Estimating Total Phosphorous and Total Suspended Solids in
- Slides: 8
Estimating Total Phosphorous and Total Suspended Solids in Freshwater Streams from Turbidity Data Emily Saad EAS 4480 Oral Presentation 27 April 2010
Background • Study Site: – Little Bear River (Northern Utah) – Drains into Cutler Reservoir • Eutrophic hypoxic due to high phosphorous levels • TMDL established in reservoir’s tributaries –. 05 mg/L – Not yet met by Little Bear – Paradise (upper watershed) and Mendon (lower watershed) • Data Collection – High Frequency: In situ sensors • Physical parameters – Water level, temperature, p. H, conductivity, DO, turbidity • Some chemical species (recent technology using UV/Vis) – NO 3 -, NO 2 -, chlorophyll – Infrequent: Grab Sampling • Short term variability, diurnal trends omitted • TP digested analyzed in lab – Surrogate Measurements • Objective – In situ turbidity as a surrogate for more accurate estimation of TP concentrations and TSS
p<. 05 r=. 84 p<. 05 r=. 95 p<. 05 r=. 69
Model Used to Simulate TP and TSS -all based on linear regression -y=a 0+a 1 x 1+Z(B 0+B 1 x 1)+ei - the probability plot of the residuals are not normally distributed suggesting that the assumed parametric distribution (normal) in the regression is incorrect Jones, et al. (2008)
Site Chi Square Test SIM OBS F Test (Variances) SIM OBS TSS@ PAR Data not normally distributed Variances are statistically different TP @ PAR Data not normally distributed Variances are statistically different TSS @ MEN Data not normally distributed Variances are statistically different
Bootstrap Analyses: TSS @ PAR TP @ PAR 9. 6 E 5± 1. 9 E-4 6. 1 E-5± 3. 3 E-4 13. 8±. 11 . 027± 7. 8 E-5 TP @ MEN TSS @ MEN -. 044±. 052 -. 12±. 03 . 13±. 004 48. 4± 1. 03
Bootstrap Regression: Simulated vs. Observed
Conclusions -TP and TSS both significantly correlated with turbidity as these sites -re-evaluate model not assuming normal distributions Jones, et al. (2008).
