PARAFAC Analysis results from Stephen et al., 2016 (STE)  PARAFAC Analysis results from Zhou et al., 2016 (JGLR)  PARAFAC Analysis results from Xu and Guo., 2016 (WR) 
Abnormality: abrupt decline of some peaks



Compared with literatures#
Walker et al. 2013#

Stedmon & Markager 2005#

Stedmon et al 2003#

Hypothesis: Cut EEMs improperly
FirstOrderCut = [-40 15];
SecondOrderCut = [-40 15];
InsertedData = [NaN 0];
for i = 1:2
if FirstOrderCut(i) ~= NaN
for j = 1: length(Ex)
k = find (Em<Ex(j)-FirstOrderCut(i));
X(:,k,j) = InsertedData(i);
end
end
if SecondOrderCut (i) ~= NaN
for j = 1:length(Ex)
k = find (Em>Ex(j)*2.05+SecondOrderCut(i));
X(:,k,j) = InsertedData(i);
end
end
endDemostration:
Solution:
InsertedData = [NaN NaN]Results:
Introduction of samples#
Samples:#
Yukon River downstream
Sampling time:#
2004-2005
Sample categories:#
colloidal organic matter (>1kDa), low molecular weight organic matter (<1kDa), dissolved organic matter (<0.45e-6 m)
Before program revised:#
3 components decomposed by PARAFAC program:

4 components from PARAFAC:

After revised:#
3 components decomposed by PARAFAC program:

4 components from PARAFAC:

Comparison#
| Component | before program revised | After revised | ||
|---|---|---|---|---|
| Excitation (nm) | Emission (nm) | Excitation (nm) | Emission (nm) | |
| 3Components | ||||
| C1 | 280 | 485 | 260 | 480 |
| C2 | 250 | 480 | 250 | 420 |
| C3 | 220 | 405 | <250 | 330 |
| 4Components | ||||
| C1 | 230 | 450 | 250 | 420 |
| C2 | 310 | 410 | 270 | 500 |
| C3 | 385 | 480 | 260/375 | 480 |
| C4 | <220 | 380 | <250 | 330 |