Seasonal variation of tropical forest LAI based on field measurements in Central African Republic

The research presented in this paper focuses on the measurement over a large area of seasonal variation in foliage of a tropical forest. This effort is part of a larger experiment linking ground and space observations over time. A dedicated 8-month field campaign was carried out in Ngotto, a semi-deciduous forest in the Central African Republic to study the spatial and temporal variability of Leaf Area Index (LAI). First, the sensitivity of the LAI-2000 Plant Canopy Analyser instrument was tested. Five sensitivity variables were tested in all: measurement procedure, positioning error, Solar Zenith Angle (SZA) effects, foliage changes, and cloud cover variations. Secondly, the LAI spatial variability was assessed based on geostatistical analysis of measured transects with lags ranging from 30 cm to 10 m. The variogram analysis, showed an overall variance estimate of 0.47 (sill) and a loss of spatial correlation at a distance of 30 m (range). These values were used to calculate the sample size and the minimum distance between adjacent positions, respectively, in order to design a relevant sampling scheme for assessing forest foliage seasonality. Throughout the 8-month field campaign, the LAI was measured at least monthly at 64 positions within 4 km(2) forest sites. A seasonal change of LAI of 0.34 was measured for this tropical forest and found to be in sync with the dry season which lasts three months. (C) 2003 Elsevier B.V. All rights reserved.