Litter‐trapping tank bromeliads in five different forests: Carbon and nutrient pools and fluxes

Bromeliads are the most abundant litter-trapping plants in Neotropical forest canopies. By intercepting litter, bromeliads obtain and retain nutrients before they reach pedosphere. Here, we analyzed litter captured stored by tank (TB) five different forests along an elevation gradient Mexico. Among those forests, carbon nutrient pools nitrogen fluxes TB were estimated a mangrove (MF) semi-deciduous tropical (SDTF). The composition of trapped was similar to litterfall mainly composed leaves. Most dry season, found significant effect projected plant area interaction between month site on bromeliad capture. Moreover, increased exponentially with differed three studied species. In MF (with ca. 2,700 ha−1), barely 1% annual is these plants, but even SDTF, >10,000 ha−1, only 2.4% captured. We that small represented <1% aboveground biomass. Furthermore, residence time floor. light our results, conclude cycle negligible. Las bromelias son las plantas captadoras de hojarasca (hojas, ramas, cortezas y otros despojos vegetales) más abundantes en el dosel los bosques Neotropicales, cuales obtienen retienen nutrientes antes que estos lleguen al suelo. En este estudio, analizamos la captura tanque (BT) cinco distintos México. Además, dos bosques, un manglar (BM) bosque subcaducifolio (BTSC), estimamos reservas carbono almacenados estas plantas. Se encontró BT capturaron mayor parte estación seca. La composición capturada todos sitios fue del suelo estuvo compuesta principalmente por hojas. se efecto significativo área interacción entre sitio mes hojarasca. cantidad almacenada tanques incrementó manera exponencial con difirió tres especies bromelias. BM (aprox. ha−1) BTSC (>10,000 aproximadamente producción anual hojarasca, respectivamente. nitrógeno fueron pequeñas representaron almacenado biomasa aérea estudiados. tiempo residencia bosques. Con base nuestros resultados, concluimos ciclo es insignificante Tree one main pathways for transferring from vegetation soils represents essential link organic production–decomposition cycle, which makes bound biomass available (Attiwill & Adams, 1993; Meentemeyer et al., 1982; Vitousek Sanford, 1986). However, fraction this (0 >50% total above ground production) gets canopy does not immediately come contact terrestrial soil (Álvarez-Sánchez Guevara, 1999; Dearden Wardle, 2008; Nadkarni Matelson, 1991). This phenomenon particularly relevant tropics because presence numerous including epiphytic species special features such as modified leaves upward-growing roots effectively capture co-opt it moves trees Weissenhofer Zona Christenhusz, 2015). canopies far (Benzing, 2000; Hietz Hietz-Seifert, 1995; Zotz, 2016), densities ranging few dozen more than 70,000 individuals ha−1 (Richardson Sugden Robins, 1979). pedosphere them compartments, is, (1) tissues (Benzing Renfrow, 1971; Benzing Seemann, 1978; Richardson 2000); (2) (TL), overlapping leaf bases (Castaño-Meneses, 2016; Paoletti 1991; Richardson, 1999); (3) external (not part study), accumulated mats outside (Ortega-Solís 2017; Victoriano-Romero 2019). Carbon first second compartments have been quantified very species, limited data suggests neither large 1974; Souza Díaz 2010; 2000). Bromeliad productivity exceed 15% net primary (Oliveira, 2004; 2000), TL make up less pool top 10 cm important dead matter, can comprise half or matter (Díaz Gómez-González Nadkarni, 1984), its decomposition contributes greenhouse gas emissions (Martinson 2010). provides bromeliads, absorb directly their tanks via foliar trichomes 1970), source energy, over broad geographic range, associated heterotrophic food webs (Farjalla characterized high diversity organisms bacteria, protozoans, fungi, invertebrates, vertebrates (Aguilar-Cruz, Arenas-Cruz, 2021; Durán-Ramirez 2019; Frank Lounibos, 2009; Goffredi 2011; Grippa 2007; 1999). Although ecological importance acknowledged, served model ecosystems studying trophic interactions (Srivastava 2020), quantitative models storage surprisingly rare. published studies generally focus phytotelm fauna (Carrias 2001; Castaño-Meneses, 1999), measurements (BLC) indirect (Ospina-Bautista Estévez Varón, 2016). there knowledge gap regarding seasonal variation at ecosystem level based density while size identity ignored (Paoletti These shortcomings, do allow assessment role processes motivated present study, (a) litterfall, one-year period types gradient; (b) modelled BLC function size, site, considering interspecific variation; (c) species; lastly, (d) two contrasting forests. study performed eastern Trans-Mexican Volcanic Belt central Veracruz, Mexico, where varies sea Gulf Mexico 4,250 m Cofre Perote Volcano (Figure 1). Temperature decreases linearly elevation, moist Trade Winds flow easterly direction, passing through coastal plains until being lifted mountain ranges, resulting higher precipitation mid-elevations (1,500–2,000 m) (CONABIO, 2011). Along elevational gradient, selected sites 0–2,200 a.s.l. 1, Table S1). included (MF, 5 a.s.l.), grows low, flat, lacustrine–marshy exclusively comprised (Lara-Domínguez 2009); (SDTF, 650 a.s.l.) dominated tree oaks; oak (TOF, 1,005 deciduous oaks, cloud (CF-C, 1,650 CF-Z, 2,210 mixture temperate floristic elements located rainiest portions rain fog events frequent throughout year (Carvajal-Hernández 2020; CONABIO, Monge-González Williams-Linera, 2012). At each sites, plots 15 × established least another. All (including palms ferns) diameter breast height (DBH) ≥5 measured (height DBH) labelled. Specimens identified Herbario XAL (INECOL, A.C., Xalapa), except Quercus, deposited Nacional México (MEXU), Instituto Biología, UNAM, City. Five conical collectors (mosquito net, 0.25 m2, 56.5 diameter) placed 1 plot (n = 75). Litter collected monthly January December 2018, dried 70°C 72 h, weighed. Litterfall calculated extrapolating 3.75 collectors’ per forest, ha reported Mg yr−1. Additionally, samples sorted into following categories: epiphytes (non-bromeliads bromeliads), non-vascular epiphytes, (sorted species), flowers, seeds/fruits, twigs/bark (<2 diameter), detritus (fragmented decomposed material could be any former category). category type weighed, subsamples 3 g (5 case leaves) (except epiphytes) every other starting when <3 sample used. To determine capture, individual surroundings medium-to-large-sized (maximum length >30 cm) without inflorescence nor damage. (Aechmea bracteata, A. nudicaulis, Tillandsia fasciculata, T. heterophylla, imperialis, aff. limbata, macropetala, S1) morphology covered gradient. labelled, all materials inside removed, (PA) obtained perpendicular view photographs taken 4 digital camera (Samsung DV150F; Samsung Electronics Co., ROK). take photographs, white tripod, plastic foil protected direct sunlight. For photograph, scale (black rigid cardboard fixed maximum horizontal orientation PA then using ImageJ software (Rasband, 2018). Afterward, relocated original locality randomly assigned (DAP >20 cm, n 75) 1.5 maintaining upright position S2, details, see Aguilar-Cruz, 2021). Once during processed same way collectors. During census, nine died replaced conspecifics. After year, again. sampled 2018 2019 SDTF. end, DBHs 7 76 common (Table S1), growing them, almost accessed ladders single rope technique (Moffett Lowman, 1995). Following criteria, Avicennia germinans, Laguncularia racemosa, Rhizophora mangle chosen MF, Exothea paniculata, Quercus polymorpha, Protium copal, Zanthoxylum fagara Then, selected. recording distance laser measuring tool (Model GLM 80 professional, Bosch; Stuttgart, Germany), labelled bags avoid losses; then, lowered ground. Plants grew colonies separated field bag. When manual harvesting impossible, branches cut pulleys ropes. lost process discarded considered estimations average values. laboratory, extracted tweezers, brushes, spatulas, also vigorous shaking turned upside down tarp. deep axils, high-pressure water separated, ramets (foliar counted individuals. [in SDTF 1/3 colonies, general amount equivalent (personal observation)]. living foliage assessed sampling disks (1.3 base, center, tip four positions rosette good representation entire plants. discs h 98 2 component chemical analysis. extraction, weighed hanging scales [precision: (<1 kg, Model SR-1KG; American Weigh Scales; Georgia, USA), (>1 HDB 10K10N, Kern Sohon; Balingen, Germany)]. 36 constant weight. relationship wet vs. mass S3) allowed estimate TL, 85 belonging (A. limbata) method described (Nikon D3300, Tokyo, Japan). Samples (trapped disks) transferred Oldenburg University They fine-milled, concentrations determined CN analyzer (Flash EA 1112; Thermo Fisher Scientific, Milano, Italy) according manufacturer's recommendations copper tungsten oxide catalyst. phosphorus analysis, mg digested 200 μl conc. HNO3 30 30% H2O2 (Huang Schulte, 1985). digestion, 770 distilled added concentration 710 nm UV–VIS spectrophotometer (Specord 50; Analytik Jena, Germany). Sodium, potassium, magnesium, calcium flame atomic absorption spectrometer (AA240; Agilent Technologies, Mulgrave, Australia) recommendations. Tank PA, Gamma linear mixed-effect (GLME). used log set random incorporate dependency among observations tree. Simplification minimal adequate done above. Four influential points removed improve fit (the fitted complete presented Figure S4). tested possible (DG) TL. outlier, log-transformed LME DG covariates effect. Statistical analyses R version 3.6.3 (R Core Team, 2020); GLME lme4 package (Bates 2015) evaluated (marginal conditional R2, residual plots) sjPlot (Lüdecke, Marginal effects ggefects available, number biomass, models. Through models, predicted mean 95% CIs site. abundance, extrapolated ha. this, S5) bromeliads. Later, S6). Using model, plots. Subsequently, Finally, m2 (non-replaced SDTF) hectare forest. independent response numeric variables log-transformed. Annual highest (13 yr−1) followed CF–C (8.6 yr−1), TOF (8.5 (7.1 CF–Z (6.0 yr−1). showed strong seasonality distinct peaks (percentage 2a). TOF, occurred warm both CFs cold season April another northerly winds November Leaves accounted contributing 60–79% S7a), 67 100% stemmed dominant Typically, dominated, Twigs bark contributed 7–17% categories 13–29%. “others,” fruits seeds, somewhat minor contributions S7a). (cold warm, 57–67%) (January–May), evenly 2b). whole, 63–79%, twigs 7–22%, 10–24% S7b). (ca. 60% Quercus), mangrove, approximately three-quarters According minimum LME, had significant, positive (χ2 80.4, df p < .001, 3a) 440, 44, 3b). explained 46% variance, 75% S2). Our 94.9, .001) 41.6, 2, (TL). fastest limbata slowly bracteata 4). 65% 66%. 4), deviance significantly 40.2, .001), decreased Akaike's information criterion S3). explanatory power reduced still moderate, explaining 41% 58%. covariates, (t 21.75, no .05, .96; log-linearly (DBH, t 4.4, difference 0.94 intercepts 2.8, .01), hosting 5a). 5.1, .29, .78, 5b). patterns increasing 1.4, .18, 5c). cover 1.6 3.5% (CI 95%, 0.5–2%) 1–5%) respectively. (bromeliads TL) (MF 6. 1,000 a.s.l, value evergreen, rain-green worldwide (i.e., y−1) (Zhang 2014). region, intermediate elevations show richness abundance vascular dominate epiphyte assemblages (Guzmán-Jacob pattern result mild temperatures humidity may provide increase release (Aguilar-Cruz 2020). productive rapidly accumulate some decomposers stigmatid mites springtails would mineralized alternate flooded cycles increases Zotz Thomas, upper (>2,000 m), decrease factors probably contribute non-suitable conditions maintain strategy. Decomposition too slow scanty sustain warm-season growth, coupled sub-freezing reduce Invariably, bark), (leaves 60–79%; 7–17%) 64–73%; 17%) Considering proportions well, leaves, especially trees, understory, probably, energy Vascular however, <5% BLC, suggesting little recycling nutrients. contrasts previous findings fine Costa Rica, 35% standing origin, indicating reuse within community rooted arboreal (Nadkarni Projected frequently morphological trait predict reflects interception Ospina-Bautista variable traits sheath length, poorly variance) Unsurprisingly, results confirmed differences temporal (month) spatial (site) variables, accurately (Figures 2b 3b), variance (46%). Contrastingly, controlled non-independence effects, variance. highlight intraspecific future seek communities. Unlike 90% morphologies species-specific, has therefore should whenever possible, use require accurate estimates plant. include predictor arrive reasonable forecast pools, considered. consistent relations bromeliads: Ontogenetic changes capacity, allowing content despite species-specific (Zotz strata receive al. (2000) 12 m, Karasawa Hijii (2006) attachment accumulation bird's-nest fern Asplenium nidus either. suggest modest. trap 3.5%, intercepted. numbers low comparison plants: rainforest palm Astrocaryum mexicanum, covers 68% ground, traps >45% New Zealand crown Blechnum discolor, occupies 2% surface, retains ca.10% falling (Dearden 2008). underestimate community, substantial quantity, rosettes (tank atmospheric bromeliads) (third compartment), included. Its addition easily double observation) works abundant. <0.1% C N (López-Portillo 2020, unpublished data), <0.06% SDTFs (Jaramillo Even (living TL), P, K, Ca, Na, (<3 kg slightly values example, twice times much K dwarf Puerto Rico, <4% floor (net >7 y−1), C, N, Ca comparable lower case, tissue factor P 9, 7, discrepancy due smaller (average <0.1 0.5 faster cycling. Seemann (1978) proposed canopy, long life, modest production, modify usual mode retaining extended periods long. Assuming steady-state input, months, turnover monospecific mixed mangroves Florida (2–6 months) (Twilley Palmarejo (SDTF), 9 Nutrients absorbed retained longer periods. assuming equal inputs minus (TL biota), arrived years takes new inner Rico (2.8 years) although potentially long-lived 2005) several years, will ultimately fall decompose floor, (Matelson 1993). authors acknowledge Karina Osorio Salomón (INECOL) Susanne Töller (University Oldenburg) help fieldwork collection, Javier Tolome sorting, Carlos M. Durán Espinosa (INECOL), Claudia Gallardo Hernández Susana Valencia Avalos (UNAM) identification botanical material, Norbert Wagner Kühn analyses. Y. Aguilar-Cruz funded Consejo Ciencia Tecnología (CONACyT 408928), Deutscher Akademischer Austauschdienst (DAAD 91569882), Heinz Neumüller Stiftung. Open Access funding enabled organized Projekt DEAL. YAC GZ conceived designed study. data. led writing manuscript JGF. support openly Dryad Digital Repository. https://doi.org/10.5061/dryad.tx95x69zg García-Franco, Please note: publisher responsible functionality supporting supplied authors. Any queries (other missing content) directed corresponding author article.