A hydrogeologic model of submarine groundwater discharge: Florida intercomparison experiment

A hydrogeologic model of submarine groundwater discharge (SGD) to the near-shore environment at a site on the northeast Gulf of Mexico has been developed to provide a basis for comparison with measurements of SGD made using seepage meters, and with estimates derived from chemical tracers. The hydrogeologic model incorporates the seaward movement of fresh water and the recirculation of sea water at the fresh water–salt water interface. The hydrostratigraphy at the site includes the Surficial Aquifer, a thin confining unit known as the Intracoastal Formation, and the underlying Upper Floridan Aquifer. It is not possible to explain either the magnitude or spatial distribution of SGD recorded by the seepage meters, or the magnitude of SGD estimated using radium and radon tracers, if only steady state flow in the Surficial Aquifer is considered. Nor does it appear likely that the difference between the model-based prediction of SGD and the field-based estimates can be fully resolved by leakage across the Intracoastal Formation from a source in the Floridan Aquifer. These results suggest that processes driven by variations in fluid pressure in the marine water column, which occur on a variety of time scales, be examined to quantify their contribution to fluid circulation within and discharge from that segment of the Surficial Aquifer located beyond the low tide line.     

Submarine groundwater discharge in Osaka Bay, Japan

Submarine groundwater discharge (SGD) rates in Osaka Bay were continuously measured and analyzed to evaluate seawater–groundwater interactions. Fast Fourier transfer and power spectrum density methods were applied to analyze the dominant periods of the SGD variations. Diurnal and semidiurnal periods of SGD variation were found, and they were caused by tidal effects. According to the separation of SGD into fresh and recirculated water components using automated seepage meter measurements and terrestrial groundwater flow analyses, the fresh groundwater component in SGD was evaluated to be in the range 4%–29% at Tannowa, Osaka. Therefore, SGD rates depend mainly on the volume of recirculated seawater. Correlation analyses between SGD and sea level show that SGD is delayed by 4h after sea level changes.     

Magnitude and variations of groundwater seepage along a Florida marine shoreline

Direct groundwater inputs are receiving increasingattention as a potential source of nutrients and otherdissolved constituents to the coastal ocean. Seepageinto St. George Sound, Florida was measuredextensively from 1992 to 1994 using seepage meters. Spatial and temporal variations were documented alonga 7-km stretch of coastline and up to 1 km from shore. Measurements were made at 3 transects perpendicular toshore and 1 transect parallel to shore. The generalresults indicated that seepage decreased with distancefrom shore (2 of 3 transects), and substantialtemporal and spatial variability was observed inseepage flow from nearshore sediments. In addition,trends in mean monthly integrated seepage rates weresimilar to precipitation patterns measured at a nearbycoastal weather station. Based on these measurements, weestimate that the magnitude of groundwater seepage intothe study area is substantial, representing from 0.23 to4.4 m3 sec-1of flow through the sediments, approximately equivalentto a first magnitude spring. Although it is unknown howrepresentative this region is with respect to globalgroundwater discharge, it demonstrates thatgroundwater flow can be as important as riverine andspring discharge in some cases. Our subsurfacedischarge rates suggest groundwater is an importanthydrologic source term for this region and may beimportant to the coastal biogeochemistry as well.     

Spatial and temporal distributions of submarine groundwater discharge rates obtained from various types of seepage meters at a site in the Northeastern Gulf of Mexico

Direct measurements of submarine groundwaterdischarge (SGD) were taken by three different(continuous heat, heat pulse, and ultrasonic)types of automated seepage meters as well asstandard Lee-type manually operated meters. SGD flux comparisons and the spatial andtemporal variations in groundwater flow wereanalyzed. Seepage rates measured by thedifferent meters agree relatively well witheach other (more than 80% agreement). Comparisons of flux rates as a function ofdistance offshore using exponentialapproximations show that more than fivemeasurement locations (200 m offshore) areneeded for a precise integrated estimation ofSGD offshore within an accuracy of ±10%. Thedominant period of seepage variations isestimated to be about 12 hours, which closelymatches the semidiurnal tides in this area. Our analysis also shows that short durationmeasurement periods may cause significantunderestimates or overestimates of the dailyaveraged groundwater flow rates (±25%–±60% difference when the measurement durationis less than 12 hours). Thus, continuousmeasurements of SGD using automated seepagemeters with high time resolution should enableus to evaluate temporal and spatial variationsof dissolved material transports viagroundwater pathways. Such inputs may affectbiogeochemical phenomena in the coastal zone.     

Seepage rate variability in Florida Bay driven by Atlantic tidal height

Atlantic tidal fluctuations drive pressure head variations in shallow offshore wells drilled into the limestone subsurface on both the Florida Bay and Atlantic sides of Key Largo, Florida, USA. We tested the hypothesis that these pressure head variations influence groundwater flow and that flux rate variability is associated with tidal variability. We used an automated Rn monitor to make continuous measurements of ²²²Rn, a natural tracer of groundwater discharge, in Florida Bay waters. We also deployed three types of seepage meters, including an automated heat pulse meter to collect a continuous record of seepage from the sediments. Drum type seepage meters inserted into soft sediments and fiberglass meters cemented to the rocky bay floor were utilized with pre-filled 4-l bag collectors, and monitored on an hourly basis. Maximum Rn inventories in Florida Bay waters were associated with high tide on the Atlantic side of the island. Modeling of the Rn variation indicated variable groundwater discharge rates with maximum flux occurring at high Atlantic tide. Seepage meter results in Florida Bay were consistent with ²²²Rn modeling. Florida Bay seepage meter rates showed positive correlation with Atlantic tide, meter 1, r?=?0.63, n?=?12, p?<?0.025 and meter 2, r?=?0.67, n?=?12, p?<?0.025. A seepage meter offshore of the Atlantic side of Key Largo exhibited rates that were inversely correlated with Atlantic tide (r?=?0.87, n?=?9, p?<?0.005) showing negative rates when the tide was high, and positive rates when the tide was low. Overall, our results are consistent with the hypothesis of Reich et al. (2002), that pressure head variations driven by Atlantic tide influence groundwater seepage rate variability in Florida Bay off Key Largo. Effectively, as proposed by Reich et al. (2002), Key Largo functions as a semi-permeable dam separating Florida Bay and the Atlantic Ocean.     

Groundwater and pore water inputs to the coastal zone

Both terrestrial and marine forces drive underground fluid flows in the coastal zone. Hydraulic gradients on land result in groundwater seepage near shore and may contribute to flows further out on the shelf from confined aquifers. Marine processes such as tidal pumping and current-induced pressure gradients may induce interfacial fluid flow anywhere on the shelf where permeable sediments are present. The terrestrial and oceanic forces overlap spatially so measured fluid advection through coastal sediments may be a result of composite forcing. We thus define “submarine groundwater discharge” (SGD) as any and all flow of water on continental margins from the seabed to the coastal ocean, regardless of fluid composition or driving force. SGD is typically characterized by low specific flow rates that make detection and quantification difficult. However, because such flows occur over very large areas, the total flux is significant. Discharging fluids, whether derived from land or composed of re-circulated seawater, will react with sediment components. These reactions may increase substantially the concentrations of nutrients, carbon, and metals in the fluids. These fluids are thus a source of biogeochemically important constituents to the coastal ocean. Terrestrially-derived fluids represent a pathway for new material fluxes to the coastal zone. This may result in diffuse pollution in areas where contaminated groundwaters occur. This paper presents an historical context of SGD studies, defines the process in a form that is consistent with our current understanding of the driving forces as well as our assessment techniques, and reviews the estimated global fluxes and biogeochemical implications. We conclude that to fully characterize marine geochemical budgets, one must give due consideration to SGD. New methodologies, technologies, and modeling approaches are required to discriminate among the various forces that drive SGD and to evaluate these fluxes more precisely.     

Biogeochemical processes in the groundwater discharge zone of urban streams

The influence of biogeochemical processes on nitrogen and organic matter transformation and transport was investigated for two urban streams receiving groundwater discharge during the dry summer baseflow period. A multiple lines of evidence approach involving catchment-, and stream reach-scale investigations were undertaken to describe the factors that influence pore water biogeochemical processes. At the catchment-scale gaining stream reaches were identified from water table mapping and groundwater discharge estimated to be between 0.1 and 0.8 m³ m^?2 d^?1 from baseflow analysis. Sediment temperature profiles also suggested that the high groundwater discharge limited stream water infiltration into the sediments. At the stream reach-scale, dissolved organic carbon (DOC) and dissolved organic nitrogen (DON) concentrations were higher in stream water than in groundwater. However, DOC and DON concentrations were greatest in sediment pore water. This suggests that biodegradation of sediment organic matter contributes dissolved organic matter (DOM) to the streams along with that delivered with groundwater flow. Pore water ammonium (NH₄ ⁺) was closely associated with areas of high pore water DOM concentrations and evidence of sulfate (SO₄ ^2?) reduction (low concentration and SO₄:Cl ratio). This indicates that anoxic DOM mineralization was occurring associated with SO₄ ^2? reduction. However the distribution of anoxic mineralization was limited to the center of the streambed, and was not constrained by the distribution of sediment organic matter which was higher along the banks. Lower sediment temperatures measured along the banks compared to the center suggests, at least qualitatively, that groundwater discharge is higher along the banks. Based on this evidence anoxic mineralization is influenced by groundwater residence time, and is only measurable along the center of the stream where groundwater flux rates are lower. This study therefore shows that the distribution of biogeochemical processes in stream sediments, such as anoxic mineralization, is strongly influenced by both the biogeochemical conditions and pore water residence time.     

Groundwater seepage stimulates the growth of aquatic macrophytes

1. The impact of groundwater seepage on the growth of submerged macrophytes was investigated in experiments on the isoetid Littorella uniflora and the elodeid Myriophyllum alterniflorum both in the laboratory and in the field. Isoetids rely mostly on sediment‐derived CO₂ and nutrients via root uptake, whereas elodeids acquire their inorganic carbon and nutrients from the water column. We thus hypothesised that L. uniflora would respond positively to seeping ground water as it should improve both CO₂ and nutrient supply. 2. Laboratory experiments were conducted by percolating vegetated cores containing natural sediment or technical sand with artificial ground water of high CO₂ concentrations and with either high or low levels of nutrients. Field experiments were conducted in the oligotrophic Lake Hampen, Denmark, with custom‐built seepage‐growth chambers that permitted a near‐natural flow‐through of seeping ground water. Chambers with a solid bottom, and thus no flow‐through of seeping ground water, served as controls in both laboratory and field experiments. In the field, seepage chambers were installed at a site with relatively high seepage fluxes (ground water from forest catchment), at a site with much lower seepage fluxes but with higher nutrient concentrations (ground water from agricultural catchment) and at a reference site with no net discharge or recharge of ground water. 3. Positive growth responses were observed in the field at transects with high groundwater discharge compared to the control chambers with no seepage. No growth response was observed at the reference transect with low or alternating direction of groundwater seepage. The growth rates of L. uniflora in the field were significantly higher in seepage treatments compared to control treatments, and final plant mass was up to 70% higher than that for plants where seepage was excluded. In areas with high groundwater discharge, a strong positive correlation was found between groundwater seepage fluxes, growth rates, and final plant mass for L. uniflora, while there was no such relationship at the reference transect. The growth of M. alterniflorum was also significantly affected by groundwater seepage, but to a lesser degree than L. uniflora. Laboratory experiments generally showed the same trend for both L. uniflora and M. alterniflorum, and the positive influence of seeping ground water was apparently related to increased inorganic carbon supply and, to a lesser degree, improved nutrient availability. 4. Groundwater discharge results in enhanced growth of isoetids and to some extent elodeids inhabiting a groundwater‐fed softwater lake. We propose that the shallow dense vegetation present where most of the discharge takes place acts as a biological filter that retains nutrients that otherwise would end up in the water column and could result in increased algal growth.     

Submarine groundwater discharge estimates at a Florida coastal site based on continuous radon measurements

The direct discharge of groundwater into thecoastal zone has received increased attentionin the last few years as it is now recognizedthat this process represents an importantpathway for material transport. Assessingthese material fluxes is difficult, as there isno simple means to gauge the water flux. Weestimated the changing flux of groundwaterdischarge into a coastal area in the northeastGulf of Mexico (Florida) based on continuousmeasurements of radon concentrations over aseveral day period. Changing radon inventorieswere converted to fluxes after accounting forlosses due to atmospheric evasion and mixing. Radon fluxes are then converted to groundwaterinflow rates by estimating the radonconcentration of the fluids discharging intothe study domain. Groundwater flow was also assessed via seepagemeters, radium isotopes, and modeling duringthis period as part of an ``intercomparison''study. The radon results suggest that the flowis: (1) highly variable with flows ranging from～5 to 50 cm/day; and (2) strongly influenced bythe tides, with spikes in the flow every 12hours. The discharge estimates and pattern offlow derived from the radon model matches theautomated seepage meter records very closely.     

Effects of Changing Rainfall on the Limnology of a Mediterranean,Flowthrough‐Seepage Chain of Lakes

Relationships between groundwater and lake ecology are often overlooked, but they may be strong, particularly in seepage lakes. As a result, the nature and degree of groundwater effects on lakes are usually neglected. In this study interactions among rainfall, groundwater and surface water and their limnological effects were traced seasonally for two years of changing rainfall in a Spanish flowthrough, seepage lake complex. Cumulative rainfall dictated recharge of groundwater with delays of nine months. Groundwater discharge, in turn, increased surface discharge downstream. Mediated by the geographical setting of lakes, both fluxes impinged on lake water renewal time, but effects of the latter on limnological variables were much stronger at the district scale than at the single lake scale. These water‐renewal effects included the following: decreasing salinity, total phosphorus concentration and phytoplankton biomass and increasing water transparency and total nitrogen concentration as water renewal shortened, the nitrogen effect arising because of nitrate‐rich water entering the lakes as groundwater levels rose. This complex response of a Mediterranean lake district to water availability may also be expected in cold temperate lakes as climate change effects become stronger. (© 2006 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)     

Groundwater discharge from the superficial aquifer into Cockburn Sound Western Australia: estimation by inshore water balance

Submarine groundwater discharge (SGD) into Cockburn Sound Western Australia was quantified by applying a distributed groundwater flow model to estimate the inshore aquifer water balance. Spatially averaged SGD along the coast was estimated to be 2.5–4.8?±?0.9?m³?day^?1?m^?1. The range in estimated average SGD reflected low and high estimates of average groundwater recharge, which ranged from 0.13 to 0.24?m?year^?1 (15–28% of average annual rainfall). The error ±0.9?m³?day^?1?m^?1 was calculated by assuming arbitrary ±20% errors in groundwater pumping and inflow across boundaries. SGD varied spatially along the coastal boundary due to variation in hydraulic connection between the coastal aquifers and ocean, and spatial variability in recharge, transmissivity and pumping. Under assumptions of low and high groundwater recharge, SGD along the coastline varied in the ranges 1.4–4.6?m³?day^?1?m^?1 and 2.4–7.9?m³?day^?1?m^?1, respectively.     

Denitrification in Aquifer Soil and Nearshore Marine Sediments Influenced by Groundwater Nitrate

We estimated rates of denitrification at various depths in sediments known to be affected by submarine discharge of groundwater, and also in the parent aquifer. Surface denitrification was only measured in the autumn; at 40-cm depth, where groundwater-imported nitrate has been measured, denitrification occurred consistently throughout the year, at rates from 0.14 to 2.8 ng-atom of N g⁻¹ day⁻¹. Denitrification consistently occurred below the zone of sulfate reduction and was sometimes comparable to it in magnitude. Denitrification occurred deep (14 to 40 cm) in the sediments along 30 km of shoreline, with highest rates occurring where groundwater input was greatest. Denitrification rates decreased with distance offshore, as does groundwater influx. Added glucose greatly stimulated denitrification at depth, but added nitrate did not. High rates of denitrification were measured in the aquifer (17 ng-atom of N g⁻¹ day⁻¹), and added nitrate did stimulate denitrification there. The denitrification measured was enough to remove 46% of the nitrate decrease observed between 40- and 14-cm depth in the sediment.     

Mangroves buffer marine protected area from impacts of Bovoni Landfill,St. Thomas,United States Virgin Islands

One of the many ecosystem services that mangrove systems provide is their ability to act as buffers between the land and sea, protecting human development from storm surges while also trapping terrestrial pollutants. In St. Thomas, United States Virgin Islands, an ecologically-important mangrove system sits between Bovoni Landfill and a marine protected area, the St. Thomas East End Reserves. To characterize the physical processes driving this mangrove system, groundwater hydraulic head, sediment cores, sediment surface temperatures, and water and sediment chemistry were analyzed. Hydraulic head data from January to November 2014 were used to determine vertical and horizontal groundwater flow directions. Water and sediment samples were tested for heavy metals potentially originating from Bovoni Landfill. Stratigraphic context was provided by the sediment cores and used to infer past environmental conditions. Subsamples were taken from these cores and analyzed for dry bulk density, organic matter content (through loss on ignition), and heavy metals using electron microscopy. Vertical groundwater velocity and sediment porosity were determined by calibrating a one-dimensional finite difference heat transport model to near surface temperature data from depths of 0, 7, 14, and 21 cm. Groundwater was found to flow from the terrestrial upland, through the mangroves, and toward the ocean for the majority of the study. Flow reversal was seen after long periods of little precipitation. In the surface and shallow groundwater samples, trace metal concentrations were measured from 23 to 105 μg/L for Cr, Ni, Sn, and Zn. Sediment samples collected near the landfill contained Bi, Cr, Sn, Ti, and Zn. Very slow flushing of sediment pore water was indicated by the vertical groundwater velocities produced from the heat transport model, which ranged from ±10^–7 to ±10^–9 m/s. This study revealed that the mangrove system is an important buffer system protecting the outer lagoon of the marine protected area from terrestrial contaminants via sediment trapping and slowing of water fluxes from the upland area into the lagoon. The results presented here can be used as a baseline for future studies and are relevant to local managers and to landfill closure plans.     

The influence of tidal marshes on upland groundwater discharge to estuaries

We investigated subsurface hydrology in two fringing tidal marshes and in underlying aquifers in the coastal plain of Virginia. Vertical distributions of hydraulic conductivity, hydraulic head and salinity were measured in each marsh and a nearby subtidal sediment. Discharge of hillslope groundwater into the base of the marshes and subtidal sediment was calculated using Darcy's law. In the marshes, fluxes of pore water across the sediment surface were measured or estimated by water balance methods. The vertical distribution of salt in shoreline sediments was modeled to assess transport and mixing conditions at depth. Hydraulic gradients were upward beneath shoreline sediments; indicating that groundwater was passing through marsh and subtidal deposits before reaching the estuary. Calculated discharge (6 to 10 liters per meter of shoreline per day) was small relative to fluxes of pore water across the marsh surface at those sites; even where discharge was maximal (at the upland border) it was 10 to 50 times less than infiltration into marsh soils. Pore water turnover in our marshes was therefore dominated by exchange with estuarine surface water. In contrast, new interstitial water entering subtidal sediments appeared to be primarily groundwater, discharged from below. The presence of fringing tidal marshes delayed transport and increased mixing of groundwater and solute as it traveled towards the estuaries. Soil-contact times of discharged groundwater were up to 100% longer in marshes than in subtidal shoreline sediments. Measured and modeled salinity profiles indicated that, prior to export to estuaries, the solutes of groundwater, marsh pore water and estuarine surface water were more thoroughly mixed in marsh soils compared to subtidal shoreline sediments. These findings suggest that transport of reactive solutes in groundwater may be strongly influenced by shoreline type. Longer soil-contact times in marshes provide greater opportunity for immobilization of excess nutrients by plants, microbes and by adsorption on sediment. Also, the greater dispersive mixing of groundwater and pore water in marshes should lead to increased availability of labile, dissolved organic carbon at depth which could in turn enhance microbial activity and increase the rate of denitrification in situations where groundwater nitrate is high.     

The influence of tidal marshes on upland groundwater discharge to estuaries

We investigated subsurface hydrology in two fringing tidal marshes and in underlying aquifers in the coastal plain of Virginia. Vertical distributions of hydraulic conductivity, hydraulic head and salinity were measured in each marsh and a nearby subtidal sediment. Discharge of hillslope groundwater into the base of the marshes and subtidal sediment was calculated using Darcy's law. In the marshes, fluxes of pore water across the sediment surface were measured or estimated by water balance methods. The vertical distribution of salt in shoreline sediments was modeled to assess transport and mixing conditions at depth. Hydraulic gradients were upward beneath shoreline sediments; indicating that groundwater was passing through marsh and subtidal deposits before reaching the estuary. Calculated discharge (6 to 10 liters per meter of shoreline per day) was small relative to fluxes of pore water across the marsh surface at those sites; even where discharge was maximal (at the upland border) it was 10 to 50 times less than infiltration into marsh soils. Pore water turnover in our marshes was therefore dominated by exchange with estuarine surface water. In contrast, new interstitial water entering subtidal sediments appeared to be primarily groundwater, discharged from below. The presence of fringing tidal marshes delayed transport and increased mixing of groundwater and solute as it traveled towards the estuaries. Soil-contact times of discharged groundwater were up to 100% longer in marshes than in subtidal shoreline sediments. Measured and modeled salinity profiles indicated that, prior to export to estuaries, the solutes of groundwater, marsh pore water and estuarine surface water were more thoroughly mixed in marsh soils compared to subtidal shoreline sediments. These findings suggest that transport of reactive solutes in groundwater may be strongly influenced by shoreline type. Longer soil-contact times in marshes provide greater opportunity for immobilization of excess nutrients by plants, microbes and by adsorption on sediment. Also, the greater dispersive mixing of groundwater and pore water in marshes should lead to increased availability of labile, dissolved organic carbon at depth which could in turn enhance microbial activity and increase the rate of denitrification in situations where groundwater nitrate is high.     

Hydrogeologic Modeling of Submarine Groundwater Discharge: Comparison to Other Quantitative Methods

Submarine groundwater discharge (SGD) is approached differently by terrestrial hydrogeologists and marine scientists, including whether to incorporate recirculated seawater with freshwater in the definition. This paper focuses on the major hydrogeologic modeling/calculational methods, what component of SGD they quantify and on what scale. It then compares the modeling methods to direct measurement and geochemical techniques used by marine scientists. Hydrogeologic modeling methods focus primarily on freshwater, but recirculated seawater can be examined with density-dependent, solute transport numerical modeling. Direct physical measurements and geochemical tracers performed in the marine environment can quantify fresh, brackish, or seawater fluxes, so that they are not always comparable to the results of modeling. Because of differences in the geochemistry (nutrients and other dissolved species) of fresh and saline waters, for many applications it may be necessary to distinguish between the fresh and recirculated seawater components of SGD.     

An assessment of karstic submarine groundwater and associated nutrient discharge to a Mediterranean coastal area (Balearic Islands, Spain) using radium isotopes

Short and long-lived radium isotopes (²²³Ra, ²²⁴Ra, ²²⁶Ra, ²²⁸Ra) were used to quantify submarine groundwater discharge (SGD) and its associated input of inorganic nitrogen (NO₃ ^?), phosphorus (PO₄ ^3?) and silica (SiO₄ ^4?) into the karstic Alcalfar Cove, a coastal region of Minorca Island (Western Mediterranean Sea). Cove water, seawater and groundwater (wells and karstic springs) samples were collected in May 2005 and February 2006 for radium isotopes and in November 2007 for dissolved inorganic nutrients. Salinity profiles in cove waters suggested that SGD is derived from shallow brackish springs that formed a buoyant surface fresh layer of only 0.3 m depth. A binary mixing model that considers the distribution of radium activities was used to determine the cove water composition. Results showed that cove waters contained 20% brackish groundwater; of which 6% was recirculated seawater and 14% corresponded to freshwater discharge. Using a radium-derived residence time of 2.4 days, a total SGD flux of 150,000 m³ year^?1 was calculated, consisting of 45,000 m³ year^?1 recirculated seawater and 105,000 m³ year^?1 fresh groundwater. Fresh SGD fluxes of NO₃ ^?, SiO₄ ^4? and PO₄ ^3? were estimated to be on the order of 18,000, 1,140 and 4 μmol m^?2 day^?1, respectively, and presumably sustain the high phytoplankton biomass observed in the cove during summer. The total amount of NO₃ ^? and SiO₄ ^4? supplied by SGD was higher than the measured inventories in the cove, while the reverse was true for PO₄ ^3?. These discrepancies are likely due to non-conservative biogeochemical processes that occur within the subterranean estuary and Alcalfar Cove waters.     

Hydrologic Variability of Small, Northern Michigan Lakes Measured by the Addition of Tracers

The hydraulic residence time (or flushing rate of water) is a key variable for any aquatic ecosystem and is used in many types of models and calculations. Rather than being measured directly, the hydraulic residence time is usually inferred from estimates of watershed size, precipitation, and water yield. Such estimates can be problematic in any environment but are especially so in environments in which flat or complex topography makes delineations of mapped watershed boundaries difficult to discern. We added lithium bromide, (LiBr) to three small seepage lakes in the flat topography of the Upper Peninsula of Michigan to provide an independent estimate of the water residence time. Water residence time [volume/(outflow + evaporation)] averaged 921 ± 381 (SD) days among lakes and years and ranged from 400 to 1661 days at the extremes. This variation was not clearly related to year-to-year variation in precipitation, which was relatively constant [0.26 ± 0.06 (SD) cm day (d)^{− 1}]. The addition of the tracer (along with measurements of lake volume) enabled us to estimate, independent from other hydrologic information, the flow of water leaving the lakes in seepage plus surface outflow. This value, in conjunction with measurement of precipitation and evaporation, enabled us to calculate complete water budgets for these lakes. Among lakes and years, the groundwater input averaged 0.48 ± 0.36 cm d^{− 1} and accounted for 57%± 19% of total water input. This estimate was larger by 150% than that obtained by multiplying precipitation (minus estimated evapotranspiration) times a mapped value of the watershed areas. Our analysis enables us to calculate the relative significance of groundwater and precipitation for solutes such as phosphorus, hydrogen ion, and dissolved organic carbon. Received 17 February 1998; accepted 19 February 1998.