The blank used was deionised water and the sample was sea water. I used an Aqualog and a DT mini and both give slightly different results for the same sample but both give greater than 100% transmittance with negative absorbance.
I have not used the specific instruments that you describe, but from what you say it's likely one of two things: (1) the absorbance at your chosen wavelength(s) is so low, that you have a low signal to noise ratio, and thus may end up with a small negative (or positive) number that is essentially noise. (2) there is a matrix effect, where the salts in the seawater are influencing the absorbance. Typically other compounds (like salts) would decrease rather than increase transmittance, but you might create some artificial seawater solutions (CDOM free) or simpler salt solutions and see if there is an effect on measured absorbance.
A key point is that presumably you aren't actually getting greater than 100% transmittance, rather you are getting higher transmittance through the seawater than through the DIW, at your given wavelength range.
Thanks for the reply Ishi. I feel like its probably, as you said, just that I'm just seeing noise in the data. I understand that the transmittance is relative to the DIW I just though it was odd that there would be greater transmittance through the seawater samples. I will try running some salt solutions and see what happens.
It may also be a temperature effect. We typically correct the CDOM spectrum for instrument drift by subtracting an offset, typically the average between 700-750nm. Absorbance by water is temperature dependent, particularly in this wavelength range, so if your DIW reference and seawater samples are at different temperatures, you may well end up with an apparent offset. You may want to take the 680-700nm average for drift-correction. Like Isshi said, it's not real, it's just lower than your reference.