Production Forecasts Research Articles

Skill assessment of seasonal forecasts of ocean variables

There is growing demand for seasonal forecast products for marine applications. The availability of consistent and sufficiently long observational records of ocean variables permits the assessment of the spatial distribution of the skill of ocean variables from seasonal forecasts. Here we use state-of-the-art temporal records of sea surface temperature (SST), sea surface height (SSH) and upper 300m ocean heat content (OHC) to quantify the distribution of skill, up to 2 seasons ahead, of two operational seasonal forecasting systems contributing to the seasonal multi-model of the Copernicus Climate Change Services (C3S). This study presents the spatial distribution of the skill of the seasonal forecast ensemble mean in terms of anomaly correlation and root mean square error and compares it to the persistence and climatological benchmarks. The comparative assessment of the skill among variables sheds light on sources/limits of predictability at seasonal time scales, as well as the nature of model errors. Beyond these standard verification metrics, we also evaluate the ability of the models to represent the observed long-term trends. Results show that long-term trends contribute to the skill of seasonal forecasts. Although the forecasts capture the long-term trends in general, some regional aspects remain challenging. Part of these errors can be attributed to specific aspects of the ocean initialization, but others, such as the overestimation of the warming in the Eastern Pacific are also influenced by model error. Skill gains can be obtained by improving the trend representation in future forecasting systems. In the meantime, a forecast calibration procedure that corrects the linear trends can produce substantial skill gains. The results show that calibrated seasonal forecasts beat both the climatological and persistence benchmark almost at every location for all initial dates and lead times. Results demonstrate the value of the seasonal forecasts for marine applications and highlight the importance of representing the decadal variability and trends in ocean heat content and sea level.

Role of IT on the Effectiveness of Logistics and Supply Chain Management Systems in the Pharmaceutical Sector in Saudi Arabia Vision 2030

The main objective of this study is to evaluate how IT might improve the efficacy and efficiency of supply chain management and logistics systems in the pharmaceutical sector because the pharmaceutical industry is experiencing a transformation by the lofty objectives outlined by Saudi Arabia Vision 2030. The study was conducted using 150 executives working in the Saudi Arabian pharmaceutical industry's logistics and supply chain division. An open-ended survey questions were used to determine how IT technology affects pharmaceutical logistics and supply chain using non-probability sampling, logistics managers' LinkedIn accounts were selected for the sample, irrespective of the organization's size, location, or gender. SPSS statistical tools were used to analyse the data. It was found that the integration of IT tools and systems has emerged as a crucial enabler in this process. By harnessing the power of data analytics, IoT (Internet of Things), and automation, pharmaceutical companies can gain real-time insights into their supply chains. This, in turn, allows them to develop agile responses to changing market dynamics and customer expectations. Such strategic decisions, driven by IT, enable businesses to adapt quickly, reorganize their operations, and offer value-added services tailored to the specific needs of the market. Moreover, the creative application of IT solutions facilitates efficient inventory management, demand forecasting, and tracking of pharmaceutical products from manufacturing facilities to end consumers results also show that companies may simply increase the pace of the supply chain with the aid of innovative chances in major choices and judgments.

Forecasting Dairy Sales Estimates: A Comparative Analysis of Logistic Regression and Random Forest Algorithm

This study examines the effectiveness of logistic regression and random forest algorithms in forecasting sales estimates for dairy products. Using a comprehensive data set that includes factors such as product features, pricing dynamics, promotional efforts and consumer demographics, the models are trained to accurately predict future sales figures. Logistic regression provides a transparent framework for estimating probabilities, while random forest uses ensemble learning to capture complex relationships between variables. Through careful evaluation and comparison, the research aims to identify the strengths and weaknesses of each algorithm in producing reliable sales forecasts for dairy products. The results show that while logistic regression offers interpretability and simplicity, random forest excels in handling non-linear relationships and achieving higher prediction accuracy. Insights gathered from this analysis can help dairy industry stakeholders make informed decisions, optimize resource allocation and improve sales forecasting strategies.

A Method Based on Artificial Intelligence That Predicts Arabica Coffee Yield by Analyzing Abiotic Factors and the Prevalence of Coffee Leaf Rust

Coffee is a perennial crop that harbors infections throughout the plant, which may worsen illness under favorable circumstances. Coffee leaf rust, a widespread coffee disease and Arabica is susceptible to leaf rust. Disease incidence and severity depend on abiotic variables. Cloudy and constant South-West monsoon weather (June–September) promotes coffee leaf rust growth. In this five-model study, an Extra Tree and Gradient Boosting regression model predicted coffee crop output in Chikamagaluru, Karnataka, with the least error utilizing biotic and abiotic factors. We investigated additional tree, gradient boosting, RF, Decision Tree, and KNN models using biotic and abiotic predictors. Used the independent testing dataset's MSE, MAE, RMSE Root mean square errors, and R-squared errors to compare model performance. The extra tree (R²=0.98 kg/ha ˉ¹ and RMSE = 7.96 kg/ha ˉ¹) and gradient boosting (R²=0.96 kg/ha ˉ¹ and RMSE = 10.96 kg/ha ˉ¹) regression models used Group 1 and 2 characteristics as predictor variables and different parameter fine tuning functions to estimate coffee yield most accurately. Compared to the less precise probabilistic models utilized in this work, such as Random Forest, decision tree, and KNN models, shown in the results section. The optimum weather parameter for coffee production forecasts and biotic-CLR incidence data outperformed random forest, decision tree, and K-Nearest Neighbor models.

Rolling forecast of snowmelt floods in data-scarce mountainous regions using weather forecast products to drive distributed energy balance hydrological model

Snowmelt flood forecasting plays a crucial role in water resource management and utilization in arid regions. However, the accuracy of current snowmelt flood forecasts falls short of requirements due to complex mountainous terrain and limited observational data. In this study, We conducted a snowmelt flood simulation and forecasting experiment in a mountainous watershed in northern Xinjiang. First, we used site observations, ERA5-land reanalysis data and WorldClim data to produce historical meteorological forcing, calibrated the VIC-CAS model and improved routing model using historical observed runoff data, and conducted a detailed simulation of the hourly snowmelt flood process in the basin. Then, the latest grid forecast weather product developed by the China Meteorological Administration was integrated into the system to conduct a rolling forecast test of the snowmelt floods in the next 72 h during the snowmelt period in 2023. The remotely sensed snow cover and the freezing level measured by the sounding balloon were assimilated into the model in real time to correct the deviation of runoff simulation. The results show that during the model calibration period, the KGE of the simulated monthly runoff in the Kayertes River basin (KRB) from 2006 to 2011 was 0.83 and the NSE was 0.86. In the continuous hourly runoff simulation during the 2022 snowmelt period, the model well captured the intraday fluctuations of snowmelt runoff in the KRB, with the daily peak flow error within ± 20 % and the peak occurrence time controlled within ± 4 h. In the rolling forecast test in 2023, the daily peak flow error is within ± 20 % (the peak occurrence error during the peak period is even controlled within ± 5 %), and the peak occurrence time is within ± 4 h. Real-time assimilation of freezing level measured by sounding balloons and remotely sensed snow cover can effectively correct the deviations in simulated temperature and snow cover. This study can provide insights for runoff prediction in other high-altitude mountainous areas where data are scarce.

Post‐processing output from ensembles with and without parametrised convection, to create accurate, blended, high‐fidelity rainfall forecasts

AbstractFlash flooding is a significant societal problem, but related precipitation forecasts are often poor. To address, one can try to use output from convection‐parametrising (global) ensembles, post‐processed to forecast at point‐scale, or convection‐resolving limited area ensembles. The new methodology described here combines both. We apply “ecPoint‐rainfall” post‐processing to the ECMWF global ensemble. Alongside we use 2.2 km COSMO LAM ensemble output (centred on Italy), and also post‐process that, using a scale‐selective neighbourhood approach to compensate for insufficient members and to preserve consistently forecast local details. The two resulting scale‐compatible components then undergo lead‐time‐weighted blending, to create the final probabilistic 6 h rainfall forecasts. Product creation for forecasters, in this way, constituted the “Italy Flash Flood use case” within the EU‐funded MISTRAL project; real‐time delivery of open access products is ongoing. One year of verification shows that, of the five components (2 raw, 2 post‐processed and blended), ecPoint is the most skilful. The post‐processed COSMO ensemble adds most value to summer convective events in the evening, when the global model has an underprediction bias. In two typical heavy rainfall case studies we observed underestimation of the largest point totals in the raw ECMWF ensemble, and overestimation in the raw COSMO ensemble. However, ecPoint elevated the ECMWF maxima and highlighted best the most affected areas and merged products seemed to be the most skilful of all. Even though our LAM post‐processing does not include (or arguably need) bias‐correction, this study still provides a unique blueprint for successfully combining ensemble rainfall forecasts from global and LAM systems around the world. It also has important implications for forecast products as global ensembles move ever closer to having convection‐permitting resolution.

Beetaloo’s gas bounty: potential for LNG exports and domestic supply

Recent drilling in the Beetaloo Basin has yielded promising results for the emerging shale gas play, enabling the explorers to secure funding and establish a range of non-binding commercial agreements. Notably, Beetaloo wells have exhibited initial production rates over 30 days (IP30), spanning from 1.1 to 5.2 mmcf/day, and there is an evident trend of increasing frac stages with each new well as operators continuously optimise stimulation. This research paper presents comprehensive Beetaloo modelling results, encompassing well type curves, a comparative analysis with the renowned Marcellus shale play in the United States (frequently used as a reference point for Beetaloo), production forecasts, economic assessments, and an examination of potential commercialisation pathways. Based on our modelling findings, Beetaloo’s estimated ultimate recovery (EUR) per well ranges from 7 to 14 bcf, with deeper wells into the Velkerri B Shale demonstrating relatively higher productivity. We anticipate that, with full-scale development, Beetaloo could attain peak production rates of 1 bcf/day, supported by a peak annual drilling rate averaging 50–60 wells. Overall, the Beetaloo Basin possesses the potential to recover close to 8 tcf of low reservoir CO2 gas, opening doors to the possibility of liquefied natural gas (LNG) exports reaching up to 4.5 mtpa and domestic gas supplies of around 94 mmcf/day, limited to the existing infrastructure capacity connecting to the east coast gas markets.

Time series forecasting of wheat crop productivity in Egypt using deep learning techniques

Time series forecasting of wheat crop productivity in Egypt using deep learning techniques

Sun from Behind the Clouds: the Appeals Board of the European Centre for Medium-Range Weather Forecasts

Abstract One of the lesser-known regional international organizations in Europe is the European Centre for Medium-Range Weather Forecasts (ECMWF). It has a mandate to undertake tasks such as the production of weather forecasts, conduct scientific research to improve meteorological projections, maintain extensive datasets, and to better coordinate European meteorological infrastructure. The Centre has three sites across three European states where it has over 400 staff, and given it is immune from suit, its staff members only have an internal justice mechanism to resort to in the event of employment/pensions disputes. Little is known about the ECMWF Appeals Board given its untransparent nature. As uncovered in this article, it has undergone partial reform, and simultaneously, there has been an uptick in the number of cases. This article analyses these developments and highlights that more must be done to align the ECMWF Appeals Boards with the features of other international administrative tribunals.

The impact of exploration/appraisal activities on the recovery per well in North Perth Basin

The North Perth Basin (NPB) in Western Australia has experienced significant petroleum industry activity over the past 3 years (2021–2023), marked by the drilling of exploration and appraisal wells by various operators, along with notable merger and acquisition transactions. This flurry of drilling activity, coupled with the acquisition of new seismic data, have enhanced our understanding of the variability within the primary petroleum play levels – namely, the Permian ‘Kingia’ and High Cliff reservoirs. This understanding is crucial for assessing the future potential of domestic gas supply in Western Australia. This paper aims to elucidate how recent drilling efforts have narrowed the basin-wide portfolio of available gas prospects relative to the situation in early 2021, and how the landscape of undiscovered gas resources has evolved. This shift is attributed to both unsuccessful exploration attempts and revised estimates of well deliverability. Furthermore, an updated long-term production forecast for the NPB – considering existing gas pools and the newly assessed undiscovered resources – emphasises the urgent need to re-evaluate the narrative concerning gas availability for the domestic market. The Permian gas plays within the NPB have proven to be more intricate and unpredictable than previously believed. Nonetheless, their significance to the Western Australian economy is anticipated to be substantial for many years ahead.

Life of field emission forecast development and reduction option screening for a multi-train LNG hub

The Karratha Gas Plant (KGP) is a complex integrated facility which produces liquified natural gas (LNG), domestic gas (Domgas), condensate, propane and butane products. The multi-train facility has been operating for 40 years with several phases of expansion resulting in a range of gas turbine technology and energy efficiency levels across the facility. The challenge for the next phase of KGP operation is to manage declining North West Shelf (NWS) feed gas levels, and the addition of Other Resource Owners feed gas while reducing emissions and maintaining energy efficiency. A flexible emissions forecasting tool has been developed which allows unit and whole of facility emissions reduction and energy efficiency options to be assessed. The tool builds on current quarterly timestep production forecasts and models ambient temperature, unit turndown and turn-off strategies, and potential implementation of lower emissions technology. The ability to generate and compare life of field emissions profiles for extrapolated operating modes and proposed modifications allows unit strategies and phasing of emissions reduction modifications to be investigated and screened before further economic assessment and creation of marginal abatement cost curves (MACC).

Production Forecasting at Natural Gas Wells

In Romania, natural gas production is concentrated in two large producers, OMV Petrom and Romgaz. However, there are also smaller companies in the natural gas production area. In these companies, the deposits are mostly mature, or new deposits have low production capacity. Thus, the production forecast is very important for the continued existence of these companies. The model is based on the pressure variation in the gas reservoir, and the exponential model with production decline is currently used by gas and oil producers. Following the variation in the production of the gas wells, we found that in many cases, the Gaussian and Hubbert forecast models are more suitable for simulating the production pattern of gas wells. The models used to belong to the category of poorly conditioned models, with little data, usually called gray models. Papers published in this category are based on data collected over a period of time and provide a forecast of the model for the next period. The mathematical method can lead to a very good approximation of the known data, as well as short-term forecasting in the continuation of the time interval, for which we have these data. The neural network method requires more data for the network learning stage. Increasing the number of known variables is conducive to a successful model. Often, we do not have this data, or obtaining it is expensive and uneconomical for short periods of possible exploitation. The network model sometimes captures a fairly local pattern and changing conditions require the model to be remade. The model is not valid for a large category of gas wells. The Hubbert and Gauss models used in the article have a more comprehensive character, including a wide category of gas wells whose behavior as evolutionary stages is similar. The model is adapted according to practical observations by reducing the production growth period; the layout is asymmetric around the production peak; and the production range is reduced. Thus, an attempt is made to replace the exponential model with the Hubbert and Gauss models, which were found to be in good agreement with the production values. These models were completed using the Monte Carlo method and matrix of risk evaluation. A better appreciation of monthly production, which is an important aspect of supply contracts, and cumulative production, which is important for evaluating the utility of the investment, is ensured. In addition, we can determine the risk associated with the realization of production at a certain moment of exploitation, generating a complete picture of the forecast over the entire operating interval. A comparison with production results on a case study confirms the benefits of the forecasting procedure used.

Financial markets value skillful forecasts of seasonal climate

Scientific agencies spend substantial sums producing and improving forecasts of seasonal climate, but they do so without much information about these forecasts’ value in practice. Here we show that financial market participants value the production of seasonal forecasts: options traders price the uncertainty generated by upcoming United States National Oceanic and Atmospheric Administration Winter and El Niño Outlooks. Each outlook affects firms throughout the economy, with total market capitalization of $6 and $13 trillion, respectively. A 1% improvement in the skill of the El Niño Outlook reduces firms’ exposure to a one standard deviation shock by $18 billion and induces traders to spend an additional $2 million hedging the outlook’s news. Firms must not be able to undertake ex-ante adaptation that would eliminate their exposure to the forecasted portion of seasonal climate without imposing substantial costs of its own.

Integration of deep-learning fault segmentation, horizontal transverse isotropy analysis, and ambient microseismic methods to enhance fracture prediction in the Crisol Anticline, Colombia

The analysis of fault and fracture systems in the subsurface is crucial for identifying potential reservoirs and optimizing production. However, conventional fault interpretation methods have limitations due to the complexity and amount of data involved. In this study, a multimethod approach was used to analyze fault and fracture systems in the Crisol area of the Valley Middle Magdalena basin. We used deep-learning fault segmentation, horizontal transverse isotropy (HTI), and ambient microseismic methods to identify and characterize fault and fracture systems in the study area. The deep-learning fault segmentation method helped to identify complex fault systems, whereas the HTI method identified areas and orientations of intense fracturing that occurred at any prior period of stress in geologic time. Meanwhile, the ambient microseismic method measured active rock movements and fracturing in the subsurface that occurred during the microseismic trace recording. The application of multiple methods for analyzing fault and fracture systems in the subsurface is demonstrated to offer a more comprehensive understanding of the geology and enhance the accuracy of well placement and production forecasts.

Northeast Atlantic fish stock productivity hindcasts and forecasts from a Bayesian framework reveal pronounced climate‐induced dynamics

AbstractClimate change continues to exert pressure on ocean ecosystems. The fisheries‐related responses, such as altered body growth, recruitment and spatial distribution of the targeted stock(s), have generally been reasonably well investigated. Nevertheless, there are still important knowledge gaps in how biophysical drivers impact stock productivity and thereby sustainable harvest levels. Considering this, we investigated 26 fish stocks in the Northeast Atlantic, a region characterized by accelerated climate change effects and a diverse set of fisheries. A novel, stepwise, Bayesian framework to quantify stock productivity was established to identify shared trends and project future patterns, aiming at determining essential baselines for adaptive fishery management in the face of climate change. Despite variation among large marine ecosystems and stocks, an overall declining trend in productivity over the past four decades was observed, especially in high‐latitude areas. These hindcast results were mainly attributed to higher temperatures posing negative effects on productivity, which was dependent on the stock's thermal preference. Contrastingly, the proxy for food availability – gross secondary production – exhibited less consistent impacts. In the forecast, the applied Shared Socio‐economic Pathways (SSPs) indicated that most stocks are likely to encounter adverse effects, with the worst cases expected to occur under SSP2‐4.5 in the 2050s, and under SSP5‐8.5 in the 2090s. Thus, this study generally not only supports earlier climate vulnerability assessments (‘scorings’) of experts but also provides evidence for revised directional effects under climate change, underlining the complexity of processes affecting stock productivity.

Operational Forecasting of Global Ionospheric TEC Maps 1-, 2-, and 3-Day in Advance by ConvLSTM Model

In this paper, we propose a global ionospheric total electron content (TEC) maps (GIM) prediction model based on deep learning methods that is both straightforward and practical, meeting the requirements of various applications. The proposed model utilizes an encoder-decoder structure with a Convolution Long Short-Term Memory (ConvLSTM) network and has a spatial resolution of 5° longitude and 2.5° latitude, with a time resolution of 1 h. We utilized the Center for Orbit Determination in Europe (CODE) GIM dataset for 18 years from 2002 to 2019, without requiring any other external input parameters, to train the ConvLSTM models for forecasting GIM 1, 2, and 3 days in advance. Using the CODE GIM data from 1 January 2020 to 31 December 2023 as the test dataset, the performance evaluation results show that the average root mean square errors (RMSE) for 1, 2 and 3 days of forecasts are 2.81 TECU, 3.16 TECU, and 3.41 TECU, respectively. These results show improved performance compared to the IRI-Plas model and CODE’s 1-day forecast product c1pg, and comparable to CODE’s 2-day forecast c2pg. The model’s predictions get worse as the intensity of the storm increases, and the prediction error of the model increases with the lead time.

A novel framework for predicting non-stationary production time series of shale gas based on BiLSTM-RF-MPA deep fusion model

Shale gas, as an environmentally friendly fossil energy resource, has gained significant commercial development and shows immense potential. However, accurately predicting shale gas production faces substantial challenges due to the complex law of decline, nonlinear and non-stationary features in production data, which greatly repair the robustness of current models in predicting shale gas production time series. To address these challenges and improve accuracy in production forecasting, this paper introduces a novel and innovative approach: a hybrid proxy model that combines the bi-directional long short-term memory (BiLSTM) neural network and random forest (RF) through deep learning. The BiLSTM neural network is adept at capturing long-term dependencies, making it suitable for understanding the intricate relationships between input and output variables in shale gas production. On the other hand, RF serves a dual purpose: reducing model variance and addressing the concept drift problem that arises in non-stationary time series predictions made by BiLSTM. By integrating these two models, the hybrid approach effectively captures the inherent dependencies present in long and nonstationary production time series, thereby reducing model uncertainty. Furthermore, the combination of BiLSTM and RF is optimized using the recently-proposed marine predators algorithm (MPA) to fine-tune hyperparameters and enhance the overall performance of the proxy model. The results demonstrate that the proposed BiLSTM-RF-MPA model achieves higher prediction accuracy and demonstrates stronger generalization capabilities by effectively handling the complex nonlinear and non-stationary characteristics of shale gas production time series. Compared to other models such as LSTM, BiLSTM, and RF, the proposed model exhibits superior fitting and prediction performance, with an average improvement in performance indicators exceeding 20%. This innovative framework provides valuable insights for forecasting the complex production performance of unconventional oil and gas reservoirs, which sheds light on the development of data-driven proxy models in the field of subsurface energy utilization.

New methodology for estimating the influence of Cu content of Ferrous Scrap materials in the Total Cost of Operation (TCO) of scrap mix in Electric Arc Furnace steelmaking

Scrap is the main raw material of the electric arc furnace (EAF) route of steelmaking and an important one in the primary route. The decarbonisation commitment acquired by the most relevant steelmaking companies, will only increase the amount of scrap used by both routes, making it a critical component to succeed in the worldwide challenge of reducing CO2 emissions. While being a very relevant material, scrap characterisation is complex due to its heterogeneity, but furthermore, its quality is worsening as its demand increases. This includes the sterile content, such as copper. Higher sterile content scrap comes at a lower price, while scrap with a lower content in undesired elements is most costly. Being able to balance the cost and benefit of purchasing scrap with a higher or lower content of copper with a respective, lower or higher purchase cost would be of interest to optimise the mix and produce the required quality steel at the lowest possible cost. Nowadays, this cost benefit analysis can only be done by performing Total Cost of Operation (TCO) calculations that require a considerable amount of information, including but not limited to the complete chemical composition of scrap, market parameters, production forecast and other relevant data; moreover, involving an arduous optimization process. Based on a parametric study of the TCO calculation, this work proposes a mathematical model to estimate the impact on the TCO of a predefined scrap mix when only modifying the specifics of one given scrap type. The results would facilitate the decision making by comparing the two scenarios as a factor of cost. This mathematical model can be easily programmed and requires only 5 parameters to be run, while providing a low error result.

Manufacturing Sentiment: Forecasting Industrial Production with Text Analysis

This paper examines the link between industrial production and the sentiment expressed in natural language survey responses from U.S. manufacturing firms. We compare several natural language processing (NLP) techniques for classifying sentiment, ranging from dictionary-based methods to modern deep learning methods. Using a manually labeled sample as ground truth, we find that deep learning models partially trained on a human-labeled sample of our data outperform other methods for classifying the sentiment of survey responses. Further, we capitalize on the panel nature of the data to train models which predict firm-level production using lagged firm-level text. This allows us to leverage a large sample of "naturally occurring" labels with no manual input. We then assess the extent to which each sentiment measure, aggregated to monthly time series, can serve as a useful statistical indicator and forecast industrial production. Our results suggest that the text responses provide information beyond the available numerical data from the same survey and improve out-of-sample forecasting; deep learning methods and the use of naturally occurring labels seem especially useful for forecasting. We also explore what drives the predictions made by the deep learning models, and find that a relatively small number of words associated with very positive/negative sentiment account for much of the variation in the aggregatesentiment index.

Interpretable Deep Learning Approach for Production Forecasting of Fractured Horizontal Wells

Interpretable Deep Learning Approach for Production Forecasting of Fractured Horizontal Wells