Daytime-only-mean data can enhance our understanding of land-atmosphere coupling

The major concern of land-atmosphere interactions (L-A) is the evolutionary process between the land surface and the planet boundary layer during the daytime, however many relevant studies had to use entire-day-mean daily time series to perform investigation due to lack of sub-daily data. Yet it is unclear whether the inclusion of nighttime data would alter the results or obscure the L-A interactive processes. To address this question, we generated daytime-only-mean (D) and entire-day-mean (E) daily data based on the ERA5 (5th ECMWF reanalysis) hourly product, and evaluated the strength of L-A coupling through a two-legged metrics, which assessed the coupling strength by the causality as well as the impact magnitude through two segments (land-fluxes and fluxes-atmosphere). The results demonstrated significant differences between the D- and E-based diagnoses as large as 67% (median 20.7%), which strongly depended on the season and the region. More importantly, for the first time, two special L-A coupling mechanisms were revealed. One was the advection-dominant L-A mechanism in tropical hyper-arid regions. The other was the soil moisture and sensible heat flux coupling mechanism during the cooling process over the nighttime. Both processes may play important roles during the night, andweaken the signal of L-A coupling if E was applied. To improve our knowledge of L-A interactions, we call attention to the urgent need for more high frequency data for relevant diagnoses. Meanwhile, we propose two approaches to resolve the dilemma of huge storage for high frequency data: (1) integration of L-A metrics in Earth System Model outputs, and (2) production of daily datasets based on different averaging algorithms.