Chapter 4: TERRAIN

4.1 Purpose

 

The program that begins any complete forecast simulation in MM5 modeling system is TERRAIN (Fig. 1.1). This program horizontally interpolates (or analyzes) the regular latitude - longitude terrain elevation, and vegetation (land use) onto the chosen mesoscale domains (see Fig. 4.1). If the land-surface model (LSM) will be used in the MM5 model, additional fields such as soil types, vegetation fraction, and annual deep soil temperature will also be generated.

 

 

 

 

 

4.1.1 Tasks of TERRAIN

 

There are essentially two tasks the program TERRAIN performs:

 

1. Set up mesoscale domains: coarse and fine grids (except for moving nests);

2. Produce terrestrial data fields for all of the mesoscale domains, which will first be used by REGRID, and later by MM5 (optionally) and NESTDOWN. The program also computes a few constant fields required by the modeling system: latitude and longitude, map scale factors, and Coriolis parameter.

 

4.1.2 Overview of TERRAIN

 

The TERRAIN program is composed of four parts (Fig. 4.2):

 

1. Source data input;

2. Interpolation from lat/long source data to mesoscale grid;

3. Nest interface adjustment and feed back; and

4. Output terrain elevation, land use and other terrestrial data in MM5 format.

 

 

 

 

 

4.2 Input Data

 

4.2.1 Source Data

 

The data available as input to the program TERRAIN include terrain elevation, landuse/vegetation, land-water mask, soil types, vegetation fraction and deep soil temperature. Most data are available at six resolutions: 1 degree, 30, 10, 5 and 2 minutes, and 30 seconds. Here is the list of available data:

 

1. Elevation data at six resolutions from USGS: 1-degree, 30-, 10-, 5-, 2-minutes (5 files) and 30-second (33 tiles directly from USGS). All lower resolution data (1 degree to 2 minutes) are created from the 30 seconds USGS data.

 

2. Three types of source vegetation/land-use data available:

 

(a) 13-category, global coverage with the resolution of 1-degree, 30- and 10-minute (3 files);

(b) 17-category, North-American coverage with the resolution of 1-degree, 30-, 10-, 5-, 2-minutes and 30 seconds (6 files);

(c) 25-category, global coverage with the resolution of 1-degree, 30-, 10-, 5-, 2-minutes and 30-seconds (6 files; all lower resolution data are created from 30 sec data from USGS version 2 land cover data).

 

3. Two types of land-water mask data:

 

(a) 17-category, North-American coverage with the resolution of 1-degree, 30-, 10-, 5-, 2-minutes and 30seconds (6 files);

(b) 25-category, global coverage with the resolution of 1-degree, 30-, 10-, 5-, 2-minutes and 30-seconds (6 files).

 

4. For LSM option in MM5, the soil, vegetation fraction, and annual deep soil temperature are needed. The source data files are:

 

(a) 17-category, six resolutions of global soil data (6 files);

(b) 12 monthly, 10-minute, global vegetation fraction data (1 file);

(c) 1-degree, global annual deep soil temperature (1 file).

 

More description of the data is available in section 4.2.3.

 

4.2.2 Data Format

 

Since the original data come from different sources, they have different formats and layouts. These data sets are translated to a standard format which is used by the TERRAIN program. The data arrangement and format in the reformatted data file are as follows,

 

• Latitude by latitude from north to south in one latitude, the data points are arranged from west to east, usually starting from 0 degree longitude (or dateline).
• Two-characters arrays are used to store the elevation and deep soil temperature data (the maximum value < 215, or 32768) (Fig. 4.3), and 1-character array to store all other data (values < 100) (Fig. 4.4).
• All source data files are direct-access, which makes data reading efficient.
• All data are assumed to be valid at the center of a grid box. Hence there are 360x180 data points for 1-degree data, and (360x2)x(180x2) for 30-minute data, and (360x120)x(180x120) data points for the 30-second data, and so on.

 

 

 

 

 

 

 

 

 

4.2.3 Input Data Sources and File Sizes

• Elevation:

Table 4.1a Terrain Height Data

Resolution	Data source*	Coverage	Size(bytes)
1 deg. (111.0 km)	USGS	Global	129,600
30 min. (55.0 km)	USGS	Global	518,400
10 min. (18.5 km)	USGS	Global	4,665,600
5 min. (9.25 km)	USGS	Global	18,662,400
2 min. (3.70 km)	USGS	Global	116,640,000
Tiled 30 sec. (0.925 km)**	GTOPO30 by U.S. Geological Survey's EROS Data Center in late 1996	Global (33 tiles: 40o lon. x 50o lat. or 60o lon. x 30o lat.)	57,600,000 or 51,840,000 for each of tiles
30 sec. (0.925 km)	USGS	Global***	1,866,240,000

* Except for the tiled 30 sec. data (GTOPO30), the data reconstruction from original source data was completed separately prior to TERRAIN. All lower resolution elevation datasets are created from the USGS global 30 second dataset since Version 3.4.

 

** For details of the GTOPO30 data, see http://www.scd.ucar.edu/dss/datasets/ds758.0.html. The tiled 30 seconds elevation data are available from the USGS EROS Data Center's anonymous ftp site edcftp.cr.usgs.gov under directory: /pub/data/gtopo30/global

 

*** This single tile global 30 second file is available through request to mesouser, or on MSS:

/MESOUSER/MM5V3/DATA/SINGLE-TILE-GLOBAL-30S-ELEVATION.gz.

 

The data reconstruction for the 30 second data is included in the ftp30s.csh which is used by the TERRAIN job deck. The reconstructing procedure contains three steps:

 

(1) determine which tiles of the elevation data are needed based on the information in namelist (data_area.exe);

(2) fetch the data from ftp site (or MSS if one runs at NCAR) (dem_read);

(3) reconstruct data in TERRAIN standard input format from the tiled data and provide the necessary information to TERRAIN (rdem.exe).

 

The outputs are new_30sdata, and new_30sdata_info, located in Data/ directory.

(1) Global 13-category data from PSU/NCAR tape

Table 4.1b PSU/NCAR Land-use Data

The 13 categories are listed in Table 4.2a. The data are represented by 13 numbers of percentages for the 13 categories at each of lat/lon grid points.

 

 

(2) North-American 17-category data used by Simple Biosphere (SiB) model (from USGS)

Table 4.1c 17-category SiB Vegetation Data

The 17 categories are listed in Table 4.2b. The 30-sec data are represented by one category-ID number at each of lat/lon grid point. The low resolution (1-deg, 30-, 10-, 5- and 2-min) data are derived from 30-sec data, and are represented by 17 numbers of percentages for the 17 categories at each of lat/lon grid points.

 

 

(3) Global 25-category data from U.S. Geological Survey (USGS)

Table 4.1d 25-category USGS Vegetation Data

The 25 categories are listed in Table 4.2c. The 30-sec data are represented by one category-ID number at each of lat/lon grid point. The low resolution (1-deg, 30-, 10-, 5- and 2-min) data are derived from 30-sec data, and are represented by 25 numbers of percentages for the 25 categories at each of lat/lon grid points.

(1) North-American Land-water mask files derived from SiB Vegetation data

Table 4.1e SiB Land-Water Mask Data

The SiB land-water mask data files are derived from SiB vegetation data files. At each of lat/lon grid points, there is one number indicating the land ( 1), water ( 0), or missing data (-1) at that point.

(2) Global Land-water mask files derived from USGS Vegetation data

Table 4.1f USGS Land-Water Mask Data

Resolution	Data source	Coverage	Size(bytes)
1 deg. (111.0 km)	USGS Vegetation	Global	64,800
30 min. (55.0 km)	USGS Vegetation	Global	259,200
10 min. (18.5 km)	USGS Vegetation	Global	2,332,800
5 min. (9.25 km)	USGS Vegetation	Global	9,331,200
2 min. (3.70 km)	USGS Vegetation	Global	58,320,000
30 sec. (0.925 km)	USGS Vegetation	Global	933,120,000

 

 

 

 

 

 

The land-water mask data files are derived from USGS vegetation data files. At each of lat/lon grid points, there is one number indicating the land ( 1), water ( 0), or missing data (-1) at that point.

  

• Soil

Table 4.1g Global 17-category Soil Data

Resolution	Data source*	Coverage	Size(bytes)
1 deg. (111.0 km)	FAO+STATSGO	Global	1,101,600
30 min. (55.0 km)	FAO+STATSGO	Global	4,406,400
10 min. (18.5 km)	FAO+STATSGO	Global	39,657,600
5 min. (9.25 km)	FAO+STATSGO	Global	158,630,400
2 min. (3.70 km)	FAO+STATSGO	Global	991,440,000
30 sec. (0.925 km)	FAO+STATSGO	Global	933,120,000

 

 

 

 

 

 

 

 

 

 

 

  The 17-Category Global Soil data files are generated by 

 

(1) Global 5-minutes United Nation FAO soil data are convert to the 17 categories same as STATSGO data (available since V3.5);

(2) North-American STATSGO 30-sec soil data

(3) Global high resolution soil data are produced from 5-min FAO data;

(4) North-American low resolution (1-deg, 30-, 10-, 5 -and 2-min) soil data are derived from the 30-sec North-American soil data;

(5) FAO and STATSGO data are merged together for each of the resolutions.

(6) Both top soil layer (0 - 30 cm) and bottom soil layer (30 - 100 cm) data are provided. Obtaining a particular dataset can be set in terrain.deck.

 

The 17 categories are listed in Table 4.2d. Similar to the vegetation data, the 30-sec data are represented by one category-ID number at each of lat/lon grid point, and the low resolution (1-deg, 30-, 10-, 5- and 2-min) data are represented by 17 numbers of percentages for the 17 categories at each of lat/lon grid points.

  

• Vegetation fraction

Table 4.1h Global Monthly Vegetation Fraction Data

Resolution	Data source	Coverage*	Size(bytes)
10 min. (18.5 km)	AVHRR	Global	27,993,600

 

 

 

The 10-min original vegetation fraction data contained 12 percentage-values for 12 months at each of lat/lon grid points, but covered only from 55oS to 75oN. To make the data file have global coverage, a zero value of vegetation fraction was assigned over the high latitude area.

 

• Soil temperature

Table 4.1i Global Annual Deep Soil Temperature Data

Resolution*	Data source	Coverage*	Size(bytes)
1 deg. (111.0 km)	ECMWF analysis	Global	129,600

 

 

 

The resolution of the 1-deg annual deep soil temperature data is rather low. For some of grid points located at small islands in the ocean, it is unable to obtain the deep soil temperature value by interpolation based on this source dataset. In this case, an annual deep soil temperature, Tg, will be assigned based on the latitude of the point, j:

 

               

where

             

and