TY - JOUR

T1 - Mapping phenology of natural vegetation in India using remote sensing derived chlorophyll index

AU - Jeganathan, C.

AU - Dash, J.

AU - Atkinson, Peter M.

PY - 2010/12

Y1 - 2010/12

N2 - Time series of MEdium Resolution Imaging Spectrometer (MERIS) Terrestrial Chlorophyll Index (MTCI) level-3 data product, with a spatial resolution of ∼4.6 km composited at 8-day intervals for the years 2003 to 2007, were used to map the phenology of natural vegetation in India. Initial dropouts and noise in the MTCI data were corrected using a temporal moving window filter, Fourier-based smoothing using the first four harmonics was applied and then the phenological variables were extracted through a temporal iterative search of peaks and valleys in the time series for each pixel. The approach was fine-tuned to extract reliable phenological variables from the complex and multiple phenology cycles. A global land cover map (GLC2000) was used as a reference to extract the spatial locations of the vegetation types to infer their phenology. The median of each phenological variable was derived and a spatial majority filter was applied to the 1° × 1° grids (representing 1:250 000 Survey of India toposheet) covering the whole of India. This study presents the results derived for the evergreen, semi-evergreen, moist deciduous and dry deciduous vegetation types of India. A general trend of earlier onset of greenness at lower latitudes than at higher latitudes was observed for the natural vegetation in India.

AB - Time series of MEdium Resolution Imaging Spectrometer (MERIS) Terrestrial Chlorophyll Index (MTCI) level-3 data product, with a spatial resolution of ∼4.6 km composited at 8-day intervals for the years 2003 to 2007, were used to map the phenology of natural vegetation in India. Initial dropouts and noise in the MTCI data were corrected using a temporal moving window filter, Fourier-based smoothing using the first four harmonics was applied and then the phenological variables were extracted through a temporal iterative search of peaks and valleys in the time series for each pixel. The approach was fine-tuned to extract reliable phenological variables from the complex and multiple phenology cycles. A global land cover map (GLC2000) was used as a reference to extract the spatial locations of the vegetation types to infer their phenology. The median of each phenological variable was derived and a spatial majority filter was applied to the 1° × 1° grids (representing 1:250 000 Survey of India toposheet) covering the whole of India. This study presents the results derived for the evergreen, semi-evergreen, moist deciduous and dry deciduous vegetation types of India. A general trend of earlier onset of greenness at lower latitudes than at higher latitudes was observed for the natural vegetation in India.

U2 - 10.1080/01431161.2010.512303

DO - 10.1080/01431161.2010.512303

M3 - Journal article

VL - 31

SP - 5777

EP - 5796

JO - International Journal of Remote Sensing

JF - International Journal of Remote Sensing

SN - 0143-1161

IS - 22

ER -