Site-scale Model Research Articles

Using airborne lidar and machine learning to predict visibility across diverse vegetation and terrain conditions

Visibility analyses, used in many disciplines, rely on viewshed algorithms that map locations visible to an observer based on a given surface model. Mapping continuous visibility over broad extents is uncommon due to extreme computational expense. This study introduces a novel method for spatially-exhaustive visibility mapping using airborne lidar and random forests that requires only a sparse sample of viewsheds. In 24 topographically and vegetatively diverse landscapes across the contiguous US, 1000 random point viewsheds were generated at four different observation radii (125 m, 250 m, 500 m, 1000 m), using a 1 m resolution lidar-derived digital surface model. Visibility index – the proportion of visible area to total area – was used as the target variable for site-scale and national-scale modeling, which used a diverse set of 146 terrain- and vegetation-based 10 m resolution metrics as predictors. Variables based on vegetation, especially those based on local neighborhoods, were more important than those based on terrain. Visibility at shorter distances was more accurately estimated. National-scale models trained on a wider range of vegetation and terrain conditions resulted in improved R2, although at some sites error increased compared to site-scale models. Results from an independent test site demonstrate potential for application of this methodology to diverse landscapes.

Integrating GIS, Remote Sensing, and Citizen Science to Map Oak Decline Risk across the Daniel Boone National Forest

Oak decline is a general term used for the progressive dieback and eventual mortality of oak trees due to many compounding stressors, typically a combination of predisposing, inciting, and contributing factors. While pinpointing individual causes of decline in oak trees is a challenge, past studies have identified site and stand characteristics associated with oak decline. In this study, we developed a risk map of oak decline for the Daniel Boone National Forest (DBNF), combining GIS, remote sensing (RS), and public reporting (citizen science, CS). Starting with ground reports of decline (CS), we developed a site-scale model (GIS and RS) for oak decline based on four previously identified predisposing factors: elevation, slope, solar radiation, and topographic wetness. We found that areas identified in the model as having a high oak decline risk also reflected areas of observed oak decline (CS). We then optimized and expanded this risk model to the entire range of the DBNF, based on both site characteristics (as piloted for the case study site) and stand inventory data. The stand inventory data (including species composition and age) further improved the model, resulting in a risk map at the landscape level. This case study can serve as a planning tool and highlights the potential usefulness of integrating GIS, remote sensing, and citizen science.

Evaluation of basin-scale hydrogeological changes induced by reservoir operation at the Xiluodu dam site

Impoundment of large-capacity reservoirs has a great impact on groundwater flow in aquifer-aquitard systems. This impact occurs not only at the local scale in the dam area, but also at the regional scale under particular site conditions, causing unexpected geohazards and deformation phenomena. In this study, basin-scale hydrogeological changes induced by reservoir filling and operation at the Xiluodu dam site in southwest China are investigated by hierarchical numerical simulations. Hierarchical groundwater flow models consisting of a basin-scale model (34 km in horizontal dimension) and a site-scale model (4 km) are developed and solved with MODFLOW and a finite element code, respectively, by applying the output of the basin-scale model on the lateral boundaries of the site-scale model. The groundwater level changes in and the groundwater exchange between the confined and unconfined aquifer systems are evaluated. The groundwater level in the confined aquifer has significantly increased by about 96 m beneath the dam site since reservoir filling, and the dimension of the influence zone with an increase over 10 m in groundwater level reaches about 7–9 km in the basin. The rate of vertical leakage from the confined to unconfined aquifer first decreases during the initial filling, and then increases gradually as the pressure propagates into the confined aquifer. It is predicted that the dynamically equilibrated state will be reached in 2027. The local vertical leakage occurring in the plunge pool area along the unplugged boreholes drilled during site investigation is confirmed by a calculation based on the Theis formula. The results are useful for decision-making related to the reservoir operation and safety assessment of the dam.

Investigations of effect of underground excavations using hydrologic and tracer transport modeling

Understanding the effect of underground excavations and barrier capability of the near field are relevant to the geologic disposal of nuclear waste. For a repository in crystalline host rock characterization of the naturally fractured rock and the excavated disturbed zone are of importance to the study of migration of radionuclides. Excavations could weaken the repository components that are part of the engineered barrier system. Permeable excavated disturbed zones could also develop around the repository walls that could create conduits for radionuclides away from the repository. Experimental hydrology and tracer transport data from the Mizunami Underground Research Laboratory in Central Japan have been used to develop a site-scale model for the study area. Upscaled permeability and porosity data of a previously developed discrete fracture network model were used. In this study development and utilization of the model is presented. The model was used to reproduce hydrology and non-reactive transport in the modeling area. The modeling work includes prediction of inflow of water into the research tunnel during excavation of tunnel parts and related draw-down. The modeling work also includes prediction of pressure recovery and changes in tracer distribution as a result of water filling of a tunnel chamber. Modeling results were successfully compared with the project hydrology and tracer concentration histories. Modeling results of inflow and recovery were complementary and together with fracture characterization combined to give a better constraint on the system. The study provides experimental and modeling methods to study the effect of excavations in an underground research laboratory.

Development of Site-Scale Conceptual Model Using Integrated Borehole Methods: Systematic Approach for Hydraulic and Geometric Evaluation

Understanding the physical flow mechanisms in aquifer systems is essential in effectively protecting groundwater resources and preserving subsurface environments from a wide range of contaminants. A conceptual model is a simplified representation of a groundwater system and gaining knowledge about the geological features and parameters controlling the flow and transport processes is a crucial first step towards properly constructing a site-scale conceptual model. In this study, we present a multi-step workflow that involves integrated borehole techniques to gain information concerning groundwater flow. Measurements from core-scale to field-scale enable us to better build a subsurface geological structure divided into the unconsolidated layer and the fractured bedrock. In addition, neutron logging and mercury injection capillary pressure techniques allow for the development of vertical porosity distribution in the alluvial layer. For fracture characterization, the fracture geometry is delineated using a series of borehole imaging techniques and single-hole tests to differentiate the individual permeable fractures from other hydraulically inactive fractures. Combining the hydraulic and geometric evaluations, the presence of large-scale connective fracture networks is identified. Our high-resolution three-dimensional (3D) site-scale conceptual model is expected to contribute to improving the reliability and availability of numerical groundwater models.

1 to 1000 Policy: Controlling Phosphorous Pollution from Tea Farms with Bioretention Cells

When implementing nonpoint source pollution control plans, the size or number of required controlling facilities is a very practical issue. However, quantifying nonpoint source pollution is difficult because it is generated by variable and random rainfall events. This study develops a two-stage optimization process to demonstrate the determination of the optimal bioretention cell size for tea farm pollution control. The optimization process was based on a verified watershed-scale model and a verified site-scale model. The verified watershed model was used to obtain total phosphorous (TP) reduction loads. Once the goal of watershed management was decided, the reduction loads were then allocated and the unit reduction loads were determined. Using the unit reduction loads, the verified tea farm model was used to assess the optimal bioretention cell size for tea farms. A case study using the Jinggualiao stream in the Feitsui Reservoir watershed, Taipei, Taiwan was presented. The results showed that the unit tea farm TP reduction loads were 270 g/ha-year and 326 g/ha-year to reach two water quality goals, and a total of 350 m2 and 600 m2 of bioretention cells were needed, respectively. A 1 to 1000 ratio of the standard bioretention cell area to the tea farm area is recommended as a general control rule.

Competitive Effects of Permeability and Gravity on the Drying-Out Process during CO2 Geological Sequestration in Saline Aquifers

Abstract Salt precipitation from the drying-out process has a profound effect on the well injectivity during the storage of carbon dioxide (CO2) in deep saline aquifers. Both gravity and reservoir heterogeneity have a significant impact on CO2-plume behavior and CO2 storage capacities. The collective effect of gravity and heterogeneity on the drying-out process by site-scale numerical simulation based on the Sleipner project had been investigated. Three site-scale permeability heterogeneous models and a fracture model had been built; simulation results showed that the gravity effect significantly increased the solid saturation at the injection well in the homogeneous model; changing the position of the injection well can change the distance that gravity can act and affect the amount of salt precipitation near the injection well. A novel conclusion is gravity and heterogeneity showed a mutual resistance relationship when considering the collective effect of gravity and heterogeneity on solid saturation. Gravity effects reduced the amount of salt deposited in the fracture model; at low CO2 injection rate, gravity force dominated CO2 flow; increased rock heterogeneity suppressed the production of salt precipitates; at high CO2 injection rate, viscous force dominated flow; and increased heterogeneity increased salt precipitation. This research is of important guiding significance for the design of site screening and injection schemes from the perspective of avoiding a large amount of salt precipitation and pressure build-up.

Site-scale modeling of surface ozone in Northern Bavaria using machine learning algorithms, regional dynamic models, and a hybrid model.

Ozone (O3) is a harmful pollutant when present in the lowermost layer of the atmosphere. Therefore, the European Commission formulated directives to regulate O3 concentrations in near-surface air. However, almost 50% of the 5068 air quality stations in Europe do not monitor O3 concentrations. This study aims to provide a hybrid modeling system that fills these gaps in the hourly surface O3 observations on a site scale with much higher accuracy than existing O3 models. This hybrid model was developed using estimations from multiple linear regression-based eXtreme Gradient Boosting Machines (MLR-XGBM) and O3 reanalysis from European regional air quality models (CAMS-EU). The binary classification of extremely high O3 events and the 1- and 24-h forecasts of hourly O3 were investigated as secondary aims. In this study thirteen stations in Northern Bavaria, out of which six do not monitor O3, were chosen as test sites. Considering the computational complexity of machine learning algorithms (MLAs), we also applied two recent MLA interpretation methods, namely SHapley Additive exPlanations (SHAP) and Local interpretable model-agnostic explanations (LIME). With SHAP, we showed an increasing effect of temperature on O3 concentrations which intensifies for temperatures exceeding 17°C. According to LIME, O3 concentration peaks are mainly governed by meteorological factors under dry and warm conditions on a regional scale, whereas local nitrogen oxide concentrations control base O3 concentrations during cold and wet periods. While recently developed MLAs for the spatial estimation of hourly O3 concentrations had a station-based root-mean-square error (RMSE) above 27μg/m3, our proposed model significantly reduced the estimation errors by about 66% with an RMSE of 9.49μg/m3. We also found that logistic regression (LR) and MLR-XGBM performed best in the site-scale classification and 24-h forecast of O3 concentrations (with a station-averaged accuracy and RMSE of 0.95 and 19.34μg/m3, respectively).

LiDAR-derived topography and forest structure predict fine-scale variation in daily surface temperatures in oak savanna and conifer forest landscapes

In mountain landscapes, surface temperatures vary over short distances due to interacting influences of topography and overstory vegetation on local energy and water balances. At two study landscapes in the Sierra Nevada of California, characterized by foothill oak savanna at 276–481 m elevation and montane conifer forest at 1977–2135 m, we deployed 288 near-surface (5 cm above the surface) temperature sensors to sample site-scale (30 m) temperature variation related to hillslope orientation and vegetation structure and microsite-scale (2–10 m) variation related to microtopography and tree overstory. Daily near-surface maximum and minimum temperatures for the 2013 calendar year were related to topographic factors and vegetation overstory characterized using small footprint LiDAR imagery acquired by the National Ecological Observatory Network (NEON) Airborne Observation Platform (AOP). At both landscapes we recorded large site and microsite spatial variation in daily maximum temperatures, and less absolute variation in daily minimum temperatures. Generalized boosted regression trees were estimated to measure the influence of tree canopy density, understory solar radiation, cold-air drainage and pooling, ground cover and microtopography on daily maximum and minimum temperatures at site and microsite scales. Site-scale models based on indices of understory solar radiation and landscape position explained an average of 61–65% of daily variation in maximum temperature; site-scale models based on tree canopy density and landscape position explained 65–83% of variation in minimum temperatures. Models explained <15% of variation in microsite-scale maximum temperatures but within-site heterogeneity was significantly correlated with within-site heterogeneity in modeled understory radiation at both landscapes. Tree canopy density and slope explained 33% of microsite-scale variation in minimum temperatures at savanna sites. Our results demonstrate that it is feasible to model site-scale variation in daily surface temperature extremes and within-site heterogeneity in surface temperatures using LiDAR-derived variables, supporting efforts to understand cross-scale relationships between surface microclimates and regional climate change. Improved understanding of topographic and vegetative buffering of thermal microclimates across mountain landscapes is key to projecting microclimate heterogeneity and potential species’ range dynamics under future climate change.

Impact of Local Groundwater Flow Model Errors on Transport and a Practical Solution for the Issue.

A groundwater flow model is typically used to provide the flow field for conducting groundwater solute transport simulations. The advection term of the mass conserved formulation for groundwater transport assumes that the flow field is perfectly balanced and that all water flowing into a numerical grid cell is exactly balanced by outflows after accounting for sources/sinks or internal storage. However, in many complicated regional or site-scale models, there may be localized flow balance errors that may be difficult to eliminate through tighter flow convergence tolerances due to simulation time constraints or numerical limits on convergence tolerances. Thus, if water is erroneously gained or lost within a grid cell during the flow computation, the solutes within it will also be numerically affected in the associated transport simulation. Transport solutions neglect this error in groundwater flow as the transport equations that are solved assume no error in flow. This flow imbalance error can however have consequences on the transport solution ranging from unnoticeable errors in the resulting concentrations to spurious oscillations that can grow in time and hinder further solution. An approach has been suggested here, to explicitly handle these flow imbalances during mass conserved advective transport computations and report them in the corresponding transport mass balance output, as corrections that are needed to handle errors originating in the flow solution. Example problems are provided to explain the concepts and demonstrate the impacts.

Experimental process parameterization of a 3D site-scale model on effects of a compressed air intrusion into a shallow aquifer

Gases stored in the subsurface for balancing power fluctuations potentially escape from their reservoirs and affect protected groundwater resources. Here we presented an integrated experimental and modeling approach examining the site-scale effects of a compressed air leakage into a shallow aquifer. A main consequence may be pyrite oxidation, which can be strongly inhibited by surface passivation in carbonate-buffered aquifers. Based on laboratory results, predictions for changes in groundwater composition at site-scale and suggestions for a leakage monitoring were presented. Our study found this workflow consisting of experiments and numerical simulations to be valuable for predicting geochemical consequences of gas leakages.

Site scale modeling of slow-moving landslides, a 3D viscoplastic finite element modeling approach

This paper presents an advanced 3D numerical methodology to reproduce the kinematics of slow active landslides, more precisely, to reproduce the nearly constant strain rate (secondary creep) and the acceleration/deceleration of the moving mass due to hydrological changes. For this purpose, finite element analyses are performed in a large area covering a long time-span (12 years), in order to exhibit different interacting slope movements. First, we perform a stability analysis using the shear strength reduction (SSR) technique with a Mohr-Coulomb failure criteria. It is done in order to compute factors of safety (FS) and to identify two different scenarios, the first one being stable (FS > 1) and the second one being unstable (FS < 1). In the studied test case, the Portalet landslide (Central Spanish Pyrenees), the first scenario corresponds to an initial stable configuration of the slope and the second one to an unstable excavated configuration. Second, taking the first scenario as an initial condition, a time-dependent analysis is performed using a coupled formulation to model solid skeleton and pore fluids interaction, and a simplified ground water model that takes into account daily rainfall intensity. In this case, a viscoplastic constitutive model based on Perzyna’s theory is applied to reproduce soil viscous behavior and the delayed creep deformation due to the excavation. The fluidity parameter is calibrated to reproduce displacements measured by the monitoring systems. Our results demonstrate that 3D analyses are preferable to 2D ones for reproducing in a more realistic way the slide behavior. After calibration, the proposed model is able to simulate successfully short- and medium-term predictions during stages of primary and secondary creep.

Northern Bobwhite Nest Site Selection in Field Borders

Field borders are used to supplement early successional habitat critical for northern bobwhite (Colinus virginianus) nesting that is lost to modern intensive agricultural practices. The suitability of field border habitat for nesting may be affected by microhabitat characteristics at the site and patch scale and placement relative to various land-cover types at the landscape scale. We sought to determine whether bobwhite select nest locations at site, patch, and landscape scales. We collected microhabitat data (stem density, percent cover, and ground composition) and distance to land-cover type data (woody edge, crop, ditch, and road) from 26 bobwhite nests and 26 control sites in field borders in North Carolina, USA, during 2010 and 2011. We modeled nest site selection by comparing nests with random locations using conditional logistic regression at the site scale and logistic regression at the combined patch–landscape scale. We performed model selection using the small sample Akaike’s Information Criterion (AICc). The top site-scale model showed that bobwhite selected for the presence of woody cover and avoided open soil at the nest. There was no clear top model at the combined patch–landscape level. In an agriculture-dominated landscape, managers should focus on microhabitat characteristics of field borders to improve suitability for bobwhite nesting.

Sedimentological characterization of Antarctic moraines using UAVs and Structure-from-Motion photogrammetry

AbstractIn glacial environments particle-size analysis of moraines provides insights into clast origin, transport history, depositional mechanism and processes of reworking. Traditional methods for grain-size classification are labour-intensive, physically intrusive and are limited to patch-scale (1 m2) observation. We develop emerging, high-resolution ground- and unmanned aerial vehicle-based ‘Structure-from-Motion’ (UAV-SfM) photogrammetry to recover grain-size information across a moraine surface in the Heritage Range, Antarctica. SfM data products were benchmarked against equivalent datasets acquired using terrestrial laser scanning, and were found to be accurate to within 1.7 and 50 mm for patch- and site-scale modelling, respectively. Grain-size distributions were obtained through digital grain classification, or ‘photo-sieving’, of patch-scale SfM orthoimagery. Photo-sieved distributions were accurate to &lt;2 mm compared to control distributions derived from dry-sieving. A relationship between patch-scale median grain size and the standard deviation of local surface elevations was applied to a site-scale UAV-SfM model to facilitate upscaling and the production of a spatially continuous map of the median grain size across a 0.3 km2area of moraine. This highly automated workflow for site-scale sedimentological characterization eliminates much of the subjectivity associated with traditional methods and forms a sound basis for subsequent glaciological process interpretation and analysis.

Hydrogeology and design of groundwater heat pump systems

Groundwater can serve as an environmentally friendly and renewable energy source when utilized in a groundwater heat pump system. Installation of such systems has been increasing worldwide in recent years. Performance of a groundwater heat pump system depends on local hydrogeological and thermo geological conditions. It is important that aquifer properties be determined to be representative of the site when designing a system in order to maximize efficiency, reduce installation costs, and minimize any environmental impacts. Factors that need to be considered are the groundwater temperature profile to the depth at which the temperature becomes relatively stable; hydraulic and thermal conductivities; diffusivity of the soil and rock layers; groundwater levels; and groundwater quality. Selection of groundwater and heat pump coupling techniques relies on, to a large extent, an accurate conceptual site scale model of groundwater conditions and understanding of environmental and legal constraints.

Uncertainty and variability of infiltration at Yucca Mountain: Part 1. Numerical model development

[1] The U.S. Nuclear Regulatory Commission investigated climate and infiltration at Yucca Mountain to (i) understand important controls and uncertainties influencing percolation through the unsaturated zone on multimillennial time scales and (ii) provide flux boundary conditions for up to 1 million years in performance assessment models of the proposed Yucca Mountain repository. This first part of a two-part series describes a procedure for abstracting the results from detailed numerical simulations of local-scale infiltration into a site-scale model considering uncertainty and variability in distributed net infiltration. Part 2 describes site-scale model results and corroboration. A detailed one-dimensional numerical model was used to estimate bare-soil net infiltration at the scales of hours and meters for 442 soil, bedrock, and climate combinations. The set of results are abstracted into three parametric response functions for decadal-average bare-soil infiltration given hydraulic and climatic parameters. The three abstractions describe deep soil, shallow soil over a coarser layer, and shallow soil over a finer layer. The site-scale model considers spatial variability and uncertainty of the input parameters on a 30 m grid, using the abstractions independently in each cell. Two additional abstractions account for overland flow and vegetation. The model uses Monte Carlo simulation, with all input parameters uncertain and spatially variable, to calculate the mean and standard deviation of net infiltration in each grid cell for selected climate states. Using abstractions rather than detailed simulations speeds calculation of infiltration realizations by many orders of magnitude relative to a detailed simulation.

Uncertainty and variability of infiltration at Yucca Mountain: Part 2. Model results and corroboration

[2] The U.S. Nuclear Regulatory Commission actively investigated climate and infiltration at Yucca Mountain for two decades to (i) understand important controls and uncertainties influencing percolation through the unsaturated zone on multimillennial time scales and (ii) provide flux boundary conditions for up to 1 million years in performance assessment models of the proposed Yucca Mountain repository. This second part of a two-part series describes site-scale model results for present and potential future conditions and confirmatory analyses for present-day conditions. At both the grid-cell and site-average scale, the calculated uncertainty distribution for net infiltration is approximately lognormal, and the coefficient of variation decreases with increasing net infiltration. Smaller relative but larger absolute responses to climate change occur where net infiltration is large. Comparisons of distributed model estimates with temperature and geochemical observations from the unsaturated zone suggest that average estimates are generally consistent but exhibit significant variability. An observed seepage event in the South Ramp of the Exploratory Studies Facility, combined with related subsurface observations across the site, suggests that subsurface spreading from zones of high infiltration to zones of low infiltration may occur in stratabound fractures, laterally extensive discontinuities, or at transitions between welded and nonwelded tuff units. Two conceptual models for unsaturated-zone flow each explain the subsurface observations, collectively providing bounding estimates for net infiltration. Model-predicted uncertainty distribution for decadal-average site-scale net infiltration is generally consistent with estimated percolation fluxes using the bounding hypotheses, suggesting that the model-calculated uncertainty is reasonably consistent with the uncertainty in interpreting site observations.

LandscapeDNDC: a process model for simulation of biosphere–atmosphere–hydrosphere exchange processes at site and regional scale

We present a new model system, which facilitates scaling of ecosystem processes from the site to regional simulation domains. The new framework LandscapeDNDC—partly based on the biogeochemical site scale model DNDC—inherits a series of new features with regard to process descriptions, model structure and data I/O functionality. LandscapeDNDC incorporates different vegetation types and management systems for simulating carbon, nitrogen and water related biosphere–atmosphere–hydrosphere fluxes in forest, arable and grassland ecosystems and allows the dynamic simulation of land use changes. The modeling concept divides ecosystems into six substates (canopy air chemistry, microclimate, physiology, water cycle, vegetation structure, and soil biogeochemistry) and provides alternative modules dealing with these substates. The model can be applied on the site scale, as well as for three-dimensional regional simulations. For regional applications LandscapeDNDC integrates all grid cells synchronously forward in time. This allows easy coupling to other spatially distributed models (e.g. for hydrology or atmospheric chemistry) and efficient two-way exchange of states. This paper describes the fundamental design concept of the model and its object-oriented software implementation. Two example applications are presented. First, calculation of a nitrous oxide emission inventory from agricultural soils for the State of Saxaony (Germany), including data preprocessing of the regional model input data. The computational effort for the LandscapeDNDC preprocessing and simulation could be speed up by a factor of almost 100 compared to the approach using the original DNDC version 9.3. Calculated N2O emissions for Saxony with LandscapeDNDC (2693 t N2O–N/a) were compared with the original DNDC model (2725 t N2O–N/a), the IPCC Tier I methodology (1107 t N2O–N/a), and the German National Inventory Report (equal to IPCC Tier II, 2100 t N2O–N/a). The second example illustrates the capabilities of LandscapeDNDC for building a fully coupled three-dimensional model system on the landscape scale. Therefore we coupled the biogeochemical and plant growth calculations to a hydrological transport model and demonstrate the transport of nitrogen along a virtual hillslope and associated formation of indirect nitrous oxide emissions.

Initial geological considerations before installing ground source heat pump systems

Abstract The performance of an open- or closed-loop ground source heat pump system depends on local geological conditions. It is important that these are determined as accurately as possible when designing a system, to maximize efficiency and minimize installation costs. Factors that need to be considered are surface temperature, subsurface temperatures down to 100–200 m, thermal conductivities and diffusivities of the soil and rock layers, groundwater levels and flows, and aquifer properties. In addition, rock strength is a critical factor in determining the excavation or drilling method required at a site and the associated costs. The key to determining all of these factors is an accurate conceptual site-scale model of the ground conditions (soils, geology, thermogeology, engineering geology and hydrogeology). The British Geological Survey has used the modern digital geological mapping of the UK as a base onto which appropriate attributes can be assigned. As a result it is possible to generate regional maps of surface and subsurface temperatures, rock strength and depth to water. This information can be used by designers, planners and installers of ground source heat pump systems. The use of appropriate geological factors will assist in creating a system that meets the heating or cooling load of the building without unnecessary overengineering.

Multicomponent Simulations of Contrasting Redox Environments at an LNAPL Site

A two-dimensional multicomponent reactive transport modeling approach was used to simulate contaminant transport and the evolution of redox processes at a large-scale kerosene-contaminated site near Berlin, Germany. In contrast to previous site-scale modeling studies that focused either on one or two contaminants or on steady-state redox conditions, multiple contaminants and electron acceptors, including mineral phase Iron (III), were considered with an evolving redox zonation. Inhibition terms were used to switch between the different electron acceptor processes in the reaction scheme. The transient evolution of redox zones and contaminant plumes was simulated for two separate transects of the site, which have different geology and ground water recharge distributions and where quite different downstream contaminant and terminal electron-accepting process (TEAP) distributions are observed. The same reaction system, calibrated to measured concentrations along one of the transects, was used in both cases, achieving a reasonable match with observed concentrations. The differences between the two transects could thus to some extent be attributed to the different hydrological and hydrogeological conditions, in particular ground water recharge distributions. Long-term simulations showed that the distribution of TEAPs evolves as Fe(III) becomes depleted, with conditions becoming increasingly methanogenic, leading to changes in contaminant plume lengths. The models were applied to assess the potential effects of planned changes in land use at the site that may affect the ground water recharge distribution. The simulated redox zonation responded strongly to changes in recharge, which in turn led to changes in the contaminant plume lengths.