Quantifying the global environmental niche of an underutilised tropical fruit tree (Tamarindus indica) using herbarium records

The importance of identifying the environmental adaptation of underutilised species such as tamarind (Tamarindus indica L.) and their ecogeographic distribution is widely recognised. However the lack of physiological or quantitative yield/growth data does not permit the use of more traditional methods of crop modelling. In this study a representative dataset of tamarind's global distribution appropriate for modelling the species distribution is produced mainly from herbarium records; while minimising the effects of error and bias inherent in such data. The exploratory analysis of the clean dataset showed that both at the bioregional and sub-regional scale tamarind is experiencing varying conditions in different regions. This indicates the existence of spatial non-stationarity, which due to the broad distribution and semi-domesticated nature of the species, could be caused by evolution of the niche. Spatial niche variation in tamarind indicates that its range may have expanded from its naturalised area in Africa and India eastward into South East Asia and Latin America, progressively moving into regions with a less pronounced dry period and wetter conditions. Difference in environmental niche between East African and West African populations may be caused by the Rift Valley preventing gene flow between the two regions of the continent. Such information can be related to genetic variation and structuring within such species, providing information useful for conservation and selection of plant material for adaptation to future climate. The implication of spatial non-stationarity and spatial variation within the niche, when considering if global or regional/multi-scale models should be used to predict the world distribution of such species under current and future climate scenarios, are discussed.