Journal of Hydrologic Engineering

A common phenomenon observed in natural and constructed wetlands is short-circuiting of flow and formation of stagnant zones that are only indirectly connected with the incoming water. Biogeochemistry of passive areas is potentially much different than that of active zones. In the research reported in this paper, the spatial resolution of a previously developed wetland nutrient cycling model was improved in order to capture the spatial variability of concentrations and reaction rates regarding nitrogen and carbon cycles throughout active and passive zones of wetlands. The upgraded model allows for several compartments in the horizontal domain, with all neighboring compartments connected through advective and dispersive/diffusive mass transport. The model was applied to data collected from a restored wetland in California that was characterized by the formation of a large stagnant zone at the southern end of the wetland due to close vicinity of the inlet and outlet structures in the northern end. Mass balance analysis revealed that over the course of the research period, about $23.4\pm 3.9\%$ of the incoming total nitrogen load was removed or retained by the wetland. It was observed that mass of all exchanges (physical and biogeochemical) regarding nitrogen cycling decreased along the activity gradient from active to passive zones. Model results also revealed that anaerobic processes become more significant along the activity gradient towards passive areas.

At the fringe of Everglades National Park in southwest Florida, United States, the Ten Thousand Islands National Wildlife Refuge (TTINWR) habitat has been heavily affected by the disruption of natural freshwater flow across the Tamiami Trail (U.S. Highway 41). As the Comprehensive Everglades Restoration Plan (CERP) proposes to restore the natural sheet flow from the Picayune Strand Restoration Project area north of the highway, the impact of planned measures on the hydrology in the refuge needs to be taken into account. The objective of this study was to develop a simple, computationally efficient mass balance model to simulate the spatial and temporal patterns of water level and salinity within the area of interest. This model could be used to assess the effects of the proposed management decisions on the surface water hydrological characteristics of the refuge. Surface water variations are critical to the maintenance of wetland processes. The model domain is divided into 10 compartments on the basis of their shared topography, vegetation, and hydrologic characteristics. A diversion of $+10\%$ of the discharge recorded during the modeling period was simulated in the primary canal draining the Picayune Strand forest north of the Tamiami Trail (Faka Union Canal) and this discharge was distributed as overland flow through the refuge area. Water depths were affected only modestly. However, in the northern part of the refuge, the hydroperiod, i.e., the duration of seasonal flooding, was increased by 21 days (from 115 to 136 days) for the simulation during the 2008 wet season, with an average water level rise of 0.06 m. The average salinity over a two-year period in the model area just south of Tamiami Trail was reduced by approximately 8 practical salinity units (psu) (from 18 to 10 psu), whereas the peak dry season average was reduced from 35 to 29 psu (by 17%). These salinity reductions were even larger with greater flow diversions ($+20\%$). Naturally, the reduction in salinity diminished toward the open water areas where the daily flood tides mix in saline bay water. Partially restoring hydrologic flows to TTINWR will affect hydroperiod and salinity regimes within downslope wetlands, and perhaps serve as a management tool to reduce the speed of future encroachment of mangroves into marsh as sea levels rise.

This study introduces two artificial neural network (ANN)-based methodologies to predict hourly water levels (WLs) in wetlands characterized by water tables at or near the surface that respond rapidly to precipitation. The first method makes use of hourly precipitation data and WL data from nearby sites. The second method is a combination of ANN, recursive digital filter, and recession curve method and does not require any nearby site. The proposed methods were tested at two headwater wetlands in coastal Alabama. Site 17 had two nearby sites whose WLs were highly correlated with Site 17’s. The root-mean-square error and Nash–Sutcliffe efficiencies were 2.9 cm and 0.98, respectively, when the first method was applied to Site 17. The second method was tested at Site 32. For this, the WL time series was separated into quick- and slow-response components. A combination of ANN and base-flow separation methods proved to be very efficient for WL prediction at this site, especially when the duration of quick-response components of individual events was less than 6 h. The proposed methodologies, therefore, proved useful in predicting WLs in wetlands dominated by both surface water and groundwater.

Understanding the spatio-temporal distribution of treatment wetland functions is critical for optimizing contaminant removal performance. This study presents the development of a spatially distributed, coupled hydrodynamic-biogeochemical model to simulate phosphorus (P) dynamics in Stormwater Treatment Area 1 West (STA1W), a large cell-network treatment wetland in South Florida. The STAs receive most of their inflow from P-enriched agricultural drainage waters. The model dynamically simulates internal hydrology, and P transport and transformation within treatment cells that are connected by hydraulic structures. The modeled biogeochemical transformations are based on primary mechanisms regulating P behavior in soils and the water column. The model was compared to measured water levels (percent model error, $\mathrm{PME}=14.4\%$), discharge ($\mathrm{PME}=7.3\%$), chloride and P concentrations in the water column ($\mathrm{PME}=13\%$ and 18.6%), and soil P at multiple locations of each treatment cell. Then, the model was used to characterize the sensitivity of total P concentrations in the water column to uncertainties in habitat-specific P cycling model parameterizations. Of the parameters tested, P settling rate by submerged aquatic vegetation has the largest impact on the water column total P (TP) concentrations. The calibrated model was further applied to explore the effects of changes in vegetation pattern in response to various inflows and TP loadings on the water column P concentrations and soil P storage. Improved treatment effectiveness was found when areal vegetation coverage was complete, with emergent and submerged vegetation in upstream and downstream treatment cells, respectively. This reduced the average outlet TP concentration 60%, from 35.9 ppb under existing operational conditions to 14.5 ppb. Similar reductions ($\sim 66\%$) in outlet TP concentration were also obtained by further reducing hydraulic and TP loadings after five years of operation. The modeling approach and findings of this study could be of interest to water managers and the wetland modeling community.

In this article, extension and application to variably saturated wetland conditions of a process-based wetland model is demonstrated. The new model described in this article is an improved version of an earlier model, which was only capable of capturing nutrient dynamics in continuously ponded wetlands. The upgraded model is capable of simulating nutrient cycling and biogeochemical reactions in both ponded and unsaturated wetland zones. To accomplish this goal, a comprehensive module for tracking water content in wetland soil was implemented in the model, and biogeochemical relationships were added to explain cycling of nitrogen (N) and carbon (C) in variably saturated zones of wetlands. The developed model was applied to a small, restored wetland receiving agricultural runoff, located on Kent Island, Maryland. On average, during the two-year study period, the ponded compartment of the study wetland covered 65% of the total 1.2 ha area. Through mass balance analysis, it was revealed that the mass of nitrogen lost to denitrification at the variably saturated compartment of the study wetland was about three times higher than that of the ponded compartment ($32.7\pm 29.3\mathrm{kg}$ versus $9.5\pm 5.5\mathrm{kg}$), whereas ammonia volatilization at the variably saturated compartment was a fraction of that of ponded compartment ($1.2\pm 1.9\mathrm{kg}$ versus $11.3\pm 11.8\mathrm{kg}$). Sensitivity analysis showed that cycling of carbon-related constituents in variably saturated compartment had high sensitivity to temperature and available soil moisture.

Rapid development in a watershed affects both surface water and groundwater resources. Urbanization results in an increase in surface runoff and volume of water collected downstream of the watershed. Low-impact development (LID) best management practices (BMPs) such as bioretentions, vegetated rooftops, rain barrels, vegetative swales, permeable pavements, and stormwater wetlands have been implemented to diminish adverse effects of urbanization such as flood by reducing peak flows on the surface and thus managing stormwater runoff. The objective of this study is to investigate the hydrological performance of bioretentions, which may also be considered as a small-size local constructed wetland, by conducting experimental analyses under various rainfall intensities and durations. For this purpose, an experimental setup called a rainfall-watershed-bioretention (RWB) system is constructed, which involves an artificial rainfall system, drainage area, and four bioretention columns with different soil textures and organic content. Inflow rate at the entrance of bioretentions and ponding depth on bioretention columns are measured during selected artificial rainfall events and durations. In addition, the outflow rate at the exit of bioretention columns is measured to understand the role of soil types in hydrological behavior of bioretention columns. Experimental results show that different soil types and organic content under various rainfall intensities and durations affect hydrological performance of bioretentions considerably.

Wetland ecosystems are controlled by their hydrology. Recent experimental and numerical investigations have suggested that flow pulses are needed to preserve sediment redistribution in some wetlands. In this study, the authors investigate the effect of pulsed-flow conditions on the hydrologic regime of low-gradient densely vegetated wetlands using a fine-resolution, two-dimensional depth-averaged numerical flow model. The model was applied to simulate flow depth and velocity within the Loxahatchee Impoundment Landscape Assessment (LILA) wetland located in Boynton Beach, Florida. Two pulsed-flow conditions with low-pulse and high-pulse flow magnitude were considered. The simulation results of low-pulse flow conditions reveal the areas within deep sloughs where flow velocities and directions change continuously, creating enhanced mixing areas within the deep slough. These mixing areas may have the potential to affect processes such as sediment redistribution and nutrient transport. Simulation of high-pulse flow magnitude, however, results in more uniform flow velocity inside deep slough. It also indicates that a pulse can only be detected when inflow discharge is at least $3.0{\mathrm{m}}^3/\mathrm{s}$. Lower inflow discharge values are too weak in magnitude to generate substantial changes in water surface elevation and velocity and they may not exhibit a flow wave propagation into the study area.

Tidal freshwater wetlands are one of the most vulnerable ecosystems to climate change and rising sea levels. However salinification within these systems is poorly understood, therefore, productivity (litterfall, woody biomass, and fine roots) were investigated on three forested tidal wetlands [(1) freshwater, (2) moderately saline, and (3) heavily salt-impacted] and a marsh along the Waccamaw and Turkey Creek in South Carolina. Mean aboveground (litterfall and woody biomass) production on the freshwater, moderately saline, heavily salt-impacted, and marsh, respectively, was 1,061, 492, 79, and $0{\mathrm{g}\mathrm{m}}^{-2}{\mathrm{year}}^{-1}$ versus belowground (fine roots) 860, 490, 620, and $\mathrm{2,128}{\mathrm{g}\mathrm{m}}^{-2}{\mathrm{year}}^{-1}$. Litterfall and woody biomass displayed an inverse relationship with salinity. Shifts in productivity across saline sites is of concern because sea level is predicted to continue rising. Results from the research reported in this paper provide baseline data upon which coupled hydrologic/wetland models can be created to quantify future changes in tidal forest functions.

Time-invariant contributing area and full hydrologic connectivity are commonly assumed in traditional watershed delineation and modeling. In reality, however, contributing area of a depression-dominated prairie basin varies and undergoes a progressive evolution process that is characterized by unique threshold behaviors associated with the depression filling–spilling–merging–splitting–depleting dynamics. The research reported in this paper aims to precisely delineate prairie potholes and the associated wetlands, effectively characterize their dynamic hydrotopographic properties, and further quantify the resultant threshold behaviors and spatiotemporal variability in hydrologic connectivity. Delineation and modeling for a wetland site in the Prairie Pothole Region (PPR) highlighted the crucial role of topographic characteristics in the formation, evolution, and connectivity of prairie potholes. Particularly, the modeling tests for real rainfall demonstrated the significant impacts of complex rainfall patterns on the behavior of a PPR wetland system and the dynamic filling, spilling, merging, splitting, and depleting processes of potholes. The research reported in this paper provides an effective tool for dynamic pothole wetland delineation and hierarchical modeling, which can be potentially used for ecohydrologic modeling and assessment.

The influence of wetlands on basin hydrology and channel hydraulics has been investigated for a small Canadian basin using the MIKE 11/NAM model of the Danish Hydraulic Institute (DHI). The Black Creek basin has an area of $93.58{\mathrm{km}}^2$, 15.6% of which consists of wetlands. The methodology involved long-term continuous simulation of the basin with and without the wetlands. The present analysis shows that on an annual basis, wetlands reduce the total runoff by less than 1% and increase the evapotranspiration by about the same amount. Overland flow decreases 21% and base flow increases by 15% due to wetland presence. Groundwater recharge also increases by 15%. During individual runoff events, wetlands reduce runoff by about 5% and 1% in terms of peak and volume, respectively. The 1:100-year flood under current conditions would become a 1:35-year flood if wetlands were removed. Wetlands can suppress the snow-melt-generated flood peaks by about 15–20% during spring freshet and summer floods by 1–5%. Wetlands also reduce the water level by 3–5 cm and 5–8 cm during spring freshet and summer storms respectively. Channel flow, water level, and velocity are influenced by the wetlands only during rare flood events.

Temporal and spatial variations of nitrogen (N) soil storages and fluxes were examined at two 1-ha created riverine wetlands in the U.S. Midwest. Soil N content (total N, organic-N, ${\mathrm{NO}}_3\text{-}\mathrm{N}$, and ${\mathrm{NH}}_4\text{-}\mathrm{N}$), N accumulation rates, and soil C:N ratios were compared between the two wetlands constructed 15 years earlier (one was planted and the other left to naturally colonize). Differences in wetland soil N content and accumulation were also examined in relation to proximity of river input and relative to a range of topographic features related to wetland water depth. The planted and naturally colonized wetlands showed similar rates of N accumulation. However, differences were detected related to the content of mineralized forms of N (${\mathrm{NO}}_3\text{-}\mathrm{N}$ and ${\mathrm{NH}}_4\text{-}\mathrm{N}$) that may relate to the history of vegetation communities at these wetlands. Significant spatial variation of N accumulation was detected within the wetlands, with the highest rates found in the deeper open water communities compared to shallow emergent marsh/edge vegetation communities ($23.5\pm 2.0$ versus $17.3\pm 1.3/18.2\pm 1.4\mathrm{g}\mathrm{N}{\mathrm{m}}^{-2}{\mathrm{y}}^{-1}$). Nitrogen budgets comparing two ages of the wetlands illustrate higher N accumulation rates (by 19%), higher N reduction in the surface water (47 to 52% reduction), and increased denitrification rates (by 13%) from year 10 to year 15. We also found out that nitrogen accumulation in the soil was 7.1 to 7.5% higher than were denitrification rates in these young wetlands. We discuss the importance of efforts such as this to support better understanding of N pathways in both created and natural wetlands, while providing critical data needed to improve modelling efforts and assess the long-term effectiveness of wetlands for improving water quality.