Plant Adaptation in a Changing Climate: From Molecular Mechanisms to Ecosystem Resilience

Compound climate extremes expose plants to environmental constraints that overlap, follow one another, or interact across time, rather than acting as isolated stresses. This critical narrative review evaluates how plants integrate such compound challenges from molecular signalling to crop performance, with emphasis on heat-drought interactions, multifactorial stress combinations, stress memory, systemic signalling, reproductive resilience, and translation to climate-resilient agriculture. Literature was selected through transparent searches of accessible scholarly sources for work published from 1 January 2000 to 15 July 2026, with older conceptual material considered only where necessary. The evidence shows consistently that responses to combined stresses cannot be inferred by adding responses to the component stresses. Early sensing converges on redox, calcium, electrical, hydraulic, hormonal and metabolic networks, yet the resulting state is highly dependent on stress identity, sequence, intensity, tissue and developmental stage. Abscisic acid, jasmonate, reactive oxygen species, heat-shock systems and growth-control pathways function as recurrent hubs, but no single pathway explains resilience across combinations. Stress memory can persist through transcriptional, chromatin and proteostatic mechanisms, particularly after heat or dehydration priming, although direct evidence that these mechanisms improve tolerance to realistic compound extremes remains limited and is concentrated in Arabidopsis. Across biological scales, reproductive organs, source-sink relations and whole-plant systemic communication emerge as critical bottlenecks that are poorly represented by seedling assays. Crop and field studies confirm strong genotype-by-environment dependence and show that tolerance to one stress does not guarantee tolerance to a combination. A productive research agenda therefore requires experiments that reproduce event sequence and recovery, connect mechanistic states to reproductive yield, quantify the costs of memory and defence, and validate network-informed traits across environments. Compound-stress resilience should be treated as an emergent, dynamic phenotype rather than a collection of single-stress tolerances.