A long-standing obstacle in the use of machine-learnt surrogates with larger PDE systems is the onset of instabilities when solved numerically. Efforts towards ameliorating these have mostly concentrated on improving the accuracy of the surrogates or imbuing them with additional structure, and have garnered limited success. In this article, we study a prototype problem and draw insights that may help with more complex systems. In particular, we focus on a viscous Burgers’-ML system and, after identifying the cause of the instabilities, propose strategies to stabilize it. To improve the accuracy of the stabilized system, we next explore methods based on the Mori–Zwanzig formalism. We show that the memory-based corrections from this approach help considerably in yielding accurate results. Finally, we draw analogies with more complex systems and how these strategies may generalize to those settings.
In October 2023 an underground chemical explosive test, Physics Experiment 1-A, was executed in a tunnel at the Nevada National Security Site. This test was part of a series of multi-physics experiments focused on enhancing nuclear explosion monitoring techniques. Part of the experiment included monitoring ga movement. As part of the facility operations, there were two events during the first five days of the experiment where the ventilation system was operated to purge gases from the facility. These events afforded an opportunity to evaluate the mass of gases emitted to the atmosphere during these ventilation events and compare to the total mass produced during the experiment. This provides a basis for estimating the fraction of gas produced in an underground test that might be expected to be released through natural and engineered pathways. Using several approaches, an estimate of between 1 and 9% of the total gas produced by the experiment was captured and discharged to the atmosphere during these events. The percentage of gas captured was nominally equal to the percentage of the plume volume that was bisected by the facility tunnels and represents a similar fraction to that assumed as potential releases for underground nuclear tests.
Acoustic Doppler Current Profilers were deployed in four tidal creeks within forested swamps and herbaceous marshes in the lower Columbia River. Measurements were collected over 12 to 20-day periods during high and low river stage, and then analyzed using wavelet tidal analysis software (CWT_Multi) to characterize temporal variability of water elevation, velocity, and discharge. While subject to similar fluvial and tidal forcing, nuances in creek geometry led to large differences in their hydraulic response. Most obviously, creeks in the higher elevation forested swamps were hydraulically isolated, which limited subtidal discharge but increased tidal monthly variability of discharge in comparison to the lower elevation marsh sites. Discharge in the marshes was ebb-dominant (3:1 ebb: flood ratio), while discharge at the forested sites was slightly flood dominant during neap tides and ebb dominant during spring tides. Net discharge over a 12-day period exceeded 1,000,000 m3 for the marsh sites but was less than 200,000 m3 for the forested sites. These transport patterns set limits on sediment dynamics, fish ingress, and material fluxes from these habitat types. Comparison to the literature on tidal creeks highlights the distinctive nature of tidal freshwater wetlands in mixed-tide environments of the lower Columbia River, which are more ebb-dominant (due to diurnal-semidiurnal tidal phase relationships), fluvially influenced, and temporally variable (due both to fluvial input and non-stationary tidal asymmetry) than saltmarsh counterparts in semidiurnal environments. Finally, we demonstrate how wavelet analysis of water levels can be used to extend velocity and discharge records, thereby facilitating long-term monitoring of wetlands.
The atmospheric response to the solar eclipse of 8 April 2024 in North America is investigated with a specific focus on the marine atmospheric boundary layer (MABL). We leverage measurements collected during the Third Wind Forecast Improvement Project (WFIP3), including Doppler lidars, sonic anemometers, and thermodynamic profiler data to investigate the atmospheric response across sites that experienced partial eclipse conditions with nearly 90 1.2^∘C to 1.4^∘C in coastal regions and from 0.3^∘C to 0.5^∘C over the ocean. This analysis suggests that the MABL’s higher thermal inertia compared to coastal regions moderates the temperature decrease during the eclipse. Wind speed exhibits a more complex behavior, as it is influenced by both the MABL and preexisting synoptic conditions. Although a reduction in wind speed is observable up to approximately 140 m above ground level (AGL) at more inland sites, at other locations closer to the coast, this reduction is constrained to the lowest 100 m AGL. Turbulence parameters retrieved from sonic anemometers, such as turbulence kinetic energy, turbulent heat flux, and friction velocity, decrease during the eclipse at coastal sites, accompanied by a brief transition of atmospheric stability from unstable to neutral or weakly stable conditions. For the open-ocean sites, the variability in turbulence statistics and atmospheric stability is minimal during the occurrence of the eclipse.
Increased aerosol concentrations can brighten low-level clouds and extend their lifetimes, but aerosol-cloud interactions (ACI) remain highly uncertain and difficult to quantify. We show that part of this uncertainty is caused by topographical influences on clouds, that is, those arising from land-water contrasts. This is demonstrated using satellite retrievals in regions with extensive river networks, such as the Amazon Basin. 15 years of MODerate resolution Imaging Spectroradiometer (MODIS) satellite data show cloud formation over the Amazon River basin is suppressed by 26% with warm low clouds above the river exhibiting a 22% smaller droplet effective radius and 18% higher droplet concentration () compared to adjacent land clouds. Thus, clouds above the river may appear polluted but are actually influenced by river-breeze circulations driven by the thermal contrast between the river and the surrounding land. These responses are robust in both wet and dry seasons, and tests using an improved MODIS retrieval product show cloud differences are unlikely due to retrieval artifacts. In situ measurements from the Green Ocean Amazon Experiment (GoAmazon) confirm that is elevated above rivers and are also higher when carbon monoxide concentrations are elevated near the large city of Manaus. Lagrangian airmass tracking over Manaus shows that regional-scale river-breeze circulations impact as much as the urban aerosol plume, complicating ACI attribution and highlighting the need to isolate land-surface effects to assess ACI in continental regions.
