Large Biases in Regression‐Based Constituent Flux Estimates: Causes and Diagnostic Tools

It has been documented in the literature that, in some cases, widely used regression‐based models can produce severely biased estimates of long‐term mean river fluxes of various constituents. These models, estimated using sample values of concentration, discharge, and date, are used to compute estimated fluxes for a multiyear period at a daily time step. This study compares results of the LOADEST seven‐parameter model, LOADEST five‐parameter model, and the Weighted Regressions on Time, Discharge, and Season (WRTDS) model using subsampling of six very large datasets to better understand this bias problem. This analysis considers sample datasets for dissolved nitrate and total phosphorus. The results show that LOADEST‐7 and LOADEST‐5, although they often produce very nearly unbiased results, can produce highly biased results. This study identifies three conditions that can give rise to these severe biases: (1) lack of fit of the log of concentration vs. log discharge relationship, (2) substantial differences in the shape of this relationship across seasons, and (3) severely heteroscedastic residuals. The WRTDS model is more resistant to the bias problem than the LOADEST models but is not immune to them. Understanding the causes of the bias problem is crucial to selecting an appropriate method for flux computations. Diagnostic tools for identifying the potential for bias problems are introduced, and strategies for resolving bias problems are described.     

Weighted Regressions on Time,Discharge, and Season (WRTDS), with an Application to Chesapeake Bay River Inputs1

Hirsch, Robert M., Douglas L. Moyer, and Stacey A. Archfield, 2010. Weighted Regressions on Time, Discharge, and Season (WRTDS), With an Application to Chesapeake Bay River Inputs. Journal of the American Water Resources Association (JAWRA) 46(5):857-880. DOI: 10.1111/j.1752-1688.2010.00482.x Abstract: A new approach to the analysis of long-term surface water-quality data is proposed and implemented. The goal of this approach is to increase the amount of information that is extracted from the types of rich water-quality datasets that now exist. The method is formulated to allow for maximum flexibility in representations of the long-term trend, seasonal components, and discharge-related components of the behavior of the water-quality variable of interest. It is designed to provide internally consistent estimates of the actual history of concentrations and fluxes as well as histories that eliminate the influence of year-to-year variations in streamflow. The method employs the use of weighted regressions of concentrations on time, discharge, and season. Finally, the method is designed to be useful as a diagnostic tool regarding the kinds of changes that are taking place in the watershed related to point sources, groundwater sources, and surface-water nonpoint sources. The method is applied to datasets for the nine large tributaries of Chesapeake Bay from 1978 to 2008. The results show a wide range of patterns of change in total phosphorus and in dissolved nitrate plus nitrite. These results should prove useful in further examination of the causes of changes, or lack of changes, and may help inform decisions about future actions to reduce nutrient enrichment in the Chesapeake Bay and its watershed.     

DETECTABILITY OF STEP TRENDS IN THE RATE OF ATMOSPHERIC DEPOSITION OF SULFATE1

A method is presented to assist policy makers in determining the combination of number of sampling stations and number of years of sampling necessary to state with a given probability that a step reduction in atmospheric deposition rates of a given magnitude has occurred at a pre-specified time. This pre-specified time would typically be the time at which a sulfate emission control program took effect, and the given magnitude of reduction is some percentage change in deposition rate one might expect to occur as a result of the emission control. In order to determine this probability of detection, a stochastic model of sulfate deposition rates is developed, based on New York State bulk collection network data. The model considers the effect of variation in precipitation, seasonal variations, serial correlation, and site-to-site (cross) correlation. A nonparametric statistical test which is well suited to detection of step changes in such multi-site data sets is developed. It is related to the Mann-Whitney Rank-Sum test. The test is used in Monte Carlo simulations along with the stochastic model to derive statistical power functions. These power functions describe the probability of detecting (α=0.05) a step trend in deposition rate as a function of the size of the step-trend, record length before and after the step-trend, and the number of stations sampled. The results show that, for an area the size of New York State, very little power is gained by increasing the number of stations beyond about eight. The results allow policy makers to determine the tradeoff between the cost of monitoring and time required to detect a step-trend of a given magnitude with a given probability.