The Mackenzie basin, a 1.8 million km2 region of Northwest Canada, displays significant spatial
variability in atmospheric and surface characteristics. Due to the large area and northern location
of the region, the variability of some surface features that affect the energy and water balance
need to be studied with remote sensing techniques. For this purpose, NOAA AVHRR data were collected
and processed at 1-km resolution within the framework of the Mackenzie GEWEX Study (MAGS).
The AVHRR thermal bands can serve in particular in studies of the land, water, snow, and cloud top
temperatures. Only the warm season water and land temperatures were studied with this thermal data
so far. Some preliminary results are presented here in two parts. First, an empirical method to
document spatial and temporal variations of the temperatures of the basin's many thousand water
bodies is briefly described. Second, the variability of land surface temperature (LST) and its
influence on the energy and water budgets are examined. In particular, a map of maximum LST was
constructed from the 1994 AVHRR data. The spatial variability of maximum LST over the basin, and
its relation to the spatial distribution of water bodies, land cover and elevation, were analyzed.
The coldest values in this map were found over mountains and water, with Great Bear Lake temperatures
not exceeding 8oC to 11oC. Maximum LST tends to decrease with increasing vegetation density.
Maximum LST values decrease with elevation at a rate of -4.5oC/km. Surface moisture cannot not be
measured at the basin scale, but we find that it might be possible to infer its distribution from
its influence on the maximum LST patterns. We also suggest that the analysis of the spatial
distribution of maximum LST and its annual variation is useful for cross-validations with numerical
model results, as well as a useful method to characterize various basins. Finally, more
work remains to be done with the use of the AVHRR thermal data to derive snow and cloud temperatures.