Introduction
Macroscale changes in landscapes, such as the creation of islands or movement of continents, as well as the stability of habitats and climate over various intervals of time, produce the template upon which assembly processes play out (Currie and Paquin 1987, Brown et al. 1996, Ricklefs 2004, Mittelbach et al. 2007, Morley 2018). Investigation of the drivers of present day species’ distributions must therefore account for both contemporary and historical processes, especially because climate change is affecting the distributions of diversity on earth, including along elevational gradients (Rahbek et al. 2019, Feeley et al. 2020). However, evaluating climate change’s effects on species’ elevational distributions requires quantification of underlying factors that determine how richness is distributed or why it is restricted, with consideration of biogeographic history (McCain and Grytnes 2010, Rehm and Feeley 2015, Pálinkás 2018, Freeman et al. 2018).
The distribution of diversity across a landscape is a function of range-sizes of the species present (Rapoport 1982). Species’ elevational distributions often exhibit asymmetry of average range-sizes for communities at either end of the gradient (e.g. smaller elevational range-sizes at lowland habitats as compared to montane areas), which is commonly referred to as Rapoport’s rule (Rapoport 1982, Stevens 1992) or abbreviated as ERR (Zhou et al. 2019). Variation in range-sizes across environmental gradients has been attributed to extrinsic factors such as climate and paleoclimate, habitat stability, geographic area or landmass positioning, and metrics of species richness (Morueta-Holme et al. 2013, Sheth et al. 2020, Sundaram and Leslie 2021, Guo et al. 2022), as well as intrinsic factors such as species’ response to competition, dispersal ability, life-history strategies, edaphic specialization, and growth forms (Morin and Lechowicz 2013, Grossenbacher et al. 2015, Xu et al. 2018, Saupe et al. 2019, Whitman et al. 2021, Freeman et al. 2022). Stochastic processes or geometric constraints add complexity when differentiating biological versus non-biological causes of observed range-size patterns (Colwell and Hurtt 1994, Lyons and Willig 1997, Šizling et al. 2009). However, despite the influence of historical factors on diversity, relatively few studies have examined variation in ERR slopes with respect to higher-level taxonomy or biogeography.
Among plants, a study in Nepal found stronger ERR support for species with tropical compared to temperate origins, especially within sections of the gradient without natural barriers that can limit range-size potential (Feng et al. 2016). Within the tropics, a study on Mt Kenya found that variation in richness and ranges-sizes was attributed to rates of endemism and growth form (Zhou et al. 2019). A study on Mt. Kinabalu reinforced the complexity of species’ distributions on equatorial mountains, finding that stress-tolerant species had larger and relatively consistent range-sizes, whereas most tropical species had smaller range-sizes with pronounced variation in ERR slopes that coincided with vegetation zone boundaries (Whitman et al. 2021). Tests of macroecological hypotheses are most robust when there is sufficient replication of analyses, often encompassing different geographic areas or taxonomic groups (Ruggiero and Werenkraut 2007, Hawkins et al. 2011, McCain and Bracy Knight 2013, Sheth et al. 2020, Guo et al. 2022). Thus, equatorial habitats with high biodiversity offer a unique opportunity to test drivers of species range-sizes.
The goal of this study was to quantify variation in elevational range-sizes and distributions of major Malesian plant families, and to investigate ecological and biogeographical drivers of these patterns. Malesia is a biodiversity hotspot that encompasses a 2,000,000 km2 area and collectively represents the largest archipelago on Earth, straddling the equator and spanning from the Malay Peninsula in the west to New Guinea in the east (Myers et al. 2000). Malesia includes a vast elevational gradient with mountain summits that are subject to freezing events (Hope 1976). Notable Malesian mountains include Mt. Kinabalu (4,095 m a.s.l.), Mt. Wilhelm (4,509 m a.s.l.), and Puncak Jaya (4,884 m a.s.l.). Surprisingly little research on ERR has been conducted within Malesia, except for localized studies (Whitman et al. 2021). There is growing evidence that equatorial mountains represent a unique environment type, with distinct richness and altitudinal zonation patterns governed by more than just temperature (Ashton et al. 2022). For these reasons, Malesia offers an exceptional study system to investigate potential correlates of variation in ERR slopes.
Delineation of biogeographic boundaries for plants is a point of ongoing investigation (van Welzen et al. 2011, Webb and Ree 2012, Kooyman et al. 2019, Joyce et al. 2021). Biogeographic boundaries within Malesia are often based on Wallace’s line, named after Alfred E. Wallace and his observations of islands with highly dissimilar terrestrial vertebrate composition (Richardson et al. 2012). However for plants, Malesia can be divided into three sections (i.e. Sundaland, Wallacea, and Sahul) in reference to geological history and major waterways (Richardson et al. 2012) as well as consideration of differences in species’ response to environmental characteristics (van Welzen et al. 2011).
Millions of years ago (Mya), major landmasses of Malesia were positioned farther apart as compared to current day orientations (Hall 1998, Morley 2018). During the Eocene (~ 49 – 45 Mya), Sundaland and portions of Wallacea (mainly the Philippines) straddled the equator, whereas Sahul (New Guinea) was partially submerged and positioned ~20 degrees southwards in latitude (Hall 1998, Zhang et al. 2022). During this time period the Indian Plate collided with Asia, changing regional patterns of sediment flow and climate within SE Asia as well as introducing Gondwanan flora to the western portion of Malesia (Hall 1998, Morley 2018). The Australian shelf, which Sahul represents the leading edge of, broke apart from Antarctica and began drifting northwards and upwards (Hall 1998). During the Miocene (~ 15 Mya) the mountains of Borneo became wider and higher and the adjacent Philippine Sea Plate began rotating closer to mainland Asia, with volcanic activity creating new islands (Hall 1998). Sahul began rising higher above sea-level, while simultaneously increasing the area of lowland habitats. The Pliocene (~ 5 Mya) was marked by the collision of Sundaland and Australian plates, with landforms positioned in a similar orientation as today (Hall 1998).
Plant families of Malesia have established across the archipelago at different intervals of time based on a complex history of speciation, local extinctions, and migration (Webb and Ree 2012, Richardson et al. 2012, Ashton 2014, Kooyman et al. 2019, Xia et al. 2022), with most families originating elsewhere (Morley 2018). Plant clades of Sundaland have historically had tropical association, with greater interchange with the mainland Indochina (Kooyman et al. 2019, Zhang et al. 2022). Conversely, plant clades positioned eastwards on the Sahul continental shelf have greater Australasian associations, especially for montane flora specific to mountain ranges that have remained above sea-level for extended periods of time (Morley 2018, Brambach et al. 2020). Wallacea is centrally located, likely mediating a greater interchange with neighboring areas (Ashton 2014), and is comprised of many smaller, younger, or recently uplifted islands with relatively shorter mountains. The oldest Gondwanan clades can occur on both western and eastern sides of Malesia (Hall 1998, Richardson et al. 2012), as can taxa with long-distance or stochastic seed dispersal abilities (Webb and Ree 2012, Baker and Couvreur 2012).
Present day species’ distributions within Malesia are heavily influenced by the last 1-15 million years of tectonic events (Hall 1998, Ashton 2014, Morley 2018, Zhang et al. 2022). The recent interchange of species (< 5 Mya) is facilitated by landmasses that are positioned closer together and the emergence of islands which act as stepping-stones for range expansion (Webb and Ree 2012, Ashton 2014). Volcanos and mountaintops further act as “sky islands” to facilitate speciation and the interchange of montane flora restricted to cooler conditions (Webb and Ree 2012, Culmsee and Leuschner 2013, Morley 2018). Land bridges, which are ephemeral under changing sea-levels, act as pathways for land-animal dispersed species (Ashton 2014, Morley 2018). Climate stability further promotes species interchange, with asymmetrical migrations in favor of tropical Asian routes as compared to Australian pathways (Ashton 2014, Zhang et al. 2022). Variation in regional conditions, such as seasonally dry periods, coincided with glacial periods before widespread establishment of perhumid conditions towards the end of the Oligocene (Morley 2018). A lingering unknown is whether the contemporary species interchange may homogenize macroecological range-size patterns (e.g. Rapoport’s rule) between biogeographic areas, or if older artifacts of geographic separation and paleoclimatic conditions continue to shape species distributions.
In order to investigate potential mechanisms underlying variation in plant elevational range-sizes, we used a database of herbarium records for 60 of the most species-rich plant families in Malesia (representing a total of 1799 genera and 18535 species) to estimate species occurrence across a 4500 m elevational gradient and by biogeographic regions. For each family, we estimated the change in average elevational range-sizes with increasing elevation (ERR slope). We evaluated the correlation between family-level ERR slopes and 15 metrics that reflect the ecology, evolution, and biogeography of each family, including richness; evolutionary age; latitudinal or elevational extent; biogeographical distribution with respect to Sundaland, Sahul, and Wallacea; degree of endemism versus cosmopolitanism; and inferred habitat specialization versus generalization.
Our predictions are rooted in the concept that “mountain passes are higher in the tropics” (Janzen 1967). When applied to determinants of species’ elevational range-sizes (Stevens 1992), this statement means that a greater shift in species’ ecological niche space is required to occupy each incremental increase in elevation. Janzen’s statement rests on at least two assumptions. First, evolutionary adaptations to reduced environmental variation promotes specialization and a reduction of species’ physiological tolerances, as well as an increase in interspecific biotic competition that leads to narrower niches, both of which would be associated with smaller elevational range-sizes. Second, evolutionary adaptations to variation in growing conditions, from factors such as higher diurnal temperature fluctuations or seasonality, would result in larger range-sizes, similar to the climate variability hypothesis (Dobzhansky 1950).
As applied to families that evolved within biogeographic regions with longer-term climate stability, based on proximity to the equator over millions of years, we would predict more positive ERR slopes. However, for taxa on landmasses that moved towards the equator, or that were newly created or uplifted, we predicted weakly positive, or neutral, ERR slopes, with the same prediction for evolutionarily younger families, recent colonizers, or those with montane or temperate origins. We anticipated that the way richness is spatially distributed would also affect these predictions. For families with more restricted distributions (e.g., narrower familial extent; concentrated centers of richness; higher endemism; adjacency to physical boundaries such as sea-level) we predicted steeper positive ERR slopes based on a combination of ecological factors and geometric constraints. For a subset of communities nearest mountain summits, we would predict negative ERR slopes. For widespread families with fewer barriers for expansion (e.g., cosmopolitanism; dispersed richness, broad familial extent) we predicted shallower or neutral ERR slopes.