Land cover controls the land-atmosphere exchange of water and energy through the partitioning of solar energy into latent and sensible heat. Observations over all land cover types at the regional scale are required to study these turbulent flux dynamics over a landscape. Here, we aim to study how the control of daily and midday latent and sensible heat fluxes over different land cover types is distributed along three axes: energy availability, water availability and exchange efficiency. To this end, observations from 19 eddy covariance flux tower sites in the Netherlands, covering six different land cover types located within the same climatic zone, were used in a regression analysis to explain the observed dynamics and find the principle drivers. The resulting relative position of these sites along the three axes suggests that land cover partly explains the variance of daily and midday turbulent fluxes. We found that evaporation dynamics from grassland, peatland swamp and cropland sites could mostly be explained by energy availability. Forest evaporation can mainly be explained by water availability, urban evaporation by water availability and exchange efficiency, and open water evaporation can almost entirely be explained by exchange efficiency. We found that the sensible heat flux is less sensitive to land cover type. This demonstrates that the land-atmosphere interface plays an active role in the shedding of sensible heat. Our results contribute to a better understanding of the dynamics of evaporation over different land cover types and may help to optimize, and potentially simplify, models to predict evaporation.

We study the controls on open water evaporation of a large lowland reservoir in the Netherlands. To this end, we analyse the dynamics of open water evaporation at two locations, Stavoren and Trintelhaven, at the border of Lake IJssel (1100ĝ€¯km2); eddy covariance systems were installed at these locations during the summer seasons of 2019 and 2020. These measurements were used to develop data-driven models for both locations. Such a statistical model is a clean and simple approach that can provide a direct indication of (and insight into) the most relevant input parameters involved in explaining the variance in open water evaporation, without making a priori assumptions regarding the process itself. We found that a combination of wind speed and the vertical vapour pressure gradient can explain most of the variability in observed hourly open water evaporation. This is in agreement with Dalton's model, which is a well-established model often used in oceanographic studies for calculating open water evaporation. Validation of the data-driven models demonstrates that a simple model using only two variables yields satisfactory results at Stavoren, with R2 values of 0.84 and 0.78 for hourly and daily data respectively. However, the validation results for Trintelhaven fall short, with R2 values of 0.67 and 0.65 for hourly and daily data respectively. Validation of the simple models that only use routinely measured meteorological variables shows adequate performance at hourly (R2Combining double low line0.78 at Stavoren and R2Combining double low line0.51 at Trintelhaven) and daily (R2Combining double low line0.82 at Stavoren and R2Combining double low line0.87 at Trintelhaven) timescales. These results for the summer periods show that open water evaporation is not directly coupled to global radiation at the hourly or daily timescale. Rather a combination of wind speed and vertical gradient of vapour pressure is the main driver at these timescales. We would like to stress the importance of including the correct drivers of open water evaporation in the parametrization in hydrological models in order to adequately represent the role of evaporation in the surface-Atmosphere coupling of inland waterbodies.