Berkeley Lab

Sustainable Remediation of Complex Environmental Systems: Key Recent Technical Advances

Simulated uranium plume (concentration>1×10-6mol/L) in 3D at the Savannah River Site F-Area in 2050. The sky-blue region is the low permeable Tan Clay Zone, which separates the upper and lower aquifers.

This book chapter provides an overview of the key recent scientific advances to support sustainable remediation in complex geological systems demonstrated at the Savannah River Site F-Area, including site characterization techniques, hydrological and geochemical model developments and numerical simulations.

We have developed various subsurface characterization and modeling technologies to improve the predictive understanding of the groundwater contaminant plumes in complex geological systems. The technologies have been demonstrated at the Department of Energy’s Savannah River Site.


Groundwater remediation has been evolved recently with increased focus on sustainable approaches such as in situ treatments and natural attenuation. However, leaving contaminants in subsurface requires the increased burden of proof to show that plumes are stable and residual contaminants do not pose a significant health risk. At the Department of Energy’s Savannah River Site, we have developed and demonstrated various characterization and modeling technologies to provide the predictive understanding of the contaminant plume migration, including (1) a multiscale data integration method to integrate surface seismic and borehole datasets that have different resolution and spatial coverage, (2) a novel surface complexation model to describe pH-dependent uranium sorption behavior based on readily available datasets, and (3) a reactive transport modeling and uncertainty quantification framework to predict the future uranium plume behavior and to identify key parameters on the future uranium concentration. These technologies are expected to transform groundwater remediation at many other sites.


Wainwright, H. M.; Arora, B.; Faybishenko, B.; Molins, S.; Hubbard, S. S.; Lipnikov, K.; Moulton, D.; Flach, G.; Eddy-Dilek, C.; Denham, M. (2018), Sustainable Remediation in Complex Geologic Systems, The heaviest metals: Science and technology in Actinides and beyond.

Geochemical Exports to River from the Intra-Meander Hyporheic Zone under Transient Hydrologic Conditions at East River

Figure 1: Spatial distributions of dissolved geochemical species in groundwater in the meander C.

Hyporheic exchange within the intra-meander region results in the interaction of nutrient-rich groundwater and oxygen-rich river water, which leads to the formation of distinct redox gradients. These redox gradients can significantly impact the export of metals and nutrients at the local, reach, and watershed scales. Further, transient hydrologic conditions, such as groundwater flow dynamics, river-stage fluctuations, and rainfall/snowmelt events, can impact redox processes in the hyporheic zone and ultimately the geochemical exports to the river thereby affecting river water quality. Here we have used high-resolution hydrodynamic assessments of the hyporheic zone combined with detailed pore-water sampling to focus on the hyporheic exchange at the meander scale for the purpose of quantifying the subsurface exports from a single meander to the river under transient hydrological conditions.

This study is a first of its kind that examines the influence of transient hydrological conditions on the hyporheic biogeochemistry using field observations. Simulation results demonstrated that intra-meander hyporheic zones display distinct anoxic and suboxic regions, suboxic regions being localized along sides of the meander bend. Permeability within the meander has a more significant impact on biogeochemical zonation compared to the reaction pathways for transient hydrologic conditions. Here we have also demonstrated the outsized implications of micro-topographic features such as gullies on redox processes using the high-resolution LiDar data.

Figure 2. Net cumulative geochemical export of TIC, DOC, and iron(II) from a single meander. River-stage is shown in green on the right y-axis, whereas the net export is shown in golden color on the left y-axis.


Hyporheic zones perform important ecological functions by linking terrestrial and aquatic systems within watersheds. Hyporheic zones can act as a source or sink for various metals and nutrients. Transient hydrologic conditions alter redox conditions within an intra-meander hyporheic zone thus affecting the behavior of redox-sensitive species. Here we investigated how transient hydrological conditions control the lateral redox zonation within an intra-meander region of the East River and examined the contribution of a single meander on subsurface exports of carbon, iron, and other geochemical species to the river. The simulation results demonstrated that the reductive potential of the lateral redox zonation was controlled by groundwater velocities resulting from river-stage fluctuations, with low-water conditions promoting reducing conditions. The sensitivity analysis results showed that permeability had a more significant impact on biogeochemical zonation compared to the reaction pathways under transient hydrologic conditions. The simulation results further indicated that the meander acted as a sink for organic and inorganic carbon as well as iron during the extended baseflow and high-water conditions; however, these geochemical species were released into the river during the falling limb of the hydrograph. This study demonstrates the importance of including hydrologic transients, using a modern reactive transport approach, to quantify exports within the intra-meander hyporheic zone at the riverine scale.


Dwivedi, D., C.I. Steefel, B. Arora, M. Newcomer, J.D. Moulton, B. Dafflon, B. Faybishenko, P. Fox, P. Nico, N. Spycher, R. Carroll, and K.H. Williams (2018), Geochemical Exports to River from the Intra-Meander Hyporheic Zone under Transient Hydrologic Conditions: East River Mountainous Watershed, Colorado, Water Resources Research, 10.1029/2018WR023377.

Using strontium isotopes to evaluate how local topography affects groundwater recharge

Figure 1. 87Sr/86Sr vs 1/[Sr] showing the mixing relationships between vadose zone porewater and groundwater. Blue line shows the mixing model of vadose zone water with upgradient groundwater with bold numbers representing the percentage of vadose zone water in the local aquifer.

A key component of understanding the connection between groundwater quality and the vadose zone (the water unsaturated region above the water table) is the movement of water from the surface to the aquifer (recharge). Measurements of the natural isotopic composition of Sr were used to assess the effect of local topography on groundwater recharge across a semi-dry riparian floodplain in the Upper Colorado River Basin.

This work demonstrates the use of 87Sr/86Sr (Sr isotopes) to measure groundwater recharge through analysis of porewater and groundwater samples from the vadose zone. The study resulted in an understanding how the microtopography of the Rifle Site affects the hyper-local variation in the downward movement of vadose-zone porewater that may carry nutrients and contaminants to groundwater.

Figure 2. “Heat” map of the percentage contribution of vadose water to the Rifle floodplain aquifer based on the Sr isotopic mixing model.


Over time, loose sand, clay, silt, gravel or similar unconsolidated, or “alluvial” material is deposited by water into alluvial aquifers. Recharge of alluvial aquifers is a key component in understanding the interaction between floodplain vadose zone biogeochemistry and groundwater quality. The Rifle Site (a former U-mill tailings site) adjacent to the Colorado River is a well-established field laboratory that has been used for over a decade for the study of biogeochemical processes in the vadose zone and aquifer. This site is exemplary of both a riparian floodplain in a semiarid region and a post-remediation U-tailings site. The authors use Sr isotopic data for groundwater and vadose zone porewater samples to build a mixing model for the fractional contribution of vadose zone porewater (i.e. recharge) to the aquifer and to assess its distribution across the site. The vadose zone porewater contribution to the aquifer ranged systematically from 0% to 38% and appears to be controlled largely by the microtopography of the site. The area-weighted average contribution across the site was 8%, corresponding to a net recharge of 7.5 cm. Given a groundwater transport time across the site of ~1.5 to 3 years, this translates to a recharge rate between 5 and 2.5 cm/yr, and with the average precipitation to the site, implies a loss from the vadose zone due to evapotranspiration of 83% to 92%.


Christensen, J. N., et al. (2018), Using strontium isotopes to evaluate the spatial variation of groundwater recharge, Sci Total Environ, 637-638, 672-685, DOI: 10.1016/j.scitotenv.2018.05.019.

Unexpected High Carbon Fluxes from the Deep Unsaturated Zone in a Semi-Arid Region

Figure: Estimated carbon fluxes and balance along the measurement transect, showing 30% of annual CO2 emissions from deeper than 1.0 m, contrary to predictions of < 1% by the Earth System Models.

Understanding of terrestrial carbon cycling has relied primarily on studies of top soils that are typically characterized to depths shallower than 0.5 m. We found and quantified 30% of CO2 annual efflux to atmosphere (60% in winter) originates from below 1 m, contrary to prediction of <1% by the ESM land models.

We contend that ESM land models need to incorporate these deeper soil processes to improve CO2 flux predictions in semi-arid climate regions.


Understanding of terrestrial carbon cycling has relied primarily on studies of topsoils that are typically characterized to depths shallower than 0.5 m. At a semi-arid site, instrumented down to 7 meters, we measured seasonal- and depth-resolved carbon inventories and fluxes, and groundwater and unsaturated zone flow rates. We identified an unexpected high dissolved organic carbon (DOC) flux from the rhizosphere into the underlying unsaturated zone. We measured that ~30% of the CO2 efflux to atmosphere (60% in winter) originates from below 1 m, contrary to prediction of <1% by Earth System Model land models. The seasonal DOC influx and favorable temperatures, moisture and oxygen availability in deeper unsaturated zone sustained the respirations of deeper microbial communities and roots. These conditions are common characteristics of many subsurface environments; thus we contend that ESM land models need to incorporate these deeper soil processes to improve CO2 flux predictions in semi-arid climate regions.


Wan, J., Tokunaga, T.K., Dong, W., Williams, K.H., Kim, Y., Conrad, M.E., Bill, M., Riley, W.J., Susan S.H., Deep unsaturated zone contributions to carbon cycling in semiarid environments. Journal of Geophysical Research: Biogeosciences, 123.

Factors Controlling Seasonal Groundwater and Solute Flux from Snow-Dominated Basins

Fig. 1: Mixing diagram in U-space with Pumphouse stream discharge plotted in rainbow. Seasonal end-member means are connected with grey connecting lines while end-member variability (black error bars) determined using ±1sd of observed tracer concentrations.

Research identifies critical zone influences on hydrologic partitioning and tracer variability as reflected in seasonal stream concentration-discharge (C-Q) relationships. The method is applied across scale and within topographically complex, snow-dominated basins. First-order controls on seasonal streamflow generation are isolated and hydro-chemical conceptual model development is initiated.

The role of groundwater contribution to snow-dominated, low-order streams residing in basins of large relief is found significant; with recharge increasing in the upper sub-alpine where maximum snow accumulation is coincident with reduced conifer cover and lower canopy densities. Error in estimated stream concentrations is attributed to differences in water partitioning, source rock, seasonal shifts in flow path and sulfate reduction in floodplain sediments.

Fig. 2: Topographic and vegetation controls on snow water equivalent (SWE) and fraction of streamflow that is groundwater (fGW).


To isolate first-order controls on seasonal streamflow generation within highly heterogeneous, snow-dominated basins of the Colorado River, authors developed a multivariate statistical approach of end-member mixing analysis (EMMA) using a suite of daily chemical and isotopic observations. Mixing models are developed across 11 nested basins (0.4 km2 to 85 km2) spanning a gradient of climatological, physical and geological characteristics. Hydrograph separation using rain, snow and groundwater as end-members indicates that seasonal contributions of groundwater to streams is significant. Mean annual groundwater flux ranges from 12% to 33% while maximum groundwater contributions of 17% to 50% occur during baseflow. Groundwater recharge is found to increase in basins of high relief and within the upper sub-alpine where maximum snow accumulation is coincident with reduced conifer cover and lower canopy densities (Fig. 1). The mixing model developed for the furthest downstream site did not transfer to upstream basins. The resulting error in predicted stream concentrations points toward weathering reactions as a function of source rock and seasonal shifts in flow path. Additionally, the potential for microbial sulfate reduction in floodplain sediments along a low gradient, meandering portion of the river is sufficient to modify hillslope contributions and alter mixing ratios in the analysis. Soil flushing in response to snowmelt is not included as an end-member but is identified as an important mechanism for release of solutes from these mountainous watersheds.


R. W. H. Carroll, L. A. Bearup, W. Brown, W. Dong, M. Bill, and K. H. Williams, “Factors Controlling Seasonal Groundwater and Solute Flux from Snow-Dominated Basins.” Hydrological Processes, Special Issue: Water in the Critical Zone 32 (14), 2187-2202 (2018). [].

An early warning system for tracking groundwater contamination

Figure 1. Kalman-filter based in situ real-time monitoring system

A real-time, in-situ, ‘smart’ monitoring and early warning system for migrating and reacting contaminant plumes was developed using a Kalman filter-based framework and successfully tested at the contaminated Savannah River Site.

Figure 2. System validation using historical, cheap proxy variables (specific conductance and pH) to estimate U-238

As this framework enables easy integration of networked, autonomous and inexpensive wellbore measurements with cloud computing, the approach is expected to reduce groundwater monitoring cost, increase confidence in the efficacy of monitored natural attenuation, and ensure early response if needed.

  • A Kalman filter method was used to estimate contaminant concentrations continuously and in real-time by coupling data-driven concentration decay models with data correlations.
  • The approach was successfully demonstrated using historical groundwater data from the Uranium and Tritium-contaminated Savannah River Site F-Area
  • Specific conductance and pH were used as proxy variables to estimate tritium and uranium concentrations over time.
  • Results show that the developed method can estimate contaminant concentrations based on in-situ, easily measured variables

Schmidt, F., Wainwright, H. M., Faybishenko, B., Denham, M., & Eddy-Dilek, C. (2018). In Situ Monitoring of Groundwater Contamination Using the Kalman Filter. Environmental Science & Technology, DOI: 10.1021/acs.est.8b00017.

Predicting Watershed Function and Response to Disturbance

Figure: The Watershed Function project is developing and testing system-within-system, scale-adaptive approaches at the East River, CO Watershed to quantify how spatially variable hydrological-biogeochemical responses to perturbations propagate through the system and lead to an aggregated downgradient watershed discharge and concentration signature.

While watersheds are recognized as Earth’s key functional unit for managing water resources, their hydrological interactions also mediate biogeochemical processes that support all terrestrial life and can lead to a cascade of downgradient effects. The Watershed Function Scientific Focus Area is developing a scale-adaptive predictive understanding of how mountainous watersheds retain and release water, nutrients, carbon, and metals. This paper describes several recently developed approaches to interrogate, monitor and simulate transient watershed partitioning and biogeochemical responses – from genome to watershed spatial scales and from episodic to decadal timescales.

A growing demand for clean water, food, and energy — in parallel with droughts, floods, early snowmelt and other disturbances — are significantly reshaping interactions within watersheds throughout the world. This is particularly true for mountainous systems, such as the East River Watershed in CO, which is located in the Upper Colorado River Basin. This Basin supplies water to 1 in every 10 Americans and supports vast agriculture and hydropower operations along its reach. Because society is dependent on watersheds, new approaches that can accurately yet tractably predict watershed responses to disturbances are critical for resource management.


The Watershed Function project is developing new approaches to quantify and predict how disturbances impact downstream water availability and biogeochemical cycling, with a current focus on early snowmelt. The research is guided by a system-of-systems perspective and a scale-adaptive approach, where a predictive understanding of the response of archetypal watershed subsystems to disturbances is being developed as well as methods to aggregate such responses into predictions of cumulative watershed exports. The paper describes several recent advances, including above-and-below ground characterization and monitoring approaches for understanding vegetation distribution; new modeling approaches for predicting bedrock-through-canopy hillslope interactions; and coupled modeling approaches that can assimilate increasingly available streaming data into models to estimate hillslope water partitioning over time. The paper also describes new watershed function insights gained through the use of such tools, including: how historical snowmelt and monsoon characteristics influence annual discharge across the entire watershed; controls on streamflow generation; and how future changes in vegetation and temperature may influence water partitioning at different positions in the watershed. Over 30 institutions are involved in advancing watershed hydrological-biogeochemical science at the East River CO DOE-BER ‘community testbed’.


Hubbard, S.S. et al. (2018) The East River, CO Watershed: A Mountainous Community Testbed for Improving Predictive Understanding of Multi-Scale Hydrological-Biogeochemical Dynamics. Vadose Zone Journal, Special Issue on ‘Hydrological Observatories’, doi: 10.2136/vzj2018.03.0061 in press.

Spring Snowmelt Drives Transport and Degradation of Dissolved Organic Matter in a Semi-Arid Floodplain

Berkeley Lab geochemists and hydrologists who study a mountainous watershed near Rifle, CO, discovered that spring snowmelt is essential to the transport of freshly dissolved organic matter (DOM) from the top soil to the part of the Earth’s subsurface that lies above the groundwater table. Because dissolved organic matter undergoes biological humification over the year, these processes involving this deep vadose zone suggest an annual cycle of DOM degradation and transport at this semi-arid floodplain site.

Spring snowmelt transports fresh DOM from top soil into the deeper vadose zone, and then undergoes microbial humification process over the year (see depth and seasonal EEM spectra)

Characterizing the dynamics of dissolved organic matter in semi-arid regions of Earth’s subsurface is challenging. The authors obtained insights into transport and humification processes of DOM using several spectroscopic techniques on depth- and temporally-distributed pore-waters. This methodology can be applied to other subsurface environments for understanding DOM responses and feedbacks to earth system processes.


Scientists studying DOM in surface waters considered it to be the mobile fraction of natural organic matter that falls into or is washed into water bodies. Although it has been extensively studied over many decades, relatively little is known about the dynamics of DOM in the subsurface of semi-arid environments. In order to understand transport and humification processes of DOM within a semi-arid floodplain at Rifle, Colorado, the authors applied fluorescence excitation-emission matrix (EEM) spectroscopy, humification index (HIX) and specific UV absorbance (SUVA) for characterizing depth and seasonal variations of DOM composition. They found that late spring snowmelt leached relatively fresh DOM from plant residue and soil organic matter down into the deeper vadose zone (VZ). More humified DOM is preferentially adsorbed by upper VZ sediments, while non- or less-humified DOM was transported into the deeper VZ. Interestingly, DOM at all depths undergoes rapid biological humification processes as evidenced by the products of microbial by-product-like matter in late spring and early summer, particularly in the deeper VZ, resulting in more humified DOM at the end of year. The finding indicates that DOM transport is dominated by spring snowmelt, and DOM humification is controlled by microbial degradation. It is expected that these relatively simple spectroscopic measurements (e.g., EEM spectroscopy, HIX and SUVA) applied to depth- and temporally-distributed pore-water samples can provide useful insights into transport and humification of DOM in other subsurface environments as well.


Dong, W., J. Wan, T. K. Tokunaga, B. Gilbert, and K. H. Williams (2017). Transport and Humification of Dissolved Organic Matter within a Semi-Arid Floodplain. Journal of Environmental Sciences 57, 24-32, DOI: 10.1016/j.jes.2016.12.011

SFA Research Identifies New Microbial Players in the Global Sulfur Cycle

Sulfate is ubiquitous in the environment, and sulfate reduction – a key control on anaerobic carbon turnover – impacts a number of other processes such as carbon oxidation and sulfide production. Until now, sulfate reduction was believed to be restricted to organisms from select bacterial and archael phyla. But scientists at UC Berkeley have now found this ability to be more widespread. They used genome-resolved metagenomics to discover roles in sulfur cycling for organisms from 16 microbial phyla not previously associated with this process.

DsrAB protein tree showing the diversity of organisms involved in dissimilatory sulfur cycling using the dsr system.
Lineages in blue contain genomes reported in this study. Phylum-level lineages with first report of evidence for sulfur cycling are indicated by blue letters.

Sulfate-reducing bacteria are anaerobic microorganisms essential to sulfur and carbon cycling. Sulfate reduction drives other key processes and produces hydrogen sulfide, an important but potentially toxic gas present in sediments, wetlands, aquifers, the human gut, and the deep-sea. The discovery of novel microbes connected to sulfur cycling is relevant in biogeochemistry, ecosystem science and engineering, and fundamentally reshape our understanding of microbial function and capabilities associated with phylogenetic information.


Phylogenetic information shapes our expectations regarding microbial capabilities. In fact, this is the basis of currently used methods that link gene surveys to metabolic predictions of community function. Sulfate Reduction, an important anaerobic metabolism, impacts carbon, nitrogen, and hydrogen transformations in numerous environments across our planet and is known to be restricted to organisms from selected bacterial and archaeal phyla. The authors used genome-resolved metagenomic analyses to determine the metabolic potential of microorganisms from six complex marine and terrestrial environments. By analyzing >4000 genomes, they identified 123 near-complete genomes that encode dissimilatory sulfite reductases involved in sulfate reduction. They discovered roles in sulfur cycling for organisms from 16 microbial phyla not previously known to be associated with this process. Additional findings include some of the earliest-evolved sulfite reductases in bacteria, identification of a novel protein unique to sulfate reducing bacteria, and a key sulfite reductase gene in putatively symbiotic Candidate Phyla Radiation (CPR) bacteria. This study fundamentally reshapes expectations regarding the roles of a remarkable diversity of organisms in the biogeochemical cycle of sulfur.


Anantharaman, K., B. Hausmann, S.P. Jungbluth, R.S. Kantor, A. Lavy, L.A. Warren, M.S. Rappé, M. Pester, A. Loy, B.C. Thomas, and J.F. Banfield (2017). Expanded diversity of microbial groups that shape the dissimilatory sulfur cycle. The ISME journal 12, 1715–1728, DOI: 10.1038/s41396-018-0078-0

Fundamental Understanding of Engineered Nanoparticle Stability in Aquatic Environments

Figure. Photographs showing sedimentation of CdSe-MUA NPs as a function of dilution.

It is commonly true that a diluted colloidal suspension is more stable over time than a concentrated one, because dilution reduces collision rates, so delays formation of aggregates. However, we observed the opposite relationship between stability and concentration for some engineered ligand-coated nanoparticles.

Because the stability of NPs determines their physicochemical and kinetic behavior including toxicity, dilution induced instability needs to be understood to realistically predict the behavior of engineered ligand-coated nanoparticles in aqueous systems.


It is commonly true that a diluted colloidal suspension is more stable over time than a concentrated one, because dilution reduces collision rates of the particles, therefore delays formation of aggregates. However, this generalization does not apply for some engineered ligand-coated nanoparticles (NPs). We observed the opposite relationship between stability and concentration of NPs. We tested four different types of NPs; CdSe-11-mercaptoundecanoic acid, CdTe-polyelectrolytes, Ag-citrate, and Ag- polyvinylpirrolidone. The results showed that dilution alone induced aggregation and subsequent sedimentation of the NPs that were originally monodispersed at very high concentrations. Increased dilution caused NPs to progressively become unstable in the suspensions. The extent of the dilution impact on the stability of NPs is different for different types of NPs. We hypothesize that the unavoidable decrease in free ligand concentration in the aqueous phase following dilution causes detachment of ligands from the suspended NP cores. The ligands attached to NP core surfaces must generally approach exchange equilibrium with free ligands in the aqueous phase, therefore ligand detachment and destabilization are expected consequences of dilution. More studies are necessary to test this hypothesis. Because the stability of NPs determines their physicochemical and kinetic behavior including toxicity, dilution induced instability needs to be understood to realistically predict the behavior of engineered ligand-coated nanoparticles in aqueous systems.


Wan, J., Y. Kim, M.J. Mulvihill, and T. K. Tokunaga (2018). Dilution destabilizes engineered ligand-coated nanoparticles in aqueous suspensions. Environmental Toxicology and Chemistry. doi: 10.1002/etc.4103.