Modeling Lake Recovery Lag Times Following Influent Phosphorus Loading Reduction.

Internal feedback of nutrients may impede timely improvement in lake water quality. We describe a parsimonious, mechanistic framework for modeling lag times to recovery of phosphorus-enriched lakes, given decreases in external loading. The approach assumes first-order kinetics in a two-compartment system taking account of phosphorus storage in and loading from benthic sediments. Bayesian parameter modeling, published sediment phosphorus release rates, and a prior dynamic calibration for one lake are used to derive estimates of key parameters. Applications are developed for an example lake, as are maps displaying estimated times to attainment of a phosphorus criterion in lakes across a midwestern state, and lag time estimates for fractional water column concentration decrease averaged over HUC-8s. Mean lag times to 50 and 75% declines in water column phosphorus concentration were estimated as 13.1 and 39.0 years respectively, across more than 70,000 lentic water bodies in the continental United States.

Assessing Evidence of Phosphorus Concentration Trends in North American Fresh Waters.

The U.S. EPA's National Aquatic Resource Surveys (NARS) documented evidence of widespread, unexplained total phosphorus (TP) concentration increases in lakes and streams across the United States during the 2000 - 2012 time period. To examine the robustness of evidence for this trend, we used additional monitoring datasets to calculate rates of TP change in thousands of individual waterbodies across the U.S. during the same time frame, and compared them against TP change rates calculated in the same manner for waterbodies that were resurveyed under NARS in different years. For the additional datasets, median rates of TP change were substantially lower than median rates calculated using NARS data. To further examine differences between NARS and non-NARS results in specific waterbodies, we assembled composite datasets for 52 predominantly northern lakes that by chance had been sampled under both NARS and other sampling programs during the same time frame. Using only NARS data, the median calculated TP change rate for this set of lakes was positive, and similar to that for the larger set of 401 resurveyed NARS lakes. However, when additional sample data were included, the median calculated TP change rate for these lakes was much lower. Results suggest that increasing TP concentrations in waterbodies may not have been as ubiquitous as suggested. They also illustrate a need to supplement randomized continental-scale monitoring with detailed, site-focused investigations.

An analysis of performance models for free water surface wetlands.

Although treatment wetlands are intended to attenuate pollutants, reliably predicting their performance remains a challenge because removal processes are often complex, spatially heterogeneous, and incompletely understood. Although initially popular for characterizing wetland performance, plug flow reactor models are problematic because their parameters exhibit correlation with hydraulic loading. One-dimensional advective-dispersive-reactive models may also be inadequate when longitudinal dispersion is non-Fickian as a result of pronounced transverse gradients in velocity (preferential flow). Models that make use of residence time distributions have shown promise in improving wetland performance characterization, however their applicability may be limited by certain inherent assumptions, e.g. that transverse mixing is nil. A recently-developed bicontinuum (mobile-mobile) model that addresses some of these weaknesses may hold promise for improving wetland performance modeling, however this model has yet to be tested against real-world wetland data. This paper examines the state of the science of free water surface wetland hydrodynamics and transport modeling, discusses the strengths and weaknesses of various steady state models, and compares them to each other in terms of each model's ability to represent data sets from monitored wetlands.

Ecosystem modeling applied to nutrient criteria development in rivers.

Threshold concentrations for biological impairment by nutrients are difficult to quantify in lotic systems, yet States and Tribes in the United States are charged with developing water quality criteria to protect these ecosystems from excessive enrichment. The analysis described in this article explores the use of the ecosystem model AQUATOX to investigate impairment thresholds keyed to biological indexes that can be simulated. The indexes selected for this exercise include percentage cyanobacterial biomass of sestonic algae, and benthic chlorophyll a. The calibrated model was used to analyze responses of these indexes to concurrent reductions in phosphorus, nitrogen, and suspended sediment in an enriched upper Midwestern river. Results suggest that the indexes would respond strongly to changes in phosphorus and suspended sediment, and less strongly to changes in nitrogen concentration. Using simulated concurrent reductions in all three water quality constituents, a total phosphorus concentration of 0.1 mg/l was identified as a threshold concentration, and therefore a hypothetical water quality criterion, for prevention of both excessive periphyton growth and sestonic cyanobacterial blooms. This kind of analysis is suggested as a way to evaluate multiple contrasting impacts of hypothetical nutrient and sediment reductions and to define nutrient criteria or target concentrations that balance multiple management objectives concurrently.