The Role of DYNAMO In-situ Observations in Improving NASA CERES-like Surface Atmosphere Radiation Estimation through Improving GMAO Reanalysis

Cloud-radiative feedback is one of the key processes for the MJO existence. The daily surface and atmospheric radiative fluxes from NASA CERES SYN1deg Lite Ed3A are one of the most commonly used data to assess surface energy budget and study observed cloud-radiative processes during MJO. The CERES SYN1deg data are, however, based on Fu-Liou radiative transfer calculations that use air temperature and specific humidity from NASA GMAO reanalysis as inputs, and are therefore subject to the quality of GMAO reanalysis used. Recently, DYNAMO in-situ observations over central tropical Indian Ocean (October 1, 2011-March 31, 2012) have been shown to considerably improve local air temperature and specific humidity, particularly their vertical profiles, in GMAO reanalysis. Notably, the assimilation of DYNAMO observations significantly moistens and warms model lower troposphere and upper troposphere, and dries and cools model middle troposphere, bringing the reanalysis closer to the observed. This motivates us to further assess the role of DYNAMO observations in improving CERES-like daily surface atmosphere radiation estimation through improving temperature and humidity in GMAO reanalysis.

Following the standard CERES SYN1deg Lite daily data production procedure, we perform two sets of Fu-Liou radiative transfer calculations for the DYNAMO period. They respectively use 3-hourly temperature and specific humidity from a 1-degree global GMAO reanalysis that assimilates global observations including DYNAMO observations and a parallel reanalysis that removes DYNAMO observations, with 3-hourly cloud and various surface fields taken from CERES SYN1deg_Lite Ed3A. The comparison of these results shows that the improvement of CERES-like surface atmosphere radiation estimation mainly occurs in longwave fluxes over central tropical Indian Ocean. Over these regions, the daily surface downward longwave estimation increases by about 5 Wm-2, which are primarily contributed by the warming and moistening in the lower troposphere; the daily top-of-atmosphere outgoing longwave radiation increases by 2-3 Wm-2 during dry periods only, which are mainly due to the drying in the middle troposphere. As a result, the estimated atmospheric longwave cooling over tropical Indian Ocean enhances by about 5 Wm-2 during deep convective periods and 7-8 Wm-2 during suppressed periods. The implications of these improvements of longwave estimation for the depiction of observed cloud-radiative processes during MJO initiation over tropical Indian Ocean will be discussed.