Search for: All records

Computational fluid dynamics (CFD) and its uncertainty quantification are computationally expensive. We use Gaussian Process (GP) methods to demonstrate that machine learning can build efficient and accurate surrogate models to replace CFD simulations with significantly reduced computational cost without compromising the physical accuracy. We also demonstrate that both epistemic uncertainty (machine learning model uncertainty) and aleatory uncertainty (randomness in the inputs of CFD) can be accommodated when the machine learning model is used to reveal fluid dynamics. The demonstration is performed by applying simulation of Hagen-Poiseuille and Womersley flows that involve spatial and spatial-tempo responses, respectively. Training points are generated by using the analytical solutions with evenly discretized spatial or spatial-temporal variables. Then GP surrogate models are built using supervised machine learning regression. The error of the GP model is quantified by the estimated epistemic uncertainty. The results are compared with those from GPU-accelerated volumetric lattice Boltzmann simulations. The results indicate that surrogate models can produce accurate fluid dynamics (without CFD simulations) with quantified uncertainty when both epistemic and aleatory uncertainties exist. 

The tropics exert enormous influence on global climate. Despite the importance of tropical regions, the terrestrial temperature history in the Indo‐Pacific Warm Pool (IPWP) region during the last deglaciation is poorly constrained. Although numerous sea surface temperature (SST) reconstructions provide estimates of SST warming from the Last Glacial Maximum to the Holocene, the timing of the onset of deglacial warming varies between records and inhibits determining the forcings driving deglacial warming in the IPWP. We present a 60,000‐year long temperature reconstruction based on branched glycerol dialkyl glycerol tetraethers (brGDGTs) in a sediment core from Lake Towuti, located in Sulawesi, Indonesia. BrGDGTs are bacterial membrane‐spanning lipids that, globally, become more methylated with decreasing temperature and more cyclized with decreasing pH. Although changes in temperature are the dominant control on brGDGTs in regional and global calibrations, we find that the cyclization of the brGDGTs is a major mode of variation at Lake Towuti that records important changes in the lacustrine biogeochemical environment. We separate the influence of lake chemistry changes from temperature changes on the brGDGT records, and develop a temperature record spanning the last 60,000 years. The timing of the deglacial warming in our record occurs after the onset of the deglacial increase in CO₂concentrations, which suggests rising greenhouse gas concentrations and the associated radiative forcing may have forced deglacial warming in the IPWP. Peaks in temperature around 55 and 34 ka indicate that Northern Hemisphere summer insolation may also influence land surface temperature in the IPWP region.