University of Leicester
U122467.pdf (36.99 MB)

Understanding large area tropical forest phenology using remotely sensed and ground data sources - Volume 1

Download (36.99 MB)
posted on 2014-12-15, 10:38 authored by Simon D. Jones
Remotely sensed (spectral) phenological data has often been used to characterise and investigate tropical forest ecosystems. The basic assumptions, linking real (biological) and remotely sensed (spectral) phenology, have however rarely been scrutinised and little empirical information from synchronous datasets exists.;This research selected three tropical vegetation communities; each characterised by a different climate, biological phenology and each hypothesised to exhibit a differing spectral phenology. An extensive verification campaign was then initiated to collect phenological data from all three of these sites for an entire phenological cycle. The verification dataset comprised; forest structural parameters, meteorological measurements, litterfall weights, phenological observations (of leaf flushing, senescence and abscission), quantifications of canopy openness and LAI (using hemispherical photography and Ceptometry) and overpass-synchronous radiometry readings.;Large area (1 km2), remotely sensed spectral data was then acquired, for all sites, from the NOAA-14 AVHRR and ERS-2 ATSR-2 satellite sensor systems. An evaluation of the biological significance of the spectral phenological data as then undertaken using two basic methodologies. First, the ground verification data were compared to several commonly used spectral vegetation indices. Next, textural changes, in the spectral landscape, attributed to each verification site were assessed. At two of the monitored sites, spectral phenology was shown to have a strong physiological basis at the scale of the vegetation community. This was attributed to, the pronounced seasonality in the climate at these locations, and, the relative structural simplicity of the vegetation formations. At the third site (a structurally more complex, seasonally-inundated tropical forest) the association between biological and spectral phenology was less conclusive. Clearly further work is required before the scientific community can be certain that all temporal trends, derived from 1km spatial resolution image data, are providing accurate insights into the biological processes, of humid tropical forests, but in general the association between spectral and biological phenology is a strong one.


Date of award


Author affiliation


Awarding institution

University of Leicester

Qualification level

  • Doctoral

Qualification name

  • PhD



Usage metrics

    University of Leicester Theses


    No categories selected



    Ref. manager