This project involves using field data, modeling and remote sensing to investigate tropical phenology.
The phenological dynamics of terrestrial ecosystems reflect the response of the Earth's biosphere to inter- and intra-annual dynamics of the Earth's climatic and hydrological regimes (Myneni, 1997). Some Global Dynamic Global Vegetation Models (DGVMs) have suggested that by 2050 the Amazon rainforest will begin to dieback (Cox et al., 2000, Salazar et al., 2007). One of the major components in DGVMs is the simulation of vegetation phenology, however modellers are challenged with the estimation of phenology, which as shown by satellite monitoring and mapping of vegetation is highly complex. Current modelled phenology is based on observations of vegetation in the temperate zones and accurate representation of phenology of the tropical zones is long overdue.
Remote sensing can assist the modelling of phenology and assessment of correlations with environmental variables. A number of recent studies based on satellite remote sensing have reported seasonal variation in the phenology of the Amazon rainforest, with enhanced "greenness" in the dry season (Huete et al., 2006), and further enhanced greenness during the drought of 2005 (Saleska et al., 2007). These studies have been interpreted as evidence of resilience of tropical rainforests to seasonal and interannual drought. The studies have been entirely satellite-based however, where “greenness” is expressed through vegetation indices (NDVI or EVI) or vegetation index based Leaf Area Index estimates, and thus far there has been little corroboration with on-the-ground observations of the phenology of tropical forests.
It is very likely that leaves vary in their spectral reflectance properties as they age, and where their life cycle is strongly synchronised it is likely to be linked to the seasonal variation in climate and/or hydrology. Hence, there is a distinct possibility that the seasonal variation in vegetation indices (VI) is driven by the leaf aging as well as by the shedding or appearance of new leaves and the corresponding changes in leaf shadowing patterns (Anderson, LO, et al., 2010). The fundamental objective of this NERC funded project is to investigate the influence that age-related variation in the spectral reflectance properties of leaves and changes in leaf showing patterns may have on apparent "greenness" of a tropical forest canopy. The requested equipment is essential in achieving this as it will allow us to conduct near-ground spectro-radiometry. Ground observations will be used to produce spectral signatures of leaves at different stages in their lifecycle sampled from a range of physiognomically different tropical forest vegetation. The spectral signatures will be used (1) to establish which physical and physiological changes in the leaves cause the most significant changes in the leaves’ spectrum and (2) as input into a light canopy interaction model FLIGHT (North, 1996) to produce canopy vegetation indices that will be evaluated against satellite-derived indices of vegetation greenness.
This study will contribute to a better mechanistic understanding of tropical phenological dynamics and help to clarify the origin of the changes in the remotely sensed vegetation indices. This study will exploit on-going measurements of vegetation phenology being conducted by Oxford-supported field researchers at two sites in French Guiana and Peru, conduct targeted field campaigns on leaf reflectance, physical and physiological properties, and utilise a vegetation canopy model to scale to whole vegetation canopies and to the wider Amazon region. Also, at a time when large-scale numerical models that simulate the interactions between changing global climate and terrestrial vegetation predict substantial carbon loss from tropical ecosystems (Friedlingstein et al., 2006), including the drought-induced collapse of the Amazon forest and conversion to savannah (Betts et al., 2004) it is critical that tropical forest vulnerability and/or resilience be assessed using models that have been improved and validated by field observations.
