Fog-Fed Oases Reveal Hidden Rules of Plant Diversity in South America’s Coastal Desert

Along the western edge of South America, where the Peruvian and Chilean coastal deserts stretch more than 3,500 kilometers under some of the driest skies on Earth, life clings to the slopes in scattered vegetation oases known as coastal lomas. These fog-fed islands of green, sustained by advective fog deposition and brief El Niño rains, have long puzzled biogeographers because their plant diversity does not appear to follow the tidy latitudinal gradients seen in wetter ecosystems. Now, the most comprehensive survey of these ecosystems ever compiled has revealed that water availability, topography, and geography—not latitude itself—are the primary forces shaping the taxonomic and phylogenetic diversity of seed plants across these remarkable habitats.

The study, published in Frontiers of Biogeography, assembled floristic data for 71 coastal lomas distributed from 7° to 30°S, drawing on herbarium records from Peru and Chile, scientific publications, theses, online datasets, and the Global Biodiversity Information Facility. After taxonomic standardization, the researchers documented 1,363 accepted plant species, placing 1,305 of them—95.7 percent of the dataset—onto a seed plant phylogeny built on the backbone of Smith and Brown’s 2018 phylogeny. This allowed the team to evaluate not just how many species live in each loma, but how evolutionarily distinct those species are from one another.

To disentangle the ecological processes structuring these communities, the researchers calculated three complementary phylogenetic diversity indices: Faith’s phylogenetic diversity, which sums all branch lengths on the tree and captures the total evolutionary history represented in an assemblage; mean pairwise distance (MPD), which averages the phylogenetic distances between every pair of species; and mean nearest taxon distance (MNTD), which measures the distance to each species’ closest relative. Because raw phylogenetic diversity is strongly correlated with species richness, the team used null models that maintain species richness constant while randomizing phylogenetic relationships, generating 1,000 null communities per loma to produce standardized effect sizes. Negative standardized values indicate phylogenetic clustering—species more closely related than expected by chance—while positive values indicate overdispersion, where co-occurring species are more distantly related than random assembly would predict.

The results overturned several expectations. Plant species richness, adjusted for area through the residuals of a species-area power-law model, showed no significant association with latitude. Standardized MNTD likewise showed no significant latitudinal trend. In contrast, standardized phylogenetic diversity and standardized MPD both exhibited statistically significant non-linear relationships with latitude, peaking unexpectedly in the hyper-arid core of the Atacama Desert around 23°S. This finding is striking because the Atacama’s extreme aridity, thought to have intensified after the late Pliocene around 5 million years ago, would seem to be the least likely place for evolutionary diversity to reach its maximum.

Using generalized linear models with a suite of climatic, soil, topographic, and geographic predictors, the team found that aridity—the ratio of annual precipitation to potential evapotranspiration—was the only variable retained in every model, underscoring the central role of water availability in structuring these assemblages. Species richness increased with aridity and cloud cover, suggesting that lomas in highly arid regions can still support rich plant communities when advective fog delivers sufficient moisture. Standardized Faith’s phylogenetic diversity decreased with aridity and distance to the coast while increasing with human footprint and northness, the orientation of slopes toward the drier equatorial-facing side. Standardized MNTD increased with human footprint and northness but decreased with rising aridity and mean annual temperature. Standardized MPD increased with slope, temperature, northness, and aridity, while decreasing with cloud cover, isothermality, and distance from the ocean.

These patterns suggest that environmental filtering—the process by which harsh conditions select for species with particular trait values—is not the only macroecological process acting on plant diversity across coastal lomas. Instead, it is often counterbalanced at different points in evolutionary time by factors associated with environmental heterogeneity, possibly reflecting the presence of climate refugia and microhabitats that promote species coexistence. Small-scale topographic variability creates microclimatic conditions that buffer communities from the direct effects of aridity, while nurse plants mediate the climate that seedlings directly experience, facilitating coexistence in extreme environments. The negative relationship between standardized MNTD and mean annual temperature indicates that warmer desert ecosystems tend to be structured by recent radiations of closely related taxa, implying that environmental filtering due to aridification operates over longer evolutionary timescales.

The team’s cluster analysis of floristic similarity, using Hellinger distances and complete linkage hierarchical clustering, revealed that Peruvian and Chilean coastal lomas form separate assemblages organized into six distinct clusters. A major separation between northern Chilean and southern Peruvian clusters occurs near 18.21°S, coinciding with a climatic barrier long hypothesized to restrict plant dispersal. The division between northern and southern Peruvian lomas near 13.5°S aligns with the Abancay deflection, a prominent geomorphological feature in the central Andes that likely functions as a dispersal barrier. Two localities—Cerro Reque in Peru and Flamenco in Chile—emerged as independent branches, suggesting that site-specific environmental conditions or local evolutionary differentiation have produced singular floristic compositions.

The study also found that standardized MPD peaked near 22.5°S and 13.5°S and reached its lowest values near 17.5°S, positions that coincide with distinct geographic barriers including the hyper-arid core of the Atacama Desert and the Lima, Chillón, and Lurín hydrographic basins in central Peru. These barriers could have influenced species dispersal and divergence, producing lomas with mixtures of distantly related taxa. Meanwhile, the decline of standardized phylogenetic diversity and MPD with increasing distance from the coast reinforces the importance of coastal fog for maintaining species coexistence, as steeper coastal hills intercept more fog than flatter inland areas where the stratus layer dissipates.

The researchers acknowledge several limitations. At least 30 percent of coastal lomas lack published species checklists, reflecting collection bias that the team addressed by consulting herbaria, online checklists, and GBIF records. Most plant diversity data for coastal lomas is limited to conservation areas, potentially underestimating the impacts of human activity on unprotected sites. The lack of high-resolution climate data on fog and low clouds along the South American coast also constrains understanding of these factors’ roles, and the 1 square kilometer resolution used to estimate environmental conditions may introduce edge effects for small lomas where adjacent desert areas fall within the same grid cell.

Despite these constraints, the findings carry broad implications. Drylands host 20 percent of global plant diversity and provide critical ecosystem services to more than 38 percent of the world’s human population, yet their biodiversity is highly vulnerable to desertification and land-use change. Understanding the macroecological and macroevolutionary processes that generate and maintain biodiversity in the most arid ecosystems on Earth may therefore provide insights into the potential impacts of aridification on other ecosystems. For coastal lomas—home to endemic tree species like Myrcianthes ferreyrae and monospecific Tillandsia formations that evolved as recently as 2 million years ago—the study makes clear that adaptation and planning strategies must be developed in response to increasing aridification and land degradation if their unique biodiversity is to be conserved.

Subject of Research: Environmental and geographic predictors of seed plant diversity in South American coastal lomas

Article Title: Environmental and geographic predictors of seed plant diversity across South American terrestrial coastal lomas

Article References: Gonzales, F. N., Kreft, H., Arapa, L., & Craven, D. (2026). Environmental and geographic predictors of seed plant diversity across South American terrestrial coastal lomas. Frontiers of Biogeography, 19, Article e142560. https://doi.org/10.21425/fob.19.142560

Image Credits: AI Generated

DOI: 10.21425/fob.19.142560

Keywords: coastal lomas, Atacama Desert, phylogenetic diversity, aridity, biogeography, plant diversity, Peru, Chile, fog oases, macroecology, Environmental, geographic