Ryan D. Stewart

A resonating rainfall and evaporation recorder

We propose a novel, accurate quantification of precipitation and evaporation, as needed to understand fundamental hydrologic processes. Our system uses a collection vessel placed on top of a slender rod that is securely fixed at its base. As the vessel is deflected, either by manual perturbation or ambient forcing (for example, wind), its oscillatory response is measured, here by a miniature accelerometer. This response can be modeled as a damped mass-spring system. As the mass of water within the collection vessel changes, either through the addition of precipitation or by evaporative loss, the resonant frequency experiences an inverse shift. This shift can be measured and used to estimate the change in the mass of water. We tested this concept by creating a simple prototype which was used in field conditions for a period of 1 month. The instrument was able to detect changes in mass due to precipitation with an accuracy of approximately 1 mm.

Modeling hydrological response to a fully-monitored urban bioretention cell

Municipalities and agencies use green infrastructure to combat pollution and hydrological impacts (e.g., flooding) related to excess stormwater. Bioretention cells are one type of infiltration green infrastructure (GI) intervention that infiltrate and redistribute otherwise uncontrolled stormwater volume. However, the effects of these installations on the rest of the local water cycle is understudied; in particular, impacts on stormwater return flows and groundwater levels are not fully understood. In this study, full water cycle monitoring data was used to construct and calibrate a two-dimensional Richards equation model (HYDRUS-2D/3D) detailing hydrological implications of an unlined bioretention cell (Cleveland, Ohio) that accepts direct runoff from surrounding impervious surfaces. Using both pre- and post-installation data, the model was used to: 1) establish a mass balance to determine reduction in stormwater return flow, 2) evaluate GI effects on subsurface water dynamics, and 3) determine model sensitivity to measured soil properties. Comparisons of modeled versus observed data indicated that the model captured many hydrological aspects of the bioretention cell, including subsurface storage and transient groundwater mounding. Model outputs suggested that the bioretention cell reduced stormwater return flows into the local sewer collection system, though the extent of this benefit was attenuated during high inflow events that may have exhausted detention capacity. The model also demonstrated how, prior to bioretention cell installation, surface and subsurface hydrology were largely decoupled, whereas after installation, exfiltration from the bioretention cell activated a new groundwater dynamic. Still, the extent of groundwater mounding from the cell was limited in spatial extent, and did not threaten other subsurface infrastructure. Finally, the sensitivity analysis demonstrated that the overall hydrological response was regulated by the hydraulics of the bioretention cell fill material, which controlled water entry into the system, and by the water retention parameters of the native soil, which controlled connectivity between the surface and groundwater.

Modeling multidomain hydraulic properties of shrink‐swell soils

Shrink-swell soils crack and become compacted as they dry, changing properties such as bulk density and hydraulic conductivity. Multidomain models divide soil into independent realms that allow soil cracks to be incorporated into classical flow and transport models. Incongruously, most applications of multidomain models assume that the porosity distributions, bulk density, and effective saturated hydraulic conductivity of the soil are constant. This study builds on a recently derived soil shrinkage model to develop a new multidomain, dual-permeability model that can accurately predict variations in soil hydraulic properties due to dynamic changes in crack size and connectivity. The model only requires estimates of soil gravimetric water content and a minimal set of parameters, all of which can be determined using laboratory and/or field measurements. We apply the model to eight clayey soils, and demonstrate its ability to quantify variations in volumetric water content (as can be determined during measurement of a soil water characteristic curve) and transient saturated hydraulic conductivity, Ks (as can be measured using infiltration tests). The proposed model is able to capture observed variations in Ks of one to more than two orders of magnitude. In contrast, other dual-permeability models assume that Ks is constant, resulting in the potential for large error when predicting water movement through shrink-swell soils. Overall, the multidomain model presented here successfully quantifies fluctuations in the hydraulic properties of shrink-swell soil matrices, and are suitable for use in physical flow and transport models based on Darcys Law, the Richards Equation, and the advection-dispersion equation.

Transport of a neonicotinoid pesticide, thiamethoxam, from artificial seed coatings

Neonicotinoid insecticides coat the seeds of major crops worldwide; however, the high solubility of these compounds, combined with their toxicity to non-target organisms, makes it critical to decipher the processes by which they are transported through soils and into aquatic environments. Transport and distribution of a neonicotinoid (thiamethoxam, TMX) were investigated by growing TMX-coated corn seeds in coarse-textured and fine-textured soil columns (20 and 60cm lengths). To understand the influence of living plants, corn plants were terminated in half of the columns (no plant treatment) and allowed to grow to the V5 growth stage (33days of growth) in the other half (with plant treatment). TMX was analyzed in leachate 12 times over 33days and in bulk soil after 8, 19, and 33days of corn growth. All 20cm columns leached TMX at levels exceeding the United States Environmental Protection Agency benchmark for aquatic invertebrates (17.5μgL-1). TMX migrated from seeds to adjacent bulk soil by the eighth day and reached deeper soil sections in later growth stages (e.g., 30-45cm depth by Day 33). Fine-particle soils transported over two orders of magnitude more TMX than coarse-textured soils (e.g., 29.9μg vs 0.17μg, respectively), which was attributed to elevated evapotranspiration (ET) rates in the sandy soil driving a higher net retention of the pesticide and to structural flow occurring in the fine-textured soil. Living plants increased TMX concentrations at depth (i.e., 30-60cm) compared to the no plant treatment, suggesting that corn growth may drive preferential transport of TMX from coated seeds. Altogether, this study showed that neonicotinoid seed coatings can be mobilized through soil leachate in concentrations considered acutely toxic to aquatic life.

Evaluation of Infiltration Discharge as a Strategy to Meet Effluent Temperature Limits

AbstractRecognizing that elevated stream temperatures can harm aquatic organisms, regulatory agencies have begun to enforce thermal total daily maximum load (TMDL) limits on wastewater discharges. Infiltration discharge, where treated wastewater is allowed to infiltrate into the soil and then percolate toward the stream or river, represents a potential method for meeting temperature requirements and for supplementing flow. This study combines observed and simulated temperature data to assess the efficacy of using infiltration discharge to meet effluent temperature limits. Observational data collected during operation of a 0.15-ha pilot-scale infiltration discharge system revealed the pattern of groundwater heating beneath the site by the wastewater, with the largest temperature increases occurring near the infiltration point. Numerical simulations used to examine the long-term groundwater response to operation of a 5.5-ha large-scale infiltration wetland system confirmed this trend and demonstrated how mu...