Tracking climate vulnerability across spatial distribution and functional traits in Magnolia gentryi from the Peruvian tropical montane cloud forest

Description

Methods

Study species— Magnolia gentryi A.Vázquez (Section Talauma, Magnolioideae, Magnoliaceae), is an endangered-evergreen Peruvian TMCF (Rivers et al. 2016) relict-tree species which only occurs in the Southeastern Tropical Andes (Vázquez-García et al. 2012). This species inhabits moderate steep-slope ravines and in valley bottoms and occasionally hilltops in association with Hyeronima macrocarpa Müll. Arg., Juglans neotropica, Cecropia spp., Clusia spp., Inga spp., Hedyosmum scabrum (Ruiz & Pav.) Solms, Guatteria sp., Socratea sp., and Podocarpus sp.

Site characteristics— The study forest is in a TMCF fragment close to the Tsachopen native community in Oxapampa Province, Oxapampa District, Perú (10° 32' 17''S, 75° 27' 05''W; from 1800 to 2000 m asl; 765 ha;). Nevertheless, TMCF is influenced by human activities such as logging, corn, coffee, and granadilla (Passiflora ligularis Juss.) plantations. The study forest has a temperate climate (Cwb; Peel et al. 2006), with rainy and cool summers, followed by mild and wet winters. Moisture levels of 75-90% are the result of frequent drizzles and fog throughout the year (Myster 2021). Soil is Andosol (Th) with silicate-rich material under acid weathering (IUSS Working Group WRB. 2015). The studied forest represents an isolated distribution of endangered Magnolia gentryi.

Magnolia gentry fieldwork survey— We conducted field surveys in the study forest from 2020 to 2022 to investigate the presence of the uncommon Magnolia gentryi in Peruvian TMCFs. All the Magnolia trees (20 individuals) were geo-referenced with a GPS (Garmin eTrex® 10; Garmin Ltd., Olathe, KS, USA) (Figure 1B).

Anatomical functional trait assessment— we measured dasometric, wood anatomical, and leaf traits (Table 1). (i) Three specific dasometric traits were measured in 20 Magnolia individuals (DBH ≥ 50 cm). 1) Tree height (measured with Nikon Forestry PRO II Laser Rangefinder/Hypsometer); 2) diameter at breast high (DBH; measured at 1.3 m tree height with a diameter tape; Forestry Suppliers, Model 283D/20F) (Pretzsch 2009); and 3) bark thickness (measured with the Swedish bark gauge; SUUNTO, 60 mm, Vantaa, Finland). The bark thickness was taken from the cardinal positions (e.g., north, south, east, and west) around the stem (Wilms et al. 2021; https://doi.org/10.3390/ f12101386).

(ii) We measured six wood anatomical traits (tree-ring width, hydraulic diameter, vulnerability index, fiber lumen diameter, fiber length, and fiber wall thickness). To obtain tree-ring width (TRW) data, we cored twice 20 Magnolia individuals at 1.5 m (breast height) with a borer of 5 mm internal diameter (Häglof^®, Langsele, Sweden). The cores were dried at room temperature (48 h), glued to wooden supports, and polished with coarse (100, and 360 grit) and fine (400, 600, 800, 1000, 1200, 1500, 2000, and 2500 grit) sandpaper for best visualization of annual growth rings and vessel structure in cross-section (Rodríguez-Ramírez et al. 2023). Wood dust and vessel lumina were removed using a compressed air device. Wood core samples were preserved according to Speer (2010) and deposited at the Dendrochronology Lab, Universidad Continental, Huancayo, Peru. We measured the TRW using a stereoscopic microscope (OMAX®), and Velmex Tree Ring Measuring System (Velmex, Bloomfield, NY, USA) with an accuracy of 0.001 mm using TSAP-Win v.4.67c (Rinn 2003). The Southern Hemisphere Convention (Schulman, 1956) was used to assign the annual growth rings.

On the other hand, to obtain hydraulic diameter (D_H), vulnerability index (V_I), fiber lumen diameter (F_D), fiber length (F_L), and fiber wall thickness (F_WT) data, we collected stem discs from 10 Magnolia individuals containing bark and sapwood (~3 cm thick). The wood samples were obtained using a handsaw. Wood samples were immediately fixed in a formalin-acetic acid-ethanol solution (10:5:85) (McCracken and Johansen, 1940). After 48 h, the samples were washed with tap water and stored in a glycerin-ethanol-water solution (1:1:1; Sass, 1958) for 60 days. Microcore samples were paraffin-blocked, trimmed, and sectioned in a rotary microtome (8 μm thick). Sections were stained with safranin-fast green (Ruzin, 2000), mounted in synthetic resin, and then oven-dried in an oven at 60 °C for 24 h. Before staining, paraffin was removed using D-limonene, and the sections were rehydrated with alcohol solutions with decreasing concentration (van der Werf et al. 2007). Additionally, we prepared macerations using Jeffrey's solution (Berlyn and Miksche, 1976) to collect data on vessel elements and fiber lengths.

Wood microsections were prepared to obtain high-resolution anatomical images (1.3 µm per pixel resolution) following IAWA recommendations (IAWA Committee, 1989) and Piermattei et al. (2020). To obtain digital images mounted slides were observed under a light microscope (Olympus SZ61), associated with a digital camera (Leica DFC 490), using the software ImagePro-Plus v. 4 (Media Cybernetics) was utilized to measure the vessel traits. We considered hydraulic diameter (D_H, mm), vulnerability index (V_I, 1 to > 3), fiber lumen diameter (F_D, mm), fiber length (F_L, mm), and fiber wall thickness (F_WT, mm) because these traits could provide an approximation of environmental susceptibility to how Magnolia wood is affected by climatic variation (Rodríguez-Ramírez et al. 2020, 2021).  

(iii) We measured from 40 specimens four leaf vein traits related to ecohydraulic vulnerability. These included vein density (mm∙mm⁻²), vein length (cm), leaf base angle, and leaf shape (qualitative) (Rodríguez-Ramírez et al. 2021). We digitalized 120 Magnolia leaves from herbarium collections (MO, MOL, USM, and HOXA herbarium. Each Magnolia specimen was placed directly on the glass of a high-resolution (700 dpi; accuracy 0.1 mm) flatbed scanner (HP OfficeJet Pro 7740, Hewlett-Packard Development Company, L.P., Houston, TX, USA). The measurements were performed using the software ImagePro-Plus v. 4 (Media Cybernetics).

Climate data— We used the monthly precipitation (Prec), evapotranspiration (EvT) in mm, mean maximum, and minimum temperature (T_max and T_min) in °C, from the CHELSA database (http://chelsa-climate.org/; Karger et al. 2021), the layer resolution was c. 1 km², with records spanning from 1980 to 2018.