Predictive ecological niche modelling of an important bio-control agent: Trichoderma harzianum (Rifai) using the MaxEnt machine learning tools with climatic and non-climatic predictors

Ecological niche model (ENM) pertains to a class of methodologies that utilise occurrence data alongside environmental data to formulate a correlative model of the environmental circumstances that satisfy a species' ecological requirements. In the current study, ENM was employed to ascertain the types of habitat for Trichoderma harzianum using machine learning algorithm known as MaxEnt Entropy. Our line of reasoning posits that the efficacy of T. harzianum as a bio-control agent can be enhanced, alongside the advancement of host/crop development and metabolic processes, through its deliberate introduction into geographically appropriate habitats. ENM was performed on 92 spatially thinned presence points of this species across India, considering three bio-climatic time periods (present, 2050, and 2070) and four greenhouse gas scenarios (known as representative concentration pathways RCPs). Non-bioclimatic factors include ecosystem rooting depths (ERD), total plant available water storage capacity (TPAWSC), habitat heterogeneity indices (HHI), land use land cover (LULC) and to soil variables at four depths. Energy-related factors, like Isothermality and minimum temperature of coldest month, were shown to be the most essential for the habitat appropriateness of this species during the current bio-climatic period. Future climate predictions and their associated RCPs revealed that water-related variables, like precipitation of wettest quarter, were the most influential. Non-climatic elements that were shown to have significant impact included soil pH, maximum diversity indices, forest and grassland types, TPAWSC, ERD (95%). Our analysis showed that this species will always find optimal suitability sites in northern eastern India with almost all predictors except root zone variables.