water quality sampling, load calculation
This study investigated the precision and accuracy of the two load calculation techniques. The study compared total phosphorus loads calculated by integration of Arkansas Water Resources Center (AWRC) intensive sampling data to loads calculated by a regression technique (rating curve) using fewer data. The 1998 AWRC dataset from the Illinois River at Arkansas Highway 59 was sub-sampled in a manner to simulate fixed period monitoring schemes supplemented with storm sampling. The ESTIMATOR software program was used to calculate loads. These loads were compared to the integrated load. The error of the integrated load when the variation in concentration between samples is not linear and the sensitivity of the integrated load to sampling interval were also investigated. The results show that the central tendency of the ESTIMATOR loads is accurate when storm data are included, but that the 95% confidence interval represents up to +/- 30-40% difference from the integrated load for individual estimates. More frequent sampling and more samples lead to more accurate loads. The results indicate that the central tendency of load estimates would be accurate for a method that uses a regression model with 32 or more samples including storm samples.
Soerens, T. S. and Nelson, M. A.. 2003. Evaluation of Sampling Strategies on Load Estimation For Illinois River at Highway 59. Arkansas Water Resources Center, Fayetteville, AR. MSC306. 10