Preferential flow, diffuse flow, and perching in an interbedded fractured-rock unsaturated zone

Abstract

Layers of strong geologic contrast within the unsaturated zone can control recharge and contaminant transport to underlying aquifers. Slow diffuse flow in certain geologic layers, and rapid preferential flow in others, complicates the prediction of vertical and lateral fluxes. A simple model is presented, designed to use limited geological site information to predict these critical subsurface processes in response to a sustained infiltration source. The model is developed and tested using site-specific information from the Idaho National Laboratory in the Eastern Snake River Plain (ESRP), USA, where there are natural and anthropogenic sources of high-volume infiltration from floods, spills, leaks, wastewater disposal, retention ponds, and hydrologic field experiments. The thick unsaturated zone overlying the ESRP aquifer is a good example of a sharply stratified unsaturated zone. Sedimentary interbeds are interspersed between massive and fractured basalt units. The combination of surficial sediments, basalts, and interbeds determines the water fluxes through the variably saturated subsurface. Interbeds are generally less conductive, sometimes causing perched water to collect above them. The model successfully predicts the volume and extent of perching and approximates vertical travel times during events that generate high fluxes from the land surface. These developments are applicable to sites having a thick, geologically complex unsaturated zone of substantial thickness in which preferential and diffuse flow, and perching of percolated water, are important to contaminant transport or aquifer recharge.

Résumé

Des couches avec de forts contrastes géologiques dans la zone non saturée peuvent contrôler la recharge et le transport de contaminants vers les aquifères sous-jacents. Les écoulements diffus lents dans certaines couches géologiques et les écoulements préférentiels rapides dans d’autres compliquent la prévision des flux verticaux et latéraux. Un modèle simple est présenté, conçu pour ne requérir que des données géologiques limitées sur le site étudié, pour prédire ces processus de subsurface critiques vis-à-vis d’une source d’infiltration continue. Le modèle est développé et testé en utilisant les données spécifiques du site du laboratoire national de l’Idaho dans la plaine de la Eastern Snake River (ESRP), USA, où se trouvent des sources anthropiques et naturelles de forts volumes d’infiltration à partir des crues, déversements, fuites, rejets d’eaux usées, bassins de rétention et expériences hydrologiques de terrain. L’épaisse zone non saturée surmontant l’aquifère ESRP est un bon exemple d’une zone non saturée nettement stratifiée. Les interlits sédimentaires sont interstratifiés entre des unités de basalte massif et fracturé. La combinaison de sédiments superficiels, basaltes et interlits détermine les flux d’eau à travers le sous-sol variablement saturé. Les interlits sont généralement moins perméables, occasionnant parfois des nappes perchées au-dessus d’eux. Le modèle prédit avec succès le volume et l’extension des nappes perchées et estime les temps de transit vertical pendant les événements qui génèrent de forts flux depuis la surface du sol. Ces développements sont applicables aux sites présentant une épaisse zone non saturée, complexe géologiquement, d’une épaisseur substantielle au sein de laquelle des écoulements préférentiels et diffus et des nappes perchées sont importants en ce qui concerne le transport de contaminants ou la recharge de l’aquifère.

Resumen

Las capas de fuerte contraste geológico dentro de la zona no saturada pueden controlar la recarga y transporte de contaminantes a los acuíferos subyacentes. El lento flujo difuso en ciertas capas geológicas, y el rápido flujo preferencial en otras, complica la predicción de los flujos verticales y laterales. Se presenta un modelo simple, diseñado para utilizar la limitada información del sitio geológico para predecir estos procesos críticos del subsuelo en respuesta a una fuente de infiltración sostenida. Se desarrolló y probó el modelo usando información específica del sitio del Laboratorio Nacional de Idaho en el este de Snake River Plain (PEIS), EEUU, donde hay fuentes naturales y antropogénicas de altos volúmenes de infiltración a partir de inundaciones, vertidos, filtraciones, disposición de aguas residuales, estanques de retención, y experimentos hidrológicos de campo. El espesor de la zona no saturada sobre el acuífero PEIS es un buen ejemplo de una zona no saturada fuertemente estratificada. Las intercalaciones sedimentarias se intercalan entre las unidades masivas y fracturadas de basalto. La combinación de los sedimentos superficiales y las intercalaciones de basaltos determina los flujos de agua a través del subsuelo variablemente saturado. Las intercalaciones son generalmente menos conductoras, a veces causando que el agua se cuelgue por encima de ellos. El modelo predice con éxito el volumen y el alcance del agua colgada y aproxima los tiempos de desplazamiento verticales durante los eventos que generan altos flujos desde la superficie. Estos desarrollos son aplicables a los sitios que tienen una espesa zona no saturada, geológicamente compleja de espesor considerable en la cual el flujo preferencial y difuso, y el colgado del agua percolada, son importantes para el transporte de contaminantes o la recarga de los acuíferos.

摘要

非饱和带内强大的地质对比层可以控制对下伏含水层的补给和污染物运移。某一特定的地质层缓慢的弥散流及其他地质层的快速优先流使预测垂直和侧向通量变得更加复杂。这里展示了一个简单的模型,就是利用有限的地质场地信息预测针对持续渗入源产生响应的关键的地表以下的过程。利用美国斯内克河平原东部爱达荷国家实验室提供的现场特定信息建立了模型并对其进行了测试,斯内克河平原东部地区具有天然和人为的大流量的渗入源,包括洪水、泄漏、渗漏、废水处理、澄清池以及水文野外实验等。覆盖斯内克河平原东部含水层的厚层非饱和带是急剧分层的非饱和带一个很好的例子。巨大的和断裂的玄武岩单元之间散布着沉积互层。地表沉积层、玄武岩及互层的组合确定着通过饱和度变化着的地表下岩层的水通量。互层通常传导性较小,有时引起上层滞水在互层上方积聚。模型成功地预测了上层滞水积聚的量和范围,粗略估算了产生地表很高通量事件期间的垂直行程时间。这些进展可应用于具有很厚的、地质复杂的非饱和带的场地,在此非饱和带中,优先流、弥散流和渗透水的积聚对污染物的运移和含水层补给非常重要。

Resumo

O mecanismo de recarga e transporte de contaminantes de aquíferos sotopostos é controlado por intercalações rochosas de propriedades altamente contrastantes ao longo da zona não saturada. A previsibilidade dos fluxos verticais e laterais nessas formações é baixa por causa do escoamento difuso e lento em determinadas formações, e o escoamento preferencial e rápido em outras. Este trabalho apresenta um modelo simples para avaliação de processos subterrâneos críticos para a recarga a partir de dados geológicos de campo limitados, o qual considera uma fonte de infiltração constante. O modelo foi desenvolvido e testado utilizando-se dados de campo do Laboratório Nacional de Idaho, na Planície do Rio Eastern Snake (PRES), EUA, onde existe fontes naturais e antropogênicas de infiltração por inundações, derramamentos, vazamentos, lançamentos de água residuária, bacias de retenção e experimentos hidrológicos de campo. A espessa camada não saturada sobreposta ao aquífero PRES é um bom exemplo de formação não saturada constituída de estratificações delgadas. São observadas intercalações sedimentares entre unidades basálticas maciças e fissurais. Este padrão de distribuição de camadas, portanto, determina o padrão de escoamento através destas camadas com diferentes graus de saturação. Intercalações de rocha são normalmente hidraulicamente menos condutivas, ocasionando aquíferos suspensos sobre si. O volume e a extensão destes aquíferos suspensos são reproduzidos com sucesso pelo modelo, assim como o modelo também é capaz de reproduzir o tempo de percurso de escoamentos verticais ocasionados por eventos abruptos originados na superfície. Este modelo é válido para sítios que têm geologia espessa, complexa e insaturada em que o escoamento aconteça tanto de forma difusa quanto de forma preferencial, em cujos aquíferos suspensos e drenagem profunda sejam relevantes para o transporte de contaminantes ou para a recarga de aquíferos sotopostos.

Appendix: Evidence to support the conceptual model

This study relies strongly on data from infiltration experiments conducted at four particular ESRP locations, designated LSIT, SATT, VZRP, and INTEC (Fig. 9). This Appendix gives details concerning three of these.

Fig. 9

Sites of major surface infiltration experiments on the ESRP, noting the spatially varying trend of surficial sediment characteristics

LSIT

The Large Scale Infiltration Test was conducted in 1994 (Wood and Norrell 1996). A circular pond, area 26,300 m², was filled with water at a constant head for 5 days, allowed to drain for 11 days, and maintained at a constant head (maximum 1.4 m) for 19 more days. Water remained in the pond at declining level for 11 more days after replenishment had stopped.

Sixty-six instrumented wells within and around the pond were used to monitor flow through basalt and perching on top of the sedimentary interbed at 55 m depth. The wells were arranged in four rings, designated A, B, C, and E, in order of increasing radii (Table 6). The wells were grouped in clusters, usually of three, positioned at approximately equal spacing around the circumference of each ring. Within each cluster, except one in the A-ring, one well was drilled to the interbed and the others to shallower depths. Two wells were drilled through the interbed.

Table 6 Radial layout of LSIT site

To evaluate changes in water content above the first interbed, neutron counts over a range of depths in the wells were measured daily from June 23 until the end of August 1994. The data from A wells (within the pond) show an increase in water content during the period of ponding (Fig. 10a). Data from the B wells, however, show no increase in water content (Fig. 10b). This suggests that infiltrated water in the surficial sediments or in the basalt did not migrate laterally more than 15 m beyond the edge of the pond, where the B wells are located.

Fig. 10

Neutron counts before the experiment on June 23, 1994, and during the experiment on August 4, 1994: a well A01C11 within the pond radius, and b well B01C11 outside. More counts indicate less water, so water content increases to the left

Water from the pond began infiltrating when the experiment started on July 25, 1994, and perched water was first detected on August 8. Results from the measurements are plotted in Fig. 11. The A wells reported much higher perched water elevations than the B, C, or E wells and show no rising or falling limb because these wells were completed at levels considerably above the interbed.

Fig. 11

Perched water-table hydrograph for all wells with perched water detected during the LSIT

The A wells show no trend in perched water-table elevation over time. The B wells do exhibit a trend over time, all but two of them (B08N11 and B09N11) reaching elevations of 1,492.0 to 1,492.5 m. The two exceptions are considerably lower than the rest, possibly because of a topographical impediment that limits the flow to these positions. The elevation of perched water in the C wells is lower than that in the B wells, and in the E well is less than any others. This trend suggests that the perched water elevations decrease with distance from the infiltration basin (Porro and Bishop 1995).

VZRP

During the 2007 VZRP experiments, the average rate of water discharged into the ponds was 4,800 m³/day. Though the pond basins at the surface were larger, actual ponded water during these experiments never covered an area larger than about 1,000 m². Data from Duke et al. (2007) suggest that infiltrated water spread laterally within the relatively thick (15 m) VZRP surficial sediment layer before flowing into the underlying basalt. The data indicated this spreading extended laterally at least 100 m, but no more than 140 m. The area of percolation into basalt then was between 31,400 and 90,800 m², much larger than the ponded area at the surface.

INTEC leaks

Two wells at the INTEC facility, MW-6 and 33-3, exhibit significant responses to nearby anthropogenic leaks of water intended for fire-fighting (Fig. 12; Forbes and Ansley 2008); J. Forbes, Idaho Cleanup Project, personal communication, 2014). Figure 13 shows the hydrographs of these wells.

Fig. 12

Aerial view of wells MW-6 and 33-3 and three leaks at the INTEC facility

Fig. 13

Perched water table hydrographs for wells 33-3 and MW-6 during near-surface water leaks at INTEC in a 2007, b 2012, and c 2013

The first leak, from fire hydrant HYD-0521, began leaking on June 27, 2007, and was stopped on February 6, 2008, after releasing an estimated 5,700 m³ of water. This leak triggered a response only in well MW-6, where the water table rose about 1 m to a steady position of 1,463 m. The second leak began January 5, 2012 and was stopped on May 15, 2012 after releasing 7,400 m³ of water. The perched water table near MW-6 rose 4 to 1,465 m above sea level, and near well 33-3 rose 2 to 1,464 m. The third leak began March 2, 2013 and was stopped on May 20, 2013 after releasing 12,900 m³ of water. The perched water table near MW-6 rose 4.5 to 1,464 m above sea level, and near well 33-3 rose 1.5 to 1,464.5 m.