Dynamic Reservoir Simulation Model Research Articles

Integrating hydraulic flow unit concept and adaptive neuro-fuzzy inference system to accurately estimate permeability in heterogeneous reservoirs: Case study Sif Fatima oilfield, southern Algeria

Formation characteristics are a crucial requirement for reservoir modeling and they need to be constantly updated as new static and/or dynamic reservoir-property details become available. This improves reservoir interpretation while achieving a good history-matching for a better representation of the reservoir. This paper proposes a new method to estimate reservoir permeability in an oil-producing reservoir by applying two distinct cases: partially cored and completely non-cored wellbores. The approach involves the determination of hydraulic-flow units (HFU) using a flow-zone indicator (FZI) distribution to consider the dynamic aspect of the permeability, which strongly influences fluid flow through porous media. The FZI is clustered by normal probability analysis to identify the optimum number of HFU clusters. The estimated reservoir permeability is validated in static and dynamic reservoir models. For an accurate prediction of the FZI over the entire reservoir thickness, an adaptive-neuro-fuzzy-inference system (ANFIS) is developed using conventional well-log variables: (i) gamma-ray (GR), (ii) sonic (DT), (iii) density (RHOB), (iv) deep resistivity (DLL), and (iv) neutron porosity (NPHI) calibrated with the available core data. The static regression parameters root-mean-squared error (RMSE), mean absolute error (MAE), coefficient of determination (R2), and R2adjusted are used to evaluate the performance of the ANFIS model in predicting FZI to determine HFU for estimating reservoir permeability. To validate this approach, a matching quality check of the downhole and reservoir pressure is carried out using a dynamic reservoir simulation model. Historical data collected from wellbore OPW-1 drilled in the Sif Fatima oilfield (Algeria) are used to evaluate the accuracy of the proposed permeability-prediction framework. The ANFIS model can accurately and efficiently predict the FZI exhibiting R2, R2adjusted, RMSE, and MAE of approximately (≈0.981, ≈0.979, ≈0.014, ≈4,600E-08)), (≈0.933, ≈0.891, ≈0.032, ≈1.615E-03), and (≈0.970, ≈0.968, ≈0.018, ≈2.381E-04)) for the training, testing and complete dataset, respectively. When applied to predict reservoir permeability, the Hydraulic Flow Unit concept was able to generate accurate results from the predicted FZI data by applying the modified Kozeny-Carman equation with, R2, RMSE and MAE values of approximately ≈0.9702, ≈0.969, ≈9.604, and ≈4.398, respectively. The calculated permeability was validated successfully at well and field levels. Additionally, it generated more accurate history matches of the well and the reservoir pressure trends than the previous permeability model, and consequently improved the reservoir modelling workflow.

A new pseudo three-dimensional hydrodynamic model with Lagrangian vertical mixing and Eulerian horizontal mass exchange

The Multi-Basin DYnamic Reservoir Simulation Model (MB-DYRESM) has been developed to bridge the gap between the computationally cheap, one-dimensional lake models and the computationally expensive, three-dimensional lake models. It was extended from the classic one-dimensional lake model DYRESM by incorporating a horizontal water exchange algorithm to simulate gravitational mass transport between adjacent sub-basins. Numerical experiments with increasing complexity were carried out for capacity demonstration, in which MB-DYRESM successfully reproduced the thermal structure and mass distribution in each sub-basin with accuracy comparable to and computational speed approximately 300 times faster than the three-dimensional model applied in parallel. In situations where turbulence was suppressed by stratification it actually provided better simulation results. The combination of low computational demand, reasonable spatial resolution of mass flux path and constrained numerical diffusion makes MB-DYRESM a more applicable option for simulations where the balance between affordable computational power and proper spatial resolution is concerned.

Integrated simulation to seismic and seismic reservoir characterization in a CO2 EOR monitoring application

Understanding the behavior of CO2 injected into a reservoir and delineating its spatial distribution are fundamentally important in enhanced oil recovery (EOR) and CO2 capture and sequestration activities. Interdisciplinary geoscience collaboration and well-defined workflows, from data acquisition to reservoir simulation, are needed to effectively handle the challenges of EOR fields and envisioned future commercial-scale sites for planned and incidental geologic CO2 storage. Success of operations depends on decisions that are based on good understanding of geologic formation heterogeneities and fluid and pressure movements in the reservoir over large areas over time. We present a series of workflow steps that optimize the use of available data to improve and integrate the interpretation of facies, injection, and production effects in an EOR application. First, we construct a simulation-to-seismic model supported by rock physics to model the seismic signal and signal quality needed for 4D monitoring of fluid and pressure changes. Then we use Bayesian techniques to invert the baseline and monitor seismic data sets for facies and impedances. To achieve a balance between prior understanding of the reservoir and the recorded time-lapse seismic data, we invert the seismic data sets by using multiple approaches. We first invert the seismic data sets independently, exploring sensible parameter scenarios. With the resulting realizations, we develop a shared prior model to link the reservoir facies geometry between seismic vintages upon inversion. Then we utilize multirealization analysis methods to quantify the uncertainties of our predictions. Next, we show how data may be more deeply interrogated by using the facies inversion method to invert prestack seismic differences directly for production effects. Finally, we show and discuss the feedback loop for updating the static and dynamic reservoir simulation model to highlight the integration of geophysical and engineering data within a single model.

Using Bayesian Model Probability for Ranking Different Prior Scenarios in Reservoir History Matching

Summary This paper considers Bayesian methods to discriminate between models depending on posterior model probability. When applying ensemble-based methods for model updating or history matching, the uncertainties in the parameters are typically assumed to be univariate Gaussian random fields. In reality, however, there often might be several alternative scenarios that are possible a priori. We take that into account by applying the concepts of model likelihood and model probability and suggest a method that uses importance sampling to estimate these quantities from the prior and posterior ensembles. In particular, we focus on the problem of conditioning a dynamic reservoir-simulation model to frequent 4D-seismic data (e.g., permanent-reservoir-monitoring data) by tuning the top reservoir surface given several alternative prior interpretations with uncertainty. However, the methodology can easily be applied to similar problems, such as fault location and reservoir compartmentalization. Although the estimated posterior model probabilities will be uncertain, the ranking of models according to estimated probabilities appears to be quite robust.

Fully integrated dynamic reservoir and well simulation model: Blacktip gas field case study

This paper describes an experience of integrating a dynamic reservoir simulation model with a dynamic well simulation model, resulting in an integrated dynamic model from the reservoir to the surface that enhances reservoir and well surveillance capability for the Blacktip gas field. Multiphase transient flow simulation is used to support daily well and pipeline operations for the project. The limitation of the standalone well model using a multiphase transient simulation software was its inability to reproduce pressure build up response during shut-ins, and pressure drawdown during start-ups. The fluid inflow from the reservoir to the well bore is modelled using the Inflow Performance Relationship (IPR) and accordingly the transient pressure behaviour near the well bore is not captured. This makes it difficult to estimate an accurate static reservoir pressure during shut-ins, as the predicted pressure instantaneously builds up to reservoir pressure specified in the IPR. The integrated dynamic model overcomes this limitation. The history matching of production intervals including shut-ins and start-ups using the integrated dynamic model along with high frequency data demonstrates that this integrated modelling approach can be used as a reliable surveillance tool to understand dynamic flow conditions from the reservoir to the surface, including liquid loading and unloading scenarios. The evolution of the history match and subsequent outcomes are discussed in the paper, along with the lessons learnt. Results of a liquid loading and unloading scenario for a gas well are also discussed in the paper.

An Approach to Estimation of Gas Reserve of Narshingdi Gas Field in Bangladesh with Dynamic Reservoir Simulation

Increasing demand of fuel globally formulates gas as one of the most valuable natural resources. There is lot of uncertainties in estimating hydrocarbon volume correctly from exploration to development stage of a gas field. The accuracy and reliability of data (reservoir geological model, fluid and rock properties) make the implement very hard-hitting. So estimating and updating the gas reserve has become vital issue, as it helps the planners for drawing mid-term and long-term development plan from field development level to national level. This paper presents the study of reserve estimation of a Narshingdi Gas Field in Bangladesh. In this paper, a dynamic reservoir simulation model has been used to perform a history match ?pressure and production? using commercial simulator for reserve estimation. The result of this study is expected to provide Gas Initially in Place (GIIP) and recoverable gas volume. Simultaneously three forecast scenarios have also been investigated. There is no strong aquifer pressure support in the producing gas zone, so gas production continues from the reservoir due to pressure depletion.Journal of Chemical Engineering, Vol. 29, No. 1, 2017: 19-23

The value of fault analysis for field development planning

Faults play an important role in reservoir compartmentalization and can have a significant impact on recoverable volumes. A recent petroleum discovery in the Norwegian offshore sector, with an Upper Jurassic reservoir, is currently in the development planning phase. The reservoir is divided into several compartments by syndepositional faults that have not been reactivated and do not offset the petroleum-bearing sandstones completely. A comprehensive fault analysis has been conducted from core to seismic scale to assess the likely influence of faults on the production performance and recoverable volumes. The permeability of the small-scale faults from the core were analysed at high confining pressures using formation-compatible brines. These permeability measurements provide important calibration points for the fault property assessment, which was used to calculate transmissibility multipliers (TM) that were incorporated into the dynamic reservoir simulation model in order to account for the impact of faults on fluid flow. Dynamic simulation results reveal a range of more than 20% for recoverable volumes, depending on the fault property case applied and for a base case producer–injector well pattern. Fault properties are one of the key parameters that influence the range of cumulative recoverable oil volumes and the recovery efficiency.

Modeling climate change impacts on the thermal dynamics of polymictic Oneida Lake, New York, United States

The thermal dynamics of lakes are affected by climate change, and in turn affect most aspects of lake ecology, including primary and secondary production, biogeochemistry, and biodiversity. These effects may be particularly large in polymictic lakes which have variable durations and degrees of thermal stratification. A one-dimensional model, the Dynamic Reservoir Simulation Model (DYRESM), was used to evaluate the impacts of predicted climate change on temperature and stability in Oneida Lake, a 207km2 shallow, polymictic lake in central New York State. Field data were collected on stream temperature and discharge, weather, and lake temperatures at varying depths to calibrate and validate the model for Oneida Lake. Modeled temperatures at 2m and 10m were generally within ±1°C of observed values, but only after wind speeds from a nearby airport were decreased by 25%. Downscaled climate data from three general circulation models and two emissions scenarios were used to assess the impacts of different anticipated climate scenarios on the lake for 2041–2050 and 2090–2099. By 2050, under the higher emissions scenario, water temperatures (April–November) at 2m and 10m depths indicated an increase of 1.36°C and 1.43°C, respectively; however, 2011 simulation temperatures already aligned with mid-century predictions. By the end of the century, the prediction of the higher emissions scenario indicated an increase in 2m and 10m water temperatures (April–November) of 3.70°C and 3.37°C, respectively, and an increase of 61 consecutive days of stratification in Oneida Lake. These impacts would lead to increased periods of lack of oxygen in the bottom waters of this lake and changes in the biotic species composition, including extirpation of a coldwater fish species, burbot (Lota lota). This study contributes to our understanding of the thermal regime of polymictic lakes in a warmer world.

One-dimensional simulation of lake and ice dynamics during winter

An ice formation model, based on the solution of the heat conduction equation across blue ice, white ice and snow cover, is integrated into the Dynamic Reservoir Simulation Model (DYRESM) to allow for one-dimensional (vertical) winter simulation of lake dynamics during periods of ice cover. This is an extension of a previous three-layer snow and ice model to include two-way coupling between the ice and the water column. The process-based ice formation is suitable for application to mid-latitude regions and includes: snowmelt due to rain; formation of white ice; and variability in snow density, snow conductivity, and ice and snow albedo. The model was validated against published observations from Harmon lake, British Columbia, and new observations from Eagle lake, Ontario. The ice thickness and water column temperature profile beneath the ice were predicted with Root Mean Square Deviations (RMSD) of 1 cm and 0.38°C, respectively, during the winter of 1990-91in Harmon lake. In Eagle lake the 2011-12 year-round water column temperature profile was predicted with an RMSD of 1.8°C. Improved prediction of under-ice lake temperature, relative to published results from simpler models, demonstrates the need for models that accurately capture ice-formation processes, including ice to water column coupling, formation of both blue and white ice layers, and process-based ice and snow parameters (density, conductivity and albedo).

Climate warming alters thermal stability but not stratification phenology in a small north-temperate lake

Recent climate change represents one of the most serious anthropogenic threats to lake ecosystems in Canada. As meteorological and hydrological conditions are altered by climate change, so too are physical, chemical and biological properties of lakes. The ability to quantify the impact of climate change on the physical properties of lakes represents an integral step in estimating future chemical and biological change. To that end, we have used the dynamic reservoir simulation model, a one-dimensional vertical heat transfer and mixing model, to hindcast and compare lake temperature-depth profiles against 30 years of long-term monitoring data in Harp Lake, Ontario. These temperature profiles were used to calculate annual (June–September) thermal stability values from 1979 to 2009. Comparisons between measured and modelled lake water temperature and thermal stability over three decades showed strong correlation (r2 > 0.9). However, despite significant increases in modelled thermal stability over the 30 year record, we found no significant change in the timing of the onset, breakdown or the duration of thermal stratification. Our data suggest that increased air temperature and decreased wind are the primary drivers of enhanced stability in Harp Lake since 1979. The high-predictive ability of the Harp Lake dynamic reservoir simulation model suggests that its use as a tool in future lake management projects is appropriate. Copyright © 2013 John Wiley & Sons, Ltd.

Simulating water temperatures and stratification of a pre-alpine lake with a hydrodynamic model: calibration and sensitivity analysis of climatic input parameters

We report on the calibration of the one-dimensional hydrodynamic lake model Dynamic Reservoir Simulation Model to simulate the water temperature conditions of the pre-alpine Lake Ammersee (southeast Germany) that is a representative of deep and large lakes in this region. Special focus is given to the calibration in order to reproduce the correct thermal distribution and stratification including the time of onset and duration of summer stratification. To ensure the application of the model to investigate the impact of climate change on lakes, an analysis of the model sensitivity under stepwise modification of meteorological input parameters (air temperature, wind speed, precipitation, global radiation, cloud cover, vapour pressure and tributary water temperature) was conducted. The total mean error of the calibration results is −0.23 °C, the root mean square error amounts to 1.012 °C. All characteristics of the annual stratification cycle were reproduced accurately by the model. Additionally, the simulated deviations for all applied modifications of the input parameters for the sensitivity analysis can be differentiated in the high temporal resolution of monthly values for each specific depth. The smallest applied alteration to each modified input parameter caused a maximum deviation in the simulation results of at least 0.26 °C. The most sensitive reactions of the model can be observed through modifications of the input parameters air temperature and wind speed. Hence, the results show that further investigations at Lake Ammersee, such as coupling the hydrodynamic model with chemo-dynamic models to assess the impact of changing climate on biochemical conditions within lakes, can be carried out using Dynamic Reservoir Simulation Model. Copyright © 2012 John Wiley & Sons, Ltd.

Sensitivity analysis for complex ecological models – A new approach

A strategy for global sensitivity analysis of a multi-parameter ecological model was developed and used for the hydrodynamic-ecological model (DYRESM–CAEDYM, DYnamic REservoir Simulation Model- Computational Aquatic Ecosystem Dynamics Model) applied to Lake Kinneret (Israel). Two different methods of sensitivity analysis, RPART (Recursive Partitioning And Regression Trees) and GLM (General Linear Model) were applied in order to screen a subset of significant parameters. All the parameters which were found significant by at least one of these methods were entered as input to a GBM (Generalized Boosted Modeling) analysis in order to provide a quantitative measure of the sensitivity of the model variables to these parameters. Although the GBM is a general and powerful machine learning algorithm, it has substantial computational costs in both storage requirements and CPU time. Employing the screening stage reduces this cost. The results of the analysis highlighted the role of particulate organic material in the lake ecosystem and its impact on the over all lake nutrient budget. The GBM analysis established, for example, that parameters such as particulate organic material diameter and density were particularly important to the model outcomes. The results were further explored by lumping together output variables that are associated with sub-components of the ecosystem. The variable lumping approach suggested that the phytoplankton group is most sensitive to parameters associated with the dominant phytoplankton group, dinoflagellates, and with nanoplankton (Chlorophyta), supporting the view of Lake Kinneret as a bottom–up system. The study demonstrates the effectiveness of such procedures for extracting useful information for model calibration and guiding further data collection.

Simulation of multiannual thermal profiles in deep Lake Geneva: A comparison of one‐dimensional lake models

In this study, we report on the ability of four one‐dimensional lake models to simulate the water temperature profiles of Lake Geneva, the largest water body in Western Europe, over a 10‐yr period from 1996 to 2005, using lake models driven by a common atmospheric forcing. These lake models have already demonstrated their capability of reproducing the temperature distribution in smaller lakes and include one eddy‐diffusive lake model, the Hostetler model; a Lagrangian model, the one‐dimensional Dynamic Reservoir Simulation Model "DYRESM" a к ‐ ε turbulence model, "SIMSTRAT"; and one based on the concept of self‐similarity (assumed shape) of the temperature‐depth curve, the Freshwater Lake model "FLake." Only DYRESM and SIMSTRAT reproduce the variability of the water temperature profiles and seasonal thermocline satisfactorily. In layers in which thermocline variability is greatest, the temperature root mean square error is ≪2°C and 3°C (at the time of highest stratification) for these models, respectively. It is possible to apply certain one‐dimensional lake models that simulate the behavior of temperature to investigate the potential future warming of the water column in Lake Geneva. Importantly, the metalimnion boundary is successfully modeled, which represents an encouraging step toward demonstrating the feasibility of coupling biogeochemical modules, such as, for example, a phytoplanktonic model, to assess the possible biological responses within lakes to climate change.

Calibrating the Dynamic Reservoir Simulation Model (DYRESM) and filling required data gaps for one‐dimensional thermal profile predictions in a boreal lake

One‐dimensional vertical heat transfer and mixing models, such as the Dynamic Reservoir Simulation Model (DYRESM), have seldom been applied to lakes in the boreal region even though this region houses the majority of global freshwater lakes. In order to employ DYRESM to predict the thermal structure of a boreal lake located near Sudbury, Ontario, Canada, we overcame two methodological challenges. First, we developed models to predict the vertical light extinction coefficient (Kd) from dissolved organic carbon (DOC) concentrations and hydraulic retention time. We also developed models to predict stream temperatures from local meteorology, and to predict the discharge of lake inflows and the lake outflow from runoff per unit area at gauged streams nearby. We then re‐calibrated several DYRESM parameters which had been tested previously primarily in the Southern Hemisphere, and explored the sensitivity of the re‐calibrated model to all of the remaining uncalibrated inputs implicated in heating and mixing processes. The mean difference between values predicted with the re‐calibrated model and field measurements (± 1 standard deviation), 1.09 m (± 0.89 m) for thermocline depth and 1.98°C (± 1.58°C) for bottom water temperature, was relatively small compared with other North American studies, and likely due to the model rather than our parameterization. Our calibration of DYRESM for Clearwater Lake, and supplementary models, provide a demonstration for and guidance to those wishing to simulate changes in thermal regimes of boreal lakes in response to climate change or other broad‐scale environmental stressors of importance both to local fisheries and freshwater resource management.

Modeling spatial and temporal variations in temperature and salinity during stratification and overturn in Dexter Pit Lake, Tuscarora, Nevada, USA

This paper examines the seasonal cycling of temperature and salinity in Dexter pit lake in arid northern Nevada, and describes an approach for modeling the physical processes that operate in such systems. The pit lake contains about 596,200 m 3 of dilute, near neutral (pHs 6.7–9) water. Profiles of temperature, conductivity, and selected element concentrations were measured almost monthly during 1999 and 2000. In winter (January–March), the pit lake was covered with ice and bottom water was warmer (5.3 °C) with higher total dissolved solids (0.298 g/L) than overlying water (3.96 °C and 0.241 g/L), suggesting inflow of warm (11.7 °C) groundwater with a higher conductivity than the lake (657 versus 126–383 μS/cm). Seasonal surface inflow due to spring snowmelt resulted in lower conductivity in the surface water (232–247 μS/cm) relative to deeper water (315–318 μS/cm). The pit lake was thermally stratified from late spring through early fall, and the water column turned over in late November (2000) or early December (1999). The pit lake is a mixture of inflowing surface water and groundwater that has subsequently been evapoconcentrated in the arid environment. Linear relationships between conductivity and major and some minor (B, Li, Sr, and U) ions indicate conservative mixing for these elements. Similar changes in the elevations of the pit lake surface and nearby groundwater wells during the year suggest that the pit lake is a flow-through system. This observation and geochemical information were used to configure an one-dimensional hydrodynamics model (Dynamic Reservoir Simulation Model or DYRESM) that predicts seasonal changes in temperature and salinity based on the interplay of physical processes, including heating and cooling (solar insolation, long and short wave radiation, latent, and sensible heat), hydrologic flow (inflow and outflow by surface and ground water, pumping, evaporation, and precipitation), and transfers of momentum (wind stirring, convective overturn, shear, and eddy diffusion). Inputs to the model include the size and shape of the lake, daily meteorological data (short wave radiation, long wave radiation or cloud cover, air temperature, vapor pressure, wind speed, and rainfall), rates for water inputs and outputs, the composition of inflowing water, and initial profiles of temperature and salinity. Predicted temperature profiles, which are influenced by seasonal changes in the magnitude of solar radiation, are in good agreement with observations and show the development of a strong thermocline in the summer, erosion of the thermocline during early fall, and turnover in late fall. Predicted salinity profiles are in reasonable agreement with observations and are affected by the hydrologic balance, particularly inflow of surface and groundwater and, to a lesser degree, evaporation. Defining the hydrodynamics model for Dexter pit lake is the first step in using a coupled physical – biogeochemical model (Dynamic Reservoir Simulation Model-Computational Aquatic Ecosystem Dynamics Model or DYRESM-CAEDYM) to predict the behavior of non-conservative elements (e.g., dissolved O 2, Mn, and Fe) and their effect on water quality in this system.

Modelling of bubble plume destratification using DYRESM

North Pine Dam in Queensland, like other Australian reservoirs, stratifies in summer. Thermal stratification prevents the vertical transfer of heat and oxygen and deteriorates the water quality. Since North Pine Dam is a source of drinking water supply, bubble plume destratification system was chosen to counter the effects of stratification. A numerical model called Destrat was used to design a bubble plume system which destratifies the strongest observed stratification in the reservoir within 3 weeks. Field observations, however, showed that the destratification of North Pine Dam occurred slower than predicted by Destrat. Therefore, Destrat was modified for future applications by including the effect of solar radiation as a source of stratification. The Dynamic Reservoir Simulation Model (DYRESM) was then run to model the destratification of North Pine Dam. The results of this simulation were not in agreement with the field data as DYRESM overestimated the destratification caused by the bubble plume. A comparison of the field and simulated results indicated that the volume of water entrained by bubble plume in the DYRESM is overestimated. Reducing the entrainment coefficient of air bubbles produced a closer agreement between the field data and the simulation results.

NUMERICAL ANALYSIS OF EUTROPHICATION AND THERMOSOLUTAL STRATIFICATION IN RESERVOIRS

Reservoir eutrophication brings deoxygenation of hypolimnetic water and high concentration of metal, nutrients and dissolved materials are released from the bed sediments. Production of dissolved materials then increases fluid density along the reservoir bed. On a sloping boundary, the heavy water mass generates an inclined thermosolutal plume that transports the polluted water and heat energy from the shallow layer to the deep layer. The authors found the plume is responsible for generation of stable salty water with inverse thermal stratification in the bottom layer.The present study is to develop a model that can reproduce eutrophication as well as seasonal variation of the thermal and salinity structure produced by this mechanism. The model is Lagrange-type based on a one-dimensional DY namic Reservoir Simulation Model, DYRESM. The original model was modified so that the thermosolutal convection and chemical and biological processes can be taken into consideration. The present model was verified through comparison between the simulation results and the field data.

Development of an Integrated Reservoir Model for a Naturally Fractured Volcanic Reservoir in China

Summary This paper presents the development of an integrated, multidiscipline reservoir model for dynamic flow simulation and performance prediction of a geologically complex, naturally fractured volcanic reservoir in the Shang 741 Block of the Shengli field in China. A static geological model integrates lithological information, petrophysics, fracture analysis, and stochastic fracture network modeling with Formation MicroImage (FMI) log data and advanced 3D seismic interpretations. Effective fracture permeability, fracture-matrix interaction, reservoir compartmentalization, and flow transmissibility of conductive faults are obtained by matching various dynamic data. As a result of synergy and multiple iterations among various disciplines, a history-matched dynamic reservoirsimulation model capable of future performance prediction for optimum asset management is constructed.

Numerical simulations of degassing scenarios for CO2‐rich Lake Nyos, Cameroon

A catastrophic outgassing of carbon dioxide (CO2) from Lake Nyos, Cameroon, occurred in 1986. Concentrations of dissolved CO2 have been increasing in the bottom waters since 1986, prompting an international effort to engineer controlled degassing that would maintain lake stability by reducing the gas concentration. A new model, based on the one‐dimensional Dynamic Reservoir Simulation Model has been used to simulate the evolution and distribution of CO2 in Lake Nyos. Initial conditions for the simulations are taken from published observations and integrated forward with synthesized, daily meteorological data for 5 years. After model results were found to be in agreement with observed trends, a hypolimnetic piped withdrawal system for degassing the lake was then modeled to simulate its effect on the CO2 concentration profile over time. In the standard degassing operation, gas‐rich bottom water continuously flows up through pipes, discharging as a fountain at the lake surface. Simulation results indicate that the lake density structure will most likely remain stable and CO2 concentrations will decrease at all depths above the pipe intake. Seven alternative operational strategies for the degassing system were also investigated, including a submerged outlet, an intake raised above the lake bottom, higher flow rate, a disrupted operation schedule, employment of a stabilization pond, multiple intake depths, and release of the withdrawn water from the lake. Multiple intake depths best maintain lake stability. A tentative conclusion is that degassing may need to be carried out indefinitely, albeit intermittently or at a reduced flow rate.

Potential effects of global climate change on small north‐temperate lakes: Physics, fish, and plankton

A 10‐yr record of the thermal characteristics of four lakes at the North‐Temperate Lakes Long‐Term Ecological Research site was analyzed and used to validate simulations of lake physics with the dynamic reservoir simulation model. Simulations of cool, warm, and intermediate years were rerun with meteorology from four general circulation models with a doubling of CO2. In all simulations with doubled CO2 there is an earlier onset of stratification, increased summer epilimnetic temperature (1–7°C), and an increased intensity and longer duration of stratification. Maximum surface temperatures at times may exceed upper lethal limits of warm and cool water fish in some scenarios. Suitable thermal habitat for cold water, cool water, and warm water fish generally increases in all scenarios after climate change.Changes in the vertical migration of Daphnia, however, are expected to vary depending on the interaction of thermal stratification and fish habitat use. In northern Wisconsin lakes with cold water planktivores, habitat overlap between fish and zooplankton is expected to decrease, while in southern Wisconsin lakes habitat overlap is expected to increase. Although most physical responses of lakes to climate change are consistent among all climate scenarios, biological responses will likely be more variable owing to the complex nature of factors determining ecological interactions in lakes.