Remotely-Sensed Data

Remotely-sensed data is data collected from some sensor above the earth. The platforms these sensors ride on can include satellites, airplanes, even the space shuttle. Our database considers only satellite mounted sensors. The sensor types used include thermal detectors, cameras, radars, and other devices.

We have implemented a high performance global database for storing Advanced Very High Resolution Radiometer (AVHRR) data. AVHRR is a thermal sensor mounted on the NOAA series of satellites. As it scans the surface of the earth in a polar orbit, the AVHRR sensor measures and records data for five thermal bands. Each measurement is referred to as an instantaneous field of view (IFOV) and represents the thermal reflectance for a region on the surface of the earth.

These IFOV's are accumulated by the sensor into two different level 1 products: Local Area Coverage (LAC) and Global Area Coverage. We will only consider the global GAC product.

Level 1 Data Products

Level 1 data products are accumulations of raw sensor radiometry. There are two types: level 1A and level 1B. The difference is that level 1B products include geo-location data for IFOV's and some meta-data such as sensor corrections, platform identifiers, etc.

The AVHRR sensor sweeps perpendicular to its orbital track. Each sweep of the detector yields 409 IFOV's and is referred to as a scan line. AVHRR GAC Level 1B data sets are an accumulation of 110 minutes of scanning, or approximately 12,000 scan lines. Each scan line record in a GAC Level 1B data set includes five bands of thermal data for each of the 409 IFOV's, 51 latitude/longitude pairs for navigating (geo-locating) the IFOV's, 51 solar zenith angles, and various other data quality indicators for the scan line. Each GAC Level 1B data file represents just over a single orbit of the sensor and occupies approximately 40MB of space.

Level 2 Data Products

Level 1 products are notoriously difficult to work with. They are quite large and cover a relatively small spatio-temporal domain. Earth-scientists looking to do global trend analysis require a more compact data product.

Go back to Remotely-Sensed-Database page.

Go back to Grand Challenge page.

This page maintained by:

ctso@umiacs.umd.edu