Remote Sensing

The use of satellite data is central to our investigation for providing the required environmental forcings. There are six primary forcing variables: 1) incoming solar (shortwave and near-infrared) radiation; 2) surface air temperature; 3) surface humidity; 4) incoming longwave (thermal) radiation from the atmosphere; 5) surface wind; 6) precipitation.

In addition, landscape variables such as soils, topography, and fractional vegetation cover are required, some of which are obtainable from remote sensing data. Derivation of some variables, such as longwave radiation and air temperature, is difficult, especially under cloudy conditions, and requires innovative approaches that may combine modeling and data assimilation.

Satellite data may be used to update hydrologic model state variables so that they match the observed conditions. For example, satellite observation of surface skin temperature, a state variable calculated by VIC-2L, is relatively straight-forward to measure from space, as shown above in the AVHRR image of the Mississippi basin. The model may then be adjusted based on these observed temperatures. Compare this image with the AVHRR air temperature image (for the same time) shown below.

Solar Radiation: solar radiation is a key variable in surface water and energy budgets. Algorithms for deriving solar radiation are well-established and based on imagery from synoptic weather satellites, such as GOES, which combines coarse spatial resolution (1 km nominal at the equator) with fine temporal resolution (30 minutes).

Solar radiation as derived from GOES observations. The image on the left is visible reflectance obtained from the satellite and used to determine cloud and ground reflectances. These data are then used in a radiative transfer algorithm that calculates surface solar irradiance, shown in image at right. The range of values (black-blue-green-yellow-red)) is from about 50 W/m² to 1000 W/m². Note the effect of clouds through the central U.S. and off the Pacific Coast.

Air Temperature: new techniques have been developed that enable the derivation of near-surface air temperature from satellite data. For example, AVHRR vegetation and surface (skin) temperature observations may be used to infer air temperature at overpass times. Data from TOVS, on-board the same NOAA polar orbiters as AVHRR, are also used to indirectly infer air temperature. The combination of these and EOS-era sensors may allow the estimation of air temperature over large areas where ground observations are scarce or non-existent.

TOVS and AVHRR observations may be used to estimate the diurnal variabiliy of air temperature. The TOVS data are from the NOAA-10 platform and the AVHRR data from NOAA-9. The diurnal curve (for 35.50N, -93.50W) of air temperature uses ground station data interpolated to 1 degree by 1 degree values (TOVS data are courtesy of J. Susskind and V. Lakshmi, NASA GSFC).

Surface air temperature fields derived from the TOVS and AVHRR sensors. Left image is early morning temperature modeled using TOVS. At right is the temperature field derived from the mid-afternoon over-pass of the AVHRR sensor. Temperatures are given in Kelvin degrees. The sharp contrast in the TOVS temperature field from west to east results from different orbital swaths: the western swath was imaged on hour later thatn the eastern one.