Enhancing lithofacies machine learning predictions with gamma-ray attributes for boreholes with limited diversity of recorded well logs

Derivative and volatility attributes can be usefully calculated from recorded gamma ray(GR)data to enhance lithofacies classification in wellbores penetrating multiple lithologies.Such attributes extract information about the log curve shape that cannot be readily discerned from the recorded well log data.A logged wellbore section for which 8911 data records are available for the three recorded logs(GR,sonic(DT)and bulk density(PB))is evaluated.That section demonstrates the value of the GR attributes for machine learning(ML)lithofacies predictions.Five feature selection configurations are considered.The 9-var configuration including GR,DT,PB and six GR attributes,and the 7-var configuration of GR and the six GR attributes,provide the most accurate and reproducible lithofacies predictions.The other three feature configurations evaluated do not include the GR attributes but just one to three of the recorded log features.The results of seven ML models and two regression models reveal that K-nearest neighbor(KNN),random forest(RF)and extreme gradient boosting(XGB)are the best performing models.They generate between 14 and 23 misclassification from 8911 data records for the 9-var model.Multi-layer perceptron(MLP)and support vector classification(SVC)do not perform well with the 7-var model which lacks the PB feature displaying the highest correlation with facies class.Annotated confusion matrices reveal that KNN,RF and XGB models can effectively distinguish all facies classes for the 9-var and 7-var configurations(that includes the GR attributes),whereas none of the models can achieve that outcome for the 3-var configuration(that excludes the GR attributes).Accurately distinguishing lithofacies using well-log data in sedimentary sections is an important objective in applied geoscience.The straightforward,GR-attribute method proposed works to improve confidence in ML-lithofacies classifications based on limited recorded well-log data.

Synthetic polymers:A review of applications in drilling fluids

With the growth of deep drilling and the complexity of the well profile,the requirements for a more complete and efficient exploitation of productive formations increase,which increases the risk of various complications.Currently,reagents based on modified natural polymers(which are naturally occurring compounds)and synthetic polymers(SPs)which are polymeric compounds created industrially,are widely used to prevent emerging complications in the drilling process.However,compared to modified natural polymers,SPs form a family of high-molecular-weight compounds that are fully synthesized by undergoing chemical polymerization reactions.SPs provide substantial flexibility in their design.Moreover,their size and chemical composition can be adjusted to provide properties for nearly all the functional objectives of drilling fluids.They can be classified based on chemical ingredients,type of reaction,and their responses to heating.However,some of SPs,due to their structural characteristics,have a high cost,a poor temperature and salt resistance in drilling fluids,and degradation begins when the temperature reaches 130℃.These drawbacks prevent SP use in some medium and deep wells.Thus,this review addresses the historical development,the characteristics,manufacturing methods,classification,and the applications of SPs in drilling fluids.The contributions of SPs as additives to drilling fluids to enhance rheology,filtrate generation,carrying of cuttings,fluid lubricity,and clay/shale stability are explained in detail.The mechanisms,impacts,and advances achieved when SPs are added to drilling fluids are also described.The typical challenges encountered by SPs when deployed in drilling fluids and their advantages and drawbacks are also discussed.Economic issues also impact the applications of SPs in drilling fluids.Consequently,the cost of the most relevant SPs,and the monomers used in their synthesis,are assessed.Environmental impacts of SPs when deployed in drilling fluids,and their manufacturing processes are identified,together with advances in SP-treatment methods aimed at reducing those impacts.Recommendations for required future research addressing SP property and performance gaps are provided.

Integrating petrophysical data into efficient iterative cluster analysis for electrofacies identification in clastic reservoirs

Efficient iterative unsupervised machine learning involving probabilistic clustering analysis with the expectation-maximization(EM)clustering algorithm is applied to categorize reservoir facies by exploiting latent and observable well-log variables from a clastic reservoir in the Majnoon oilfield,southern Iraq.The observable well-log variables consist of conventional open-hole,well-log data and the computer-processed interpretation of gamma rays,bulk density,neutron porosity,compressional sonic,deep resistivity,shale volume,total porosity,and water saturation,from three wells located in the Nahr Umr reservoir.The latent variables include shale volume and water saturation.The EM algorithm efficiently characterizes electrofacies through iterative machine learning to identify the local maximum likelihood estimates(MLE)of the observable and latent variables in the studied dataset.The optimized EM model developed successfully predicts the core-derived facies classification in two of the studied wells.The EM model clusters the data into three distinctive reservoir electrofacies(F1,F2,and F3).F1 represents a gas-bearing electrofacies with low shale volume(Vsh)and water saturation(Sw)and high porosity and permeability values identifying it as an attractive reservoir target.The results of the EM model are validated using nuclear magnetic resonance(NMR)data from the third studied well for which no cores were recovered.The NMR results confirm the effectiveness and accuracy of the EM model in predicting electrofacies.The utilization of the EM algorithm for electrofacies classification/cluster analysis is innovative.Specifically,the clusters it establishes are less rigidly constrained than those derived from the more commonly used K-means clustering method.The EM methodology developed generates dependable electrofacies estimates in the studied reservoir intervals where core samples are not available.Therefore,once calibrated with core data in some wells,the model is suitable for application to other wells that lack core data.

利用井筒稳定性分析确定砂岩安全钻井液密度窗口

以伊朗西南部某油田Asmari组砂岩储集层为研究对象,研究其中一口垂直井的井筒稳定性,利用FLAC3D软件,根据钻遇地层地质力学特征建立井筒的有限体积模型,监测井壁岩石塑性状态的形成以确定砂岩层安全钻井液密度的上限值和下限值。评估了岩石强度特性、井筒周围主要地应力和孔隙压力对该井安全钻井液密度窗口的影响,敏感性分析结果表明,井壁岩石内聚力和内摩擦角的减小会导致安全钻井液密度窗口大幅变窄;孔隙压力和最大水平应力与最小水平应力之比的减小则会使安全钻井液密度窗口显著增大。此模型便于量化安全钻井液密度窗口的变化,可作为一种油气井钻井方案设计和监测工具。

The impacts of gas impurities on the minimum miscibility pressure of injected CO_2-rich gas–crude oil systems and enhanced oil recovery potential

An effective parameter in the miscible-CO_2 enhanced oil recovery procedure is the minimum miscibility pressure(MMP)defined as the lowest pressure that the oil in place and the injected gas into reservoir achieve miscibility at a given temperature. Flue gases released from power plants can provide an available source of CO_2,which would otherwise be emitted to the atmosphere, for injection into a reservoir. However, the costs related to gas extraction from flue gases is potentially high. Hence, greater understanding the role of impurities in miscibility characteristics between CO_2 and reservoir fluids helps to establish which impurities are tolerable and which are not. In this study, we simulate the effects of the impurities nitrogen(N_2), methane(C_1), ethane(C_2) and propane(C_3) on CO_2 MMP. The simulation results reveal that,as an impurity, nitrogen increases CO_2–oil MMP more so than methane. On the other hand, increasing the propane(C_3)content can lead to a significant decrease in CO_2 MMP, whereas varying the concentrations of ethane(C_2) does not have a significant effect on the minimum miscibility pressure of reservoir crude oil and CO_2 gas. The novel relationships established are particularly valuable in circumstances where MMP experimental data are not available.

Machine learning and data-driven prediction of pore pressure from geophysical logs:A case study for the Mangahewa gas field,New Zealand

Pore pressure is an essential parameter for establishing reservoir conditions,geological interpretation and drilling programs.Pore pressure prediction depends on information from various geophysical logs,seismic,and direct down-hole pressure measurements.However,a level of uncertainty accompanies the prediction of pore pressure because insufficient information is usually recorded in many wells.Applying machine learning(ML)algorithms can decrease the level of uncertainty of pore pressure prediction uncertainty in cases where available information is limited.In this research,several ML techniques are applied to predict pore pressure through the over-pressured Eocene reservoir section penetrated by four wells in the Mangahewa gas field,New Zealand.Their predictions substantially outperform,in terms of prediction performance,those generated using a multiple linear regression(MLR)model.The geophysical logs used as input variables are sonic,temperature and density logs,and some direct pore pressure measurements were available at the reservoir level to calibrate the predictions.A total of 25,935 data records involving six well-log input variables were evaluated across the four wells.All ML methods achieved credible levels of pore pressure prediction performance.The most accurate models for predicting pore pressure in individual wells on a supervised basis are decision tree(DT),adaboost(ADA),random forest(RF)and transparent open box(TOB).The DT achieved root mean square error(RMSE)ranging from 0.25 psi to 14.71 psi for the four wells.The trained models were less accurate when deployed on a semi-supervised basis to predict pore pressure in the other wellbores.For two wells(Mangahewa-03 and Mangahewa-06),semi-supervised prediction achieved acceptable prediction performance of RMSE of 130—140 psi;while for the other wells,semi-supervised prediction performance was reduced to RMSE>300 psi.The results suggest that these models can be used to predict pore pressure in nearby locations,i.e.similar geology at corresponding depths within a field,but they become less reliable as the step-out distance increases and geological conditions change significantly.In comparison to other approaches to predict pore pressures,this study has identified that application of several ML algorithms involving a large number of data records can lead to more accurate prediction results.

Brittleness index predictions from Lower Barnett Shale well-log data applying an optimized data matching algorithm at various sampling densities

The capability of accurately predicting mineralogical brittleness index (BI) from basic suites of well logs is desirable as it provides a useful indicator of the fracability of tight formations.Measuring mineralogical components in rocks is expensive and time consuming.However,the basic well log curves are not well correlated with BI so correlation-based,machine-learning methods are not able to derive highly accurate BI predictions using such data.A correlation-free,optimized data-matching algorithm is configured to predict BI on a supervised basis from well log and core data available from two published wells in the Lower Barnett Shale Formation (Texas).This transparent open box (TOB) algorithm matches data records by calculating the sum of squared errors between their variables and selecting the best matches as those with the minimum squared errors.It then applies optimizers to adjust weights applied to individual variable errors to minimize the root mean square error (RMSE)between calculated and predicted (BI).The prediction accuracy achieved by TOB using just five well logs (Gr,ρb,Ns,Rs,Dt) to predict BI is dependent on the density of data records sampled.At a sampling density of about one sample per 0.5 ft BI is predicted with RMSE~0.056 and R^(2)~0.790.At a sampling density of about one sample per0.1 ft BI is predicted with RMSE~0.008 and R^(2)~0.995.Adding a stratigraphic height index as an additional (sixth)input variable method improves BI prediction accuracy to RMSE~0.003 and R^(2)~0.999 for the two wells with only 1 record in 10,000 yielding a BI prediction error of>±0.1.The model has the potential to be applied in an unsupervised basis to predict BI from basic well log data in surrounding wells lacking mineralogical measurements but with similar lithofacies and burial histories.The method could also be extended to predict elastic rock properties in and seismic attributes from wells and seismic data to improve the precision of brittleness index and fracability mapping spatially.

Mathematical model for iron corrosion that eliminates chemical potential parameters

Iron corrosion in acidic media is a natural phenomenon that converts elemental iron to a more chemically-stable form,i.e.its oxide and hydroxide.In this study,the iron corrosion process is modeled as a completely implicit problem,solved by a novel finite difference model to provide insight into the ionic aspects of corrosion behavior.This new mathematical model eliminates the chemical potential parameters from the corrosion process equations,thereby reducing the need for experimental determination of chemical potentials.The eliminatedchemical-potential-parameters model predicts and quantifies key parameters(concentrations of conjugate base ion,iron(Ⅱ)ion,hydrogen ion,anodic and cathodic potentials,and the electrical current density)associated with the iron corrosion process in acidic solutions.The rigorous derivation and novel application of the eliminated-chemical-potential-parameters model and its results provide new insights into the iron corrosion process.The present model is also applicable in any industrial process which is associated with metal corrosion.The model helps to guide the design of future corrosion resistant systems,and various experimental studies pertaining to corrosion inhibition techniques.

Wellbore stability analysis to determine the safe mud weight window for sandstone layers

The wellbore stability of a vertical well through the sandstone reservoir layers of the Asmari oil-bearing formation in south-west Iran is investigated.The safe drilling-fluid density range for maintaining wellbore stability is determined and simulated using FLAC3 D software and a finite volume model established with drilled strata geomechanical features.The initiation of plastic condition is used to determine the safe mud weight window(SMWW)in specific sandstone layers.The effects of rock strength parameters,major stresses around the wellbore and pore pressure on the SMWW are investigated for this wellbore.Sensitivity analysis reveals that a reduction in cohesion and internal friction angle values leads to a significant narrowing of the SMWW.On the other hand,the reduction of pore pressure and the ratio between maximum and minimum horizontal stresses causes the SMWW to widen significantly.The ability to readily quantify changes in SMWW indicates that the developed model is suitable as a well planning and monitoring tool.

A critical review of self-diverting acid treatments applied to carbonate oil and gas reservoirs

Carbonate reservoirs generally achieved relatively low primary resource recovery rates.It is therefore often necessary to clean those reservoirs up and/or stimulate them post drilling and later in their production life.A common and basic carbonate reservoir cleanup technique to remove contaminating material from the wellbore is acidizing.The efficiency of acid treatments is determined by many factors,including:the type and quantity of the acid used;the number of repeated treatments performed,heterogeneity of the reservoir,water cut of the reservoir fluids,and presence of idle zones and interlayers.Post-treatment production performance of such reservoirs frequently does not meet design expectations.There is therefore much scope to improve acidizing technologies and treatment designs to make them more reliable and effective.This review considers acid treatment technologies applied to carbonate reservoirs at the laboratory scale and in field-scale applications.The range of acid treatment techniques commonly applied are compared.Differences between specific acid treatments,such as foamed acids,acid emulsions,gelled and thickened acid systems,targeted acid treatments,and acid hydraulic fracturing are described in terms of the positive and negative influences they have on carbonate oil production rates and recovery.Opportunities to improve acid treatment techniques are identified,particularly those involving the deployment of nanoparticles(NPs).Due consideration is also given to the potential environmental impacts associated with carbonate reservoir acid treatment.Recommendations are made regarding the future research required to overcome the remaining challenges pertaining to acid treatment applications.

Machine Learning and Regression Analysis Reveal Different Patterns of Influence on Net Ecosystem Exchange at Two Conifer Woodland Sites

Variations in net ecosystem exchange(NEE)of carbon dioxide,and the variables influencing it,at woodland sites over multiple years determine the long term performance of those sites as carbon sinks.In this study,weekly-averaged data from two AmeriFlux sites in North America of evergreen woodland,in different climatic zones and with distinct tree and understory species,are evaluated using four multi-linear regression(MLR)and seven machine learning(ML)models.The site data extend over multiple years and conform to the FLUXNET2015 pre-processing pipeline.Twenty influencing variables are considered for site CA-LP1 and sixteen for site US-Mpj.Rigorous k-fold cross validation analysis verifies that all eleven models assessed generate reproducible NEE predictions to varying degrees of accuracy.At both sites,the best performing ML models(support vector regression(SVR),extreme gradient boosting(XGB)and multi-layer perceptron(MLP))substantially outperform the MLR models in terms of their NEE prediction performance.The ML models also generate predicted versus measured NEE distributions that approximate cross-plot trends passing through the origin,confirming that they more realistically capture the actual NEE trend.MLR and ML models assign some level of importance to all influential variables measured but their degree of influence varies between the two sites.For the best performing SVR models,at site CA-LP1,variables air temperature,shortwave radiation outgoing,net radiation,longwave radiation outgoing,shortwave radiation incoming and vapor pressure deficit have the most influence on NEE predictions.At site US-Mpj,variables vapor pressure deficit,shortwave radiation incoming,longwave radiation incoming,air temperature,photosynthetic photon flux density incoming,shortwave radiation outgoing and precipitation exert the most influence on the model solutions.Sensible heat exerts very low influence at both sites.The methodology applied successfully determines the relative importance of influential variables in determining weekly NEE trends at both conifer woodland sites studied.

Medium-term Air Quality Benchmarking for Ecosystem Monitoring and Sustainability Planning: Case Study Dallas County (U.S.A.) 2015 to 2020

Medium-term air quality assessment,benchmarking it to recent past data can usefully complement short-term air quality index data for monitoring purposes.By using daily and monthly averaged data,medium-term air quality benchmarking provides a distinctive perspective with which to monitor air quality for sustainability planning and ecosystem perspectives.By normalizing the data for individual air pollutants to a standard scale they can be more easily integrated to generate a daily combined local area benchmark(CLAB).The objectives of the study are to demonstrate that medium-term air quality benchmarking can be tailored to reflect local conditions by selecting the most relevant pollutants to incorporate in the CLAB indicator.Such a benchmark can provide an overall air quality assessment for areas of interest.A case study is presented for Dallas County(U.S.A.)applying the proposed method by benchmarking 2020 data for air pollutants to their trends established for 2015 to 2019.Six air pollutants considered are:ozone,carbon monoxide,nitrogen dioxide,sulfur dioxide,benzene and particulate matter less than 2.5 micrometres.These pollutants are assessed individually and in terms of CLAB,and their 2020 variations for Dallas County compared to daily trends established for years 2015 to 2019.Reductions in benzene and carbon monoxide during much of 2020 are clearly discernible compared to preceding years.The CLAB indicator shows clear seasonal trends for air quality for 2015 to 2019 with high pollution in winter and spring compared to other seasons that is strongly influenced by climatic variations with some anthropogenic inputs.Conducting CLAB analysis on an ongoing basis,using a relevant near-past time interval for benchmarking that covers several years,can reveal useful monthly,seasonal and annual trends in overall air quality.This type of medium-term,benchmarked air quality data analysis is well suited for ecosystem monitoring.

Machine Learning and Pattern Analysis Identify Distinctive Influences from Long-term Weekly Net Ecosystem Exchange at Four Deciduous Woodland Locations

A methodology integrating correlation,regression(MLR),machine learning(ML),and pattern analysis of long-term weekly net ecosystem exchange(NEE)datasets are applied to four deciduous broadleaf forest(DBF)sites forming part of the AmeriFlux(FLUXNET2015)database.Such analysis effectively characterizes and distinguishes those DBF sites for which long-term NEE patterns can be accurately predicted using the recorded environmental variables,from those sites cannot be so delineated.Comparisons of twelve NEE prediction models(5 MLR;7 ML),using multi-fold cross-validation analysis,reveal that support vector regression generates the most accurate and reliable predictions for each site considered,based on fits involving between 16 and 24 available environmental variables.SVR can accurately predict NEE for datasets for DBF sites US-MMS and US-MOz,but fail to reliably do so for sites CA-Cbo and MX-Tes.For the latter two sites the predicted versus recorded NEE weekly data follow a Y≠X pattern and are characterized by rapid fluctuations between low and high NEE values across leaf-on seasonal periods.Variable influences on NEE,determined by their importance to MLR and ML model solutions,identify distinctive sets of the most and least influential variables for each site studied.Such information is valuable for monitoring and modelling the likely impacts of changing climate on the ability of these sites to serve as long-term carbon sinks.The periodically oscillating NEE weekly patterns distinguished for sites CA-Cbo and MX-Tes are not readily explained in terms of the currently recorded environmental variables.More detailed analysis of the biological processes at work in the forest understory and soil at these sites are recommended to determine additional suitable variables to measure that might better explain such fluctuations.

Optimized feature selection assists lithofacies machine learning with sparse well log data combined with calculated attributes in a gradational fluvial sequence

Machine learning(ML)to predict lithofacies from sparse suites of well-log data is difficult in laterally and vertically heterogeneous reservoir formations in oil and gas fields.Meandering,braided fluviatile depositional environments tend to form clastic sequences with laterally discontinuous layers due to the continuous shifting of relatively narrow sandstone channels.Three cored wellbores drilled through such a reservoir in a large oil field,with just four recorded well logs available,are used to classify four lithofacies using ML models.To augment the well-log data,six derivative and volatility attributes were calculated from the recorded gamma ray and density logs,providing sixteen log features for the ML models to select from.A novel,multiple-optimizer feature selection technique was developed to identify high-performing feature combinations with which seven ML models were used to predict lithofacies assisted by multi-k-fold cross validation.Feature combinations with just seven to nine selected log features achieved overall ML lithofacies accuracy of 0.87 for two wells used for training and validation.When the trained ML models were applied to a third well for testing,lithofacies ML prediction accuracy declined to 0.65 for the best performing extreme gradient boosting model with seven features.However,an accuracy of~0.76 was achieved by that model in predicting the presence of the pay bearing sandstone and siltstone lithofacies in the test well.A model using only the four recorded well logs was only able to predict the pay-bearing lithofacies with~0.6 accuracy.Annotated confusion matrices and feature importance analysis provide additional insight to ML model performance and identify the log attributes that are most influential in enhancing lithofacies prediction.

Re-Establishing the Merits of Thermal Maturity and Petroleum Generation Multi-Dimensional Modeling with an Arrhenius Equation Using a Single Activation Energy

Thermal maturation and petroleum generation modeling of shales is essential for suc- cessful exploration and exploitation of conventional and unconventional oil and gas plays. For basin- wide unconventional resource plays such modeling, when well calibrated with direct maturity meas- urements from wells, can characterize and locate production sweet spots for oil, wet gas and dry gas. The transformation of kerogen to petroleum is associated with many chemical reactions, but models typically focus on first-order reactions with rates determined by the Arrhenius Equation. A miscon- ception has been perpetuated for many years that accurate thermal maturity modeling of vitrinite re- flectance using the Arrhenius Equation and a single activation energy, to derive a time-temperature index （~TTIARa）, as proposed by Wood （1988）, is flawed. This claim was initially made by Sweeney and Burnham （1990） in promoting their ＂EasyRo＂ method, and repeated by others. This paper dem- onstrates through detailed multi-dimensional burial and thermal modeling and direct comparison of the ~TTIARR and ＂EasyRo＂ methods that this is not the case. The ~TTIA^R method not only provides a very useful and sensitive maturity index, it can reproduce the calculated vitrinite reflectance values derived from models based on multiple activation energies （e.g., ＂EasyRo＂）. Through simple expres- sions the ~TTIAaa method can also provide oil and gas transformation factors that can be flexibly scaled and calibrated to match the oil, wet gas and dry gas generation windows. This is achieved in a more-computationally-efficient, flexible and transparent way by the ~TTIARR method than the ＂EasyRo＂ method. Analysis indicates that the ＂EasyRo＂ method, using twenty activation energies and a constant frequency factor, generates reaction rates and transformation factors that do not realisti- cally model observed kerogen behaviour and transformation factors over geologic time scales.

Kerogen Kinetic Distributions and Simulations Provide Insights into Petroleum Transformation Fraction (TF) Profiles of Organic-Rich Shales

Two hundred and fifty single first-order Arrhenius reactions are simulated to generate S2 pyrograms at three heating rates 25,15,and 5°C·min-1.The activation energy(E)and pre-exponential factor(A)of the reactions simulated follow a long-established trend of those variable values displayed by shales and kerogens.The characteristics of the transformation fraction(TF)profiles(product generation window temperatures)of the simulated single reactions are compared to the TF profiles of recorded shale pyrograms generated by multiple reactions with different E-A values lying near the defined E-A trend.Important similarities and differences are observed between the TF profile values of the two datasets.The similarities support the spread of E-A values involved in shale pyrogram best fits.The differences are most likely explained by the complexity of the multiple kerogen first-order and second-order reactions contributing to the recorded shale pyrograms versus the simplicity and crispness of the single first-order reactions simulated.The results also justify the validity of using the previously described“variable E-A pyrogram-fitting method”of multi-heating-rate shale pyrograms enabling optimizers to choose multiple reactions from an unlimited range of E-A values.In contrast,further doubt is cast on the validity of the constant-A pyrogram-fitting method used by the Easy%Ro technique,in that a distribution of reactions with a single A value is unlikely to represent the complex variety of kerogen macerals observed in shale formations.TF profiles generated by the variable E-A pyrogram-fitting method lie close to the established E-A trend and are likely to provide more realistic TF generation window temperatures than TF profiles generated by the constant-A pyrogram-fitting method.

Weeks-Ahead Epidemiological Predictions of Varicella Cases From Univariate Time Series Data Applying Artificial Intelligence

Background:"Chickenpox"is a highly infectious disease caused by the varicella-zoster virus,influenced by seasonal and spatial factors.Dealing with varicella-zoster epidemics can be a substantial drain on health-authority resources.Methods that improve the ability to locally predict case numbers from time-series data sets every week are therefore worth developing.Methods:Simple-to-extract trend attributes from published univariate weekly case-number univariate data sets were used to generate multivariate data for Hungary covering 10 years.That attribute-enhanced data set was assessed by machine learning(ML)and deep learning(DL)models to generate weekly case forecasts from next week(t0)to 12 weeks forward(t+12).The ML and DL predictions were compared with those generated by multilinear regression and univariate prediction methods.Results:Support vector regression generates the best predictions for weeks t0 and t+1,whereas extreme gradient boosting generates the best predictions for weeks t+3 to t+12.Long-short-term memory only provides comparable prediction accuracy to the ML models for week t+12.Multi-K-fold cross validation reveals that overall the lowest prediction uncertainty is associated with the tree-ensemble ML models.Conclusion:The novel trend-attribute method offers the potential to reduce prediction errors and improve transparency for chickenpox timeseries.

Characterization of Organic-Rich Shales for Petroleum Exploration & Exploitation: A Review-Part 2: Geochemistry, Thermal Maturity, Isotopes and Biomarkers

As shale exploitation is still in its infancy outside North America much research effort is being channelled into various aspects of geochemical characterization of shales to identify the most prospective basins, formations and map their petroleum generation capabilities across local, regional and basin-wide scales. The measurement of total organic carbon, distinguishing and categorizing the kerogen types in terms oil-prone versus gas-prone, and using vitrinite reflectance and Rock-Eval data to estimate thermal maturity are standard practice in the industry and applied to samples from most wellbores drilled. It is the trends of stable isotopes ratios, particularly those of carbon, the wetness ra- tio （C1/~＇（C2＋C3））, and certain chemical biomarkers that have proved to be most informative about the status of shales as a petroleum system. These data make it possible to identify production ＂sweet- spots＂, discriminate oil-, gas-liquid- and gas-prone shales from kerogen compositions and thermal ma- turities. Rollovers and reversals of ethane and propane carbon isotope ratios are particularly indica- tive of high thermal maturity exposure of an organic-rich shale. Comparisons of hopane, strerane and terpane biomarkers with vitrinite reflectance （Ro） measurements of thermal maturity highlight dis- crepancies suggesting that Ro is not always a reliable indicator of thermal maturity. Major and trace element inorganic geochemistry data and ratios provides useful information regarding provenance, paleoenvironments, and stratigraphic-layer discrimination. This review considers the data measure- ment, analysis and interpretation of techniques associated with kerogen typing, thermal maturity, sta- ble and non-stable isotopic ratios for rocks and gases derived from them, production sweet-spot identi- fication, geochemical biomarkers and inorganic chemical indicators. It also highlights uncertainties and discrepancies observed in their practical application, and the numerous outstanding questions as- sociated with them.

Characterization of Organic-Rich Shales for Petroleum Exploration & Exploitation: A Review-Part 1: Bulk Properties, Multi-Scale Geometry and Gas Adsorption

Shales, the most abundant of sedimentary rocks, are valued as the source-rocks and seals to porous petroleum reservoirs. Over the past-twenty years, organic-rich shales have also emerged as valuable petroleum systems （reservoir, seal, and source rocks contained in the same for- mation）. As such they have become primary targets for petroleum exploration and exploitation. This Part 1 of a three-part review addresses the bulk properties, multi-scale geometry and gas adsorption characteristics of these diverse and complex rocks. Shales display extremely low permeability, and their porosity is also low, but multi-scale. Characterizing the geometry and interconnectivity of the pore-structure frameworks with the natural-fracture networks within shales is essential for establish- ing their petroleum exploitation potential. Organic-rich shales typically contain two distinct types of porosity： matrix porosity and fracture porosity. In addition to inter-granular porosity, the matrix po- rosity includes two types of mineral-hosted porosity： inorganic-mineral-hosted porosity （1P）; and, organic-matter-hosted （within the kerogen） porosity （OP）. Whereas, the fracture porosity and per- meability is crucial for petroleum production from shales, it is within the OP where, typically, much of the in-situ oil and gas resources resides, and from where it needs to be mobilized. OP increases signifi- cantly as shales become more thermally mature （i.e., within the gas generation zones）, and plays a key role in the ultimate recovery from shale-gas systems. Shales＇ methane sorption capacities （MSC） tends to be positively correlated with their total organic carbon content （TOC）, thermal maturation, and mi- cropore volume. Clay minerals also significantly influence key physical properties of shale related to fluid flow （permeability） and response to stress （fracability） that determine their prospectivity for pe- troleum exploitation. Clay minerals can also adsorb gas, some much better than others. The surface area of the pore structure of shales can be positively or negatively correlated with TOC content, de- pending upon mineralogy and thermal maturity, and can influence its gas adsorption capacity. Part 2 of this three-part review considers, in a separate article, the geochemistry and thermal maturity cha- racteristics of shale; whereas Part 3, addresses the geomechanical attributes of shales, including their complex wettability, adsorption, water imbibition and ＂fracability＂ characteristics. The objectives of this Part 1 of the review is to identify important distinguishing characteristics related to the bulk properties of the most-prospective, petroleum-rich shales.

Characterization of Organic-Rich Shales for Petroleum Exploration & Exploitation: A Review-Part 3: Applied Geomechanics, Petrophysics and Reservoir Modeling

Modeling geomechanical properties of shales to make sense of their complex properties is at the forefront of petroleum exploration and exploitation application and has received much re- search attention in recent years. A shale＇s key geomechanical properties help to identify its ＂fracibility＂ its fluid flow patterns and rates, and its in-place petroleum resources and potential commercial re- serves. The models and the information they provide, in turn, enable engineers to design drilling pat- terns, fracture-stimulation programs and materials selection that will avoid formation damage and op- timize recovery of petroleum. A wide-range of tools, technologies, experiments and mathematical techniques are deployed to achieve this. Characterizing the interconnected fracture, permeability and porosity network is an essential step in understanding a shales highly-anisotropic features on multiple scales （nano to macro）. Weli-log data, and its petrophysical interpretation to calibrate many geome- chanical metrics to those measured in rock samples by laboratory techniques plays a key role in pro- viding affordable tools that can be deployed cost-effectively in multiple well bores. Likewise, micro- seismic data helps to match fracture density and propagation observed on a reservoir scale with pre- dictions from simulations and laboratory tests conducted on idealised/simplified discrete fracture net- work models. Shales complex wettability, adsorption and water imbibition characteristics have a sig- nificant influence on potential formation damage during stimulation and the short-term and long-term flow of petroleum achievable. Many gas flow mechanisms and models are proposed taking into ac- count the multiple flow mechanisms involved （e.g., desorption, diffusion, slippage and viscous flow op- erating at multiple porosity levels from nano- to macro-scales）. Fitting historical production data and well decline curves to model predictions helps to verify whether model＇s geomechanical assumptions are realistic or not. This review discusses the techniques applied and the models developed that are relevant to applied geomechanics, highlighting examples of their application and the numerous out- standin~ questions associated with them.