HR: 16:35h
AN: H42K-03    [PDF]
TI: Evaluation of EMMA Results Using End-Members Measured from a Reference Period, Green Lakes Valley, Colorado Front Range
AU: Caine, N
EM: cainen@spot.colorado.edu
AF: Institute of Arctic and Alpine Research, University of Colorado, Boulder, CO 80309 United States
AU: Caine, N
EM: cainen@spot.colorado.edu
AF: Department of Geography, University of Colorado, Boulder, CO 80309 United States
AU: * Liu, F
EM: fengjing@snobear.colorado.edu
AF: Institute of Arctic and Alpine Research, University of Colorado, Boulder, CO 80309 United States
AU: * Liu, F
EM: fengjing@snobear.colorado.edu
AF: Department of Geography, University of Colorado, Boulder, CO 80309 United States
AU: Williams, M W
EM: markw@snobear.colorado.edu
AF: Institute of Arctic and Alpine Research, University of Colorado, Boulder, CO 80309 United States
AU: Williams, M W
EM: markw@snobear.colorado.edu
AF: Department of Geography, University of Colorado, Boulder, CO 80309 United States
AB: Isotopic and geochemical tracing, including end-member mixing analysis (EMMA), has greatly improved the understanding of streamflow sources and pathways during the last two decades. A continuing problem with hydrograph separations, however, is the inability to adequately measure all end-members over time. It would be advantageous if end-members identified from expensive sampling campaigns can be used in different years or in different sites. Here we evaluate the use of data rank analysis as a diagnostic tool to determine if end-members and their ionic concentrations identified using EMMA in 1996 can be used for 1997 at the 8-ha Martinelli and the 220-ha Green Lake 4 (GL4), alpine catchments in the Green Lakes Valley, Colorado Front Range. Using principal component analysis (PCA), eigenvectors and eigenvalues were extracted for the streamflow chemistry of 1996 and 1997 from the correlation matrices of 6 conservative tracers (ANC, Ca$^{2+}$, Mg$^{2+}$, Na$^{+}$, SO$_{4}$$^{2-}$, and Si). The results from the diagnostic tool indicated that rank (mixing subspace) of the streamflow chemistry data was 2 (rank + 1 = number of end-member) based on random patterns of the residuals between observation and prediction using their own eigenvectors for both 1996 and 1997 at both catchments. Also, the eigenvectors derived from 1996 worked as well as those from 1997 in the data rank analysis for the 1997 data at both catchments, indicating that ionic ratios in the end-members of 1997 were consistent with those of 1996. Using the EMMA end-members from 1996, three flowpaths were determined for 1997. Measured values in streamflow were reproduced reasonably well, with R$^{2}$ values close to 1 for all conservative tracers. Thus, it is believed that the processes controlling the streamflow chemistry were similar for 1996 and 1997 and all end-members from 1996 could be simply used in EMMA for 1997 at the Martinelli and GL4 catchments.
DE: 1871 Surface water quality
SC: Hydrology [H]
MN: 2003 Fall Meeting