What drives the Amazon green up phenomenon?

The seasonality of light and moisture availability governs net ecosystem production in tropical forests. Whether tropical forest productivity is more limited by light or moisture remains an area of active debate. Recent satellite-based studies support the light limitation hypothesis, as Amazon forests respond with higher productivity through leaf turnover or more leaf area during dry season months, measured as increases in the Enhanced Vegetation Index (EVI). Here, we used two novel approaches to evaluate the role of canopy phenology for this "green up" of Amazon forests during the dry season. First, we evaluated the consistency of dry season increases in near-infrared reflectance (NIR) over Amazon forests using LiDAR and optical remote sensing observations. ICESat GLAS LiDAR data provide estimates of both apparent reflectance and height metrics of returned energy from each laser pulse. We analyzed ICESat data from June (Laser Periods 3c, 3f) and October (Laser Periods 3a, 3i) to evaluate whether changes in forest canopy drive seasonal increases in NIR, measured as an increase in waveform centroid relative height. We then compared seasonal changes in NIR from MODIS and ICESat at ICESat footprint locations. Second, we tested competing hypotheses for the Amazon green up phenomenon using FLIGHT, a ray-tracing model capable of simulating both optical and LiDAR radiative transfer. Our sensitivity study compared observed and modeled ICESat and MODIS data based on seasonal increases in leaf area index, leaf reflectance, litter reflectance, and solar zenith angle for a simulated Amazon forest. The integration of observations from LiDAR and optical sensors with sophisticated radiative transfer simulations provides new constraints on the role of phenology for changes in MODIS EVI.