Modern pollen rain reveals differences across forests, open and mosaic landscapes in Madagascar

Palaeoecological investigations are crucial in understanding millennial to centennial land use and land cover change. By analysing the modern pollen rain from four main vegetation types across Madagascar, this research provides baselines for improving the interpretation of pollen records in palaeoecological studies. This study determines the modern pollen‐vegetation relationship in Madagascar and gives a better understanding of the island's landscapes. Despite some spatial limitations, this approach contributes to resolving the debated topic related to the origin of Madagascar's open ecosystems. Knowing the vegetation history prior to and after human settlement would help guide biodiversity management and its associated ecosystem services. Modern pollen rain contributes to calibrating and defining palaeo‐records relative to the current vegetation of diverse landscapes. Quantifying past vegetation change is important to better frame sustainable and resilient management methods of biodiversity considering human needs. In this paper, we aim to assess modern pollen assemblages in different vegetation types in Madagascar and inform how these could contribute to the interpretation of palaeo‐records. We identified modern pollen assemblages from 21 samples across four major vegetation types in Madagascar, including seven (N = 7) newly generated samples from the tropical dry forest in the Northwest region and 14 datasets downloaded from the African Pollen Database website. The downloaded datasets are from core tops representing modern vegetation previously analysed in the tropical dry forests and spiny thickets, the wooded grassland–bushland mosaic and the littoral humid forest. We evaluated the frequency of different taxa in the samples and conducted multivariate analyses to evaluate the similarities and ecological affinities between samples. Pollen rain reflected the main vegetation types across Madagascar and provided information on the current structure and state of the landscape. Regarding the interpretation of pollen records, pollen rain samples revealed that: (i) Open mosaic ecosystems should contain a percentage of at least 40% grass pollen, informing on the origin of the vegetation in older samples and the state of degradation in more recent vegetation reconstructions; and (ii) Anthropogenic vegetation such as plantations are marked by the abundance of pioneer taxa such as Pinus and/or Eucalyptus spp. with a frequency >20%. Our findings improve the interpretations of pollen records and help differentiate past distribution of forest, open and mosaic vegetation, in Madagascar. It will contribute to further research on human practices, land use and into scenarios of conservation planning not only in Madagascar but also across the tropics.