|
<br />74
<br />
<br />(GARCIA-HERNANDEZ ET AL.
<br />
<br />
<br />~~- --------
<br />
<br />----F--r------ ...
<br />
<br />
<br />~
<br />
<br />. > 2.5 uglg Se cone, in 8M
<br />e < 2,5 uglg Se cone. in 8M
<br />I"V Streams
<br />. Wetland area
<br />
<br />~
<br />
<br />10
<br />
<br />o
<br />
<br />10
<br />
<br />20 Kilometers
<br />
<br />Figure 2. Distribution of Se concentrations in bottom material (BM) cores ( < 63 Ilm) from 41
<br />sampling sites in the Colorado River delta.
<br />
<br />For comparisons, we grouped the sites according to their principal so~rce of water,
<br />which was: (1) Agricultural runoff: Cienega de Santa Clara, E1 IndlO, Zacatecas
<br />drain, Campo Rafael, Ayala drain, El Mayor, Hardy River and Cucapa north; (2)
<br />Colorado River water: Colorado River, Bocana and Cucapa south; and (3) other: EI
<br />Doctor and geothermal lagoons. .' .
<br />EI Doctor and geothermal lagoons group had small~r concentraoons o~ selemum m
<br />bottom material compared to sites influenced by nver waters (one-sIded p-value
<br />< 0.001 from one-sample z-test, z = 4.8, df. = 20) or agricultural drains (one-sided
<br />p-value < 0.001 from one-sample z-t~st, ~ = 5.6, df. = 23).. Concentrati?n of selenium
<br />in bottom material was greater at sItes mfluenced by agncultural dramage (n = 20,
<br />geom. mean = 1.811g g-I) than at sites influenced by river water (n = 19, geom. mean
<br />= 1.3 I1g g-I) (Fig. 3) (one-sided p-value = 0.03 from two-sample I-test, t = 2,1,
<br />df. = 37).
<br />
<br />Dynamics of selenium in Ihe colorado river delta wellands
<br />
<br />Redox potential (Eh in mV) was higher (positive) in bottom material from river water
<br />sources (n = 19, mean = 45 mV) than from bottom material derived from agricultural
<br />runoff (n = 20, mean = - 118 mV) (one-sid~d p-value < 0.?0~1 from tw~-sample
<br />I-test, I = 5.2, df. = 37). Concentration of Se In bott?m ~atenal In~~eased WIth lower
<br />(negative) values of redox potential, and decreased WIth higher (pOSI~ve) redo~ poten-
<br />tials (Fig. 4) four outliers (not shown in Fig. 4) from the C:olorad? River are d!scussed
<br />below (If = 0.53, FI.34 = 39.5; p-value < 0.0001 from a SImple linear regressIon),
<br />
<br />
<br />SELENIUM, SELECTED INORGANIC ELEMENTS FROM COLORADO RIVER DELTA 75
<br />
<br />1.5
<br />
<br />I
<br />t
<br />~
<br />i
<br />I
<br />Vi
<br />t
<br />I
<br />I
<br />!
<br />I
<br />!
<br />t
<br />~
<br />1
<br />I
<br />I
<br />t
<br />!
<br />
<br />C:.'
<br />c.'l
<br />W
<br />
<br />
<br />.. 1,0
<br />'C
<br />III
<br />ii
<br />E
<br />E
<br />o
<br />::: 0'5
<br />.8
<br />.5
<br />..,
<br />CI)
<br />:5
<br />0.0
<br />
<br />E.'.~')
<br />
<br />
<br />.
<br />
<br />I. -',
<br />GO
<br />
<br />Percentiles:
<br />
<br />_ 90th
<br />: 75th
<br />
<br />
<br />B~
<br />
<br />
<br />--'- 10th
<br />
<br />-0'5
<br />
<br />Agriculture influenced sites River influenced sites
<br />
<br />~igure 3. Comparison between the concentration of selenium in bottom material from sites
<br />mfluenced by agricultural runoff to sites influenced by Colorado River water.
<br />
<br />Water pH was high~r (~ore basic) at si~es influenced by agricultural drains (n = 20,
<br />mean = 8.4) than at sItes mfluenced by nver waters (n = 19, mean = 8,1) (one-sided
<br />p-value < 0.002 from two-sample I-test, t = 3.3, elf. = 37). Concentration of selenium
<br />in bottom material increased with water pH, excluding two outliers from the Colorado
<br />River with high pH and low selenium concentration (FI 36 = 8.3; p-value = 0.006 from
<br />a simple linear regression). .
<br />Other explanatory variables were the clay, silt and sand content of the bottom
<br />material. Selenium concentration increased with the clay (F138 = 5.6; p-value = 0.02
<br />from a simple linear regression) and silt content of the sample (F1,38 = 4.4;
<br />p-value = 0.04 from a simple linear regression). Selenium decreased with the sand
<br />content of the sample (F1,38 = 6.1; p-value = 0.02 from a simple linear regression). The
<br />am?unt of c1a~ was greater in bottom material collected from sites influenced by
<br />agncultura1 drains (n = 20, mean = 26%) compared with sites influenced by river water
<br />; (n = 19, mean = 10%) (one-sided p-value = 0.001 from two-sample t-test, t = 3.6
<br />I df. = 37). This was also true for silt, which was greater in agricultural runoff site~
<br />'I (n = 20, mean = 40:-) compared to sites influenced by river waters (n = 19,
<br />mean = 13%) (one-sIded p-value < 0.0001 from two-sample I-test, 1= 5.0, df. = 37).
<br />I The opposite occurred with sand. Sites influenced with river water had more
<br />! sand percentage (n = 19, mean = 77%) than sites influenced by agricultural drains
<br />I ~~. =: ;o'h~ean = 34%) (one-sided p-value < 0.0001 from two-sample t-test, t = 5.0,
<br />
<br />! Percent organic carbon in bottom material was also related with selenium concentra-
<br />tion. Higher selenium concentrations were detected in samples with a high organic
<br />i carbon content (FI.38 = 6.5; p-value = 0.01 from a simple linear regression). More
<br />I... organic carbon was detected in bottom material from agriculture runoff sites
<br />(n = 20, mean = 1,3%) than from bottom material from river sites (n = 19,
<br />mean = 0.5%) (one-sided p-value = 0.0002 from two-sample I-test, I"" 4.2, elf. = 37).
<br />
<br />I
<br />
|