Land Surface Temperature (LST) is a key indicator of land surface states, and can provide information on surface-atmosphere heat and mass fluxes, vegetation water stress, and soil moisture. A daily, day and night, LST data set for continental Africa, including Madagascar, was derived from Advanced Very High Resolution Radiometer (AVHRR) Global Area Coverage (GAC; 4-km resolution) data for the 6-year lifetime of the NOAA-14 satellite (from 1995 to 2000) using a modified version of the Global Inventory Mapping and Monitoring System (GIMMS) (Tucker et al., 1994). The data were projected into Albers Equal Area and aggregated to 8-km spatial resolution. The data were cloud-filtered with CLAVR-1 algorithm (Stowe et al., 1999). The LST values were estimated with a split-window technique (Ulivieri et al., 1994) that takes advantage of differential absorption of the thermal infrared signal in bands 4 and 5. The emissivity of the surface was generated using a land cover classification map (Hansen et al., 2000) combined with the FAO soil map of Africa (FAO-UNESCO, 1977), and additional maps of tree, herbaceous, and bare soil percent cover (DeFries et al., 2000). Collateral products include cloud mask, time-of-scan, latitude and longitude, and land/water mask files.
The data are in flat binary files. Each data file contains 1152 columns and 1152 rows, in signed integer format (2 bytes), with 8 km by 8 km spatial resolution. A unique map exists for each day and each night of the 6-year NOAA-14 lifetime. The data are best used to infer broad temporal and spatial trends rather than pixel-by-pixel values.
Cite this data set as follows:
Pinheiro, A. C. 2005. SAFARI 2000 AVHRR-derived Land Surface Temperature Maps, Africa, 1995-2000. Data set. Available on-line [http://daac.ornl.gov/] from Oak Ridge National Laboratory Distributed Active Archive Center, Oak Ridge, Tennessee, U.S.A. doi:10.3334/ORNLDAAC/826.
DeFries, R. S., M. C. Hassen, and J. R. G. Townshend. 2000. Global continuous fields of vegetation characteristics: a linear mixture model applied to multi-year 8 km AVHRR data. Int. J. Remote Sensing, 21(6): 1389-1414.
Hansen, M. C., R. S. DeFries, J. R. G. Townshend, and R. Sohlberg. 2000. Global land cover classification at 1 km spatial resolution using a classification tree approach. Int. J. Remote Sensing, 21(6): 1331-1364.
Pinheiro, A. C., J. L. Privette, R. Mahoney, and C. J. Tucker. 2004. Directional effects in a daily AVHRR land surface temperature dataset over Africa. IEEE Trans. Geosci. Remote, 42(9): 1941-1954.
Pinheiro, A. C. T., R. Mahoney, J. L. Privette, and C. J. Tucker. 2005 (submitted). A Daily Long-Term Record of NOAA-14 AVHRR Land Surface Temperature over Africa. Remote Sensing of Environment.
Stowe, L. L., P. A. Davis, and E. P. McClain. 1998. Scientific basis and initial evaluation of the CLAVR-1 global clear/cloud classification algorithm for the Advanced Very High Resolution Radiometer. J. Atmospheric and Oceanic Technology, 16: 656-681.
Tucker, C. J., W. E. Newcomb, and. H. E. Dregne. 1994. Improved data sets for determination of desert spatial extent. Int. J. Remote Sensing, 15: 3519-3545.
Ulivieri, C., M. M. Castronuovo, R. Francioni, and A. Cardillo. 1994. A split window algorithm for estimating land surface temperature from satellites. Adv. Space Research, 14(3): 59-65.
The data consist of a unique daily gridded land surface temperature map of continental Africa and Madagascar for each day and each night of the 6-year NOAA-14 lifetime (1995-2000). The data files are organized by year, with separate folders for daytime and nighttime data (e.g., AVHRR_1995_DAY and AVHRR_1995_NIGHT). Each folder contains a compressed (.zip) data file, a compressed cloud mask file, and a compressed local solar time file for that year. The data files uncompress to flat binary files.
Each data file contains 1152 columns and 1152 rows, in signed integer format (2 bytes), with 8 km by 8 km spatial resolution.
LST_UL and CLD data fields for the first 20 days (DOY 1 through DOY 20) of Year 1995 should be discharged. This period corresponds to the beginning of NOAA-14 acquisition and resulted in unreliable data, or no data, for those granules.
Data in the LST_UL files correspond to land surface temperature retrieved using the Ulivieri et al. (1994) split-window algorithm, in units of degrees Kelvin. The temperature values are scaled by a factor of 10. A value of -999 is used for cells where channels 4 and 5 reach saturation (323 K and 330 K, respectively). A value of -888 is assigned if the brightness temperatures in channels 4 and/or 5 are below the processing threshold (230 K) or for pixels with no data (e.g., at the edge of the image).
Data in CLD files correspond to a cloud mask generated using the CLAVR algorithm (Stowe et al., 1999). No scaling factor was used. The following flags are used in the cloud mask:
Over land:
| Cloud Mask Flag | Conditions |
| 3 | Clear |
| 4 | Clear and DDV (dense dark vegetation) |
| 6 | Cloudy or mixed |
| 8 | Shadow |
Over ocean:
| Cloud Mask Flag | Conditions |
| 1 | Clear |
| 2 | Clear and glint |
| 5 | Cloudy or mixed |
| 7 | Shadow |
| 8 | Shadow |
Data in the LSTIME files correspond to local solar time of observation for each pixel, scaled by a factor of 1000.
All data are in Albers Equal Area Projection. The upper left corner of the data value column is centered at longitude 24.60 W and latitude 43.71 N, the upper right corner of the data value column is centered at longitude 64.52 E and latitude 43.71 N, the lower left corner of data value column is centered at longitude 23.48 W and latitude 42.24 S, and the lower right corner of data value column is centered at longitude 63.41 E and latitude 42.24 S.
There are also some additional files that help with data use. These include a latitude file (africa_albers.lat), a longitude file (africa_albers.lon), and a land/water mask file (africa_albers_8bit.mask). The data values in the latitude and longitude files are scaled by a factor of 100.
These files are in flat binary and do not require any special software (such as Arc-info) to access the data. They can be opened with any general Fortran, C, IDL, etc. code once the dimensions and data type are defined.
These files are also in flat binary and do not require any special software (such as Arc-info) to access the data. They can be opened with any general Fortran, C, IDL, etc. code once the dimensions and data type are defined.
ENVI description = {File Imported into ENVI.}
Samples (columns): 1152
Lines (rows): 1152
Bands: 1
No header offset
Generic binary bsq file
2 bit data type.
Signed or unsigned.
Additional information about the data collection, theory of data processing, data processing, and data quality is found in the companion file: http://daac.ornl.gov/daacdata/safari2k/remote_sensing/AVHRR_LST/comp/AVHRR_LST_readme.pdf.
2005/2/22
2005/2/22
webmaster@www.daac.ornl.gov