MONITORING RICE REFLECTANCE AT FIELD LEVEL FOR ESTIMATING BIOMASS ANDLAI

The present study aims to monitor rice crop status during the growing season by estimating its aboveground biomass and leaf area index (LAI) from field reflectance measurements taken with a hand-held radiometer . First, vegetation indices (PVI, NDVI, WDVI, PVI) were calculated fro m rice crop reflectance. The fraction of intercepted photosyntheticall y active radiation (PAR), f(PAR), is calculated based on a physical re flectance model from the vegetation indices, WDVI and PVI, indices tha t correct for soil reflectance, show a more Linear and less-scattered relation than NDVI and RVI. The NDVI relationship with f(PAR), gave a good prediction during the vegetative stage but saturated at f(PAR) gr eater than or equal to 0.4. The soil reflectance needed for PVI and WD VI could be easily standardized for continuously flooded fields (10.2% = red reflectance; 7.0% = near-infrared reflectance). Two procedures are discussed: (a) Estimation of leaf area index (LAI) and (b) estimat ion of biomass. The first links f(PAR) With LAI by means of the extinc tion coefficient (K) which varies during the growing season. The dynam ic behaviour of K is analyzed and calibrated for different development stages. LAI of the rice crop was estimated directly from the calculat ed f(PAR) once K was established, This procedure accounted for only 67 % of the variance in LAI. The second uses the Monteith model that requ ires one crop specific parameter, the conversion factor (alpha) for in tercepted PAR into dry matter. Values of cu for direct-seeded paddy ri ce were 2.25 g MJ(-1) for total aboveground biomass over the growing p eriod, but were smaller towards the end. Estimation of biomass from re mote sensing data appears to be more reliable than estimation of LAI. (C) 1998 Elsevier Science B.V.