Search for a command to run...
Against the backdrop of accelerated terrestrial hydrological cycling and the increasing concurrence of drought-heatwave compound extremes under global warming, regional land-atmosphere coupling has emerged as a central mechanism shaping climate feedbacks and trajectories of ecosystem carbon uptake. However, prior studies spanning climatic regimes, observational scales, and data sources have often yielded contradictory conclusions. Here, we challenge these fragmented perspectives by constructing an integrated LST-SM-ET-GPP chain that jointly represents land surface temperature, soil moisture, evapotranspiration, and gross primary productivity, thereby linking water availability, surface energy balance, and plant physiological processes within a unified framework. We synthesize a conceptual diagnostic roadmap for interpreting land-atmosphere coupling across observations and models. When ecosystems operate in humid, energy-limited environments, radiative and advective controls should be prioritized to diagnose system forcing. By contrast, as the system becomes water-depleted, attribution must shift to a nonlinear regime transition framework governed by a critical soil moisture threshold. This threshold mechanism implies that, once the system enters the moisture-limited regime, even modest declines in soil moisture can trigger a rapid weakening of evaporative cooling, substantially amplifying LST anomalies and strongly suppressing GPP. The competitive regulation of stomatal conductance by atmospheric demand (vapor pressure deficit, VPD) and terrestrial supply (rootzone soil moisture) further explains why the “dominant” controlling factor can dynamically reverse across hydrothermal states, timescales, and stages of extreme-event evolution. Notably, the steady-state coupling assumption may break down under flux “flooring” during extreme drought, or when structural buffering such as deep root water uptake is present, delineating strict applicability bounds for existing diagnostic frameworks. Finally, current assessments remain constrained by multiple uncertainties, particularly the lack of ET partitioning constraints, representativeness biases arising from clear-sky observations and sampling-depth limitations, and systematic errors in Earth system model simulations during the warm season.