EVI index became extremely popular due to its ability to eliminate background and atmosphere noises, as well as its non-saturation in high biomass regions while NDVI asymptotically saturates in high biomass regions.
If EVI has such advantages over NDVI, why do we use NDVI ? In what kind of studies we should choose one over another ?
I think one of the advantage of NDVI over EVI is the data availability because of the extra blue band used in the formula, NDVI dataset can be date back from 1982 in GIMMS and GIMSS 3g. By contrast, EVI are available from the recent MODIS product. NDVI is proved to be more correlated to the leaf area index, while EVI is considered to be more accurate in GPP modeling.
Dear Deepak,
I think you can choose the index also based on site-specific conditions (vegetation abundance, climatic condition (arid to humid), etc) and performances. I found a paper by Lei Ji (USGS) and Albert J. Peters (Univ. Nebraska-Lincoln) on the performance evaluation of spectral vegetation indices, I hope this can help you.
Best!
Thanks for answering the query, but the question remain same, as we have EVI and NDVI products available since 2000 from MODIS, what should be our rationale behind choosing one over another when we want to show vegetation response to rainfall pattern? Can we just discard the idea of using EVI as NDVI is more correlated to LAI ? @Zhang
Vegetation abundance may change for a cropped field from season to season @Abliz.
Hi,
NDVI itself have much disadvandages, though by its easy way of calculation, it can be used in some global-scale works and some applied research. Now NDVI maps are mostly used as a intermediate products for calculation of more complex indices, modelling or there can be some upgrades like MNDVI as Dr. Hayder Dibs mentioned or GA-NDVI. Recently I was very impressed by the papers of Dr. Peyman Kabiri dealt with this approach.
"To compensate for the effects of NDVI saturation over high biomass areas, EVI tends to present relatively low values in all biomes and also lower ranges over semiarid sites.
The correction for aerosol impact on the final index makes use of reflectances measurements within VIS blue, not always available (the case of AVHRR or SEVIRI sensors)."
http://www.eumetrain.org/data/3/36/navmenu.php?page=4.2.0
Theoretically long time series of MODIS EVI should better relate to Vegetation response than NDVI as EVI incorporates corrections for aerosol - which may add significant noise to the signal. Besides, EVI has equally good response to vegetation with high biomass/LAI. Practically, my experience for agricultural landscape shows that when we do correlation studies between VI and rainfall pattern, the significance and value of correlation is more or less same for the two. What is more important is how good you filter the NDVI or EVI time series. The VI-Rainfall relation depend more on the type / method of filtering than NDVI or EVI per-se.
Thanks for clarification Dr. Sehgal, could you attach some relevant articles to refer, if you have any.
Hi,
The NDVI has been around for over 40 years, has a simple formula and requires only red and NIR data, all of which contribute to its ongoing popularity. However, it is highly sensitive to soil background and atmospheric effects, while the EVI was designed to be much less sensitive to these extraneous effects and, therefore, to provide a better quantification of vegetation abundance and physiological activity. The paper below reviews the development of some of the most important post-NDVI indices and provides a modified EVI equation that does not require blue band data:
http://vip.arizona.edu/documents/dataviewer/EVI2_Paper_Huete.pdfI hope it helps.
JP
NDVI has lagging time. Therefore, it is better to use EVI since it is normalizes the condition better that NDVI.
Dear Deepak,
I agree that NDVI itself is hardly useful for many research. And dont forget it also strongly tend to decrease in the presence of background high amounts of dry biomass. though it can be applicable you may think about the use of its application as an intermediate product for advanced analysis. There are many publication dealt with this. (see for instance the last link)
In any cases you shouldn't use only one index for your research. Every index and every approach has, in more or less degree, its own special features, advantages and disadvantages, so dont stop using the only one.
The following links may be useful for you
I think these two indices both useful but depend on what you intend to assess, such as different land covers, abundance or climate. Even the improvement of EVI, it still reveals time-lag with precipitation particularly.
EVI is better. It used more bands for calculation. EVI of MODIS is in access.
