Mapping grass communities based on multi-temporal Landsat TM imagery and environmental variables

Information on the spatial distribution of grass communities in wetland is increasingly recognized as important for effective wetland management and biological conservation. Remote sensing techniques has been proved to be an effective alternative to intensive and costly ground surveys for mapping grass community. However, the mapping accuracy of grass communities in wetland is still not preferable. The aim of this paper is to develop an effective method to map grass communities in Poyang Lake Natural Reserve. Through statistic analysis, elevation is selected as an environmental variable for its high relationship with the distribution of grass communities; NDVI stacked from images of different months was used to generate Carex community map; the image in October was used to discriminate Miscanthus and Cynodon communities. Classifications were firstly performed with maximum likelihood classifier using single date satellite image with and without elevation; then layered classifications were performed using multi-temporal satellite imagery and elevation with maximum likelihood classifier, decision tree and artificial neural network separately. The results show that environmental variables can improve the mapping accuracy; and the classification with multitemporal imagery and elevation is significantly better than that with single date image and elevation (p=0.001). Besides, maximum likelihood (a=92.71%, k=0.90) and artificial neural network (a=94.79%, k=0.93) perform significantly better than decision tree (a=86.46%, k=0.83).

[1]  Russell G. Congalton,et al.  Assessing the accuracy of remotely sensed data : principles and practices , 1998 .

[2]  G. Hill,et al.  Vegetation mapping of a tropical freshwater swamp in the Northern Territory, Australia: A comparison of aerial photography, Landsat TM and SPOT satellite imagery , 2001 .

[3]  I. Vogiatzakis,et al.  Environmental factors and vegetation composition, Lefka Ori massif, Crete, S. Aegean , 2003 .

[4]  P. Adam,et al.  A review of wetland inventory and classification in Australia , 1995, Vegetatio.

[5]  S. A. Samson,et al.  Two indices to characterize temporal patterns in the spectral response of vegetation , 1993 .

[6]  Jennifer A. Miller,et al.  Modeling the distribution of four vegetation alliances using generalized linear models and classification trees with spatial dependence , 2002 .

[7]  D. Roberts,et al.  A comparison of methods for monitoring multitemporal vegetation change using Thematic Mapper imagery , 2002 .

[8]  Ross S. Lunetta,et al.  Application of multi-temporal Landsat 5 TM imagery for wetland identification , 1999 .

[9]  Thierry Toutin,et al.  Geometric Correction of Remotely Sensed Images , 2003 .

[10]  V. Radeloff,et al.  Phenological differences in Tasseled Cap indices improve deciduous forest classification , 2002 .

[11]  W. G. Howland.  Multispectral aerial photography for wetland vegetation mapping. , 1980 .

[12]  A. Haraguchi,et al.  Effects of scale-dependent factors on herbaceous vegetation patterns in a wetland, northern Japan , 2004, Ecological Research.

[13]  M. J. Dallwitz,et al.  Grass Genera of the World , 1992 .

[14]  Robert A. Schowengerdt,et al.  A detailed comparison of backpropagation neural network and maximum-likelihood classifiers for urban land use classification , 1995, IEEE Trans. Geosci. Remote. Sens..

[15]  A. Hirzel,et al.  Which is the optimal sampling strategy for habitat suitability modelling , 2002 .

[16]  N. Mathews,et al.  Floodplain vegetation phenology in the Southeast USA: Optimizing the timing of aerial imagery acquisition , 1990, Wetlands Ecology and Management.

[17]  Russell G. Congalton,et al.  Classification of multi-temporal SPOT-XS satellite data for mapping rice fields on a West African floodplain , 1998 .

[18]  J. Campbell.  Introduction to remote sensing , 1987 .

[19]  Peter T. Wolter,et al.  Improved forest classification in the northern Lake States using multi-temporal Landsat imagery , 1995 .

[20]  S. Prasher,et al.  Classification of hyperspectral data by decision trees and artificial neural networks to identify weed stress and nitrogen status of corn , 2003 .

[21]  Yong Wang,et al.  Delineation of inundated area and vegetation along the Amazon floodplain with the SIR-C synthetic aperture radar , 1995, IEEE Trans. Geosci. Remote. Sens..

[22]  E Brown de Colstoun,et al.  National Park vegetation mapping using multitemporal Landsat 7 data and a decision tree classifier , 2003 .