Aim: Colonisation of new areas may entail shifts in diversification rates linked to biogeographic movement (dispersification), which may involve niche evolution if species were not pre-adapted to the new environments. Scleria (Cyperaceae) includes c. 250 species and has a pantropical distribution suggesting an extraordinary capacity for long-distance dispersal and colonisation. We investigate patterns of diversification in Scleria, and whether they are coupled with colonisation events, climate niche shifts or both.

Location: Tropics and subtropics.

Taxon: Nutrushes Scleria (Cyperaceae).

Methods: We used molecular data from three DNA regions sequenced for 278 accessions representing 140 Scleria taxa (53% of species) to develop a chronogram, model ancestral ranges, and measure rates of diversification. Integrating data from 12,978 digitised and georeferenced herbarium records, we investigated niche evolution.

Results: High dispersal rates in Scleria, a genus with multiple dispersal syndromes, make reconstruction of ancestral areas at deep nodes in the phylogeny highly equivocal. Main dispersal and colonisation events involve movements from South to Central America (c. 19), from Africa to Madagascar (c. 12), from Asia to Oceania (c. 7), from Africa to South America (c. 7) and Central America to South America (c. 6). Two main shifts in diversification rates happened during the warm period of the Miocene.

Main conclusions: Dispersification from South America to Africa without climate niche shift seems to explain the diversification shift in section Hypoporum implying that species were pre-adapted. Shifts in climate niche evolution predate the second shift in diversification rates suggesting lineages were pre-adapted prior to biogeographic movements. Within subgenus Scleria, colonisations of Asia and Madagascar by sections Elatae and Abortivae, respectively, are coupled with niche shifts suggesting that these colonisations involved climate niche adaptation.

DNA sequence data of three markers (ITS, ndhF, rps16) generated in previous studies (Bauters et al., 2016, 2018; Galán Díaz, 2017; Semmouri et al., 2019) are used in this study. The sampling includes four species of tribe Bisboeckelereae, sister to tribe Sclerieae, and 140 accepted Scleria taxa (representing 53% of Scleria species). Approximately 48 of 112 (43%) American species (areas: South, Central and North America), 72 of 105 (69%) African species (Africa and Madagascar), and 20 of 58 (34%) Asian and Oceanian species (Eurasia and Oceania). Accessions sequenced per marker, ITS: 137, ndhF: 136, rps16: 135.

All georeferenced entries of Scleria available on the Global Biodiversity Information Facility (GBIF.org, 2018) were downloaded. GBIF is the largest repository of digitised occurrences information, however, it is necessary to apply certain filtering steps to minimize error in posterior analyses (Spalink et al., 2016). First, we eliminated all duplicate records and corrected the taxonomy following Bauters et al. (2016, 2018) and Galán Díaz et al. (2019). Second, for each species, all occurrences were plotted and points falling outside its known range were manually excluded. Third, because spatial clustering as a result of sampling bias can influence climatic niche analyses, we randomly retained one point per species and per cell of a 2.5-minute spatial resolution raster (about 4.5 km at the equator). Fourth, points were eliminated that were clearly outside the climatic range of the species. This was done by extracting the value of annual mean temperature and temperature annual range from WorldClim Global Climate Dataset v.2.0 (Fick & Hijmans, 2017) for every point. Then, for every species and climatic variable, we retained points that were within 1.5 times the interquartile range. For steps three and four, cell-size and climatic threshold were established after several trials in order to retain as many points as possible while eliminating aberrant observations. Finally, the database was supplemented with records for collections not yet available on GBIF from herbaria such as BR, K, GENT, L, MO, NY, P, US and WAG (Thiers, continuously updated), which were georeferenced using Google Earth. A total of 12,978 records were used for posterior analyses. The average number of records per species is c. 93 (1–1306; Table S3).

ITS: Larridon_Scleria_biogeography_ITS.fas

ndhF & rps16: Larridon_Scleria_biogeography_plastid.fas

Occurrence data: all_observations_phylogeny.txt

GBIF occurrence data download reference: GBIF.org (November 2018) GBIF Occurrence Download: https://doi.org/10.15468/dl.nzlqqq