2. FIELD EXPERIMENT 
612 
Site description - The field measurements were conducted in the West Central supersite of the HAPEX-Sahel area, 
centered on the village of Fandou Béri (13 0 31’N; 2°33'E). The supersite (approximately 20 km by 20 km) includes the 
five principle Sahelian landscape components which are a grassland (a more or less continuous layer of herbs made of 
grass and forbs), a degraded bushland (a discontinuous woody layer made of sparse bushes over bare soil), a 
bush/grassland (a combination of a herb layer with a woody layer), a millet field (a discontinuous crop, sown in 
clumps), and a tiger bush (a stripes of dense vegetation over laterite plateaux). Except the tiger bush, each vegetation 
canopy was represented by a vegetated plot (10 m by 10 m) selected in the site area (300 m by 300 m). On the plots, 
radiative measurements were continuously collected while biological properties occasionally measured. 
Radiative measurements - PARIR sensors manufactured by SOLEMS Industry were used to measure radiative fluxes. 
This sensor is composed by two silicon detectors mounted on the same frame. The amorphous silicon detector is 
sensitive to PAR, while the crystalline silicon detector quantifies the global shortwave radiation (400-1100 nm), 
referred further is this paper as VNIR (NIR signal is taken from the difference between PAR and VNIR signals). A 
thin layer of polymetacrylate material covering the detectors filters the ultra-violet radiation, assures a good 
mechanical protection, shows a satisfactory cosine response up to 70°, and prevents saturation up an incoming 
radiation of 1100 W/m 2 . The sensors are particularly adapted to resist to tropical conditions. 
The number of sensors per plot to measure transmitted radiation varies according to the homogeneity scale of the 
canopy (see Table 1). Also, two pairs of 4 sensors were mounted on a 7-meters mast above the canopy to measure the 
incident and reflected radiations. Further the plot, a 2-sensors set measured the soil radiance. In addition on the 
bush/grassland, two pyranometers measured the global and diffuse radiations in the 0.3-3 pm wavelength range. The 
radiative instruments were connected to Campbell 21X data loggers. The data were averaged and stored every 10 mn, 
from sunrise to sunset. Before and after the experiment, in-situ absolute calibrations were carried out between the 
incident set of PAR detectors and a PAR quantum detector (LI-191, Licor), and between the incident set of VNIR 
detectors and a pyranometer. 
The interception and absorption efficiencies of radiation by the canopy are calculated on hourly and daily basis. The 
interception efficiency e ( of the canopy is derived simply from the hemispherical transmittance x as, 
£j=l-X (1) 
while the calculation of the absorption efficiency £ a , a more "true" quantity in terms of production, makes appear the 
hemispherical reflectances (albedos) a of the canopy and a s of the bare soil, viz: 
e a = e i - a + C 1 - e i) a s ( 2 ) 
Biological measurements - Those measurements describe the structure of the vegetation on the plots and on the sites 
(Table 2). The LAI was determined through the 'specific leaf area' (surface area:dry matter ratio) of the species and 
biomass measurements. The destructive measurements were made bi-weekly on the sites and only operated at the end 
of the experiment on the plots. For the herb layer, once assessed the floristic composition, five samples of 0.5 m 2 were 
harvested on each plot, oven-dried and weighed. The time profile is obtained by analogy with the seasonal evolution 
of the herb layers on the site. For the woody layer, the bushes inside the biological sampling area were measured 
(height and diameter) then harvested. The temporal evolution of the leaves dry matter was obtained through allometric 
relationships derived between volume and biomass of bushes sampled over the site (12-30 bushes were harvested 
every 2 weeks). For the millet field, the height and diameter of the clumps were measured twice a month while the 
senescent leaves were regularly clipped off during the maturation stage. Leaf dry matter was obtained through 
allometric relationships derived between volume and biomass of 12-30 clumps sampled over the sub-site at the same 
dates. At the end of the maturation, all the clumps of the plots were harvested, dried and weighed. 
Radiometric measurements - Nadir ground and airborne reflectance data were collected on the four sites with Exotecb 
100AX radiometers (Landsat/TM channels) with 15° field of view (FOV) lenses. At the ground level, weekly transect 
measurements were made around 9:00 UT while the radiometer was mounted at a heigh of 3 m (target diameter is 70 
cm). Bi-weekly aircraft (Piper Saratoga) radiometric measurements were made around 10:00 UT, with pixels of 60 m. 3 
3. RESULTS AND ANALYSIS 
Daily results - The climatic efficiency (ratio PAR to global incident radiation) keeps a value around 0.475 for the 
whole season, in agreement with previous studies. This constancy can be explain by the fact that Sahelian aerosols are 
poorly absorbant, and by consequent nearly spectrally neutral, and also the précipitable water watour is still important, 
making it mostly unsensitive to small variations (Tanré, 1993, personal communication). 
Daily efficiencies variability in time for the four sites are plotted in Figures 1 and 2. Those figures reveal that we did