3rd, Sep.

Recently, I have a tough time to deal with SiF data.

In the past, I used FLD and 3FLD methods to retrieve sif value. As I always dont care about issue of computing methods, I think FLD method, as a classic and original method to retrieve sif, should work well in our case, too.

But, you know, the relationship between sif with GPP and APAR is not satisfying.

As I have obsession in some degree and under the instruction of my supervisor, firstly, I spent a lot of time removing outliers from raw DN value to clean my data. However, it's time consuming and totally useless.

Maybe you would say, "possibly, that's the truth, and by analyzing their new relationship you could find something novel". But I tend to believe that it proves, in some degree, my retrieval of sif is not correct.

After that, I start to considering my sif retrieval method. I changed the wavelengths used to retrieve SiF. It seems that I got better results and relationships became better. But something wired.

Anyway, I dont want to talk too much details here.

Today, I read a paper about SiF calculation method. I find it's very useful. I gona share it here.

Luis Alonso, Luis Gómez-Chova, Joan Vila-Francés, etc. Sensitivity analysis of the Fraunhofer Line Discrimination method for the measurement of chlorophyll fluorescence using a field spectro radiometer. Geoscience and Remote Sensing Symposium (pp. 3756–3759). IGARSS 2007.

This paper proposed an improved method to retrieve SIF value based on standard FLD method.

Two points of this novel method are very attractive to me.

Firstly, this method shows strong robustness. this study proved that choice of 'out' band has almost no influence on sif value.

Secondly, this picture is really attractive to me. in this picture, sif value calculated by standard FLD is not coherent. My results have similar issue. However, sif value by the novel method is coherent and linear. So I'm thinking maybe by this or similar method I can get more accurate sif value~