Abstract. Low energy balance closure (EBC) at a particular eddy-covariance flux site increased the uncertainties of carbon, water and energy measurements and thus hampered the urgent research of scaling up and modeling analysis through site combinations. A series of manipulative experiments were conducted in this study to explore the role of net radiation (R_n) in the EBC in relation to spatial variability of vegetation characteristics, source area, sensor type, and dome condition in the Inner Mongolian grassland of Northern China. At all three sites, the daytime peak residual fluxes of EBC were consistently about 100 W m⁻² regardless of radiometers (i.e., REBS Q7.1 or CNR1). The spatial variability in net radiation was 19 W m⁻² (5% of R_n) during the day and 7 W m⁻² (16%) at night, with an average of 13 W m⁻² (11%) from eight plot measurements across the three sites. Net radiation results were affected more by measurement source area in unclipped heterogeneous system than in clipped homogeneous vegetation. Large area measurement significantly (P<0.0001) increased by 9 W m⁻² during the day and decreased by 4 W m⁻² at night in unclipped treatments. With an increase in clipping intensity, net radiation decreased by 25 W m⁻² (6% of R_n) at midday and 81 MJ m⁻² (6%) during a growing season with heavier regular clipping than that in unclipped treatments. Additional effort in EBC between 9:00 and 15:00 LT is needed for future research because of high variation. Using this method, the EBC difference derived from the two types of net radiometers was only 6 W m⁻². Results from Q7.1 with new domes were higher during the day but lower at night than those with used domes. Overall, the inclusion of the uncertainty in available energy accounted for 60% of the 100 W m⁻² shortfalls in the lack of closure. Clearly, the unclosed energy balance at these three grassland sites remains significant, with unexplored mechanisms for future research.