkit them monitoring these populations. Along well-established tools such camera traps fecal samples, use emerged viable technique gain insight into distributions vertebrates (Drinkwater et al., 2019). Jawed land leeches, annelid family Haemadipsidae particular, gaining traction biodiversity assessments. The encompasses over 80% species, including species-rich genus Haemadipsa (Fahmy 2019; G?siorek & Ró?ycka, 2017). have recently featured several studies invertebrate-derived DNA (iDNA) tool survey populations detect pathogens (Abrams Alfano 2021; Baker Drinkwater 2020; Fahmy Schnell 2012; 2015; Siddall Tessler 2018; Weiskopf 2018). Other work investigated role vectors animal human (Kang 2016). Despite this abundance ecosystems, A complete understanding ecology promises maximize utility emerging techniques. characterization natural can place disturbed systems appropriate context. Early research focused primarily taxonomy, anatomy, history, providing only simple characterizations geographic (Mann, 1962; Moore, 1929; Sawyer, 1981; Smythies, 1959). One exception was Fogden Proctor's (1985) experimental which demonstrated (H. picta, leech, H. zeylanica, complex) survive months provided humidity levels kept (Fogden Proctor, 1985). Based findings, authors predicted would likely absent habitats humidity, logged forests. Only empirical begun shed light Kendall (2012) changes picta zeylanica response modifications. showed overall dropped rising temperature increased moisture, although sites standing water. Contrary predictions, than primary forests, but fell zero oil-palm plantations. solely due significant increase In addition, decreasing those declined, changing affected differently (Kendall, 2012). Within degraded habitats, structure microclimatic conditions Forests canopy possibly humid microclimates (Jucker 2018) small occupancy wet dry seasons However, association seen season sumatrana (of increasing height season. Further, less occur heterogenous or heavily during season, microclimate quality foraging behavior, Ró?ycka (2017) wider range microhabitats subagilis complex). Specifically, aggregated along trails subagilis, where they hunted ground by climbing onto vegetation thereby number potential hosts. contrast, specialized habitat, preferring hunt leaf litter (G?siorek This preference bushes also been noted some earlier fauna Borneo (Smythies, To date, different elevational gradient, nor forested landscape. Such information critical understand respond especially important given biomonitoring knowledge preferences avoid biases (Baker Cabang Panti Research Site (CPRS) West Borneo, spans ranging 5 1100 m a.s.l. covers seven distinct (Marshall 2021). As such, it ideal determining patterns local CPRS home least terrestrial leeches: (hereafter leech) spp., complex leech), studied eight year(s). While there confusion pertaining taxonomic identification not possible distinguish among using morphological alone. do molecular evidence confidently identify area, limit level. Tiger behavioral characteristics allow us tell apart field. features prominent yellow stripes notable bite, uniformly color painless bite (Figure 1). study, ecological year(s) collected 14 vertebrate census surveys total sampled distance 7748.7 km (Table whether influenced (a) (altitude, rainfall history), (b) (forest species). gathered Gunung Palung National Park, Indonesia (1°13? S, 110°7? E) 2). area 34 km2 distinct, contiguous differ geology, drainage, elevation, plant composition, structure, phenology These peat swamp nutrient-poor, bleached white soils overlain variable amounts organic matter (5 10 a.s.l.); freshwater nutrient-rich, seasonally flooded, poorly drained gleyic (c) alluvial bench rich sandstone-derived deposited upstream sandstone granite parent material 50 (d) lowland well-drained clay content sparse patches shale (20 200 (e) well-drained, granite-derived (200 400 (f) upland (350 800 (g) montane largely (750 a.s.l.). 2000, AJM established transects 2021) mean length 3.4 ± SD 0.24 km. Some contained entirely single type, others spanned multiple analysis, divided segments varied 550 m. measured altitude each its approximate midpoint Suunto Vector Altimeter 2014). walked per sampling occasion entire duration listed Table 1. From November 2012 September 2020, systematically counted transects. transect constant, slow speed (~1 hour) twice month (starting opposite ends), beginning 0530 h. Surveys normally carried out observer, occasionally observers walk same (e.g., when new staff trained). Every 500 segment, paused examine themselves thoroughly period minutes. Any attached bodies counted, identified, removed. When surveyors present, individuals searched recorded combined observers. addition counts, made note history. examined dampness categorized “wet,” “moist,” “dry.” history whole day “Raining now,” “Rained last night,” yesterday,” before 48 hours ago” based station. Rainfall sometimes highly localized site, so segment's assessment station may inaccurate. reason believe rare mischaracterizations biased any particular direction, however. conducted preliminary exploration subsequent analysis program R (R Core Team, RStudio 2020). Histograms species-specific over-dispersed, indicating base S1). Next, checked observer bias comparing / transect; (JK) consistently observers, warranted S2). Although most count long, variation segment halting change transect). therefore next see S3). Leech length, habitat-specific variables perhaps driving S4). plotted histories, (Figures S5–S8). All four predictors appeared thus included candidate models. and/or reliable ran function GLMER Lme4 Package (version 1.1.26; 2020), following predictors–altitude, rain fixed effects. ID random account pseudo-replication unit repeated times. models, top produced non-convergence warnings, large (n = 197). Nevertheless, did produce interpretable plausible estimates sizes standard errors. confirm results, reran without simpler convergence warnings comparable multi-level S9). feel confident presenting basing inferences deem preferable because permit control segments. parse spatially partitioned result competition best first predictor again predictor. Akaike Information Criterion (AIC) S1a,b) Base 2020) visualize S11). considered 95% confidence intervals overlap zero. Prior fitting, pair-wise correlation plots S10) ensure correlated (which define r > 0.75) issues convergence. Thus, together model. continuous (altitude species) centered standardized direct comparison size magnitude. prior “peat swamp,” “dry,” “rained reference respective categorical predictors. Of 58 (coded JK Rest) essentially all weight (Akaike 1) S1a,b). except three highest reliably counts. 3). presented below back transformed (exponentiated) beta coefficients facilitate interpretation scale. Soil odds counting 4.99 (±1.08 SE) 4.20 fold moist conditions, respectively, Similarly, moist, 5.62 (±1.09 3.85 times, 2.10 (±1.06 1.96 times rained previous night day, ago 1.87 1.72 Freshwater occurrence, 3.32 (±1.25 3.29 (±1.23 swamp, level granite, montane, 0.27 (±1.24 SE), 0.22 (±1.30 0.13 every additional 1.21 (±1.04 34.72 5.09 (±1.19 Again, 1.58 assess abovementioned model, predictor, excluded came up 1 shows low-lying whereas 4). similarities direction specifically bench, frequently forest. Our findings consistent suggested (Enguang, 2001; 1985; Sket Trontelj, 2008). demonstrate larger recent events rainfall. controlling still albeit smaller assessment, few empirically activity. hainana rubber plantations China abundant precipitation Enguang Chuanjing, 2000). Rubber characterized intensity, temperatures regenerating (Hidayat Measures opposed current/recent better capture properties retention drainage. Higher soils, generally retain experience periodic flooding 2008; Marshall mainly overlapped significantly observation, paired showing suggests fluctuations limited retention. zeylanica) decreased temperature, leech. explanation Kendall's took explain denser understorey creates newer niches occupation inhabit vegetation, contrast lives litter. al. (2019) sumatrana) humidity. impacts invertebrate (Boyle Luke highlight absence between accounted type. despite specializing inter- intra-specific exhibiting behaviors. interspecific reduced areas subagilis) via horizontal vertical partitioning aggregating closer off-trail locations distributed itself vertically hunt, potentially accessing host preferred avoidance temporal spatial case 2017; Miler 2019) Larger ambush preferentially hunting morning, juveniles active throughout order competition, future should focus scale differences behavior. offers resolution previously available, limitations. First, 1130 h, dataset diurnal detectability mentioned above, described intra- inter-species feeding confound merits further exploration. second limitation indicate absolute latched walking Therefore, reflective There means actual abundance. instance, remain dormant Proctor blood meal exhibited dormancy sluggishness proximity. disparately latch counted. make species' directly modeled occurrences separately. third able analysis. 2017) subagilis). highlights gaps suggest foci research. past Kendall, 2012) degradation, plantations, ours investigates More data, however, needed clarify relationship forest-dwelling (i.e., prey) densities relatively unaffected factors. investigations axes enrich arboreal, well crepuscular nocturnal, vertebrates. conceptualized study. HUW acquired funding. AJM, ES, TWS, TMS administered project. RR, J, DA, SL, SMJ, EJB, ZJ field investigation. curated data. SN, ABB, GRE, BJF, formal HUW, wrote original manuscript draft. reviewed edited manuscript. supported funding University Michigan, Victoria Wellington, Leakey Foundation, Idea Wild, Mohamed bin Zayed Species Conservation Fund, Disney AZA Ape TAG Initiative, Anne Nacey Maggioncalda Jeffery Maggioncalda. sincere thanks counterparts supporting institutions: State Ministry Research, Technology, Education, Forestry's Directorate General Forest Protection Nature Conservation, Institute Sciences, Universitas Nasional, Tanjungpura, Park Bureau. No conflict interest reported authors. applicable institutional national guidelines care animals followed. Fieldwork under approval Technology Republic Indonesia, Jakarta, Balai Taman Nasional Palung, Ketapang, Kalimantan Barat, support openly Dryad Digital Repository: https://doi.org/10.5061/dryad.r4xgxd2g9 (Nelaballi 2022), https://github.com/andrewjohnmarshall/leech_distribution. Appendix S1 Please note: publisher responsible functionality supplied queries (other missing content) directed corresponding author article.