DETECTING LEAF-AREA AND SURFACE-RESISTANCE DURING TRANSITION SEASONS

In this study we seek empirical relationships among canopy resistance to water vapor transport, the time-varying leaf area index (LAI), in s itu radiative flux observations, and a satellite-based estimate of lea f state (NDVI, the normalized difference vegetation index) from a leaf less deciduous forest to a covered canopy and vice versa. These relati onships can be used in numerical models such as verification in global climate models. They also can be useful tools for developing remote s ensing techniques. LAI was found through analysis of frequent video im ages of canopy evolution in spring and autumn during 1992 and 1993 at a deciduous forest in central Massachusetts. We examined the impact of leaf presence on water Vapor transport during spring and autumn using an LAI time series during leaf emergence and leaf fall for the four s tudy seasons. The canopy resistance to water vapor transport (r(c)) de creased abruptly at leaf emergence in each year but then also continue d to decrease slowly during the remaining growing season, owing to slo wly increasing LAI. One remarkable result is that a single linear rela tionship between r(c) and LAT during leaf emergence can be used to est imate the minimum seasonal r(c) associated with the maximum foliage co ver. Canopy resistance and PAR-albedo (albedo from photosynthetically active radiation (PAR) instruments) began to increase about 1 month be fore leaf fall with the diminishment of CO2 gradient above the canopy as well, at which time evaporation began to be controlled as if the ca nopy were leafless. We present empirical linear regressions relating N DVI, r(c), and PAR-albedo. The NDVI linear regressions with surface me asurements indicate that tower-based measurements can represent at lea st a satellite pixel region. These results reinforce the notion that r elationships among these parameters are scale independent from tower-b ased measurements spatial scale to a satellite pixel resolution (1.1 k m x 1.1 km area), at least. (C) 1997 EIsevier Science B.V.