UNDERWATER RADIATION MEASUREMENTS: CONSEQUENCES OF AN INCREASED UV-B RADIATION - Environmental UV Radiation: Impact on Ecosystems and Human Health and Predictive Models - page 192

4. Discussion
It has been shown in several papers that by far the most successful predictor of
UV attenuation in aquatic systems is dissolved organic carbon (DOC) concentration 8 .
Models have been developed to establish this dependence 7 . However, it is not clear how
general they are, mostly because the chemical nature of the DOC varies tremendously 9 .
An empirical model developed in one system might not be valid for another water
system. For temperate lakes a linear relationship between UV attenuation and DOC
fluorescence was found, whereas UV attenuation was a power function of DOC
concentration, but the reason for that was not clear 10 . Going to the extremes, highly
coloured ponds, it was shown that UVB attenuates in the top 10-20 cm of the pond. A
lack of reliable relationship between DOC and attenuation resulted from differences
throughout the season in the fraction of DOC capable of absorbing radiation and the
fraction capable of fluorescing 11 . Thus, seasonal variability or organic substances
strongly influence the UV doses at different depths.
Attenuation coefficients in the UV-B vary from 20 m -1 in strong coloured ponds
or lakes to less than 0.2 m -1 for the clearest ocean water. Likewise, for these examples,
the 10% light level for UV-B ranges more than two order of magnitude between
temperate lakes and the clearest ocean water 12 . Even between Fjords in Northern Europe
there might be one order of magnitude difference in maximum 10% level for 310 nm,
from 6.1 m at Kongsfjord (Svalbard, Norway) to 0.72 m at Inner Oslo Fjord (Norway) 2 .
There are few studies as the one performed in Trondheim Fjord where the UV
attenuation has been studied at the same site throughout the season for several years.
The variability shown in Trondheim Fjord was about one order of magnitude, which can
be compared with variability between different Fjord systems. The high level of TOC
during summer reduces the UV irradiance at certain depths compared with irradiances
in spring at the same depth. Simulations of monthly mean irradiances assuming clear
sky conditions showed that mean June irradiances at 3m depth is lower than mean
irradiance in Mars at the same depth (data not shown). Mars is the time of the year when
the Fjord was most transparent to UV.
The slope S found from the spectral dependence of k d did not change much
during the year in Trondheim Fjord, ranging from 0.012 to 0.015, with one exception.
These values are similar to what is found in other marine systems and even northern
high latitude lakes 7 . For an overview see Markager and Vincent 13 .
Uncertainty in estimating k d becomes highlighted when different
spectroradiometers are used at the same site. Results from Samnanger Fjord in Norway
with four different radiometers revealed uncertainties up to 30% in water with high UV
attenuation. An intercomparison performed by Kirk et al. 14 showed uncertainties in k d ,
specially for short UV wavelengths in optical thick lakes, of the same order of
magnitude as the Norwegian study. The same type of instrumentation participated in
both intercomparisons.
All these data indicates how important it is to take into account the attenuation of
UV in the water column and the annual variability in different aquatic systems when
consequences of climate change is discussed.