Impact of bidirectional reflectance distribution function on modis vegetation indices in southeast Asia tropical forests

Copyright © 2017 ISRS, All Rights Reserved. Tropical forests play important roles on global climate and biodiversity. The Moderate Resolution Imaging Spectroradiometer (MODIS), with high temporal resolution, provide a useful tool to study tropical forest dynamics, including seasonality and inter-annual variation. However, optical satellite data have cloud, aerosol and bidirectional reflectance distribution function (BRDF) effects, that create uncertainty in tropical forest studies. In the Amazon, some researchers demonstrated the difficulties in separating true forest dynamics from BRDF artefacts and seasonal cloud and aerosol influences. Lastly, optical reflectance saturation in dense tropical forests may restrict the retrieval of phenology information. In this study, we investigated the impact of BRDF effects on MODIS vegetation indices (VI) in Southeast Asia (SEA) tropical forests, the least studied area compared to other major tropical forests (South America and Central Africa). Moreover, unlike Amazon tropical forests, VI seasonality in SEA forests is not synchronous with sun-sensor geometries. We used 10-year data of daily MODIS BRDF (MCD43A1) collection 6 product, a kernel-driven model product that allows us to retrieve VI values for a range of fixed solar zenith angles (SZA). We compared these with the standard VI products (MOD13A1, MYD13A1) to analyse BRDF influences. The results show significant BRDF effects in all forest sites. Generally, smaller SZA yielded higher VI signals in forests. We found tradeoff's between VI robustness to BRDF effects and saturation that impacted upon the retrievals of phenology parameters