pure and applied geophysics最新文献

The article investigates the reliability of moment tensor (MT) inversion in time domain with use of first P-wave amplitude, a method used to determine the source mechanisms of earthquakes, across four different seismic networks. The study compares the synthetic tests results of MT inversion for two underground mining and two artificial reservoir monitoring seismic networks. The analysis was performed to assesses how consistency and accuracy of the results depend on different factors like: network configuration, events depth, velocity model, focal mechanism of event and applied noise. The findings highlight the impact of network configuration compared to other variables and data quality on the reliability of moment tensor inversion and provide insights into different factors which have to be considered to enhance MT accuracy. The significance of events depth in P-wave amplitude MT inversion and the necessity to consider velocity model influence, especially presence of high velocity gradient, is highlighted by the presented results.

Changes in the integral characteristics of the global water exchange, at climatic time scales, are considered as random functions (processes). “Trajectories” obtained as the results of numerical calculations on various, from 34 to 43, climate models (participating at the CMIP-6 “historical” experiment covering the period from 1850 through 2014) are taken as realizations of these processes. Temporal variations of following annually averaged parameters are studied: (1) average evaporation from the ocean surface, (2) precipitation over the ocean, (3) “effective evaporation” from the ocean (difference “evaporation minus precipitation”, on average equal to the water transport from the ocean to land), (4) precipitation over land, (5) evaporation (evapotranspiration) from the land surface, (6) “effective precipitation” over land (or “climatic runoff”: precipitation minus evaporation), and (7) river runoff. It is shown that precipitation over the ocean and evaporation from land largely suppress the monotonous trends in the mean values of evaporation from the ocean and precipitation over land, respectively, at secular time scales. At the same time, this damping does not extend to the trends of the last few decades, which may be due to a combination of a sharp increase in global temperature with explosive volcanic eruptions that preceded this period. An analysis of the time divergence in the model trajectories of each of the components of the global water exchange, as well as the very existence of such divergences, indicates an increase in the uncertainty of processes that is not associated with anthropogenic impact on the climate system.

Seismic inversion has been in use for the last two decades to measure inverted impedances using an integrated data set approach. This research focuses on the application of multi-attribute seismic inversion and the geostatistical probabilistic neural network (PNN) approach for determining rock properties and litho-fluid classification in the Mehar-Mazarani Field of the Lower Indus Basin (LIB), Pakistan. The study compares five different inversion techniques, including model-based inversion (MBI), colored inversion (CI), linear sparse spike inversion (LSSI), band-limited inversion (BLI), and maximum likelihood sparse spike inversion (MLSSI). The inverted outputs, such as acoustic P-impedance (Zp), density (ρ), porosity (φ), and shale volume (Vsh), were analyzed in Paleocene and Cretaceous geological complex reservoirs to identify gas-bearing zones. The results indicated the existence of gas between 1630 and 1700 ms (ms) and corresponding depth ranges from approximately 3200 m up to 4200 m with varying thickness. Amongst the inversion techniques, MBI demonstrated greater accuracy, with inverted density volumes showing a strong correlation coefficient of 0.98 and the lowest root mean square error (RMSE) and relative error of 0.10 m/s * g/cc. A geostatistical PNN approach was employed to estimate variations in Vsh and φ within the sand reservoir. MBI again yielded more reliable results, with a strong correlation between the measured and inverted attributes. High φ and low Vsh were observed in predetermined low-impedance zones. Overall, MBI is proven to be the most accurate and reliable technique, providing clear identification of the gas occurrence. This research highlights the effectiveness of seismic inversion, particularly the application of MBI, in determining rock properties and identifying gas-bearing zones within the Mehar-Mazarani gas field.

Surface air temperature (SAT) affects both natural systems and human activities, impacting health, agriculture, energy demand, and so on. To investigate and analyze SAT over the region of interest, it is crucial to choose suitable climate models. The study commenced with the evaluation of 42 Coupled Model Intercomparison Project phase 6 (CMIP6) models’ simulations of SAT over India for annual and all four seasons (summer, southwest monsoon, northeast monsoon, and winter) during the historical period 1985 to 2014 with respect to the gridded SAT datasets obtained from the India Meteorological Department (IMD). Multi Model Mean (MMM) of 42 models was included in the evaluation. The evaluation was performed with various statistical metrics such as root mean squared error (RMSE), mean bias error (MBE), correlation coefficient (R), mean squared error (MAE), Taylor skill score (TSS), Brier skill score (BSS), and Interannual variability skill score (IVSS). By the method of estimating Comprehensive Rating Index (CRI), the top-ranking models were identified to be CMCC-CM2-SR5 for the annual and summer season, MIROC6 for the winter season, ACCESS-ESM-1-5 for the southwest monsoon, and NorESM2-LM for northeast monsoon. The novelty of this study lies in the approach of identifying the best ensemble. For each season, statistical metric-wise top-ranked models were picked to develop the best ensemble. Again, the overall ranking of the models along with the best ensemble for each season is determined by estimating CRI. It was observed that for all seasons, the best ensemble falls within the top 3 models’ category. Future projections of SAT under four shared socio-economic pathways (SSP-2.6, 4.5, 7.0, and 8.5) were also analyzed with the best ensemble obtained for each season. The results convey that, the country will witness, especially during the summer season, there will be a 1.160 °C, 1.288 °C and 2.368 °C rise in the mean SAT between historical (1985–2014) and near future (2021–2040), near and mid future (2041–2060), mid and far future (2081–2100) if the pathway, SSP5-8.5 is followed.

In the E&P industry, accurate lithology classification is an essential task for successful exploration and production. Geophysical logs provide high-resolution petrophysical properties, but core logging is expensive and traditional techniques may not accurately classify lithologies. We demonstrated a comparative analysis of six ML algorithms: k-nearest neighbor (kNN), support vector machine (SVM), decision tree (DT), random forest (RF), extreme gradient boosting (XGBoost) and artificial neural network (ANN) for the prediction of lithologies from geophysical logs. Here we analysed the wireline logs of eight wells associated with the petroliferous Lakadong-Therria formation of the Bhogpara oil field of the Assam-Arakan Basin. This formation contains eight typical lithologies, namely clay stone, sand stone, calcareous sandstone, shale, calcareous shale, carbonaceous shale, coal and limestone. Performance of the ML algorithms were evaluated through accuracy, precision, recall, F1-score and receiver operating characteristic (ROC) curve. During the training and test phases, the computed overall accuracy of the predicted ML modes exceeded 82% and 71%, respectively. The model accuracy hierarchy was ANN > XGBoost > RF > SVM > DT > kNN during training, and ANN/XGBoost > kNN > DT/RF > SVM during testing. This approach allows interpreters to select the most accurate ML model based on training phase performance. This study provided a clear insight towards generating a supplement for litholog sequence and improving the accuracy and efficiency of lithology prediction in a geologically complex petroleum reservoir using pre-received core derived litholog information at few wells.