Spectral Indices are nothing but the ratio between two or more wavelengths. It converts the multiple spectral band values to a single value which is easy to correlating to biophysical variables (Tucker 1979). Their sensitivity to the biophysical properties is higher than the individual spectral bands and it can be used to study the earth system that includes physical, chemical and biological process. The advantage of the indices are easy to compute because of the simple algorithms and requires minimum computation power. It can be done over the large area with minimal ancillary data to classify and extract the information what the user required. The best example is NDVI which is used over the decades to do the qualitative as well as the quantitative study of biophysical variables. The application of spectral indices in landscape ecology is quite enormous. Vegetation indices were created and validated on their usefulness in vegetation characteristics studies. The parameters like phenology, spatial distribution, extent, vegetation decomposition, vegetation fraction, productivity and stress has been studied with the help of remote sensing indices. The temporal indices could be used to quantify the forest fire, drought, desertification, deforestation and flooding.
The selection of indices depends on the parameters to be studied as well as the landscape where it presence. Example, the indices like Soil Adjusted Vegetation Index (SAVI) has a soil background adjustment factor which could reduce the soil influence and enhance the vegetation signals (Huete 1987). This is specially designed for rangeland and grassland mapping. Another example is ARVI, a modified NDVI best suitable for deserted area studies, because of its usefulness on aerosol presence. Generally the ARVI is four times less sensitive to the atmospheric effect than the NDVI (Kaufman 1992). Indices like TSAVI, MSAVI and tasseled cap index are the better indices for wetland vegetation mapping because the soil brightness and wetness plays an important role (Tiner et al 2015).
The modern day development of sensors have increased the choices of wavelength selection to study the earth. The hyperspectral sensors having narrow wavelengths which could give more accurate results on this kind studies. Even though the spectral indices have more benefits than individual wavelengths, but it has their own limitations too. The ranges of values for a single parameter may differ at different locations and seasons. Also, the accuracy of the results depends on the data quality, image preprocessing, complexity of the landscape and the quality of field data.
References
- Tiner, Ralph W., Megan W. Lang, and Victor V. Klemas. "Remote Sensing of Wetlands: Applications and Advances." (2015).
- Tucker, Compton J. "Red and photographic infrared linear combinations for monitoring vegetation."Remote sensing of Environment 2 (1979): 127-150.
- Huete, A. R., and R. D. Jackson. "Suitability of spectral indices for evaluating vegetation characteristics on arid rangelands."Remote sensing of environment2 (1987): 213-IN8.
- Kaufman, Yoram J., and Didier Tanre. "Atmospherically resistant vegetation index (ARVI) for EOS-MODIS."Geoscience and Remote Sensing, IEEE Transactions on 2 (1992): 261-270.