|Published in||Environmental Research Letters, v.2(10)|
Rankine, C., Sánchez-Azofeifa, G. A., Antonio Guzmán, J., Espirito-Santo, M. M., Sharp, Iain.
Centre for Earth Observation Sciences, Department of Earth and Atmospheric Sciences, University of Alberta, Edmonton, Alberta, T6G 2E3, Canada Departamento de Biologia Geral, Universidade Federal de Minas Gerais, 30161-970, Belo Horizonte, MG, Brazil
|Comparing MODIS and nearsurface vegetation indexes for monitoring tropical dry forest phenology along a successional gradient using optical phenology towers.pdf|
Tropical dry forests (TDFs) present strong seasonal greenness signals ideal for tracking phenology and primary productivity using remote sensing techniques. The tightly synchronized relationship these ecosystems have with water availability offer a valuable natural experiment for observing the complex interactions between the atmosphere and the biosphere in the tropics. To investigate how well the MODIS vegetation indices (normalized difference vegetation index (NDVI) and the enhanced vegetation index (EVI)) represented the phenology of different successional stages of naturally regenerating TDFs, within a widely conserved forest fragment in the semi-arid southeast of Brazil, we installed several canopy towers with radiometric sensors to produce high temporal resolution near-surface vegetation greenness indices. Direct comparison of several years of ground measurements with a combined Aqua/Terra 8 day satellite product showed similar broad temporal trends, but MODIS often suffered from cloud contamination during the onset of the growing season and occasionally during the peak growing season. The strength of the in-situ and MODIS linear relationship was greater for NDVI than for EVI across sites but varied with forest stand age. Furthermore, we describe the onset dates and duration of canopy development phases for three years of in-situ monitoring. A seasonality analysis revealed significant discrepancies between tower and MODIS phenology transitions dates, with up to five weeks differences in growing season length estimation. Our results indicate that 8 and 16 day MODIS satellite vegetation monitoring products are suitable for tracking general patterns of tropical dry forest phenology in this region but are not temporally sufficient to characterize inter-annual differences in phenology phase onset dates or changes in productivity due to mid-season droughts. Such rapid transitions in canopy greenness are important indicators of climate change sensitivity of these already endangered forest ecosystems and should be further monitored using both ground and satellite approaches.