Evolution of land surface feedbacks on extreme heat: Adapting existing coupling metrics to a changing climate

Abstract

Episodes of extreme heat are increasing globally, and dry land surface states have been implicated as an amplifying factor in several recent heat waves. Metrics used to quantify land-heat coupling in the current climate, relating sensible heat fluxes to near-surface air temperature, are applied to multimodel simulations of the past, present, and future climate to investigate the evolving role of land–atmosphere feedbacks in cases of extreme heat. Two related metrics are used: one that describes the climatological state of land-heat coupling and one that gives an episodic estimate of land feedbacks, here defined as the metric’s value at the 90th percentile of monthly mean temperatures. To provide robust statistics, seasonal multimodel medians are calculated, with the significance of changes determined by the degree of model consensus on the sign of the change. The climatological land-heat coupling mirrors other metrics of land–atmosphere interaction, peaking in transition regions between arid and humid climates. Changes from preindustrial to recent historical conditions are dominated by decreased land surface controls on extreme heat, mainly over the broad areas that have experienced expanded or intensified agriculture over the last 150 years. Future projections for increased atmospheric CO2 concentrations show a waning of areas of weakened land-heat feedbacks, while areas of increasing feedbacks expand over monsoon regions and much of the midlatitudes. The episodic land-heat metric is based on anomalies, which creates a quandary: how should anomalies be defined in a nonstationary climate? When the episodic coupling is defined relative to the means and variances for each period, a broadly similar evolution to the climatological metric is found, with historically dominant decreases giving way to widespread moderate increases in future climate scenarios. Basing all statistics on preindustrial norms results in huge increases in the coupling metric, showing its sensitivity to the definition of anomalies. When the metric is reformulated to isolate the impact of changing land and temperature variability, the tropics and Western Europe emerge as regions with enhanced land feedbacks on heatwaves, while desert areas and much of the remainder of the midlatitudes show reduced land-heat coupling.