Rainfall Effects Acknowledgements Dr. Radek Fucik - Czech Technical University Dr. Kathleen Smits - Colorado School of Mines Dr. John Christ - US Air Force Academy Alex Maul - Colorado School of Mines Paul Schulte - Colorado School of Mines Michael Glatthar - US Air Force Academy Effects of Water Table Fluctuation and Rain Infiltration on Soil Vapor Concentrations for the Vapor Intrusion Exposure Pathway Benjamin Petri 1 , Tissa Illangasekare 1 , Carolyn Sauck 1 , Toshihiro Sakaki 2 1 Center for Experimental Study of Subsurface Environmental Processes (CESEP), Colorado School of Mines, Golden, CO , 2 National Cooperative for Disposal of Radioactive Waste, Wettingen, Switzerland Introduction Objectives Water Table Fluctuation Effects The intrusion of contaminated vapor originating from groundwater plumes and source zones into homes and businesses is a major driver of remedial action at today's contaminated sites. However, decision-making regarding the vapor intrusion (VI) pathway is frequently complicated by wide variability observed in both space and time in indoor air sampling data. Much of this variability is well documented to originate from background contamination and occupant activities, but some of this variability may also originate from vapor transport processes in the subsurface. This study is focused on evaluating two specific scenarios that may impart transient variability, which include water table fluctuation and rain events. Rainfall and water table fluctuations have been acknowledged as possible processes that may affect vapor intrusion observations in the field (e.g. Fitzpatrick and Fitzgerald 2002, EPA 2012) but at present a full understanding of how these processes affect the vapor intrusion pathway is not available. To help address this knowledge gap, an experimental and modeling study was initiated at the Colorado School of Mines to determine the full transient effect of rainfall events and water table fluctuations on the VI pathway and build an improved conceptual model. Experimental studies offer the advantage of known and tightly controlled conditions that are often absent from field sites; this enables the exploration of fundamental physical processes and the validation of models to ensure that all relevant physics are considered. The validated model can then be used to make predictions for VI based on physically realistic scenarios. Schematic of vapor intrusion processes GC Vacuum supply Left Boundary: Air: No flow Water: Rising and falling water table Constant hydraulic gradient Constant TCE conc. Water table Groundwater direction Top Boundary: Air: Atmospheric pressure Water: No flux Right Boundary: Air: Constant vacuum Advective TCE flux Water: Rising and falling water table Constant hydraulic gradient Air Flowlines Displaced Vapor plume Dissolved plume Rain Reduced Airflow Infiltration front Displaced vapor plume Dissolved plume as source Stable, open atmosphere Water table fluctuation Building vacuum Experimental Apparatus GC Vacuum supply Left Boundary: Air: No flow Water: Steady water table Constant TCE conc. Groundwater direction Top Boundary: Air: Atmospheric pressure Zero TCE conc. Water: Uniform rain flux for 5 min. Zero aqueous TCE conc. Right Boundary: Air: Constant vacuum Advective TCE flux Water: Steady water table Advective TCE flux Air Flowlines Wetting front 0 5000 10000 15000 20000 25000 30000 72 73 74 75 76 77 78 79 Concentration (μg/m 3 ) Time (days) 5 Minute rain application 0 10000 20000 30000 0 2 4 6 8 Concentration (μg/m 3 ) Time (hours) 0 50000 100000 150000 200000 250000 49 51 53 55 57 59 61 63 65 Concentration (μg/m 3 ) Time (days) Rising water table Stable high water table Falling water table Objective: Evaluate the transient effects of rainfall and water table fluctuations on the vapor intrusion pathway using an experimentally validated numerical model Approach: Use a combination of physical experiments at multiple scales and mathematical models to determine the behavior of the vadose zone in response to rain and water table fluctuation events ● Evaluate transport fundamentals in small-scale experimental systems ● Develop a physically-based numerical model that captures complex processes ● Perform upscaled experiments combining complex transport in an large system more representative of field conditions to validate model ● Simulate scenarios representing revised conceptual model and develop guidance and recommendations based on results A large scale laboratory experiment was conducted to generate data to validate the conceptual and numerical model developed by the project. Vapor plume dynamics were simulated in a large sand tank apparatus simulating volatilization from a groundwater plume. The apparatus is a large 4.8m x 1.2 m x 0.05m (16ft x 4ft x 2.2in) laboratory sand tank. ● Both homogeneous and heterogeneous (layered) sand packs are tested ● Top boundary represents an atmospheric soil boundary with free airflow ● Rainmakers in 4 segments for rainfall experiments ● Aqueous TCE is introduced to the tank as a groundwater plume via constant head devices ● A vacuum is applied at the right boundary simulating airflow to a foundation crack and this airflow (which simulates “subslab” air) is sampled continuously. ● An array of air pressure, flow and moisture sensors monitor flow physics. Tank Schematic Procedure: ● Groundwater heads and building vacuum rates are set to constant values and the tank system is allowed to come to steady-state (~1 week). ● Upon steady state conditions, the tank is perturbed with either a water table fluctuation or rain event and the dynamic response observed ● Experiments ran a total of 106 days for both homogeneous and heterogeneous tank experiments ● A total of 3 different water table fluctuations and 3-4 different rain events were conducted for both homogeneous and heterogeneous tanks Tank rear panel showing sensors, data acquisition system Atmospheric chambers Example moisture data from a rainfall experiment (Strategic Environmental Research and Development Program - Sponsor) Model Development Results from the water table fluctuation experiments showed complex effects. In each water table fluctuation, the water table is first raised at a steady rate, then held constant at a new higher level, and the reduced back to the initial level at a steady rate. ● Increases in concentrations were observed for both the rising water table and the falling water table in all fluctuation experiments ● Spike from falling table is likely due to exposure of new residual contaminant mass in capillary fringe as table drops ● Cause of spike under rising water table is under investigation, but may be a result of contaminant pressed from capillary fringe as the fringe is compressed ● The model does NOT capture the behavior of the water table fluctuation (either rising or falling) ● Suspicion is that flow hysteresis may explain why the model and experiment do not match well ● Compared to rain effects, the effect of water table fluctuation is slower and longer lasting ● Majority of mass transfer resistance is within the capillary fringe. Based on steady-state modeling, higher water tables are expected to produce more VI and lower water tables less VI but the transient stages between water table positions can produce unexpected results 0.00 0.04 0.08 0.12 0.16 0 10 20 30 40 50 60 Water Content Time (min.) Sensor 1 (top) Sensor 2 (middle) Sensor 3 (bottom) Water Table Fluctuation A B Conceptual model of volatilization from capillary fringe Dissolved contaminants Air flow Air flow Retention curve NAPL Source Spill source Water table fluctuation Rain / Snow Vapor plume Wind Vapor intrusion Basement Infiltration front Drained fine media Irrigation Groundwater plume Volatilization Volatilization Evapo- Transpiration Undrained fine media Future work GC Vacuum supply Left Boundary: Incoming groundwater flow with dissolved TCE Water table Groundwater direction Top Boundary: 4 flux chambers with rainmakers, flowmeters Right Boundary: Constant vacuum for air, groundwater flow out Air Flowlines Tank: 16 ft x 4 ft x 2.5 inches Groundwater Zone Vadose Zone Atmospheric Zone Aqueous Advection Aqueous Diffusion Gaseous Advection Gaseous Diffusion Mass transfer (volatilization) Aqueous Advection Building Zone Aqueous Diffusion Aqueous Phase Aqueous Phase Gas Phase Gas Phase Gaseous Advection Gaseous Diffusion Gas Phase Mass transfer (Henry’s Law) Air exchange Aqueous Boundary Fluxes - Infiltration / evaporation A model was developed to explore the response of the vapor intrusion pathway to the dynamic effects of infiltration and water table fluctuation. This model is validated against the experimental data and then used to simulate field scale scenarios of VI and explore sensitivity to determine the practical effect for VI. The model is uses the commercially available finite element simulator COMSOL Multiphysics (v3.5a) to solve multiphase flow and solute transport. g d S w d p c p c t + S g M a RT p g t + S g 1 M a M n C g t + • g k rg μ g K i p g + g 2 k rg μ g K i g = 0 w d S w d p c p c t + • w k rw μ w K i p g + w k rw μ w K i p c + w 2 k rw μ w K i g = 0 S g + S w H T C g t + C g 1 H T 1 S w p c p c t + D g C g + u g C g D w C g H T + u w C g H T = 0 S w, e = S w S w, r 1 S w, r S g, r = 1 + p c ( ) n m Schematic of processes included within the model Governing Equations Gas phase flow equation Water phase flow equation Gas + Water phase VOC transport (Local equilibrium assumption) Constitutive Relationships Van Genuchten - Mualem capillary pressure / saturation / relative permeability relations k rg = 1 S w, e ( ) 1 3 1 S w, e ( ) 1 m ( ) m 2 k rw = S w, e 1 1 S w, e ( ) 1 m ( ) m 2 H T = C g C w Henry’s Law (Local equib) D w = ( L T ) u w u w T u w + T u w + D m, w w 10/3 2 ij D g = ( L T ) u g u g T u g + T u g + D m, g g 10/3 2 ij Millington Quirk, Bear diffusion-dispersion models 0 5000 10000 15000 20000 25000 30000 72 73 74 75 76 77 78 79 Concentration (μg/m 3 ) Time (days) Air pressure, soil moisture sensor arrays Tank front panel showing sand pack, water table Results from the rain experiments in the homogeneous tank showed three phases of behavior from rainfall ● An short term rapid spike in concentrations of intruding vapor due primarily to capping and downward advection ● A rapid drop in concentrations as wetting front passes the point of intrusion due to washout ● Slow rebound as infiltration front dissipates and original concentration gradient reestablishes itself. ● Results agree well with Shen et al. (2012) - Science of the Total Environment Results from the heterogeneous (single fine media layer) tank suggest that underneath the fine layer the effect of rainfall on the vapor plume is damped and less rapid, unless the rainfall is significant enough to propagate a wetting front beneath the fine layer. Modeling Result Experimental Result Initial Condition t = 5 min t = 15 min t = 60 min t = 1 day t = 3 day Initial Condition t = 5 min t = 15 min t = 60 min t = 1 day t = 3 day Vapor plume concentrations (expressed as C gas x gas saturation, red = high concentrations Water table source is blue due to low gas saturation) Water content distribution (red = dry soil, blue = wet soil or water table Streamlines represent gas flow to the “building”) ● Complete model validation (no empirical curve fitting) ● Test hysteresis to determine if this may explain the mismatch with the water table fluctuation ● Simulated field scale scenarios (examples at left) to determine practical implications for VI ● Sensitivity analysis to determine important parameters ● Write papers 100000 150000 200000 250000 300000 49 51 53 55 57 59 61 Concentration (μg/m 3 ) Time (days) Rising water table Stable high water table Falling water table Experimental Result Modeling Result Model does not represent data well. Are additional physical processes going on? 0% 20% 40% 60% 80% 100% 50 51 52 53 54 55 56 57 58 59 60 61 62 Water Saturation Time (days) Mean of experimental data Upper and lower 95% confidence interval of data Model prediction Testing of volatilization from a dropping water table in a small tank Plot of water saturation at bottom array of sensor showing possible evidence of hysteresis Gap between model and data Conceptual model Conceptual model Vacuum due to ventilation Constant Uniform TCE plume Land surface Static water table Basement Homogeneous soil horizons Vapor plume Flowlines Infiltration front propagation Rain Static water table Low permeability layer Vapor plume Narrowed infiltration front Irrigation / rain