Comparation of observation methods of sensible heat flux in Hulunbuir meadow steppe

Sensible heat flux is one of the most important parameters to represent the surface turbulent motion. Accurate and quantitative observation can greatly contribute to the patterns of water and heat balance within the region. It is also of great significance to climate observation, agricultural production, and water resources management. Eddy covariance (EC) and large aperture scintillometer (LAS) are the primary measurement techniques for the surface water and heat fluxes in the ecosystems at present. However, there are great differences in the spatial scales of the observation between the two techniques. Taking the Hulunbuir meadow steppe ecosystem as the research object, this study aims to clarify the variation characteristics of sensible heat flux observed by LAS and EC techniques. The contributing factors were also identified to determine the difference between sensible heat flux. A systematic analysis was finally made to quantitatively explore the influence of various factors on the difference between the two techniques. The results indicated that there was all the same trend of sensible heat flux observed by LAS and EC, which was predominantly driven by net radiation. There was also a positive correlation between the difference and net radiation at various temporal scales. The slope of 1.13 was found in the linear regression equation between the sensible heat flux observed by LAS and EC, indicating the better correlation (the goodness of fit was 0.81). The root mean square difference (RMSD) was 26.0 W/m², while the mean absolute percentage error (MAPE) was 26.9%. The difference fell within the range of −50 to 50 W/m² in the majority of cases, accounting for 94.2 % of all samples. The results showed that the LAS shared excellent applicability in the meadow steppe region. The energy balance closure of the EC system then posed some influence on the difference between sensible heat flux observed by LAS and EC. The difference between LAS and EC was reduced after the removal of the data with an energy balance ratio (EBR) less than 0.8. Therefore, the difference between LAS and EC tended to be smaller in the higher energy balance closure. The primary meteorological environmental factors were determined as the net radiation, wind speed (WS), Bowen ratio (Bowen), and saturated water vapor pressure difference (VPD), indicating the significant correlations with the difference between LAS and EC. The explanatory power of individual factors was ranked in the descending order of the net radiation, wind speed, EC energy balance closure ratio, saturated water vapor pressure difference, and Bowen ratio. Both net radiation and WS shared the greatest influence on the difference between LAS and EC, with an explanatory power of over 50%. There was a two-factor enhancement between any two distinct influencing factors, particularly with the interactive factors of net radiation/wind speed and wind speed/Bowen ratio, indicating the high explanatory power of over 70%. A more accurate understanding was gained for the observation of sensible heat flux at the spatial scale by LAS and EC. The findings can also provide a strong reference for the quality control of flux data during regional-scale extension and ground validation using remote sensing.