Title: Comparison of vegetation indices to determine their accuracy in predicting spring phenology of Swedish ecosystems
Abstract:Phenological observations of terrestrial ecosystems are useful in monitoring the
changes in the local climate due to their relatively long time series and high temporal
resolution. However performin...Phenological observations of terrestrial ecosystems are useful in monitoring the
changes in the local climate due to their relatively long time series and high temporal
resolution. However performing field based phenological observations can be a labour
intensive and time consuming process. Using satellite-based remotely sensed data can
make the process much more efficient. Although the satellites do not measure the
plant phenology directly they can be used to observe seasonal changes on a landscape
scale and to estimate the dates of a number of phenological events, such as the onset
of greenness or the beginning of the leaf senescence. The launch of new Earth
observation satellites over the last decade, with improved spatial, temporal and
spectral resolutions, presented an opportunity to develop new vegetation indices (VI)
which could potentially be suited to the observation of phenological changes.
The present study compares four VIs on how accurately they can be used to estimate
the timing of spring phenological events in ecosystems in the north, centre and south
of Sweden. The indices under study are the NDVI, WDRVI, EVI2 and NDWI. The
reference data comes from tower mounted or hand held instruments which measure
the photosynthetically active radiation (PAR). The phenological events being looked
at are the onset of the green season (when green vegetation appears in spring either
through being exposed from underneath the melting snow or through fresh growth)
and the onset of the growing season (when the new vegetation, especially tree leaves,
begin to grow).
The results of the study indicate that NDWI is the only index that can estimate the
onset of the leaf growing season in deciduous forests both in the north and south of
Sweden. The other indices are only able to predict the start of the green season in this
type of ecosystem. In coniferous forests EVI2 seems to be the most appropriate index
to use to estimate the start of the growing season. In low vegetation ecosystems the
findings are more inconclusive but it appears that EVI2 also performs the best in
estimating the start of the green season. The study also found that it is necessary to
use under-the-canopy upward pointing PAR sensors to observe the start of the leaf
growing season in deciduous forests and over-the-canopy downward pointing PAR
sensors to observe the start of the growing season in coniferous forests.Read More
Publication Year: 2010
Publication Date: 2010-01-01
Language: en
Type: article
Access and Citation
Cited By Count: 12
AI Researcher Chatbot
Get quick answers to your questions about the article from our AI researcher chatbot