Monthly surface energy fluxes in ERA5-Land reanalysis: a comparison with FLUXNET observations.

Abstract

<p>We compare Earth's surface energy balance components (monthly mean radiation and energy fluxes) from the ERA5-Land reanalysis with their measured equivalents from FLUXNET stations in the period 2001-2018. The comparison (FLUXNET station vs. the nearest grid point of ERA5-Land) is focused on Central European sites with an elevation under 1000 m above sea level during the summer half of the year (April - September) that roughly corresponds to the growing season in Central Europe. FLUXNET stations are further divided into four groups according to biomes (and surface types) that they represent: grass, crop, forest or wetland. We put special emphasis on the turbulent heat fluxes: sensible (H) and latent heat (LE) heat flux and their relative share in a form of evaporative fraction (EF, a ratio of latent heat flux, LE, to the sum of both turbulent fluxes, H + LE). </p><p><br>There's a perfect fit between monthly FLUXNET observations and ERA5-Land data in case of some radiation fluxes: incoming shortwave (SWin) or outgoing longwave (LWout) radiation. Incoming longwave radiation (LWin) is underestimated by ERA5-Land, while the score of reflected shortwave radiation (SWout) is mixed. </p><p><br>Latent heat flux (LE) is overestimated by the reanalysis, while sensible heat flux (H) is lower in the reanalysis than in FLUXNET observations. This leads to overestimation of the evaporative fraction (EF) by ERA5-Land in the growing season, with monthly values of EF being mostly between 0.7 - 0.9. On the other hand, FLUXNET's EF values range from 0.2 to 1.0. When focused on individual biomes, the best fit between ERA5-Land and FLUXNET is seen for grass. The shape of "half-annual" (Apr - Sep) cycle is also well captured for grass locations. In other major biomes (cropland or forest) we see not only disagreement in terms of absolute values of H and LE fluxes (or EF) but in the shape of "half-annual" cycle too. There's no significant decline of EF in summer for cropland locations in ERA5-Land. The magnitude of a steady EF increase in forest sites of FLUXNET is weaker in ERA5-Land.</p><p><br>Acknowledgement: "SustES - Adaptation strategies for sustainable ecosystem services and food security under adverse environmental conditions (CZ.02.1.01/0.0/0.0/16_019/0000797)”.</p>