There is simply an operational fitting that worked on several remote sensing sources, probably dr Werdell can assess the relative benefit of reversing reference wavelenght and there is a reference adr in http://oceancolor.gsfc.nasa.gov/REPROCESSING/R2009/ocv6/

Thank you Guido Bongi, i refered to the link provided. So for the best fitting, these ratios are used. But is there any theoretical importance to use the ratios as it is(i.e. 443/555 and not 555/443)?? if we leave the curve fitting, can we use the reverse band ratios for retrieval of chlorophyll??

The issue is taken under scrutiny in Tin-Wei et al. and refs therein, but you should demonstrate the profit of reversing ratios by same error decrease. The figure of merit is experimental but remote sensing shares procedures for operational agreement, after international panels trials. https://www.osapublishing.org/view_article.cfm?gotourl=https%3A%2F%2Fwww.osapublishing.org%2FDirectPDFAccess%2F1240E888-0E2B-E9A3-6D798F65A78E3DBC_268864%2Foe-21-21-24929.pdf%3Fda%3D1%26id%3D268864%26seq%3D0%26mobile%3Dno&org=

At the outset, let me tell you that  such band ratios are applicable where interaction of visible light is linear. Means chl-a pigment is the only optically active sub. This means, such algorithm is applicable in open ocean. Now why blue to green ratio. If you look at the spectral absorption of chl-a, you will find two abs peaks. One at the blue wavelength and other at red wavelength with a  trough in between. This trough is at green wavelength (in the above case 555 nm). A trough indicates least absorption and peak indicate maximum absorption. Since in the open ocean chl-a is the OAC harvesting blue, you will find an inverse relation between blue and green. Means more chl-a more absorption of blue light and less absoprtion/more scattering of green light. As I said initially as the interaction of light (visible spectrum) in the open ocean is linear, a ratio between blue to green will be an index of chl-a conc. Hope I answered you.

Thank you so much Harilal Menon sir. I understood the answer completely. So, if in the blue wavelength, the absorption of Chlorophyll is more, the light reflected is less, which implies the energy recorded at the sensor ( at sensor radiance)

1. in the blue wavelength must be less

2.   in green wavelength must be more.

for same location. 

For example: Let us consider a location in open ocean, where there is high amount of chlorophyll. According to the absorption curve, let us assume a value of 200 (at sensor radiance) for blue wavelength, 300 (at sensor radiance) for green wavelength. The ratio of blue to green wavelengths = 200/300=0.666, whereas the reverse gives us 1.5. As i am mainly referring to satellite data, higher values at a pixel can be better interpreted than less values. so green to blue wavlength must be used. But they are used in reverse.

According to the previous answer provided by Guido Bongi, the ratio of 443/555 is used instead of 555/443 because it fits well with field data and operational measurements.

I send back link as requested, shall you find it in keys ocean color nasa:

 http://oceancolor.gsfc.nasa.gov/REPROCESSING/R2009/ocv6/