Abstract:
In NASA’s Modern-Era Retrospective analysis for Research and Applications, version 2 (MERRA-2) reanalysis the land is forced by replacing the modelgenerated precipitation with observed precipitation before it reaches the surface. We assess the MERRA-2land surface energy fluxes and 2 m air temperatures (T2m) and investigate whether the inserted precipitation observations result in the expected improvements to these fields. Averaged over global land, MERRA-2 overestimates the latent heat (LH) and incoming shortwave radiation (by 5-10 W/m2) and underestimates (in magnitude) the incoming and outgoing longwave radiation (by 10-15 W/m2), while the sensible heat biases are unclear.

The global average biases differ only slightly from those in NASA’s previous reanalysis, MERRA. In regions where LH is limited by surface moisture availability and hence sensitive to local precipitation, the MERRA-2 LH biases are reduced from those in MERRA, and the direction of change is consistent with the precipitation increments from inserting the observed precipitation. However, large positive biases remain in some regions where LH is not moisture-limited.

The globally averaged anomaly correlations (Ranom) with reference data were improved from MERRA toMERRA-2, for both LH (from 0.40 to 0.50 vs. GLEAM data) and the daily maximum T2m in Boreal summer (from 0.64 to 0.70, vs. CRU data). There is a strong correspondence between regions of the greatest improvement in the MERRA-2 T2m Ranom, and regions where the MERRA-2 T2m is most sensitive to the inserted observed precipitation, suggesting that the observed precipitation did in fact lead to improved T2m.