Journal of Hydrologic Engineering

It is often assumed that soils are homogeneous when designing septic systems or modeling wastewater flow from their trenches. The main objective of this study was to assess water infiltration and movement from the trenches similar to the ones commonly used for on-site wastewater dispersal by septic systems. Four separate experiments, each using a small drainfield with four parallel trenches, were conducted at three sites with different soils. In two experiments the trenches were in the coarse-textured soil above a clayey Bt horizon. In the other two experiments the trenches were in the Bt horizon. For each experiment, $50\mathrm{L}$ of a solution containing potassium bromide and brilliant blue FCF (as a tracer dye) were applied once a day to each trench for 14 or 15 days. A sampling pit was dug perpendicular to the trenches after the tracer solution application, and the distribution of the tracer dye and ${\mathrm{Br}}^{-}$ around the trenches on the two walls of the pit were assessed. Tracer solution infiltration from the trenches was not uniform in any of the experiments. Water flow in the Bt horizon was mainly through macropores. Further, water containing ${\mathrm{Br}}^{-}$ and dye moved a substantial distance from the trenches through macropores when trenches were installed in the Bt horizon. When trenches were installed in the coarse-textured soil above the Bt horizon, most of the tracer solution moved away laterally from the drainfield through the zone above the Bt horizon. Overall, the results indicate that soil morphological properties and soil horizonations, as well as the nature of water movement from trenches must be considered when modeling water flow from septic system trenches or when designing septic system drainfields.

Concerns from local health departments regarding premature septic system failure (less than 1 year from installation) has led to an investigation of septic system soil absorption field design parameters in northeast Indiana. The objective of this study was to compare the loading rate based on field measured saturated hydraulic conductivity $\left({\mathrm{LR}}_m\right)$ across a toposequence to the estimated allowable loading rate $\left({\mathrm{LR}}_e\right)$ based on soil morphological properties. Saturated hydraulic conductivity measurements were determined by a compact constant-head permeameter at five landscape positions, at four depths (surface horizon, upper argillic horizon, transition zone between the argillic horizon and till parent material, and till). Results showed that for all depths, the ${\mathrm{LR}}_m$ was smaller than ${\mathrm{LR}}_e$ . Results from this study suggest that the current method of using soil morphological properties to determine the loading rate may overestimate the ability of the soil to properly disperse septic system effluent.

A field evaluation was conducted in Minnesota to identify whether there is a statistical difference in performance between chambered and rock-filled trench systems. This was achieved by a large-scale survey of over 100 selected onsite systems of both rock-filled and chambered trenches. Each system type was studied within three major soil permeability categories (fast, medium, and slow) and the percentage of the system in use at the time of the site visit was determined. Adjusting both types of systems to a standard size datum, the ponding levels were compared. Nearly 60% of the systems visited during the study of ages 5–10 years did not have any ponding observed at the end of the first trench segment. When the amount of ponding was compared between rock-filled and chambered systems, the data did not prove the hypothesis that chambered systems of a similar age as rock-filled systems utilize 25% less area than rock systems at 10% significance level.

The goal of this study was to investigate how extreme hydraulic loading influences hydraulic pathways, and thus failure potential, of two soil absorption systems. A grid of tensiometers and piezometers were installed beneath and adjacent to the absorption trenches, and water height was manipulated to simulate different hydraulic loadings. At both sites, measured soil matric potentials along the trench sidewalls increased toward zero as water height increased over time, indicating that water was preferentially flowing laterally above the biomat zones. Saturated hydraulic conductivity $\left(K_s\right)$ of the soil and biomat, and other hydraulic parameters were calibrated using inverse modeling procedures in HYDRUS-2D. We found generally good agreement between measured and predicted matric potentials. The calibrated model was then used to predict fluxes through different infiltrative zones under typical and extreme hydraulic loads. Modeling indicated that the overall infiltration rates through the biomat and subbiomat zones fell within a relatively narrow range of $0.025–0.046\mathrm{m}∕\mathrm{day}$ at the two sites. Under extreme hydraulic loading to trenches, fluxes through the upper trench sidewalls (i.e., the exfiltration zone) were predicted to be substantially greater (i.e., 80% of total flux) than through the biomat zone. This confirms the conclusions made from field results, that a permeable exfiltration zone provides a critical buffer to surface surcharging during extreme trench loading. Conversely, the presence of a high sidewall biomat can reduce sidewall exfiltration of effluent, and therefore increase the likelihood of surface surcharging.

Little is known about how much wastewater infiltrates the soil via the trench sidewall versus the trench bottom in onsite wastewater systems. Our objectives were to develop a method of simulating trench bottom and sidewall flow using a two-dimensional numerical computer model that would include the trench within the model space and determine how much sidewall flow would occur under steady-state conditions. We used HYDRUS-2D to simulate water flow in a two-dimensional cross section of a conventional gravel-filled trench and the surrounding drainfield. Hydraulic properties of the gravel were assumed (saturated hydraulic conductivity $K_s=\mathrm{1,000}\mathrm{cm}{\mathrm{d}}^{-1}$ ) and simulations were run for drainfield soils consisting of a clay loam $(K_s=3.2\mathrm{cm}{\mathrm{d}}^{-1})$ and a sand $(K_s=41\mathrm{cm}{\mathrm{d}}^{-1})$ . Biomats were simulated at the bottom of the trench and part way up the trench sidewall ( $K_s=0.2$ to $2.8\mathrm{cm}{\mathrm{d}}^{-1}$ depending on the simulation). Typical wastewater loading rates for Georgia of 2 and $4\mathrm{cm}{\mathrm{d}}^{-1}$ for the clay loam and sand, respectively, were simulated in three daily doses of $1.4\min$ in length. Simulations were run until the water level in the trench reached a dynamic equilibrium, as indicated by a repeating pattern of water level daily changes. The method we developed predicted the water level in the trench instead of specifying it as a boundary condition, as has been done in previous modeling studies. In the clay loam soil, we found that the water level in the trench at steady conditions averaged a depth of $9.5–10.9\mathrm{cm}$ and that 29–31% of the total flow occurred through the sidewall. In the sand soil, we found that water did not pond in the trench and there was no sidewall flow. Much of the sidewall flow in the clay loam soil appeared to be in the “lip” area just above the maximum height of the sidewall biomat. Our results show that sidewall flow is important, but not as high as others have estimated.

Water table mounding warrants consideration in the design of cluster and high-density wastewater soil absorption systems (WSAS). Use of iterative analytical solutions with the design approach presented here provides a useful initial assessment of the potential for mounding beneath a WSAS, but cannot account for anisotropy, heterogeneity, regional flow, and local hydraulic boundaries. The relative influence of these factors may not be intuitive, but can be evaluated with site-specific numerical models in cases where the failure of these systems may have severe consequences. There is a greater difference of calculated mound height for isotropic analytical solutions and anisotropic numerical solutions in low hydraulic conductivity materials and thicker aquifers. In all cases, increased vertical anisotropy increases mounding, but not nearly as much as would be calculated by using the vertical hydraulic conductivity value for hydraulic conductivity in the analytical solution. Increased gradient has less influence on mound height in thicker aquifers.

On-site wastewater systems (OWS) provide a sustainable method for disposing of residential wastewater. Large “cluster” OWS systems are now frequently being used to serve multiple-unit housing developments. Groundwater perching, mounding, and lateral flow can occur when the effluent application rate to an infiltration area exceeds the soil’s capacity to infiltrate the wastewater. This study addresses mounding of wastewater on a low hydraulic conductivity $\left(K\right)$ layer in the vadose zone, and focuses on modeling tools available to design systems so that the potential for the occurrence of surface or side-slope breakout is minimized. In particular, we demonstrate how an existing analytical model developed by Khan et al. (1976) can be used for evaluating the mounding problem. The robustness and limitations of the analytical solution due to certain constraining assumptions are evaluated for a few realistic scenarios by comparing it to results of a numerical model that is not constrained by the assumptions. The height and extent of mounding depends on both the hydraulic conductivity of the soil below the infiltration field, as well as the $K$ value of the low-permeability layer. The numerical modeling demonstrates that the analytical solution is more accurate for the case of a thick low- $K$ layer, and for conditions that represent a deep water table, because these conditions are more closely aligned with the conceptual model that the analytical solution derivation is based on. The analytical solution is less accurate for cases of a shallow water table, or a thin low- $K$ layer. Mounding is smaller for thin low- $K$ layers and for deep water tables. In general, it is shown that the analytical model is a conservative predictor for mounding for several reasons. First, it does not consider the capacity of the vadose zone to conduct infiltrating water laterally. Second, nonuniformities in a low-permeability layer (e.g., breaks in the layer) tend to reduce the mounding compared to the uniform layer assumed by the analytical solution. Third, anisotropic soils, where the vertical $K$ is less than the horizontal $K$ (a typical case), exhibit a smaller mounding height and lateral extent of mounding than predicted by the analytical solution. Measurement or accurate estimation of hydraulic conductivity values for subsurface soil layers is necessary for both analytical and numerical models if an accurate assessment of mounding is desired.

This paper describes the performance of six lateral flow sand filters (LFSFs) for their treatment of septic tank effluent in Truro, Nova Scotia, Canada. This report presents LFSF performance data collected during the first year of monitoring (Sept. 2004–Sept. 2005). The objectives of this initial study were to: (i) Evaluate the performance of LFSFs in field conditions and determine the influence of temperature and external hydrologic processes on treatment processes; (ii) evaluate the effects of slope and sand characteristics on LFSF performance; and (iii) characterize the hydraulic operation of LFSF systems in field conditions. Six LFSFs were constructed according to the Nova Scotia Department of Environment and Labour’s (NSDEL) design guidelines. Fine $(d_{10}=0.15\mathrm{mm})$ , medium $(d_{10}=0.17\mathrm{mm})$ , and coarse $(d_{10}=0.30\mathrm{mm})$ sands were tested at 5 and 30% slopes. The hydraulic conductivity of these sands ranged from $1.5\times {10}^{-4}$ to $1\times {10}^{-3}\mathrm{m}{\mathrm{s}}^{-1}$ . Each LFSF was loaded with approximately $100\mathrm{L}{\mathrm{d}}^{-1}$ of septic tank effluent for 1 year and samples were collected monthly. Average removal efficiencies for all LFSFs met NSDEL requirements: biological oxygen demand $(>98.5\%)$ , total suspended solids $(>95.5\%)$ , and E. coli ( $>5.4$ log reduction). Phosphorus removal ranged from 98% in the fine sand to 71.2% in the coarse sand filter. Nitrification was favored because the filters were operating under aerobic and unsaturated conditions. Therefore, denitrification was limited causing elevated nitrate effluent concentrations. Total nitrogen removal ranged from 60 to 66%. The LFSFs provided consistent year-round treatment and did not appear to be impacted greatly by slope, temperature, or external hydrologic influences.

The depth of soil below the absorption trench of a septic system is considered an important factor in protection of groundwater. We examined the effects of depth on the ability of intermittently aerated sand-filled leachfield mesocosms to renovate domestic wastewater. Mesocosms $(n=3)$ consisted of lysimeters with a headspace ${\mathrm{O}}_2$ concentration maintained at $0.21\mathrm{mol}∕\mathrm{mol}$ and containing 7.5, 15, or $30\mathrm{cm}$ of sand that were dosed with septic tank effluent every $6\mathrm{h}$ for 328 days $(12\mathrm{cm}∕\mathrm{d})$ . Sand depth had no effect on pH, dissolved ${\mathrm{O}}_2$ , ${\mathrm{PO}}_4$ , ${\mathrm{NH}}_4$ , or ${\mathrm{BOD}}_5$ levels in percolate water. Nitrate levels in percolate water were higher for $30\mathrm{cm}$ than for 7.5 and $15\mathrm{cm}$ during the first $70\mathrm{d}$ of the experiment, after which no differences were observed. Time-averaged removal rates of N, P, fecal coliform bacteria, and ${\mathrm{BOD}}_5$ were 22–28, 13–18, 81–92, and 81–99%, respectively, and were unaffected by depth. Wastewater renovation in intermittently aerated leachfield mesocosms appears to take place in a narrow zone $(\le 7.5\mathrm{cm})$ below the infiltrative surface, with the medium below contributing little to renovation.

A conventional on-site wastewater system (OWS) is traditionally comprised of one or more septic tanks for treatment of raw wastewater followed by percolation through natural soil to achieve purification prior to groundwater recharge. Other types of engineered treatment units (e.g., sand filters, textile media filters) may be implemented where site conditions are not suitable for conventional systems or in sensitive areas, such as those with nitrogen loading concerns. The research presented here is part of a large field study conducted to evaluate the purification performance of OWS employing three different engineered treatment units: a septic tank, a septic tank with a textile filter unit (TFU), and a septic tank with a membrane bioreactor (MBR). The TFU or MBR employ treatment processes that are designed to achieve higher purification compared to a septic tank so that soil treatment can be accomplished at higher hydraulic loading rates and/or with less unsaturated soil depth. This paper describes the installation, operation, and monitoring of the effluent generated from three engineered treatment units operated for a period of 16–28 months. The three treatment units, as expected, achieved different purification efficiencies for organic matter, total suspended solids, nutrients, and bacteria with the relative removal efficiency of $\mathrm{MBR}>\mathrm{TFU}>$ septic tank. The relative degree of operational complexity, operation and maintenance requirements, energy use, and cost followed the same ranking.

Watershed assessment of nutrient loading from on-site systems is an important issue for a state such as North Carolina, where approximately 50% of the total housing units rely on these systems for household wastewater treatment. Integration of spatially incongruent data sets has always been a major obstacle for environmental studies dealing with interdisciplinary data sets. A geographical information system (GIS)-based area normalization procedure was developed to transfer demographic information from census blocks to watersheds. The GIS-based area normalization procedure improved statewide estimates by 5 and 12%, respectively, for housing units using on-site systems and resulting nitrogen loadings compared to a visualization procedure. A parametric test conducted using a univariate procedure indicated that there were significant differences between the number of housing units using on-site systems $(p=0.0014)$ , nitrogen loadings $(p=0.0012)$ , and total population $(p=0.0065)$ estimated by the visualization procedure and the normalization procedure at $\alpha =0.05$ level.

This paper presents an overview of the field test results of the La Pine National Decentralized Wastewater Treatment Demonstration Project (La Pine Project). The main objective of the field test was to identify innovative on-site wastewater treatment systems that reduce the amount of nitrogen discharged to groundwater. The aquifer underlying the La Pine subbasin of the upper Deschutes River Watershed is shallow and unconfined and provides the drinking water source for the region’s population. The project also evaluated septic tank and sand filter performance in order to provide baseline information on the impacts of conventional systems on the aquifer. The performance of nitrogen reducing treatment technologies are compared to septic tank and sand filter performance, the State of Oregon’s residential waste strength definition, and La Pine Project performance criteria. The study found that several systems can effectively reduce typical water quality parameters (5-day biochemical oxygen demand, total suspended solids, and fecal coliform) but that nitrogen reduction can be more difficult to achieve in the field. Several systems that showed high levels of nitrogen reduction in test centers did not perform as well in the field and nitrogen reduction below $10\mathrm{mg}∕\mathrm{L}$ appears to be difficult to achieve consistently without a secondary carbon source.