Hydrobiogeophysics: Linking geo-electrical properties and biogeochemical processes in shallow subsurface environments

spectral induced polarization (SIP);hydrogeophysics;biogeophysics;reactive transport modelling;biogeochemistry;geomicrobiology

Mellage, Adrianadvisor:Rezanezhad, Fereidoun ; affiliation1:Faculty of Science ; advisor:Van Cappellen, Philippe ; Van Cappellen, Philippe ; Rezanezhad, Fereidoun ;

University of Waterloo

关键词: biogeophysics;    geomicrobiology;    reactive transport modelling;    spectral induced polarization (SIP);    hydrogeophysics;    biogeochemistry;    Doctoral Thesis;   

Others  :  https://uwspace.uwaterloo.ca/bitstream/10012/14235/1/Adrian_Mellage.pdf

瑞士|英语

来源: UWSPACE Waterloo Institutional Repository

【 摘 要 】

Microbially mediated reactions drive (bio)geochemical cycling of nutrients and contaminants in shallow subsurface environments. Environmental forcings exert a strong control on the timing of reactions and the spatial distribution of processes. Spatial and temporal variations in electron acceptor and donor availability may modulate nutrient/contaminant turnover. Characterizing the preferential spatial-zonation of biologically driven reactions, and quantifying turnover rates is hindered by our inability to access the subsurface at the spatial and temporal resolution required to capture reaction kinetics. Typical subsurface sampling methods generate discrete spatial datasets as a function of the prohibitive cost and operational challenges of borehole installation and core sampling, coupled with sparse temporal datasets due to intermittent sampling campaigns. In order to improve our ability to access biogeochemical information within the subsurface, without the need for destructive and intrusive sampling, non-invasive geophysical techniques (a comparatively inexpensive alternative) have been proposed as a means to characterize subsurface reactive compartments and locate zones of enhanced microbial activity and the timing of their development. The challenge lies in linking electrical responses to specific changes in biogeochemical processes. In this thesis, I assess the potential and suitability of spectral induced polarization (SIP) and self-potential (SP) / electrodic potential (EP) derived geo-electrical signals to detect, map, monitor and quantify microbially mediated reactions in partially- and fully-saturated heterogeneous porous media (i.e., soil). I build on existing literature delineating the sensitivity of SIP, SP and EP to biogeochemical processes and both qualitatively and quantitatively link geo-electrical signal dynamics to specific microbial processes at the experimental scale. I address the monitoring of complex, coupled processes in a well-characterized near-natural system, and combine reactive transport models (RTMs) with single-process reactive experiments (reduced complexity), to isolate diagnostic signatures of specific reactions and processes of interest. In Chapter 2, I begin by monitoring biogeochemically modulated geo-electrical signals (SIP and EP), in the variably (and dynamically) saturated reactive zone within the capillary fringe of an artificial soil system. SIP and EP responses show a clear dependence on the depth-distribution of subsurface microbes. Dynamic SIP imaginary conductivity (σ'') responses are only detected in the water table fluctuation zone and, in contrast to real conductivity (σ') data, do not exhibit a direct soil moisture driven dependence. Using multiple lines of evidence, I attribute the observed σ'' dynamics to microbially driven reactions. Chapter 2 highlights that continuous SIP and EP signals, in conjunction with periodic measurements of geochemical indicators, can help determine the location and temporal variability of biogeochemical activity and be used to monitor targeted reaction zones and pathways in complex soil environments.Building on the findings from Chapter 2, that biomass distribution and activation strongly modulate SIP responses, in Chapter 3 and 4, I focus on isolating the geo-electrical contribution of microbes themselves. In Chapter 3, I couple geochemical data, a biomass-explicit diffusion reaction model and SIP spectra from a saturated sand-packed (with alternating layers of ferrihydrite-coated and pure quartz sand) column experiment, inoculated with Shewanella oneidensis, and supplemented with lactate and nitrate. The coupled RTM and geo-electrical data analysis show that imaginary conductivity peaks parallel simulated microbial growth and decay dynamics. I compute effective polarization diameters, from Cole-Cole modeling derived relaxation times, in the range 1 – 3 µm; two orders of magnitude smaller than the smallest quartz grains in the columns, suggesting that polarization of the bacterial cells directly controls the observed chargeability and relaxation dynamics.In Chapter 4, I address the lack of experimental validation of biomass concentrations in Chapter 3. I present a measurement-derived relationship between S. oneidensis abundance and SIP imaginary conductivity, from a microbial growth experiment in fully saturated sand-filled column reactors. Cole-Cole derived relaxation times highlight the changing surface charging properties of cells in response to stress. The addition of concurrent estimates of cell size allow for the first measurement-derived estimation of an apparent Stern layer diffusion coefficient for cells, which validates existing modelled values and helps quantify electrochemical polarization during SIP-based monitoring of microbial dynamics.The relaxation time results from Chapter 4 suggest that bacterial cell surface charge is modified in response to nitrite toxicity-induced stress. In Chapter 5, I present a biomass-explicit reactive transport model, which integrates nitrite-toxicity, as a key modulator of the energy metabolism of S. oneidensis, to predict the rates of nitrate and nitrite reduction. I validate the model with results from two separate experiments (at different experimental scales): (1) a well-mixed batch suspension and (2) the flow-through reactor experiment from Chapter 4. The incorporation of toxicity-induced uncoupling of catabolism and anabolism in the reactive term predicts the observed delay in biomass growth, facilitated by endogenous energy storage when nitrite is present, and consumption of these reserves after its depletion. The model is further validated by the close agreement between the trends in imaginary conductivity and simulated biomass growth and decay dynamics.Finally, in Chapter 6, I apply the RTM-SIP integrative framework from Chapters 3 and 5 to develop quantitative relationships between SIP signals and engineered nanoparticle concentrations. Therein, SIP responses measured during injection of a polymer-coated iron-oxide nanoparticle suspension in columns packed with natural aquifer sand are coupled to output from an advective-dispersive transport model. The results highlight the excellent agreement between simulated nanoparticle concentrations within the columns and SIP signals, suggesting that polarization increases proportional to increasing nanoparticle concentration. The results from Chapter 6, introduce the possibility of quantitative SIP monitoring of coated metal-oxide nanoparticle spatial and temporal distributions.Overall, my results show the applicability of SIP and EP to map and monitor the spatial zonation of biogeochemical hotspots and to detect their temporal activation. By coupling RTMs with geo-electrical datasets, I highlight the direct control that polarization of microbial cells exerts on SIP signals in biotic systems. Furthermore the measurement-derived SIP-biomass quantitative relationship provides a first attempt to directly measure in situ biomass density, using geo-electrical signals as a proxy. I show that geo-electrical signal dynamics (Cole-Cole relaxation time) can be used to inform processes within RTMs. Finally, the implementation of the combined modeling and electrical monitoring approach, to the case of engineered nanoparticles, confirms SIP’s suitability to monitor colloid transport in the environment and highlights considerations for method optimization.