Spatial and temporal variability of surface cover at BOREAS using reflectance from a helicopter platform

Helicopter-based radiometric measurements of forested sites acquired during the Boreal Ecosystem-Atmosphere Study (BOREAS) were used to examine the spatial, temporal, and spectral variability in surface reflectance and vegetation indices (VI), including the normalized difference vegetation index (NDVI) and the simple ratio (SR), and for comparison with surface cover and fluxes. In this analysis, the sensors, which were employed during all three intensive field campaigns (IFC) of 1994, consisted of an eight-channel modular multiband radiometer (MMR), ground-based Sun photometer used for atmospheric correction, and LICOR LAI-2000, hemispherical photographs, and a ceptometer for retrieval of surface biophysical variables. Means and coefficients of variation were calculated and linear regression analysis performed on reflectances, VIs, and surface variables over the entire data set and as a function of season and cover type. Surface biophysical variables included leaf area index (LAI), effective LAI, and the fraction of absorbed photosynthetically-active radiation (green fAPAR). While each dominant species displayed recognizable reflectance spectra, variability in reflectance, which was high in every channel, was most likely strongly influenced by understory and ground reflectances, and by atmospheric effects. Of the eight MMR bands, those most responsive to surface variations were the third and second middle infrared (IR) (2.08-2.37 and 1.57-1.80 μm, respectively), the red (0.63-0.68 μm), and the blue (0.45-0.52 μm). Of the two VIs examined, the authors´ results showed that the NDVI offered more predictive information than the SR regarding temporal and spatial variations in surface characteristics. Linear regression analyses between the surface biophysical measurements and the helicopter reflectances and VIs resulted in low r² values (consistently <0.5), which may be explained by the effects of incomplete canopy cover and background effects. Only among sites that were mainly vegetated by Aspen (Populus tremuloides) and observed during the summer IFC was a stronger relationship observed. Atmospheric contamination of the radiometric signal by clouds and smoke from forest fires may also have contributed to the unsatisfactory results. Improved estimates of aerosol optical depth, taken from a sun photometer mounted on the helicopter, will be available in the future