In the CSIR 2016 question, correct match of hypotheses to graph is:
The graph matching associates the rivet model with progressive loss of function, redundancy with an initial plateau, the idiosyncratic model with an irregular response, and the keystone model with an abrupt change after loss of an influential species. Those diagnostic curve shapes yield the stated mapping. Biodiversity–function hypotheses differ in how they assign ecological importance among species. Complementarity emphasizes distinct contributions, redundancy emphasizes overlap and compensation, rivet theory emphasizes cumulative weakening, keystone theory emphasizes particular influential species, and idiosyncrasy emphasizes context dependence. Real communities can display several patterns because functions differ and environmental conditions alter interactions. Richness alone therefore cannot identify mechanism; species traits, relative abundance, functional groups, and the order of species loss must also be considered. In this context, the keyed term, a-rivet, b-redundancy, c-idiosyncratic, d-keystone, identifies the relevant mechanism or quantitative relationship and links the observed pattern to its underlying ecological cause. The distinction is testable by measuring changes in organisms, resources, or process rates through time rather than relying on the label alone.
Ref: Ecology: Concepts and Applications, Molles, 9th Ed., Ch. 20