Machine Learning Global Ecological Niche Modelling of Indigofera
oblongifolia (Forssk.): A Palatable Desert Legume
Abstract
The goal of this study was to identify the global geographical
distribution patterns of a lesser known indigenous legume species,
Indigofera oblongifolia, using three bio-climatic timeframes (current,
2050, and 2070) and four greenhouse gas scenarios (RCPs 2.6, 4.5, 6.0,
and 8.5), as well as non-climatic predictors like global livestock
population, human modification of terrestrial ecosystem (GHMTE), global
fertilizers application (nitrogen and phosphorus). In addition, we
assess the degree of indigenousness using the area, habitat suitability
categories, and number of polygons, and we identify the temporal effects
of various bio-climatic variables on its fundamental and realized niche.
The AUC for models built using current climate data and RCPs for the
years 2050 and 2070 was 0.90. This research reveals that climatic
predictors outperform non-climatic predictors in terms of improving
model quality. Precipitation Seasonality is one of the most important
factors influencing this species’ optimum habitat suitability up to 150
mm for the current, 2050-RCP 8.5, and 2070-RCPs 2.6, 4.5, and 8.5. The
range of this parameter has altered from 79–176.9 to 85–196 as the
climatic conditions and RCPs have improved. Our ellipsoid niche
modelling extends the range of these bioclimatic variables to 637 mm and
26.5-31.80 degrees Celsius, respectively. India has a higher indigenous
score in the optimal class than the African region. These findings
indicate that this species inhabits more continuous areas in Africa,
whereas it is fragmented into a number of smaller meta-populations in
India (group of spatially separated population of the same species).