This is an inferential statistics question. I have attached a directed acyclic graph $-$ a causal diagram showing variables that affect CORN YIELD $($our response variable$).$ The question is to explore the impact of environmental and management factors on the yield of the five corn varieties. I know there are many linear regression models we could use.

Can you suggest:

$-$ A structured approach in selecting the models,

$-3$ models that I could potentially use, with or without interaction terms $($and the full reasons for choosing them?

GDP Country Rainfall Irrigation Variety Fertiliser_type) Temperature Yield Herbicide Soil_pH