Partitioning of sensible and latent heat fluxes in different vegetation types and their spatiotemporal variations based on 203 FLUXNET sites

Bowen ratio reflects the partitioning between sensible and latent heat fluxes and plays a crucial role in land–atmosphere interaction. In this study, the spatiotemporal variations of Bowen ratio among 12 vegetation types were analyzed using observations from 203 FLUXNET sites worldwide and compared against Community Land Model simulations. Results showed that the annual mean Bowen ratio across all sites was 1.48 ± 1.20 (mean ± SD). Sites with Bowen ratios less than 1 (39%, 80 sites) were found across all continents, and the ones with a higher Bowen ratio (>3) (7%, 14 sites) appeared in dry and warm areas. Open shrublands showed the highest Bowen ratio (3.04 ± 0.58), whereas wetlands showed the lowest (0.74 ± 0.09). In terms of seasonality, Bowen ratio generally showed a U–curve with lower values in local summer and higher in spring and autumn in the northern hemisphere; the opposite occurred in the southern hemisphere. The spatiotemporal variations in Bowen ratio can be explained by climatic, geographical, and biological factors, with climate factors having the greatest impact. The spatial correlation analyses suggested Bowen ratio increased under higher VPD (R=0.45, P<0.001) and hotter (R=0.14, P<0.05) conditions with more shortwave radiation (R=0.39, P<0.001), and decreased with higher precipitation (R=−0.34, P<0.001), albedo (R=−0.16, P>0.05), and leaf area index (R=−0.25, P<0.001). Compared to FLUXNET observations, CLM well reproduced the global annual mean Bowen ratio (1.48 for CLM vs. 1.56 for FLUXNET), but showed larger differences for certain vegetations types such as open shrublands (−1.53), woody savannas (+1.17). Our results could enhance our understanding of biotic and environmental controls on land surface energy fluxes and help improve the land surface and climate models.