Energy partitioning and water use efficiency anomalies 2018 at Eddy-Covariance sites across ecosystems.

Droughts and heat waves deeply interact with the exchange of energy and matterbetween land surface and atmosphere. The conditions associated with a combinedheat wave and drought can have positive or negative effects on the sensible heatflux (H), latent heat flux (LE), and net flux of CO2 (NEE). In return, while eachof these fluxes can exert different local feedbacks on temperature, atmospherichumidity and soil moisture, all of them result in bulk heating of the troposphere (Hand LE through direct local and indirect non-local heat transfer, and NEE throughradiative forcing). H is positively affected by increased solar irradiation, but can besupressed by advection of warm air diminishing the gradient between surface and airtemperature. Potential LE is positively affected by irradiation, temperature and watervapour pressure deficit, but actual LE can be suppressed by stomatal closure ofplants and by reduced soil moisture from low rainfall and high past LE. NEE is theresult of plant photosynthesis (GPP) and plant and soil respiration (R), both ofwhich can be enhanced by high temperatures or irradiation, and suppressed bydrought. Focusing on direct measurements mostly by a network of Eddy-Covariance (EC) stationsof the ICOS (www.icos-ri.eu), TERENO (www.tereno.net) and other networks, wehypothesize that the net effect of the 2018 event at a site, and thus its feedback on globalwarming, depends on the balance between co-existing positive and negative effects of thecombined heat wave and drought on the respective fluxes. Variables such as albedo, growingdegree days (GDD), soil moisture and ecosystem-level water use efficiency help to separatethese co-existing positive and negative effects from each other. Preliminary results indicatedifferent degrees of heterogeneity between sites for different variables. As expected, reducedprecipitation and soil moisture, as well as increased GDD, could be found at almost all sitesin the affected region. NEE was mostly less negative, indicating a weaker sink or evena source for CO2, as expected from past studies on earlier events. The networkdensity and number of site-years available now confirms that this was true for allmajor ecosystem types - forest, grassland and (rainfed) cropland, and resulted inincreased atmospheric CO2 concentrations. However, notable exceptions occurred atelevated low mountain range sites and during early stages of the event, supporting thehypothesis of a balance between positive and negative effects, where the former mayprevail at strongly energy-limited sites with a usually large water surplus. H wasmostly above-average, indicating that local heat production contributed to the eventand was typically not suppressed by warm air advection. LE as well as inferredGPP and R reacted most heterogeneously across sites, demonstrating the largediscrepancy between potentially high fluxes due to high irradiation and temperatureon the one hand, and suppression by water shortage on the other hand. Despitethe variability in LE, its relation to NEE was such that ecosystem-level water usewas less efficient than usual at the majority of sites through reduced CO2 uptake. [ABSTRACT FROM AUTHOR]