Sampling Design for Organic Matter Fluorescence Analysis
137
to precipitation during freeze/thaw. Both Spencer et al. (2007c) and Hudson et al. (2009)
concluded that from their knowledge of original sample properties no simple relationship
could be found between initial sample characteristics and the amount of change that will
occur with freezing and subsequent thawing, and as such it was not possible to apply a correction factor to their data. Guidelines have been suggested though with respect to when it
might be appropriate to freeze samples and DOC loss on freezing has been shown to be as
high as 30% in organic rich freshwater samples and to be related to initial DOC concentration with higher initial DOC concentrations exhibiting greater loss during freeze/thaw
(Fellman et al., 2008). Fellman et al. (2008) suggest that in freshwater samples with low
DOC concentrations (<5 mg L
–1
) and/or low SUVA 254 values (<3.5–4 L mg C
−1
m
−1
) freezing is potentially a practical choice for sample preservation. However, samples with higher
DOC and/or SUVA 254 values should not be frozen if possible, as freezing these samples
not only causes a loss of concentration due to precipitation but also changes the chemical
composition and thus spectrophotometric characteristics of the sample. This is further supported by the observed lack of change in DOC concentrations after freeze/thaw in marine
samples (Tupas et al., 1994).
Although the impact of freeze/thaw appears to be greater in organic-rich waters a study
by Otero et al. (2007) investigating the effect of freezing and thawing on fluorescence and
DOC properties from a number of organic-rich sediment pore waters observed no change in
pore water fluorescence characteristics on freezing and thawing. Thus, further highlighting
the variable response of DOM to freeze/thaw as observed in the Spencer et al. (2007c) and
Hudson et al. (2009) studies. This highlights the need for researchers to evaluate the impacts
of freezing if it is to be employed as a storage method on a study- and site-specific basis.
Yamashita et al. (2010a) in a study examining optical properties of DOM in Venezuelan
tropical rivers examined the differences before and after freezing for the CDOM absorption
coefficient at 350 nm (a 350 ), the spectral slope ratio (S R ) parameter (Helms et al., 2008),
and fluorescence index (FI; McKnight et al., 2001; Jaffe et al., 2008; Cory et al., 2010) and
found minor changes of 2.5 ± 6.9%, –0.4 ± 1.5%, and 2.8 ± 2.5% respectively. On the contrary, Yamashita et al. (2010a) observed significant increases and decreases in fluorescence
intensity of different PARAFAC components after freeze/thaw in their study and so chose
not to use the frozen samples in their PARAFAC data set but showed the validity of including the CDOM absorption, S R , and FI data from the samples that were frozen. Similarly,
Gao et al. (2010) examined the effect of freeze/thaw on optical properties of DOM as
samples were stored frozen in their study of Zhejiang coastal waters (southeast China) and
found that for their measurements of interest the percent change was less than 15%. Gao
et al. (2010) therefore concluded that the main conclusions of their study were not due to
measurement biases introduced from the freeze/thaw process. Finally, Spencer et al. (2010)
in an examination of DOM dynamics in a pristine organic-rich Congolese tropical river
investigated the impact of freeze/thaw as samples were stored frozen and found that after
freezing and subsequent thawing DOC, a350, S 275–295 , S 350–400 , SR, SUVA 254 , and FI values
exhibited changes typically within analytical error and always less than ± 2%. Therefore,
although extra care is warranted in organic-rich samples if freezing is to be employed, it
137
to precipitation during freeze/thaw. Both Spencer et al. (2007c) and Hudson et al. (2009)
concluded that from their knowledge of original sample properties no simple relationship
could be found between initial sample characteristics and the amount of change that will
occur with freezing and subsequent thawing, and as such it was not possible to apply a correction factor to their data. Guidelines have been suggested though with respect to when it
might be appropriate to freeze samples and DOC loss on freezing has been shown to be as
high as 30% in organic rich freshwater samples and to be related to initial DOC concentration with higher initial DOC concentrations exhibiting greater loss during freeze/thaw
(Fellman et al., 2008). Fellman et al. (2008) suggest that in freshwater samples with low
DOC concentrations (<5 mg L
–1
) and/or low SUVA 254 values (<3.5–4 L mg C
−1
m
−1
) freezing is potentially a practical choice for sample preservation. However, samples with higher
DOC and/or SUVA 254 values should not be frozen if possible, as freezing these samples
not only causes a loss of concentration due to precipitation but also changes the chemical
composition and thus spectrophotometric characteristics of the sample. This is further supported by the observed lack of change in DOC concentrations after freeze/thaw in marine
samples (Tupas et al., 1994).
Although the impact of freeze/thaw appears to be greater in organic-rich waters a study
by Otero et al. (2007) investigating the effect of freezing and thawing on fluorescence and
DOC properties from a number of organic-rich sediment pore waters observed no change in
pore water fluorescence characteristics on freezing and thawing. Thus, further highlighting
the variable response of DOM to freeze/thaw as observed in the Spencer et al. (2007c) and
Hudson et al. (2009) studies. This highlights the need for researchers to evaluate the impacts
of freezing if it is to be employed as a storage method on a study- and site-specific basis.
Yamashita et al. (2010a) in a study examining optical properties of DOM in Venezuelan
tropical rivers examined the differences before and after freezing for the CDOM absorption
coefficient at 350 nm (a 350 ), the spectral slope ratio (S R ) parameter (Helms et al., 2008),
and fluorescence index (FI; McKnight et al., 2001; Jaffe et al., 2008; Cory et al., 2010) and
found minor changes of 2.5 ± 6.9%, –0.4 ± 1.5%, and 2.8 ± 2.5% respectively. On the contrary, Yamashita et al. (2010a) observed significant increases and decreases in fluorescence
intensity of different PARAFAC components after freeze/thaw in their study and so chose
not to use the frozen samples in their PARAFAC data set but showed the validity of including the CDOM absorption, S R , and FI data from the samples that were frozen. Similarly,
Gao et al. (2010) examined the effect of freeze/thaw on optical properties of DOM as
samples were stored frozen in their study of Zhejiang coastal waters (southeast China) and
found that for their measurements of interest the percent change was less than 15%. Gao
et al. (2010) therefore concluded that the main conclusions of their study were not due to
measurement biases introduced from the freeze/thaw process. Finally, Spencer et al. (2010)
in an examination of DOM dynamics in a pristine organic-rich Congolese tropical river
investigated the impact of freeze/thaw as samples were stored frozen and found that after
freezing and subsequent thawing DOC, a350, S 275–295 , S 350–400 , SR, SUVA 254 , and FI values
exhibited changes typically within analytical error and always less than ± 2%. Therefore,
although extra care is warranted in organic-rich samples if freezing is to be employed, it