Small Satellite Humidity Sounding for Global NWP: First Evaluation of CAMS on TY-16 in the CMA-GFS 4D-Var System

Polar-orbiting microwave humidity sounders, with 6-h revisit intervals, cannot resolve rapidly evolving moisture features in the lower troposphere. Cost-effective small-satellite constellations promise shorter revisit times, yet their impact on global numerical weather prediction (NWP) using actual observations has not been quantified. From 25 March to 25 April 2024, clear-sky radiances from the Compact Atmospheric Microwave Sounder (CAMS) aboard the 500-km small satellite Tianyan-16 (TY-16; launched December 2023) were, for the first time, assimilated into the China Meteorological Administration Global Forecast System (CMA-GFS) global four-dimensional variational (4D-Var) system. Four cycling experiments were carried out: a control experiment (CTRL) assimilating conventional and operational satellite observations [excluding radiances from the Microwave Humidity Sounder-2 (MWHS-2) aboard Fengyun-3E (FY-3E)]; CTRL plus CAMS radiances (EXP1); CTRL plus MWHS-2 radiances (EXP2) for benchmark comparison; and CTRL plus both CAMS and MWHS-2 radiances (EXP3). Offline bias correction was applied to 12 out of 18 channels (7 for temperature and 5 for humidity) against CMA-GFS model backgrounds. Compared to CTRL, EXP1 reduces the standard deviation of background departures for humidity-sensitive channels of independent sounders by 0.5%–1.1% (95% confidence interval), peaking at 1.1% for FY-3D MWHS-2 channel 15. The root mean square error (RMSE) of 850-hPa specific humidity analyses decreases by 4.2%, 2.4%, and 1.1% over the Southern Hemisphere, tropics, and Northern Hemisphere, respectively; the RMSE of tropical temperature at 850 hPa falls by 1.8%. CAMS alone yields a 4%–6% reduction in RMSE for 1–3-day moisture forecasts below 700 hPawithin 20°S–20°N. Joint assimilation with MWHS-2 extends moisture improvements over the 10-day forecast and significantly reduces meridional wind errors through Day 5. These results from real satellite data confirm that CAMS on a small satellite can rival the performance of traditional large platforms, while also substantially enhancing global humidity analyses and short- to medium-range forecasts. This supports the planned deployment of a 45-satellite constellation for operational NWP.