Modelling the latent heat transport through secondary circulations

Abstract

More frequent and more severe droughts and heat waves in the context of climate change are increasingly affecting the ability of ecosystems to meet their water demand. Quantifying the water demand of different ecosystems is therefore a fundamental step in developing management strategies to maintain intact ecosystems that function as CO2 sinks. The evapotranspiration of an ecosystem can be determined as latent heat flux using the eddy covariance (EC) method. However, a commonly found gap in the energy balance indicates that the atmospheric transport of sensible and latent heat is underestimated by single-tower EC measurements. One reason for this systematic error is transport through mesoscale secondary circulations, which inherently cannot be captured by such measurements. This transport fraction, so-called dispersive fluxes, is particularly large over heterogeneous surfaces. We present a novel model that is able to predict this dispersive flux of latent heat, thereby taking into account the effect of thermal surface heterogeneity. This model has been developed by combining a machine-learning approach with a large set of idealized large-eddy simulations covering different surface-heterogeneity scales and stability regimes. We further evaluate how the model can be applied to 30-minute EC measurements without additional instrumentation at the example of the CHEESEHEAD19 field campaign. An initial application to these real-world measurements together with realistic concurrent large-eddy simulations indicate a good agreement.