Dear Svetla!

I'm not sure whether I get your question right, so in my attempt to answer you I might miss the point, but at least upon your (eventual) feedback we will clear that. Since you have mentioned "infected plants" at the end of your post, I presume that you are particulalrly interested in using RS data to estimate extent of damage in plants. To simplify to extreme we can say that all vegetation indices do the same, and that is: estimating amount of living plant biomass (based on quantity of clorophyl). Therefore, you can use same indices for estimating amount (and type - e.g. forests has higher VI values in comparison to graslands) of present vegetation, or to quantify damage on vegetation (e.g. defoliation in forests, or infection that will cause lower concetration of clorophyle in leaves). If you take a look at formula of some of those indices (in this paper you can find some of them http://journals.sfu.ca/apan/index.php/apan/article/viewFile/30/pdf_14 ) you will see that they are always dealing with near infrared and red wavelengths, because of their reflectance from plants. I don't know what is exactly your research problem, but spatial resolution of used RS data could be more critical than used vegetaion index. Hope I have helped at least a little with my answer.

All the best

Sven

I have data produced by remote sensing applied to healthy plants and to plants infected by different viruses. I know yet that the healthy plants has lower spectral characteristics in green area and know the biological explanation of this. Sum of the viruses keep it spectral characteristic higher than the healthy one other have lower in NIR area - I still have not found explanation of this.I calculated many VI which I have found in literature. But in all articles I have read, authors just say - this index increases other decreases compare to control. I would like to know - if I could estimate how much the chlorophyll decreases on the base of how much he index increases. NDVI index shows decrease of chlorophyll - I saw many articles about this. I would like to know which of the other indexes preset changes in othe physiological or other parameters of the plants growth. For me is very important to enrich my conclusions for the healthy status of observed plants.

Thank you to all of you that answeredme

Dear Svetla,

Directly linking increase or decrease in NDVI or other Indexes to Chlorophyll concentration increase or decreases is very tricky/difficult. If you are working on space borne sensors then making such a conclusion would be difficult. You would have to get exact correlation between NDVI and Chlorophyll conc for the plant you are working on 

All the best in your work

Dharmendra

Just as complement to what has been already said.

Vegetation Indexes are just a measure of the spectral response of the vegetated surface, in some cases normalized through some other to "highlight" the vegetation; as for example in the case of PVI, where it is rescaled against the response of the soil. To put it plainly it is a measure of the ammount of sun radiation of given wavelenghts (for the case of passive sensors) that is reflected. Nothing more. They do not measure directly any other physical aspect.That is why most authors only say under such and such conditions this index increases or decreases.

Of course you can try to stablish a relation between this spectral response and any other variable; in your case the chlorophyll concentration, but it can also be biomass, surface leaf area, and so on. Most likely you will get some correlation measure between your proposed variables, but it is then your job to assess if this correlation is at all meaningfull in physical / theorethical and empirical terms, the degree of confiability  and its usefullness. Hope it helps.

Rafael

Thank you to the all. I have read all this cortical that why I decided to ask you this question. I looking for some methods for early detection of the viral or some root pathogens which could help to find the stress in plants with out we to "kill" the plant. I tough if I develop a large data base with repeated measurement and calculate set of indexes, I could get some decision which of them  a representative of existing of a definite pathogen or the level of the damages caused by it. In fact you helped me, thank you

Dear Svetla,

The issue raised in your last message (detecting pathogens) is rather different, and much more specific, than your original question (meaning of vegetation indices). As pointed out by Rafael, vegetation indices are rather arbitrary (and largely obsolete) combinations of spectral bands that exhibit at best limited correlations with plant properties collected in field campaigns. And when such relations have been established, they are generally applicable only to the geographical area and time periods of those campaigns.

Coming back to your much more pointed question on the pathogens: the first thing to do would be to establish whether the presence or the effect of these pathogens has any noticeable impact on the bidirectional spectral reflectance of the affected canopies (hence in either the color or the anisotropy of the plant canopy), at the spatial scales and resolutions sampled by space borne instruments. If the canopy reflectance does not change as a result of the pathogen attack (or if the change is too small to be noticeable from space), then remote sensing is not an appropriate approach for this problem, irrespective of what index or other 'trick' you may want to use...

Assuming now that the reflectance of plants is indeed sufficiently affected by pathogens to be measurable, then the next step would be to characterize those changes, and to design a method to highlight those particular changes (not any generic change in plant properties). Hence, you should develop your own approach to address this particular problem: there is no reason to believe that any odd generic vegetation index is relevant or optimized for your specific interest.

To explore these ideas further, you may want to look at this paper:

Verstraete, M. M. and B. Pinty (1996) 'Designing optimal spectral indices for remote sensing applications', IEEE Transactions on Geoscience and Remote Sensing, 34, 1254–1265.

Good luck with your investigations. Michel.

NDVI and SR

thank to all - you were very helpfull for me. Thanks