1 Background

The Sierra Nevada de Santa Marta (SNSM) is an isolated mountain on Colombia’s Caribbean coast. Rising to 5775 m asl, this immense massif is the tallest peak in Colombia and the tallest coastal mountain in the world. The SNSM is separated from the Andes mountains by hot and dry lowland Caribbean savannas to the southeast and by a large wetland, the Ciénaga Grande de Santa Marta, to the southwest. Owing to its unique location, isolation, and complex topography, the mountain is home to a mosaic of distinct ecosystems with high levels of endemism (Adams 1973; Carbono and Lozano-Contreras 1997; Cardona and Ojeda 2010).

Unfortunately, our understanding of the mountain’s flora is incomplete. Although there have been important botanical expeditions to the SNSM (Van der Hammen and Ruiz 1984; Ayers and Boufford 1998), no study has implemented a standardized tree plot with which to compare the cloud forests of SNSM to others in the region (ForestPlots.net et al. 2021). Therefore, our knowledge of the diversity, composition, and structure of tree communities on the SNSM remains severely limited. To address this problem, we installed a permanent 1-ha forest plot on the SNSM, and we here present baseline data from the plot. Our work represents the first standardized tree plot above 600 m asl on the SNSM and will contribute to our understanding of both local and regional cloud forest diversity and composition.

2 Methods

2.1 Study site

The 1-ha plot is located on Cerro Kennedy, a high point reaching 2830 m asl on the northwest side of the SNSM. On the southern face of Cerro Kennedy, in Reserva El Dorado, is a strip of relict old-growth forest totaling just over 250 ha. The plot is located in this remnant forest at 2200-m elevation (11° 6′4.37″ N, 74° 2′22.20″ W, Fig. 1). Mean annual temperature is ~ 15 °C, and mean annual precipitation is ~ 1900 mm.

Fig. 1
figure 1

Map showing the location of the 1-ha plot in Colombia and on the Sierra Nevada de Santa Marta

Full size image

2.2 Plot demarcation

The specific location of the plot was chosen pseudo-randomly, avoiding difficult terrain and steep topography, in a location easy to access from the trail, and ensuring that each plot edge was at least 100 m from a forest edge to avoid as many edge effects as possible (Laurance et al. 2002). We demarcated the plot boundaries with PVC tubes every 10 m following cardinal directions with a compass. We then installed 100 10 × 10 m subplots following the same methods. The shape of the plot is irregular due to a cliff we encountered in the southwest corner and is steep, with an ENE to WSW slope exceeding 30°. Because each 10-m interval was demarcated using a planar projection and the ground distance was corrected with a digital clinometer, the total ground area of the plot is approximately 1.21 ha.

2.3 Tree measurements

We tagged, measured, mapped, and identified every stem ≥ 10-cm DBH following standard protocols (Lopez-Gonzalez et al. 2011; Phillips et al. 2018). For trees with buttressed trunks or other irregularities, diameter was measured above the irregularities, and measurement height was noted. We collected voucher specimens for species identifications and stored them in the herbarium at the Cartagena Botanical Garden “Guillermo Piñeres” (JBGP).

We also scored each tree for liana infestation, sun exposure, and canopy damage. Liana infestation scores were 0 for trees without lianas and 1, 2, 3, or 4 for trees with 1–25%, 26–50%, 51–75%, or 76–100% of their canopy covered by lianas, respectively (Clark & Clark 1990). Sun exposure scores were from 1 to 5, with 1 for plants lacking direct sunlight except during sun flecks and 5 for canopy emergents (Dawkins 1978). Crown damage scores were 0 for trees with intact crowns and 1, 2, 3, or 4 for trees with 1–25%, 26–50%, 51–75%, or 76–100% of the crown broken, respectively. Finally, we estimated tree height.

2.4 Access to the data and metadata description

The datasets generated during the present study are available in the SiB Colombia (Contreras et al. 2024), URL: https://ipt.biodiversidad.co/sib/resource?r=parcelasnsm#anchor-downloads, DOI: https://doi.org/10.15472/ftwol2. Metadata files are included with this dataset and also accessible at https://metadata-afs.nancy.inra.fr/geonetwork/srv/fre/catalog.search#/metadata/a4a7855f-4cb2-459b-a1a2-287bd1e6d587.

2.5 Technical validation

All variables were examined for inconsistencies. Any outliers, impossible tree scores, and mapping errors were revised by checking with field sheets and confirming with the field team. We made figures to summarize and visualize the data to ease error identification (Figs. 2–3). Species identifications that were made in the field were confirmed or refined by comparing our collections to physical material in the Jardín Botánico de Cartagena “Guillermo Piñeres” herbarium (JBPG). When physical material was not available, we reviewed digital herbarium specimens in the online repositories of the New York Botanical Garden (NY), Missouri Botanical Garden (MO), Field Museum Herbarium (F), the Royal Botanic Gardens Kew (K), and the Herbario Forestal de la Universidad Distrital “Francisco José de Caldas” (UDBC).

Fig. 2
figure 2

A Stem map of the 1-ha plot. Each dot is a stem, with the size of the dot corresponding to its diameter. The four most common species are color coded. B Bar graph of the most common species, ranked by number of individuals. The four most common species are color coded as in A. C Bar graph of the 10 most species-rich families, ranked by number of species

Full size image
Fig. 3
figure 3

Histogram of stem diameters in the 1-ha plot. The blue dashed line indicates the mean diameter (23.3 cm) at breast height

Full size image

In total, we measured, tagged, mapped, and identified a total of 924 stems representing 846 individual trees and 85 species in 41 families. Four species represent nearly half of total stems, and four families account for 32% of all species (Table 1, Fig. 2). We calculated three diversity metrics including Shannon, Simpson, and inverse Simpson indices, which were 3.36, 0.93, and 14.3, respectively. Even though each metric is calculated differently, they all illustrate relatively high diversity within the plot.

Table 1 Table summarizing the number of stems, number of individuals, maximum diameter at breast height in cm (max DBH), wood density in g/cm3 (WD), basal area in cm2, and the percent of total plot basal area for each species (or morphospecies)
Full size table

Using measured tree diameters (D), we calculated the basal area of each stem as (D/2)^2*π. We then assigned wood density for each species from a wood density database (Zanne et al. 2009). When a species-level wood density was not available, we assigned a species the genus- or family-level wood density average instead. We then estimated biomass for the entire stand using an allometric model via the R package BIOMASS (Chave et al. 2014; Réjou-Méchain et al. 2017). Tree diameters range from 10 to 123.5 cm, with a mean diameter of 23.3 cm (Fig. 3). The total basal area for the entire plot is 55.96 m2. The four species with the largest basal area are Pouteria espinae (18% of total basal area, Table 1), Ficus insipida (Moraceae, 11%), Chrysochlamys colombiana (8%), and Calatola costaricensis (7%). Total plot biomass is ~ 451 t.

2.6 Reuse and potential limits

Plot data is essential for analyses of species composition across space and through time. Our plot data will be uploaded into ForestPlots.net (Lopez-Gonzalez et al. 2011), a pantropical network of tree plots which facilitates studies that investigate patterns of tree species composition and diversity. Data on tree-level liana infestation, sun exposure, and crown damage in 1-ha plots is scarce in tropical montane cloud forests, so our plot will be useful when doing cross-site examinations of these variables. We plan to do a full recensus of the plot in 2028 and every 5 years thereafter in order to characterize compositional changes and calculate growth and mortality rates through time.

Data on the structure and composition of trees such as those presented here are important components of earth systems modeling. Indeed, forest plots help to inform and calibrate remote sensing estimations of aboveground biomass and carbon sequestration (Chave et al. 2019). Thus, our ground-based data can be used in future studies that estimate biomass on large spatial scales using remote sensing techniques.

Some species in our dataset are rare, endangered, and/or endemic (e.g., Pouteria espinae, the most common species in the plot) to the SNSM. Therefore, the occurrence data (and future growth, mortality, and recruitment data) of these species can be useful for their management, for protection, and for future assessments for the Red List of the International Union for the Conservation of Nature (IUCN).