SEARCH · Geology Reports
Results for “Reliability Engineering and System Safety (RESS)”
Search indexed USGS publications on groundwater, aquifers, geologic maps, mineral resources and earthquakes. Explore source records by subject and place.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Don’t Let Negatives Hold You Back: Accounting for Underlying Physics and Natural Distributions of Hydrothermal Systems When Selecting Negative Training Sites Leads to Better Machine Learning Predictions
Selecting negative training sites is an important challenge to resolve when utilizing machine learning (ML) for predicting hydrothermal resource favorability because ideal models would discriminate between hydrothermal systems (positives) and all types of locations without hydrothermal systems (negatives). The Nevada Machine Learning project (NVML) fit an artificial neural network to identify areas favorable for hydrothermal systems by selecting 62 negative sites where the research team had confidence that no hydrothermal resource exists. Herein, we compare the implications of the expert selection of negatives (i.e., the NVML strategy) with a random sample strategy, where it is assumed that areas outside the favorable structural ellipses defined by NVML are negative. Because hydrothermal systems are sparse, it is highly probable that, in the absence of a favorable geological structure, hydrothermal favorability is low. We compare three training strategies: 1) the positive and negative labeled examples from NVML; 2) the positive examples from NVML with randomly selected negatives in equal frequency as NVML; and 3) the positive examples from NVML with randomly selected negatives reflecting the expected natural distribution of hydrothermal systems relative to the total area. We apply these training strategies to the NVML feature data (input data) using two ML algorithms (XGBoost and logistic regression) to create six favorability maps for hydrothermal resources. When accounting for the expected natural distribution of hydrothermal systems, we find that XGBoost performs better than the NVML neural network and its negatives. Model validation was less reliable using F1 scores, a common performance metric, than comparing probability estimates at known positives, likely because of the extreme natural class imbalance and the lack of negatively labeled sites. This work demonstrates that expert selection of negatives for training in NVML likely imparted modeling bias. Accounting for the sparsity of hydrothermal systems and all the types of locations without hydrothermal systems allows us to create better models for predicting hydrothermal resource favorability.
Effective site coefficients for the 2024 International Building Code (IBC)
The U.S. National Seismic Hazard Models (NSHMs), developed by the U.S. Geological Survey (USGS), have long been the scientific foundation of seismic design guidelines and have been used to compute design ground motions for construction of new buildings and retrofit of existing buildings in the United States and its territories. The 2018 NSHM is adopted by the 2024 International Building Code (IBC). Prior to the 2018 NSHM update, hazard calculations were required only at one reference site condition defined by V S30 =760 m/s (where V S30 is the time-averaged shear wave velocity from the surface to a depth of 30 m) and three periods (peak ground acceleration, PGA, and pseudo spectral accelerations at periods of 0.2 and 1 s, S s and S 1 ). Site coefficients, F PGA , F a , and F v , were then defined by the Building Seismic Safety Council (BSSC) Provisions Update Committee (PUC) in the site-specific procedures of National Earthquake Hazard Reduction Program (NEHRP) Recommended Seismic Provisions to calculate ground motions for other site classes with different V S30 values at the given periods. The design ground motions at other periods were then estimated using a generic spectral shape that was also defined by the BSSC PUC in NEHRP provisions. In recent years, the engineering community has realized there were deficiencies with the F a and F v site coefficients and the generic spectral shape. To avoid potentially dangerous underestimations of design ground motions for long period structures on soft site conditions in high seismicity regions, the BSSC PUC recommended the use of multi-period response spectra (MPRS) in 2017. As a result, the USGS produced multi-period and multi-V S30 response spectral values in the 2018 NSHM for calculations of design ground motions and the site coefficients F a and F v were eliminated from the 2020 NEHRP Provisions. As these site coefficients were widely used inside and outside of the United States, in this study we back-calculate the “effective” site coefficients F a,eff , and F v,eff by comparing MPRS for various site classes with the MPRS for the reference site condition, and discuss the changes that are observed in the 2024 IBC compared to its previous version in 2021. The effective site coefficients are presented for test site locations and their dependence on various factors including period, ground motion intensity, and regional models are discussed. Ratio maps between the new effective site coefficients and the old ones are then presented for soft site classes and for short and long periods. For soft site classes at short periods, the new effective site coefficients are lower than the old site coefficients for high seismicity regions and higher for low seismicity regions. As it was expected, for soft site classes at long periods and high seismicity regions, the new effective site coefficients are much larger than the old site coefficients without imposing the 50% increase as a penalty that was suggested in the 2021 IBC, whereas they could be much smaller if the 50% increase would have been imposed particularly around New Madrid and Charleston high seismicity regions. For low seismicity regions, the long period effective site coefficients can be smaller or larger by 20% compared to the 2021 IBC coefficients.
Bayesian belief network model to predict human-wildlife conflict in protected areas
Human-wildlife conflict (HWC) poses a pervasive global challenge, affecting livelihoods and threatening biodiversity. To better anticipate and mitigate HWC risk, we developed a large-scale predictive model using a Bayesian Belief Network (BBN). We surveyed 1,011 park rangers across 135 terrestrial protected areas in three Andean countries, documenting recent HWC incidents involving wildlife persecution or killing, livestock depredation, crop damage, or threats to human safety and property. We identified key drivers of HWC risk, including governance, wildlife acceptance, participation, and habitat quality. A sensitivity analysis revealed that enhancing governance and improving wildlife acceptance could reduce HWC risk by > 85%. The BBN model demonstrated scalability, effectively identifying strategies to reduce HWC risk at multiple scales, from individual protected areas to national networks. Our findings highlight the importance of strengthening governance, increasing wildlife acceptance, and enhancing community participation in conservation efforts. BBNs provide a flexible, cost-effective, and data-driven tool to guide protected areas and wildlife managers in monitoring, anticipating, and making informed decisions to mitigate conflict and promote coexistence.
A synthesis engine for constructing geologic maps of the United States
The geologic history of the United States is cataloged in thousands of geologic maps produced during many decades. However, the disparate nature of these individual maps makes it challenging to assess resources, research geologic histories, or characterize natural hazards holistically across the Nation. The U.S. House of Representatives 2020 appropriations bill for the U.S. Department of the Interior (H.R. 116-100) requires the U.S. Geological Survey to “bring together detailed national and continental-resolution [two-dimensional] and [three-dimensional] information produced throughout the Survey and by [F]ederal and [S]tate partners.” In response to this directive, this report presents a compilation and synthesis of geologic maps across the United States in the form of a relational database. The synthesis database includes thematic maps that synthesize the Nation’s geology, and retains the original input maps as well as linkages to standardized vocabularies to aid the discoverability of geologic information. Specifically, the synthesis database is targeted toward producing four National-resolution maps for the conterminous United States: Quaternary geology, the geology at the Earth’s surface, pre-Quaternary geology, and Precambrian geology. In addition, the synthesis database includes the infrastructure necessary to expand to additional resolutions in the future.