How biodiversity is structured globally and which factors determine such patterns are long-standing issues in ecology. Latitudinal diversity patterns are well known for plants and several groups of vertebrates, including amphibians, reptiles, birds, and mammals (Hughes et al., 2021; Jenkins et al., 2013; Kreft and Jetz, 2007; Torres-Romero and Olalla-Tárraga, 2015). For invertebrates, global biodiversity patterns are much more scarce due to their overall richness and strong sampling limitations and biases. This knowledge gap, however, is closing for some insect groups, such as dragonflies, bees, butterflies, and ants (Economo et al., 2018; Hughes et al., 2021; Kass et al., 2022; Orr et al., 2021). Describing global species richness distribution patterns is crucial for biodiversity conservation planning (Mittermeier et al., 2004) and actions, such as delimiting conservation areas (Orr et al., 2021). Besides, understanding how biodiversity patterns are structured by abiotic factors allows us to predict how they will be affected by climate change (Komatsu et al., 2019; Orr et al., 2021).

Global biodiversity patterns are strongly determined by the abiotic environment (Hawkins et al., 2003; Ilsøe et al., 2017). Frequently, annual rainfall and mean temperature have emerged as the best predictors of global biodiversity patterns for several taxa, for example, mammals, amphibians, birds, and plants (Hughes et al., 2021; Kreft and Jetz, 2007; Torres-Romero and Olalla-Tárraga, 2015), as higher water availability and higher temperatures constitute favorable environmental conditions for many species in many places (Hawkins et al., 2003). Other abiotic factors, such as actual or potential evapotranspiration, also emerged as good predictors of biodiversity (Kaspari et al., 2000). Besides, factors promoting environmental heterogeneity, such as topographic factors (e.g., mean altitude and topographic heterogeneity), are also recognized as drivers of biodiversity on a global or regional scale (Torres-Romero and Olalla-Tárraga, 2015). In short, heterogeneous environments enable a wider range of species to be included in local species reserves (Stein et al., 2014).

Global variation in biodiversity is also influenced by historical factors such as geographic isolation and limited dispersal. Furthermore, climatic and topographic conditions are distinct between regions providing specific ecological and evolutionary contexts (Holt et al., 2013). Thus, generalizations of global patterns are risky because they ignore regional environmental variation, so the importance of approaches that include analyses at global and regional spatial scales grows (e.g., Torres-Romero and Olalla-Tárraga, 2015; Ilsøe et al., 2017; Lion et al., 2019). Investigations from bioregions, such as the zoogeographic realms proposed by Holt et al. (2013), can help us advance our understanding of the historical and ecological determinants of diversity patterns, given that zoogeographic realms have intrinsic differences related to the history of their phylogenetic communities.

Ants are social insects that are one of the most abundant organisms in terrestrial ecosystems (Dunn et al., 2009). The twenty-two thousand ant species estimated in the world dominate basically every inch of land and canopy, except when excluded by abiotic extremes. They are ferocious predators influencing the trophic web of natural and anthropogenic ecosystems, being able of altering the physical and chemical properties of the soils (Gomes et al., 2010). Also, ants are engaged in relevant mutualistic interactions with plants, including their defense role against herbivores, seed dispersal, and pollinators (Oliveira and Koptur, 2017). Also, they perform important ecosystem services, such as pest control. Recently, an effort was made to make available the world ant records on public platforms, allowing us to advance several aspects of ant ecology, evolution, and biogeography (Guénard et al., 2017). It is now apparent that the distribution of ant diversity resembles that of vertebrates, such as reptiles and mammals (Kass et al., 2022). The equatorial region exhibits higher ant biodiversity supported by a larger number of tropical ant lineages and a longer diversification time (Economo et al., 2018). However, marked differences are noted between the distribution of ant diversity and other taxa, such as low species richness in the Afrotropical realms (Gueńard et al., 2012; Kass et al., 2022). As in other taxon, abiotic conditions are important determinants of ant diversity (Costa et al., 2017). However, the global and regional determinants of ant diversity still seem uncertain. Our knowledge is limited to a few studies of communities at the global level with 50 sampling points worldwide (Dunn et al., 2009; Gibb et al., 2015), limited to regions (Bishop et al., 2014), or at the genus level (Gueńard et al., 2012). The effect of topographic conditions on ant species richness is still uncertain, with local and regional studies indicating positive, negative, and neutral effects of altitude (Bishop et al., 2014). Thus, the aim of our work was (1) to test the relative strength of abiotic and topographic gradients in determining global diversity patterns of ants, and (2) to test the generality of the importance of these factors across zoogeographic realms.

Climatic and topographic conditions are important determinants of global biodiversity patterns (Hawkins et al., 2003; Buckley and Jetz, 2007; Ilsøe et al., 2017). Our findings point to mean temperature, annual rainfall and mean altitude as the main drivers of ant biodiversity worldwide. However, the influences of abiotic factors are complex, and although we detect general patterns, differences between zoogeographic realms reveal that biodiversity is structured differently given local and regional ecological conditions.

The spatial patterns presented here are similar to global patterns of species richness of terrestrial vertebrates (Buckley and Jetz, 2007; Torres-Romero and Olalla-Tárraga, 2015), dragonflies (Hughes et al., 2021), and vascular plants (Kreft and Jetz, 2007), with the hot zones of diversity concentrated especially in the equatorial region, mostly in the Neotropical, Panamanian, Afrotropical, and Oriental realms. Plants secure resources, and shelter, and locally alter the microclimate, favoring ants (Oliveira and Koptur, 2017), so it could be expected that the general patterns of ants are partly similar to those of vascular plants. Our findings are partly at odds with global patterns of bee species richness (Orr et al., 2021), as the southern regions of the Neotropical and Australian realms show high numbers of bee species, whereas ants do not appear to be as diverse in these regions. This discrepancy highlights the importance of revealing the global patterns of other invertebrate taxa. Our results largely agree with other studies of ants on a global scale, indicating the tropical region as the main center of diversity (Dunn et al., 2009; Gibb et al., 2015; Kass et al., 2022). Here, however, we point to the Afrotropical realm as a center of ant diversity, different from that described in Kass et al. (2022).

Water availability is a strong factor affecting diversity patterns globally (Hawkins et al., 2003). Annual rainfall has previously been pointed out as a key driver of ant diversity at local, regional and global scales (Gibb et al., 2019, 2015; Vasconcelos et al., 2018). In our global analysis, annual rainfall emerged as an important driver of ant species richness at the global scale, being also relevant for several realms (i.e., Panamanian, Sino-Japanese, Afrotropical, Saharo-Arabian, Neotropical and Madagascar). Ants benefit from annual rainfall through different pathways. Areas with higher rainfall develop forests with higher, continuous, and multilayer canopy, providing additional foraging ground, shelter, and nesting sites for the ant colonies (Yanoviak and Kaspari, 2000). Also, higher rainfall is normally associated with a higher diversity of plants and, consequently, to a higher diversity of their associated arthropods, many of them used as prey by generalist and specialized ants. Under adequate amounts of nutrients, higher rainfall also leads to higher productivity, which is translated into a higher availability of food resources (Queiroz et al., 2013). Globally, higher mean rainfall was positively correlated to the rainfall of the driest month (r = 0.64; Fig. S3), which indicate a higher consistency of conditions and resources over time (Arnan et al., 2006; Gibb et al., 2015). Consistently, rainfall seasonality had a negative effect on ant estimated species richness in global scale, and in the Nearctic and in Afrotropical realms.

World regions with high temperatures host a large proportion of the global biodiversity, this being frequently associated to the higher diversity of plants and higher productivity, providing resources to other organisms in a positive trophic cascade (Kaspari et al., 2000; Liang et al., 2016). Since most ants are thermophilic, regions with higher mean temperatures are expected to lead, in general, to higher local species diversity (Dunn et al., 2009; Ilsøe et al., 2017). In addition, hottest regions, especially near the equator, harbor a greater number of ant lineages, which have had a longer time to diversify (Economo et al., 2018), and therefore harbor a greater number of species today. Our results indicate that mean temperature is a major determinant of estimated ant species richness on a global scale, as has been suggested by other large scale studies (Dunn et al., 2009; Gibb et al., 2015). Increasing temperature favors ants by allowing them to forage more and harvest more energy (Yanoviak and Kaspari, 2000). In colder regions such as Nearctic, increasing temperature it decreases individual mortality, thus preventing colony decline and reducing local extinction rates, especially by nest freezing (Sanders et al., 2007). However, in hot regions, such as the Madagascar realm, species richness decreases with mean temperature. In this realm, the mean temperature value is strongly correlated with the maximum temperature of the warmest month (r = 0.78) and the minimum temperature of the coldest month (r = 0.78) (Fig. S3), which should limit the local establishment of many species, since the maximum temperature values are probably above their ecophysiological limits.

Actual evapotranspiration has been suggested as a measure of ecosystem productivity (Rosenzweig, 1968), an important driver of diversity as has been demonstrated globally for mammals (Torres-Romero and Olalla-Tárraga, 2015) and plants (Kreft and Jetz, 2007) and, regionally, for ants (Kaspari et al., 2000). However, the global model did not support actual evapotranspiration as an important factor in determining ant diversity patterns at global scale, in contrast to what it could be expected by the energy-water hypothesis (Kaspari et al., 2000). Indeed, in contrast to what it could be expected, we found negative effects of actual evapotranspiration in five realms, Afrotropical, Madagascar, Nearctic, Palearctic and Panamanian, while we found a strong positive effect only in the Australian realm.

The effects of altitude on ant diversity patterns have been extensively investigated at various scales (Bishop et al., 2014; Longino et al., 2019; Sanders et al., 2007), but their effects are still uncertain. A number of studies indicate that species richness of ant peaks in areas of intermediate elevation, at about 500 m, and then diversity declines to near absent species in areas of high elevation (Longino et al., 2019). However, other studies report a decline in species richness with increasing average altitude, driven by decreasing temperature (Sanders et al., 2007). On a global scale, we found that increasing mean altitude elevates estimated species richness, with this effect emerging in the Australian, Neotropical, Nearctic, and Palearctic realms. Higher altitude sites have been relatively less affected by anthropogenic disturbance and are therefore better conserved (Lewis et al., 2015). In this context, the effects of altitude on ant species richness may be linked to local conditions (e.g., conservation status, soil type, or vegetation cover). In addition, the mean altitude may represent to some extent the heterogeneity of the environment, since at our grain (2 degrees), it may harbor a range of heterogeneous habitats at different levels of mean altitude, making it possible for a larger set of species to be included in a local pool. Congruently, our findings indicate that topographic heterogeneity elevates the estimated species richness in global scale and, Afrotropical and Madagascar realms. Previously, topographic heterogeneity (environmental) has been noted as a global driver of biodiversity (Stein et al., 2014), this is because, greater heterogeneity ensures that more species meet their niche needs.

Abiotic factors have a significant influence on the structure of ant communities worldwide and across realms, as discussed in this and other studies (Costa et al., 2018; Dunn et al., 2009; Gibb et al., 2015). However, historical processes of diversification and extinction strongly influence current patterns of diversity (Holt et al., 2013). The niche conservatism hypothesis, offers a useful explanation for the distribution of current ant diversity (Economo et al., 2018). According to this hypothesis, climatic niches evolve slowly, and clades have higher diversity in their ancestral climatic niche (Wiens and Donoghue, 2004). Ants emerged approximately 170 million years ago and diversified over the Cretaceous (age of angiosperms) to the late Pleistocene when the global climate was hottest and more humid (Economo et al., 2018; Moreau et al., 2006). Thus, current patterns of ant diversity are the end result of niche conservatism, given that ants are ectothermic and thermophilic(Sanders et al., 2007), the transition to cooler climates at high latitudes could be a barrier, with a tendency to occur rarely, leading high latitude regions to have a smaller set of phylogenetically clustered species from younger lineages (Economo et al., 2018). In addition, the niche conservatism hypothesis may explain the peak diversities in higher-elevation areas, currently, lowlands near sea level (<500 m altitude) are excessively hot currently (compared to the past) for many ant species, which limits the local species assemblage (Longino et al., 2019; Longino and Branstetter, 2019).

Insects, including ants, are declining at an alarming rate (Fonseca, 2009), jeopardizing important ecosystem services, such as population and pest control, nutrient cycling and seed dispersal (Leal et al., 2014; Oliveira and Koptur, 2017). Therefore, conservation protocols aimed at conserving this group are becoming increasingly important. Currently, conservation goals are generally formulated at a global scale based on the assumption that regions that are diverse for vertebrates or plants are also diverse for other taxa (Mittermeier et al., 2004). However, this assumption is unsafe since patterns of biodiversity differs among taxa, scales, and ecological and historical context. This study found that the hot zones of ant diversity resemble that of vertebrates and vascular plants, as described previously (Gueńard et al., 2012; Kass et al., 2022), but important differences also emerged. Thus, further effort should be done to describe the geographical patterns of arthropods and other invertebrates which represent the bulk of the world biodiversity.

We are currently facing an unprecedent global climate change caused by human activities. It is estimated that global temperatures will increase, in average, at least two degrees Celsius by the end of this century and that this will lead to drastic regional spatio-temporal changes in temperature and precipitation (Magnan et al., 2021). The impacts are expected to disproportionately affect different regions of the world. For example, many regions of the Southern Hemisphere are expected to experience a sharp decline in annual precipitation by the end of the century (Pontes et al., 2020). This can be exacerbated by the high rates of deforestation in many tropical rainforest areas (d’Annunzio et al., 2015). Since ants are sensitive to natural and human-induced changes (Ribeiro-Neto et al., 2016; Vasconcelos et al., 2018) and our results demonstrated their sensitivity to abiotic gradients, we predict that ant diversity will be greatly affected by climate change. Our results not only reveal natural biogeographical patterns of ants but can serve as a basis for future global and regional conservation initiatives aiming to mitigate the effects of global change.

AD and CRF contributed to the study conception, sampling and analysis design. AD ran all statistical analyses and made all tables and figures. AD and CRF contributed to the writing up of all drafts of the manuscript.

The authors fully agree with the submission of the manuscript and have no conflicts of interest

All the data used here can be retrieved on publicly accessible digital platforms, antmaps (https://antmaps.org/about.html) and antweb (https://www.antweb.org/). The R project with the scripts will be available upon request to the corresponding author.