地球科学最新文献

The physical and biogeochemical properties of the western Arctic Ocean are rapidly changing, resulting in cascading shifts to the local ecosystems. The nutrient-rich Pacific water inflow to the Arctic through the Bering Strait is modified on the Chukchi and East Siberian shelves by brine rejection during sea ice formation, resulting in a strong halocline (called the Upper Halocline Layer (UHL)) that separates the cold and relatively fresh surface layer from the warmer and more saline (and nutrient-poor) Atlantic-derived water below. Biogeochemical signals entrained into the UHL result from Pacific Waters modified by sediment and river influence on the shelf. In this synthesis, we bring together data from the 2015 Arctic U.S. GEOTRACES program to implement a multi-tracer (dissolved and particulate trace elements, radioactive and stable isotopes, macronutrients, and dissolved gas/atmospheric tracers) approach to assess the relative influence of shelf sediments, rivers, and Pacific seawater contribution to the Amerasian Arctic halocline. For each element, we characterized their behavior as mixing dominated (e.g., dCu, dGa), shelf-influenced (e.g., dFe, dZn), or a combination of both (e.g., dBa, dNi). Leveraging this framework, we assessed sources and sinks contributing to elemental distributions: shelf sediments (e.g., dFe, dZn, dCd, dHg), riverine sources, (e.g., dCu, dBa, dissolved organic carbon), and scavenging by particles originating on the shelf (e.g., dFe, dMn, dV, etc.). Additionally, synthesized results from isotopic and atmospheric tracers yielded tracer age estimates for the Upper Halocline ranging between 1 and 2 decades on a spatial gradient consistent with cyclonic circulation.

Gravity anomaly maps often contain spatially overlapping signatures from numerous sources, each with varying shapes, depths, and density contrasts. Locating these signatures using edge detection techniques is crucial for geological structural interpretation and imaging of horizontal boundaries. This paper proposes two effective edge detection tools: one combining the balanced total horizontal gradient (BHG), and the hyperbolic tangent function, abbreviated as “MTBHG”; and the other combining the tilt angle of the total horizontal gradient (TAHG) and the hyperbolic tangent function, abbreviated as “MTAHG.” Additionally, the Modified Non-Local Means (MNLM) filter was applied to suppress possible noise effects amplified by the gradient calculation process. Synthetic tests validated that the MTAHG and MTBHG detectors outperform other representative detectors. Two high-resolution gravity data sets from the Western Carpathians in Slovakia and the Witwatersrand Basin in South Africa were used to test the applicability of the modified methods. Results show that the modified detectors achieve superior edge delineation and avoid creating spurious anomalies or artifacts even in the presence of unwanted noise interference. Furthermore, by eliminating false tilt-depth (TD) solutions via the edge detection results, we enhance the accuracy of depth estimates and facilitate the credible identification of both horizontal and vertical structure distributions.

Since the weather is chaotic, it is necessary to forecast an ensemble of future states. Recently, multiple AI weather models have emerged claiming breakthroughs in deterministic skill. Unfortunately, it is hard to fairly compare ensembles of AI forecasts because variations in ensembling methodology become confounding and the baseline data volume is immense. We address this by scoring lagged initial condition ensembles—whereby an ensemble can be constructed from a library of deterministic hindcasts. This allows the first parameter-free intercomparison of leading AI weather models' probabilistic skill against an operational baseline. Lagged ensembles of the two leading AI weather models, GraphCast and Pangu, perform similarly even though the former outperforms the latter in deterministic scoring. These results are elaborated upon by sensitivity tests showing that commonly used multiple time-step loss functions damage ensemble calibration.