1. SURFACE BOUNDARY DATA

There are a number of arrays used in the MRF model to specify initial surface and subsurface conditions. All have been interpolated to the model computational grid from a base grid (table 7.1). Inspection of interpolated values is necessary to see if adjustments are needed around coastlines. Soil moisture and snow depth are predicted quantities, but the others are fixed at their initial values for the duration of the forecast. Land surface (“skin") temperature values at the initial hour were discussed in Chapter 3 and are not included here. Sea surface temperature is fixed for duration of the forecast; however, it is modified at the initial hour because spectrally-truncated terrain is not zero over oceans ( see section 7.10).

1.1. Land/Sea Specification

Specification of land/sea points is obtained from percentage of water surface area data on the U. S. Navy's 1/6 degree latitude-longitude grid. Model grid boxes, surrounding each grid point, which are less than (or equal to) 50 percent water are defined as land. The Navy data was obtained from the National Center for Atmospheric Research, where minor adjustments were made to the original Fleet Numerical Oceanography Center's data set. Land/sea specification for the rhomboidal-40 computational grid is shown in Fig. 7.1 .

1.2. Orography

The basic orography data set is the U.S. Navy 1/6 degree latitude-longitude grid of ground elevation. In order to present a realistic profile to incident winds, adjustments are made to the orography using techniques developed by F. Mesinger during a stay at NMC. To obtain this “silhouette" orography (Caplan, l985; Mesinger et al., l988), one averages the highest elevation in both the east/west and north/south directions for each model grid box. The generated mountain heights (Fig. 7.2) are comparable to the “envelope" orography being used at other NWP centers.

For the gravity wave drag parameterization ( Chapter 6 ), variance of the silhouette orography on the 1/6 degree grid is computed using the Navy mean elevations. The 1/6 degree grid values are then averaged to the desired model grid ( Fig. 7.3 ) .

1.3. Sea Surface Temperature

Data are obtained from the 15 day running mean analysis, using ship and satellite information, produced daily by the Ocean Products Center at NMC (Feit, l986). Monthly climatological sea surface temperature (Alexander and Mobley, 1974) is used in some model experiments ( Fig. 7.4 - Fig. 7.15). Seasonal changes are simulated by computing an initial sea surface temperature (SST) anomaly and adding it to a time-interpolated mean SST (obtained from the monthly climatology).

1.4. Surface Roughness

Background surface roughness (cm) data over land (Fig. 7.16) is obtained from a 2.5 degree latitude-longitude array produced by GFDL from the work of Baumgartner, et al. (l977). Values have been set to 1 cm over Antarctic and Greenland permanent land-ice points. Values over land snow points and oceans are modified during a model forecast.

1.5. Subsurface Soil Temperature

An annual mean surface temperature (Fig. 7.17) is used as a fixed deep soil temperature (TG3).

1.6. Surface Albedo

Seasonal background surface albedo (Matthews, 1985) on a 1 degree latitude-longitude grid (Fig. 7.18 - Fig. 7.21) is interpolated to the model computational grid. Monthly data are obtained by linear interpolation between the seasonal values. Albedo over land and water are altered for snow, sea ice, and solar zenith angle (water only) in the radiation parameterization -- see Chapter 3 .

1.7. Sea Ice Distribution

Monthly 2 degree latitude-longitude gridded arrays of sea ice were produced from a (5 x 4) degree NASA-GLA data set via “brute force" techniques. Sea ice is allowed to cover land masses in this base data set so that the interpolation procedure does not leave open water around coastlines where ice should exist. Monthly sea ice distribution is shown in Fig. 7.22 - Fig. 7.33 .

1.8. Soil Moisture

Monthly soil moisture is interpolated from GFDL data on 1.875 degree latitude-longitude gridded arrays. Soil moisture ranges from 0-150 mm and is set to zero over ocean on the model grid. Monthly climatologies on the base grid are shown in Fig. 7.34-Fig. 7.45.

1.9. Snow Depth

Monthly snow depth is interpolated from GFDL data on 1.875 degree latitude-longitude gridded arrays. The GFDL data is adjusted by placing 200 mm (water equivalent) snow depths at sea ice grid points prior to the interpolation. Snow depth ranges from 0-50000 mm (water equivalent) and is set to zero over open water on the model grid. Monthly climatologies on the base grid are shown in Fig. 7.46 - Fig. 7.57 .

1.10. Sea Surface Temperature Adjustment

Serious overestimates of sensible and latent heat fluxes at the ocean surface are found in coastal waters near high terrain (e.g. west coast of South America, Indonesia, etc). The error is caused by a mismatch between surface and free-atmosphere parameters in regions where the spectral truncation of terrain spreads non-zero elevations into nearby oceans. An adjustment is made to the analyzed sea surface temperature (SST) which prevents erroneous estimates of vertical gradients of temperature and moisture at these locations. The sea surface temperatures are modified by assuming a 6.5 degree K/km lapse rate between the spectrally -truncated values of the terrain and the zero elevation of the SST analysis (Fig. 7.58) . The adjustment produces SST values several degrees colder in various regions of the tropical oceans and significantly improves the precipitation forecasts.

Table 7.1

Surface Boundary Condition Datasets


Type

Primary Data Source

Processing

Frequency

Surface Roughness

2.5° lat/long grid (from GFDL)-see Baumgartner , 1977.

Bilinear interpolation

Annual

Subsurface Soil Temperature

(2.6° x 1.6° ) lat/long grid (from GFDL). This is fixed deep soil temperature.

Bilinear interpolation

Annual mean surface temperature

Surface Albedo (Land)

1° lat/long grid, seasonal-from Matthews, 1985.

Bilinear interpolation in space, linear interpolation in time

Monthly

Soil Moisture

1.875° lat/long grid (from GFDL)

Bilinear interpolation

Monthly

Snow Depth

1.875° lat/long grid (from GFDL)

Bilinear interpolation

Monthly

Sea Ice Distribution

2° lat/long array produced from NASA GLA (5° x 4° ) gridded data.

Bilinear interpolation

Monthly

Sea Surface Temperature

2° lat/long grid ; 15-day running mean (blended ship and satellite data) - Ocean Products Center (NMC), Feit (1986).

Biquadratic interpolation

Daily

Topography

1/6° lat/long grid from U.S. Navy data -Navy Fleet Numerical Oceanography Center, Monterey, CA (as adjusted at NCAR).

F. Mesinger's silhouette procedure-Mesinger,1986 and Caplan, 1985. In each model grid box take an average of highest elevation in both horizontal coordinate directions.



References

Alexander, R. C., and R. L. Mobley, 1974: "Monthly Average Sea Surface Temperatures and Ice-pack Limits on a 1 Degree Global Grid", Rand Corporation Report R-1310-ARPA, December.

Baumgartner, A., H. Mayer, and W. Metz, l977: “weltweite Verteilung des Rauhigkeitsparameters zo mit Anwendung auf die Energiedissipation an det Erdoberflache", Meteorol. Rdsch., pp. 43-48.

Caplan, P. M., l985: “Recent Changes in the New Medium-Range Forecast (MRF) Model", NMC Office Note 307.

Feit, D. M., l986: “Compendium of Marine Meteorological and Oceanographic Products of the Ocean Products Center", NOAA Technical Memorandum NWS-NMC 68, 93 pp.

Matthews, E., l985: “Atlas of Archived Vegetation, Land Use, and Seasonal Albedo Data Sets", NASA Technical Memorandum 86199, Goddard Institute for Space Studies, New York.

Mesinger, F., Z. Janjic, S. Nickovic, D. Gavrilov, and D. G. Deaven, l986: “The Step-Mountain Coordinate: Model Description and Performance for Cases of Alpine Lee Cyclogenesis and for a Case of an Appalachian Redevelopment", submitted to Monthly Weather Review.

Fig. 7.1 Land (X) and ocean points on the rhomboidal 40 Gaussian grid.

Fig. 7.2 Land surface elevation (meters). Contour interval = 1000 m .

Fig. 7.3 Variance of land surface elevations ( m2 ) , needed by the gravity wave drag parameterization . Contour interval = 200 m2 .

Fig. 7.4 January climatological sea surface temperature ( C ). Contour interval=4C. .

Fig. 7.5 SST as in Fig. 7.4 for February .

Fig. 7.6 SST as in Fig. 7.4 for March .

Fig. 7.7 SST as in Fig. 7.4 for April .

Fig. 7.8 SST as in Fig. 7.4 for May .

Fig. 7.9 SST as in Fig. 7.4 for June .

Fig. 7.10 SST as in Fig. 7.4 for July .

Fig. 7.11 SST as in Fig. 7.4 for August .

Fig. 7.12 SST as in Fig. 7.4 for September .

Fig. 7.13 SST as in Fig. 7.4 for October .

Fig. 7.14 SST as in Fig. 7.4 for November .

Fig. 7.15 SST as in Fig. 7.4 for December .

Fig. 7.16 Surface roughness ( centimeters ) . Contour interval = 25 cm .

Fig. 7.17 Deep soil temperature ( C ) . Contour interval = 5 C .

Fig. 7.18 January climatological background surface albedo ( fraction ) on 1 degree grid. Blank < 0.07 ; light shading for values > 0.07 but < 0.2 ; medium shading for values > 0.2 but < 0.4 : dark shading for values > 0.4 .

Fig. 7.19 Surface albedo as in Fig. 7.18 for April .

Fig. 7.20 Surface albedo as in Fig. 7.18 for July .

Fig. 7.21 Surface albedo as in Fig. 7.18 for October .

Fig. 7.22 January climatological sea ice ( X ) distribution on 2 degree grid .

Fig. 7.23 Sea ice as in Fig. 7.22 for February .

Fig. 7.24 Sea ice as in Fig. 7.22 for March .

Fig. 7.25 Sea ice as in Fig. 7.22 for April .

Fig. 7.26 Sea ice as in Fig. 7.22 for May .

Fig. 7.27 Sea ice as in Fig. 7.22 for June .

Fig. 7.28 Sea ice as in Fig. 7.22 for July .

Fig. 7.29 Sea ice as in Fig. 7.22 for August .

Fig. 7.30 Sea ice as in Fig. 7.22 for September .

Fig. 7.31 Sea ice as in Fig. 7.22 for October .

Fig. 7.32 Sea ice as in Fig. 7.22 for November .

Fig. 7.33 Sea ice as in Fig. 7.22 for December .

Fig. 7.34 January climatological soil moisture ( millimeters ) . Contour interval=20 mm.

Fig. 7.35 Soil moisture as in Fig. 7.34 for February .

Fig. 7.36 Soil moisture as in Fig. 7.34 for March .

Fig. 7.37 Soil moisture as in Fig. 7.34 for April .

Fig. 7.38 Soil moisture as in Fig. 7.34 for May .

Fig. 7.39 Soil moisture as in Fig. 7.34 for June .

Fig. 7.40 Soil moisture as in Fig. 7.34 for July .

Fig. 7.41 Soil moisture as in Fig. 7.34 for August .

Fig. 7.42 Soil moisture as in Fig. 7.34 for September .

Fig. 7.43 Soil moisture as in Fig. 7.34 for October .

Fig. 7.44 Soil moisture as in Fig. 7.34 for November .

Fig. 7.45 Soil moisture as in Fig. 7.34 for December .

Fig. 7.46 January climatological snow depth ( millimeters ) over land ( sea ice values of 200 mm not shown here ) . Blank < 0.01 mm ; - for values > 0.01 but < 10.0 mm ; = for values > 10.0 but < 100.0 mm ; X for values > 100.0 mm .

Fig. 7.47 Snow depth as in Fig. 7.46 for February .

Fig. 7.48 Snow depth as in Fig. 7.46 for March .

Fig. 7.49 Snow depth as in Fig. 7.46 for April .

Fig. 7.50 Snow depth as in Fig. 7.46 for May.

Fig. 7.51 Snow depth as in Fig. 7.46 for June .

Fig. 7.52 Snow depth as in Fig. 7.46 for July .

Fig. 7.53 Snow depth as in Fig. 7.46 for August .

Fig. 7.54 Snow depth as in Fig. 7.46 for September .

Fig. 7.55 Snow depth as in Fig. 7.46 for October .

Fig. 7.56 Snow depth as in Fig. 7.46 for November .

Fig. 7.57 Snow depth as in Fig. 7.46 for December .

Fig. 7.58 Illustration of mismatch between sea surface at zero elevation ( point C ) and modelled earth's surface ( point A ) due to spectral truncation of terrain . Heavy solid line is non-truncated orography and k=1 is location of lowest model atmospheric parameters. For the SST adjustment , a new SST at point A is computed from observed SST at point C using a 6.5 K/km lapse rate .

62