Increasing population and climate change are causing water managers to reassess water storage. In this context, alluvial aquifer storage and recovery (ASR), in which excess water is stored in the alluvium near a river, offers a plausible option. To investigate this option, a coupled technical–administrative analysis was conducted to investigate the feasibility of alluvial ASR in the semi-arid US state of Colorado, where water rights are governed by the doctrine of prior appropriation. A hypothetical alluvial ASR facility near Brighton, Colorado with a storage capacity of 118,500 cubic meters (96 ac-ft) was considered. This analysis comprises both technical feasibility, using a groundwater model that explicitly accounts for clogging, and administrative feasibility, using a first-of-its-kind analysis of the legal availability of water including both free river and reusable effluent water. This coupled technical–administrative analysis suggests that alluvial ASR facilities present a viable option to meet rising demand for water storage, preventing water loss due to evaporation, reducing the effect of climate stress on water resources, and avoiding the need to purchase land for above-ground water storage facilities. More generally, this study illustrates the crucial importance of placing hydrologic analysis in the broader context of policy constraints.
With increasing global population resulting in higher consumption of basic resources, the task to supply clean water is becoming more challenging, and water managers around the world must be prudent and innovative to meet the ever-growing water demand. The US state of Colorado is a case in point, where the population of approximately five million in 2013 is expected to exceed eight million by 2040 (Grigg, 2013). In particular, in northeastern Colorado, the population of the South Platte River basin and the Metro Basins of central Colorado are projected to grow from approximately 3.5 million in 2008 to six million by 2050, according to the Colorado State Demography Office (HDR Engineering & West Sage Water Consultants, 2015). Thus, population growth will significantly increase the basin's future municipal and industrial water needs.
Climate change triggers additional stress on water resources (Bates et al., 2008). Colorado's average annual precipitation of only 41 cm (16 in) and its high evaporative losses (statewide average of approximately 81%) result in a water balance deficit over most of the state, with the exception of the higher mountainous regions (Topper et al., 2003). With no projected increase in average precipitation within Colorado, and with climate variability and projected population growth, the shortfall between water supply and water demand will increase over the next few decades (Lukas et al., 2014). Climate change, in Colorado and globally, is attributed to increased greenhouse gas emissions. Dahlman (2017) reports that since 1880, global surface temperature has risen at an average pace of 0.07 °C (0.13 °F) every 10 years for a net warming of 0.95 °C (1.71 °F) through 2016, with several additional degrees of warming possibly by 2100. In Colorado, Lukas et al. (2014) estimate that the statewide average temperature will rise +1.4 °C (2.5 °F) to +2.7 °C (5 °F) by 2050, relative to a 1971–2000 baseline. These changes are expected to cause extreme weather patterns, smaller snowpack, shift in spring snowmelt with runoff 1 to 4 weeks earlier, more frequent droughts and floods, and greater evaporation with reduced precipitation globally (NOAA National Centers for Environmental Information, 2016). Each of these factors increases water demand for agricultural, municipal, and recreational use. Therefore, new water storage methods must be developed that will help solve the water storage issue, while also minimizing global losses due to evaporation.
In light of the known limitations of above-ground dams and reservoirs, including evaporative losses, sedimentation, greenhouse gas emissions resulting from anaerobic degradation of inundated vegetation, and displacement of populations (Lei et al., 2012), aquifer storage and recovery (ASR) is gaining popularity especially in arid and semi-arid climates (Topper et al., 2004; Ishida et al., 2011). ASR is carried out by recharge of water into aquifers for storage during the wet season, which is later pumped out during the dry season.
Storing water in underground aquifers is advantageous compared to storing water in conventional dams and reservoirs. Livneh et al. (2016) report that reservoir evaporation represents the only irreversible loss to the water resources system, and that preliminary estimates of evaporation from major reservoirs along the Colorado River are of the order of 10% of the system-wide natural flow. Hence, storing water underground in alluvial aquifers could potentially assist water managers in maximizing the beneficial use of water resources. The primary objectives of artificial recharge are to manage water supply, manage water quality, restore aquifers, protect the environment, and meet legal obligations (Topper et al., 2004). In particular, ASR is used to combat drought scenarios (Barker et al., 2016). To achieve some or all of these benefits, ASR has been applied widely throughout the world (Brown et al., 2006) and in several US states including Arizona, California, Florida, and Oregon (Bloetscher et al., 2014). The Prairie Waters project, implemented by the city of Aurora, Colorado, USA, uses ASR for preliminary storage and treatment of water extracted from the South Platte River. The treated water is then supplied to meet the residents’ water demand and also to help improve the city's drought resilience (Diebel & Catalano, 2011).
Despite the numerous benefits, ASRs have known limitations. Martin (2013) edited a collection of essays focused on clogging in ASR, and Bloetscher et al. (2014) estimated that 26% of the ASR projects have been functionally abandoned due to geochemical reactions, clogging, and small percentage recovery of stored water. Brown et al. (2006) identified four key lessons learned from 50 ASR projects worldwide. First, well clogging, including air binding, is problematic. However, a potential solution is to incorporate regular back flushing programs. Second, water quality can diminish the usefulness of ASR, particularly when arsenic, iron, manganese, or other metals can be released from the local geologic materials. Third, hydraulic analysis is important when evaluating multi-well clusters in order to avoid interference with other wells. Fourth, alluvial ASR facilities need to incorporate monitoring equipment, such as sampling ports on recharge or discharge lines, to allow real-time monitoring of specific conductivity and turbidity.
Addressing these technical limitations, however, is not sufficient for water resources management because water resources exist within a broader context of policy constraints. The foundations of modern water law in Europe and the New World were formulated nearly two thousand years ago by Roman jurists who were inspired by Greek philosophy of reason. Recognizing that vital natural elements such as water, air, and the sea were governed by immutable natural laws, they reasoned that these elements belonged to all humans, and therefore cannot be owned as private property. Legally, such public property was to be governed by jus gentium, the law of all people or the law of all nations. Remarkably, jus gentium continues to be relevant in our contemporary society in which science plays a pivotal role in exploiting vital resources common to all (Narasimhan, 2008). For example, in the western United States, most states operate under the doctrine of prior appropriation (McIntyre & Mays, 2017) subject to Federal regulations, for example, for protection of endangered species, and subject to interstate compacts allocating river flows between upstream and downstream states. For example, the state of Colorado is entitled to one-third of the total water available within its territories, while the remaining two-thirds is sent downstream to neighboring states. Colorado water law is based upon the recognition that the right to use decreed water, either surface water or groundwater, is the property of the owner and therefore protected by law from injury (McIntyre & Mays, 2017). Injury occurs when a decreed water right cannot be diverted in accordance with its decree, due to the actions of another, in which case, other water right holders have an opportunity to oppose a water rights application.
This study provides both technical and administrative analysis of a hypothetical ASR facility in the alluvial aquifer near the South Platte River of northeastern Colorado. The technical analysis describes a modeling approach that explicitly anticipates the widely recognized limitation of clogging, and provides an algorithm that can be implemented to effectively operate the ASR despite clogging. In addition, and critical to the operation of such a storage vessel, a detailed administrative analysis of legally available surface water was performed. The results demonstrate the feasibility of alluvial ASR in Colorado, and the discussion revisits several key assumptions used in this study, places the technical challenge of clogging in a broader framework, and provides recommendations for further research.
A recent study conducted by the Colorado Water Conservation Board estimated an expansive area of alluvial aquifer along the South Platte River that could potentially be useful for storage and recovery, specifically 6,500 km2 (2,500 mi2) from Denver downstream to the Colorado–Nebraska state line (Figure 1) (CDM Smith, 2013). Waskom (2013) compiled a table of all the alluvial aquifers present in the South Platte basin, from the Colorado Water Conservation Board's SB06-193 study on underground water storage, and concluded that an estimated 12 km3 (10 million ac-ft) of pore space is available for water storage and recovery in the alluvium. For comparison, this figure is larger than the 7.9 km3 (6.4 million ac-ft) of storage available in all the existing reservoirs across Colorado (Grigg, 2005).
Within this context, this study presents a case study for a conceptual alluvial ASR facility located adjacent to the South Platte River in northeastern Colorado, paying due attention to the relevant technical constraints (i.e., recharge and extraction rates, clogging effects, and consolidation) and administrative constraints (i.e., water availability, the doctrine of prior appropriation). The criteria for site selection were (1) available alluvium and (2) proximity to the Denver metropolitan area. Based on these criteria, the area selected for detailed study was along the main stem of the South Platte River between Denver and Greeley, Colorado. Along this reach, a candidate site was identified on the east bank immediately upstream of the Colorado Highway 7 bridge in Brighton, Colorado (Figure S1 in the Supplementary material, available with the online version of this paper). As is typical for Colorado's high plains, this site has high summer temperatures and high wind speeds, which would be problematic for surface storage. This hypothetical site is about 40 km (25 mi) northeast of Denver and has a storage capacity of 148,000 m3 (120 ac-ft).
Design of this conceptual alluvial ASR facility is shown in Figure S1, where the facility area is denoted by the box 195 m (640 ft) wide and 380 m (1,247 ft) long. Soils’ information, obtained by entering the site coordinates in the Web Soil Survey (US Natural Resources Conservation Service, 2016), shows three types of soils in and near this area: sandy alluvial land, loamy alluvial land with gravelly substratum, and wet alluvial land. All three soil types are indicative of gravelly soils, which makes the selected site reasonable for alluvial ASR. CDM Smith (2013) analyzed samples from the South Platte River alluvium, reporting median values for hydraulic conductivity K = 130 m/d (427 ft/d), transmissivity T = 1,300 m2/d (14,000 ft2/d), and specific yield Sy = 0.2. These values give a rough estimate of the saturated thickness b = T/K = 10 m (33 ft). With an assumed effective porosity of 20%, these dimensions equate to approximately 148,000 m3 (120 ac-ft) of potential storage volume. This alluvial ASR facility will be enclosed by an impermeable barrier, keyed into bedrock, which is required to exert dominion and control over the water as required by Colorado water law (Jones & Cech, 2009; C.R.S. 37-82-101; C.R.S. 37-90-137). This could be constructed, for example, from interlocking sheet piles or from a bentonite slurry placed in an excavated trench. Enclosing the alluvial ASR facility within an impermeable barrier would also obviate the legal concerns about losing control over recharged water that were recently articulated by Potyondy (2017).
Water from the South Platte River is diverted into this facility by gravity flow through a canal (Figure 2). Water storage and recovery at this alluvial ASR facility is achieved through a manifold comprising eight symmetrically placed recharge–extraction wells connected by a horizontal manifold of perforated pipes at a depth of 9.5 m (31 ft) (Figure 3). Perforated pipes were placed at a depth of 9.5 m (31 ft) rather than the assumed alluvial aquifer thickness of 10.0 m (33 ft) to provide a minimum of 0.5 m (1.6 ft) of saturated depth, which accelerates the rate at which the system approaches hydraulic equilibrium during extraction (compared to a scenario with no residual saturated depth). The rationale for designing wells connected to a horizontal manifold of perforated pipes is to prevent a scenario where a high extraction rate pumps out water from the well faster than it can be replenished by the aquifer, leading to dry pore spaces around the well and air-locking in pumps. The horizontal manifold of perforated pipes precisely helps in avoiding this scenario by moving water from around the wells and into the perforated pipe network, permitting a symmetrical drawdown in the overall system.
Extraction requires pumps, the electrical power for which is likely the most significant energy expense to operate this alluvial ASR facility. Appurtenances to facilitate in the operation of this alluvial ASR facility, such as water treatment systems and instruments to test water quality, water level, etc., would result in minor additional energy expense.
A groundwater simulation model, developed using MODFLOW (Barlow & Harbaugh, 2006), was constructed to evaluate the technical feasibility of the alluvial ASR facility described above. The model assumes a two-layered unconfined aquifer with effective porosity (i.e., specific yield) of 0.20, homogeneous hydraulic conductivity of 130 m/d (427 ft/d), 10.0 m depth-to-bedrock, and grid blocks of 5 × 5 m (16.4 × 16.4 ft). The top layer is 9.5 m (31 ft) thick. The bottom layer is 0.5 m (1.6 ft) thick and is assumed to be saturated with water at all times (which prevents dry cells during simulation). The perforated pipes were modeled with a horizontal rectangle of cells in the bottom layer, where the hydraulic conductivity was set to be very high at 2 × 10+08 m/d (6.6 × 10+08 ft/d), such that water could move through the horizontal manifold with essentially zero head loss. Wells were modeled in the top layer using MODFLOW's well package, such that the bottom of each well contacted the top of the rectangle of high hydraulic conductivity cells representing the perforated pipes. Water recharge was simulated assuming an initial water depth of 1.5 m (5 ft); water extraction was simulated assuming an initial water depth of 9.5 m. Therefore, the effective depth of water is 8 m, which translates to 80% useable storage volume, 118,500 m3 (96 ac-ft). Recharging and pump-out simulations were implemented with MODFLOW's PCG2 solver (Hill, 1990), which uses modified incomplete Cholesky and polynomial methods to check convergence conditions in each cell, thereby preventing dry cell conditions when water is being extracted through wells.
The model was used to answer three questions. First, what is a practical operating protocol to recharge and extract water without triggering surface ponding or air-locking pumps? Second, how much time is required to fill or empty the facility? Third, to account for clogging, how do results depend on reduced hydraulic conductivity?
The operating protocol for this alluvial ASR facility was designed to be practical: (1) preventing surface ponding during recharge, (2) preventing air-locking in pumps during extraction, (3) minimizing the number of times when pumps require adjustment, and (4) minimizing the need for adjustments in the night. Figure 4 shows the algorithm for selecting recharge rates of this alluvial ASR facility. The recharge protocol is determined by setting the recharge rate to an assumed maximum diversion rate of 2,500 m3/d (1 cfs) for 12 hours during the first time period and then checking the average head in the alluvial ASR facility. If the average head is below 10 m (33 ft), then recharge is continued at the same rate for an additional 12 hours. If the average head is above 10 m (33 ft), then the current time period's length is reduced by 12 hours, and a new time period is added with the recharge rate reduced to half the previous time period's rate. The algorithm continues until the average head is between 9.5 m (31 ft) and 10 m (33 ft), indicating that the facility is recharged.
The extraction protocol is similar, starting with an extraction rate of 2,500 m3/d (1 cfs) for 12 hours during the first time period and then checking the average head in the alluvial ASR facility. If no dry cells are encountered in the MODFLOW simulation, then extraction is continued at the same rate for an additional 12 hours. If dry cells are encountered, then a new time period is added with the extraction rate reduced to half the previous time period's rate. The algorithm continues until the average head is below 1.0 m (3.3 ft), indicating that the facility is drained.
The question of clogging is a critical design and operational consideration in any proposed alluvial ASR project (Martin, 2013). The alluvial ASR facility at Brighton, Colorado has an assumed unclogged hydraulic conductivity of K = 130 m/d (427 ft/d). To simulate clogging, additional simulations were performed while successively halving the hydraulic conductivity to K/2, K/4, and K/8. In addition, a geotechnical analysis was performed to determine whether repeated recharge–extraction cycles could cause subsidence (see Supplementary material, available with the online version of this paper).
Within the broad context of jus gentium, water law in Colorado is based on the doctrine of prior appropriation (Hobbs, 2007; McIntyre & Mays, 2017), which grants water rights in the order in which they were put to beneficial use, with accounting for expected return flows via surface- or groundwater. When flow is insufficient to meet demand, those holding senior water rights may place a call on the river to guarantee delivery of their full appropriation. Accordingly, feasibility of the alluvial ASR facility design depends not only on the technical considerations discussed above, but also on administrative feasibility. For surface water diversions, Colorado water law establishes a date of priority decreed by the water court and administered by Colorado Division of Water Resources, which is also known as the State Engineer. The South Platte River is over-appropriated, meaning there is seldom sufficient water to satisfy all decreed surface water rights (HDR Engineering & West Sage Water Consultants, 2015). This administrative analysis was necessary since a junior water storage right, which this storage facility would be granted in the water court, would not necessarily constitute a dependable water source.
Legal diversions to a new water storage structure require a court decree with a priority date. The alluvial ASR facility at Brighton would be granted a junior priority date, meaning more senior water rights would take priority. Therefore, an analysis of the nature of the flow in the river at Brighton is required to access if and when the vessel can fill. Two administrative classifications of surface water are available to fill the alluvial ASR facility (HDR Engineering & West Sage Water Consultants, 2015): (1) free river and (2) reusable effluent water.
The Colorado State Engineer defines free river as ‘a condition that exists for a reach of a river when there is sufficient natural supply to satisfy all water uses, whether decreed or undecreed, and State Engineer administration is unnecessary for the protection of decreed water rights’ (Wolfe, 2015). This excess water in the river is available to even the most junior water right holder, making it possible to store water in alluvial ASRs during periods of peak flow. Wolfe (2015) instructs that anyone may use water during free river conditions for an undecreed use after confirming with the State Engineer that a free river exists and that the water will be put to beneficial use. Undecreed uses need not comply with terms of a decreed water right. Conversely, decreed users must still comply with the terms of their decreed water right.
In general, free river exists when there is no call affecting the stretch of river at the time water is to be extracted for storage. Absence of call confirms that no senior water decree has been violated, so that diversions to meet decreed rights are not necessary. Under this scenario, water can be extracted from the river for any undecreed use without the intervention of the State Engineer. Free river availability at the Brighton site was analyzed based upon 14 years of data from water years 2000–2013, where each water year begins on October 1st (i.e., water year 2000 is October 1st, 1999 to September 30th, 2000).
Among the many stream gages on the South Platte River used by the US Geological Survey and the Colorado State Engineer, Henderson Gage is the closest to Brighton. It is assumed that flow on the South Platte River at the Brighton site and at the Henderson Gage are equal, since no water diversion or augmentation occurs between the two places. The timing, discharge, and volume of free river were determined using the point-flow method described in Appendix G of HDR Engineering & West Sage Water Consultants (2015). In brief, the point-flow method quantifies the free river available as the minimum streamflow between and including the point of measurement, such as the Brighton site, to the farthest downstream point of streamflow measurement within Colorado, in this case, Liddle Ditch. The South Platte River Compact (C.R.S. 37-65-101) between Colorado and Nebraska, mandates Colorado to send a minimum flow of 3.4 m3/s (120 cfs) across the state border to Nebraska from April 1st to October 15th every year. To account for this compact, during point-flow analysis, a virtual call of 3.4 m3/s (120 cfs) is placed at Liddle Ditch. In addition, the free river available, calculated using point-flow analysis, must be reduced by the daily reusable effluent data described below. Free river analysis was carried out using a script in RStudio (version 1.0.136) (RStudio Team, 2016) that is provided in Text S1 of the Supplementary material. Raw data of daily stream flow and daily call chronology were downloaded (HDR Engineering & West Sage Water Consultants, 2015), then saved into two separate comma-delimited text files provided in the Supplementary material, namely, ds01.xls and ds02.xls. Daily stream flow data, contained in ds01.xls, have units of cubic feet per second (cfs). Daily call chronology, contained in ds02.xls, follows the convention that 0 denotes no call and 1 indicates a call placed on the noted reach of the South Platte River between October 1st, 1999 and September 30th, 2013.
Reusable effluent water
The second administrative classification of surface water available to fill the alluvial ASR facility is fully reusable effluent water, which is water supply from non-tributary sources. Non-tributary sources are those that would not be present in the river in the absence of engineering intervention, such as tunnels under the continental divide that provide water from the western slope of the Rocky Mountains, or wells extracting non-tributary groundwater. At the Brighton case study alluvial ASR facility, a portion of the flow in the South Platte River is fully reusable water from the upstream cities of Denver and Aurora. This water could be exchanged, that is, have its point of diversion moved upstream, to fill the vessel. The yearly average of available fully reusable effluent water from Denver and Aurora to store in the alluvial ASR facility is found in Table S1 in the Supplementary material (HDR Engineering & West Sage Water Consultants, 2015). This reusable effluent water data were provided by Joe Frank (South Platte Basin Roundtable, unpublished data, 2015) through personal communication, and can found in the spreadsheet, ds03.xls, of the Supplementary material. Under Colorado water law, municipal water providers are entitled to capture their fully reusable effluent water for additional use. Hence, this reusable effluent water is subtracted from the free river availability calculated by the point-flow method discussed above.
Assuming a maximum diversion rate of 2,500 m3/d (1 cfs), and before the onset of clogging, this alluvial ASR facility could be filled up from 1.5 m (4.9 ft) to 9.5 m (31 ft) in approximately 6.5 days. Subsequently, this 8 m (26 ft) depth of water could be extracted in approximately 9.0 days. Figures S2 and S3, respectively, show the time taken to completely recharge and empty the alluvial ASR facility under various levels of clogging. As clogging increased, no qualitative change in hydraulic results was observed, so clogging did not change any of the assumed initial design parameters such as pipe diameter or porosity. Increasing clogging did trigger the need for additional time periods, leading to multiple pumping rates being needed to recharge and empty the alluvial ASR facility. Under a clogging factor of 8, simulations indicate that this alluvial ASR facility can be recharged in 24 days and emptied in 20 days. Figure 5 shows the relationship between pumping time and clogging factor (Ko/K) for this alluvial ASR facility, illustrating how more time is needed to recharge and extract water as the clogging factor increases. Thus, under severe clogging, a small pumping rate over a long time period may be necessary to recharge or empty the alluvial ASR facility. The geotechnical analysis predicted negligible subsidence (see Supplementary material, available with the online version of this paper).
Free river analysis
Figure S4 shows free river available daily and Figure S5 shows annual free river. Water years 2002 and 2003 occurred during a historic drought with no free river available; water years 2000 and 2009 had high flows and consequently frequent occurrence of free river. Figure 6 shows the monthly mean, minimum, and maximum discharge of free river analyzed over those 14 water years. April, May, and June have the largest number of days with free river thanks to the spring melt. By the end of summer, free river conditions are fewer and more calls are placed on the river. January and February have least free river availability. Results obtained by the point-flow method for free river analysis were validated by comparison with the water availability model for the Henderson Gage (HDR Engineering & West Sage Water Consultants, 2015). Total free river availability calculated from water years 2000 to 2012 closely matched with the validation data except for September 2013, whose flood was omitted from the validation data. Free river analysis from water years 2000 to 2013 established that free river conditions exist for an average of 103 days per year, with an average flow of 2.15 m3/s (76 cfs). This availability more than satisfies the recharge rate needed to fill the alluvial ASR facility.
Discussion and conclusion
This study has demonstrated, through technical and administrative analysis, the feasibility of alluvial ASR to provide a potentially beneficial option for engineered water storage. The hypothetical facility in the alluvium of the South Platte River near Brighton, Colorado was selected to be representative of other streams and rivers where the climate is semi-arid and water administration follows the doctrine of prior appropriation. Similar hydrologic analysis would be applicable at other facilities worldwide, but the key contribution of this study was the coupled analysis of hydrogeological and administrative aspects. The details of this analysis are limited to regions following the doctrine of prior appropriation, and specifically to Colorado, but this analysis can also be effectively extended to other parts of the globe where the Roman law of jus gentium is used as a fundamental principle in managing resources. It is one thing to design an alluvial storage and recovery (ASR) structure; it is quite another to operate it within the significant constraint of existing water law. To our knowledge, this study is one of the first to integrate free river and reusable effluent water for storage into ASR facilities, based on the analysis of historical flow data, and come up with a practical pumping protocol for ASR facilities under various levels of clogging.
There are a number of limitations to this study that would need to be addressed before deployment of a field demonstration. These limitations fall under two general categories: (1) practical considerations and (2) aquifer clogging. Turning first to the practical considerations, a field deployment would require due consideration of legal, environmental, and financial factors. Legal factors comprise both the availability of water through free river or reusable effluent and also the ability to purchase or rent the required land with protective covenants to prevent contamination stemming from incompatible land uses. The applicability of numerous federal, state, and local regulations including the US Army Corps of Engineers 404 permit, the National Environmental Policy Act, the Endangered Species Act, the Platte River Recovery Implementation Program, the South Platte River Compact, Colorado water rights administration, and local 1041 regulations make the ASR facility application process extensive (Stantec & Leonard Rice Engineers, 2017). Stantec & Leonard Rice Engineers’ (2017) final report South Platte storage, released in accordance with Colorado Legislature's HB16-1256 also mentions, ‘… the current rules and regulations for management and control of designated ground water are not explicit as to the nature of free river water that is diverted from the South Platte River and placed in the designated basin for later withdrawal and export.’ This stretches the time to acquire permits for a new ASR facility. Environmental factors comprise a historical review of prior site activities and sampling to confirm that the site is free of contamination that would trigger the need for potentially expensive treatment processes. Financial factors comprise the economic analysis required to calculate the construction and operating costs, presumably on the basis of cost per water volume stored, which would allow comparison to other storage options such as construction or expansion of in-stream or off-stream dams. The financial feasibility of alluvial ASR would also depend on the value of water stored upon delivery, where delivery could be through pipes, ditches, or other conveyances connected to the groundwater extraction wells (Figure 3). Perhaps, more likely, delivery could instead be through an augmentation plan (Topper et al., 2004) by which extracted water would be delivered to downstream uses via the South Platte itself. Each of these legal, environmental, and financial factors would need to be addressed, but none of these is expected to undermine the overall feasibility of alluvial ASR.
A potentially more challenging consideration is clogging, defined as a detrimental reduction in hydraulic conductivity resulting from physical, chemical, or biological processes. There are examples of ASR without clogging (Hemenway & Grundemann, 2002), but in general, clogging is recognized to be a major limitation for ASR technology (Pyne, 1995: §5.4; American Society of Civil Engineers, 2001), both worldwide (Martin, 2013) and specifically in Colorado (Topper et al., 2004). Clogging can result from one or more of five mechanisms: filtration of suspended solids (i.e., colloids); precipitation of dissolved minerals; growth of bacterial biofilms; air binding from entrapped gases; and consolidation of the porous media itself (Pyne, 1995: §4.2). To reduce clogging, the standard recommendation is to minimize the total suspended solids (TDS) and organic contents (measured as biochemical oxygen demand or BOD) of recharged water (American Society of Civil Engineers, 2001). However, clogging in ASR has been observed even when recharging treated drinking water (Pyne, 1995), which illustrates that clogging depends not only on the constituents of the recharged water, but also on the compatibility of the recharged water and the geochemical conditions within the aquifer. Geochemical conditions are particularly important with regard to clay minerals, which are colloids that can reduce permeability by up to three orders of magnitude when their structure changes in response to changes in groundwater chemistry. This key idea has been known for many decades within the soil science literature, for example, Quirk & Schofield (1955) showed how soil permeability depends on salinity and sodicity through their ability to control the structure of clay minerals in soils, and the American Society of Civil Engineers (2001) reported that recharge water with high sodicity resulted in ASR clogging. This clogging mechanism is important because clay minerals are ubiquitous in soils. The fundamental connection between colloid science and permeability in porous media is an area of active research (e.g., Mays, 2010, 2013; Roth et al., 2015). As a practical matter, deployment of alluvial ASR would require both pilot testing within the intended water supply and the intended geologic context, and also pretreatment to minimize TSS and BOD. The pretreatment, in turn, could be based on the extensive literature on best management practices (BMPs) for treatment of stormwater, including but not limited to detention basins for settling out TSS and engineered wetlands to consume BOD (National Research Council, 2008). Future work toward deployment of alluvial ASR should therefore emphasize the adaptation of stormwater BMPs to treatment of river water.
Despite the legal and technical constraints, ASR facilities are being widely considered for streamflow management through augmentation. Ronayne et al. (2016) highlight the benefits of managed aquifer recharge for groundwater–surface water exchange to augment streamflow during the high-water-demand summer season by capturing water during the low-water-demand months from December to March. ASR facilities are being increasingly considered as alternative storage options due to the limitations of surface reservoirs and dams (Stantec & Leonard Rice Engineers, 2017), and are seen as viable management practices for improving the summer season low flows through augmentation and for basin exchanges.
The authors would like to thank Mr Brent Schantz and Mr Colin Watson at the Office of the State Engineer in Greeley, Colorado for hosting a field trip to the South Platte Basin and for providing data on reusable effluent water. The authors give special thanks to Mr Joe Frank, Chair of the South Platte Basin Roundtable, for his valuable insights on free river analysis. The authors also thank two anonymous referees for their constructive feedback. The University of Colorado Denver provided financial support to the first author but played no role in study design; in the collection, analysis, and interpretation of data; or in the decision to submit the article for publication.