In 2006, the National Centers for Environmental Prediction (NCEP) implemented the Real-Time Mesoscale Analysis (RTMA) in collaboration with the Earth System Research Laboratory and the National Environmental, Satellite, and Data Information Service (NESDIS). In this work, a description of the RTMA applied to the 5-km resolution conterminous U.S. grid of the National Digital Forecast Database is given. Its two-dimensional variational data assimilation (2DVAR) component used to analyze near-surface observations is described in detail, and a brief discussion of the remapping of the NCEP stage II quantitative precipitation amount and NESDIS Geostationary Operational Environmental Satellite (GOES) sounder effective cloud amount to the 5-km grid is offered. Terrain-following background error covariances are used with the 2DVAR approach, which produces gridded fields of 2-m temperature, 2-m specific humidity, 2-m dewpoint, 10-m U and V wind components, and surface pressure. The estimate of the analysis uncertainty via the Lanczos method is briefly described. The strength of the 2DVAR is illustrated by (i) its ability to analyze a June 2007 cold temperature pool over the Washington, D.C., area; (ii) its fairly good analysis of a December 2008 mid-Atlantic region high-wind event that started from a very weak first guess; and (iii) its successful recovery of the finescale moisture features in a January 2010 case study over southern California. According to a cross-validation analysis for a 15-day period during November 2009, root-mean-square error improvements over the first guess range from 16% for wind speed to 45% for specific humidity.

This study assesses the impact of assimilating remote sensing observations of significant wave height (SWH) and 10 m wind speed (WSPD) into the analysis using a weakly coupled data assimilation (WCDA) framework to improve ocean surface monitoring offshore of the Contiguous US (CONUS). We assimilated SWH and WSPD data from 10 altimeters and six scatterometers into the Local Ensemble Transform Kalman Filter (LETKF) data assimilation (DA) scheme over 6 months, from 1 January to 1 July 2024, to update the SWH and WSPD variables of the 30-member Global Ensemble Forecast System wave component (GEFS-wave), which provided the background fields. We validated the SWH and WSPD analyses against assimilated remote sensing observations and found that assimilating these observations corrected the systematic background bias in WSPD values above 15 m/s to near zero, with uncertainty as low as 0.4 m/s. For SWH, the background overestimation averaged 30 cm for observed values between 6.2 and 6.4 m and was reduced to 6 cm in the analysis, with a 30 cm RMSE. These results were consistent with cross-validation experiments that assimilated data from six altimeters for the SWH analysis and three scatterometers for WSPD analysis updates, respectively, and were evaluated against independent remote sensing observations. We show that the diurnal cycle in the spatial coverage and temporal sampling of state-of-the-art altimeter and scatterometer constellations over 6 months in maritime CONUS identified regions where remote sensing data had the greatest impact on the analysis and where they were scarce and had limited impact when assimilated using a reduced 1-h window. A case study showed that assimilating altimeter data during a Pacific storm that developed on 1 February 2024 allowed the LETKF to adjust the SWH analysis spatially.