NCEP Implements Major Upgrade to Its Medium-Range Global Forecast System,
Including Land-Surface Component
Ken Mitchell, Helin Wei, Sarah Lu, George Gayno and Jesse Meng
NOAA/NWS/NCEP Environmental Modeling Center
Prediction systems often rely on physical process research to realize the model improvements needed to improve forecast skill. This article details recent improvements in the National Oceanic and Atmospheric Administration (NOAA) National Centers for Environmental Prediction (NCEP) global model that have been derived in part from GEWEX research.
On 31 May and 14 June 2005, NCEP extensively upgraded the land-surface component of its Global Forecast System (GFS), including its Global Data Assimilation System (GDAS). Other substantial GFS upgrades were also implemented on May 31, including increased horizontal resolution from about 50 km (T254) to about 35 km (T382), a new sea-ice model, enhanced mountain blocking in the gravity-wave drag, modified vertical diffusion, and upgrades to the objective analysis in the GDAS (more satellite radiance data, enhanced quality control, and improved emissivity calculations over snow and sea ice). All the upgrades and names of the many contributors are given in the NCEP technical bulletin at
This article describes the land surface model (LSM) upgrade, wherein the NCEP Noah LSM (Version 2.7.1) replaced the Oregon State University (OSU) LSM, which had been operational in the GFS since the mid-1990s. The Noah LSM embodies about 10 years of upgrades (see Chen et al., 1996; Koren et al., 1999; Ek et al., 2003) to its ancestor, the OSU LSM. These upgrades were developed and tested by the Environmental Modeling Center (EMC) Land Team, assisted by many collaborators, including the National Weather Service Office of Hydrological Development, National Environmental Satellite Data and Information Service Office of Research and Applications, National Center for Atmospheric Research, National Aeronautics and Space Administration, OSU and university principal investigators of the GEWEX Americas Prediction Project, which is sponsored by the NOAA Office of Global Programs (OGP).
The Noah LSM upgrade includes an increase from two (10, 190 cm thick) to four soil layers (10, 30, 60, 100 cm thick), addition of frozen soil physics, new formulations for infiltration and runoff (giving more runoff for unsaturated soils), revised physics of the snowpack and its influence on surface heat fluxes and albedo, tuning and adding canopy resistance parameters, allowing spatially varying root depth, revised treatment of ground heat flux and soil thermal conductivity, reformulation for dependence of direct surface evaporation on first layer soil moisture, and improved seasonality of green vegetation cover. The frozen soil physics includes soil heat sinks/sources from freezing/thawing and influences vertical transport of soil moisture, soil thermal conductivity and heat capacity, and surface infiltration. The prognostic states of snowpack depth and liquid soil moisture were added to the already present prognostic states of snowpack water-equivalent (SWE), total soil moisture (liquid plus frozen), soil temperature, canopy water, and skin temperature. SWE divided by the snowpack depth gives the snowpack density. Total soil moisture minus liquid soil moisture gives the frozen soil moisture.
To provide initial values of soil moisture/temperature, the Noah LSM land states cycle continuously in the coupled atmosphere/land global model of the GDAS. These land states respond to the global modelís predicted land-surface forcing (precipitation, surface radiation, near-surface air temperature, humidity, and wind speed). Since the land component of the GDAS is forced by model prediction rather than observed precipitation, we avoid undue drift by nudging soil moisture towards a monthly global climatology (in GDAS only, not in forecast).
For some years, the GFS had manifested two prominent biases in land-surface processes: 1) an early depletion of snowpack; and 2) a high bias in both surface evaporation and precipitation in the warm season in non-arid mid-latitude regions. The Noah LSM upgrade greatly reduces the early depletion of snowpack and the warm-season high bias in both surface evaporation and precipitation in mid-latitudes. (In the online technical bulletin cited earlier, figures 1 and 2 show the significant reduction of early snowpack depletion and figure 44 illustrates the reduced warm-season surface evaporation.) The top panel of the figure below shows the reduced high bias in precipitation over the eastern two-thirds of the Continental U.S. (CONUS) during May 2005 in the new versus old GFS for the 60-84 hour forecast range. The bottom panel also depicts an increase in the precipitation forecast skill in the new versus old GFS. Precipitation results for other GFS forecast lengths and other regions (western CONUS, Amazon River basin) are given in figures 45-48 of the online technical bulletin. While the increase in precipitation forecast skill is likely due as much to the increase in model resolution as to the land surface model upgrade, the reduction of the positive precipitation bias is largely due to the LSM upgrade (see figure below).
Extensive pre-implementation testing of the Noah LSM upgrade by itself in the GFS was carried out, but at substantially lower spatial resolution (T62) to facilitate efficient execution of multi-year tests of the continuously cycled GDAS. Given the ďlong memoryĒ nature of land states, such multi-year tests with continuous cycling are critical for proper assessment of major land-surface upgrades, to allow sufficient spin-up of the land-states under the physics/parameters of the new land model. GFS tests (not shown) with the new Noah LSM initialized from GDAS tests of insufficient length manifested significantly different land surface fluxes than those from GDAS/Noah-LSM tests that had cycled for more than one annual cycle.
Of additional note, when full resolution testing (T382) of the Noah LSM commenced in the GFS, it was found that tuning of several canopy resistance parameters (e.g., minimal stomatal resistance) was required to achieve the desired reduction in the high bias of GFS warm-season surface evaporation in mid-latitudes over non-arid regions. Figures 44-48 in the online technical bulletin cited earlier provide examples of GFS performance before and after the tuning of canopy resistance parameters. These examples show a direct link between the reduced surface evaporation and the reduced high bias in precipitation.
The Noah LSM had been previously implemented in NCEPís mesoscale Eta model and its data assimilation system (EDAS), and in NCEPís 25-year North American Regional Reanalysis. The implementation of the Noah LSM in NCEPís GFS positions the Noah LSM for inclusion in the next-generation upgrade of NCEPís ocean/atmosphere/land Coupled Forecast System (CFS) for seasonal forecasting. At NCEP, the atmosphere/land component of any future CFS upgrade for seasonal forecasting is unified with the atmosphere/land component of a recent operational version of the GFS for medium-range forecasting. Assessing the Noah LSM as the land component of the next-generation CFS is now a major thrust of the NOAA Climate Test Bed at NCEP, co-sponsored by NOAA OGP.
Chen, F., K. Mitchell, J. Schaake, Y. Xue, H.-L. Pan, V. Koren, Q.-Y. Duan, M. Ek, and A. Betts, 1996. Modeling of land-surface evaporation by four schemes and comparison with observations, J. Geophys. Res., 101, No. D3, 7251-7268.
Ek, M. B., K. E. Mitchell, Y. Lin, E. Rogers, P. Grunmann, V. Koren, G. Gayno, and J. D. Tarpley, 2003. Implementation of Noah land-surface model advances in the NCEP operational mesoscale Eta model, J. Geophys., 108, No. D22, 8851, doi:10.1029/2002JD003296, 2003.
Koren V., J. Schaake, K. Mitchell, Q.-Y. Duan, F. Chen and J. Baker, 1999. A parameterization of snowpack and frozen ground intended for NCEP weather and climate models, J. Geophys. Res., 104, No. D16, 19569-19585.
(Appeared in GEWEX News, Nov. 2005)