Browsing UA Faculty Research by Journal
Now showing items 1-2 of 2
On distinguishing snowfall from rainfall using near-surface atmospheric information: Comparative analysis, uncertainties and hydrologic importanceThe accurate estimation of precipitation phase has broad applications. In this study, we compared the skill of using various atmospheric variables and their combinations as predictors in accurately identifying surface precipitation phase, determined uncertainties associated with commonly used fixed temperature thresholds, and explored the sensitivity of hydrologic model output to uncertainty in precipitation phase using two case-studies. The results suggest that among all single predictors, wet-bulb temperature yields the highest skill score for determining precipitation phase and can reduce uncertainties due to regional differences, especially compared to the commonly used near-surface air temperature. However, addition of good-quality near-surface wind speed measurement to dew-point temperature and pressure showed slightly higher skill than wet-bulb temperature. We showed that the scale mismatch between temperature from stations and gridded products can cause large uncertainties in determining precipitation phase, especially in regions with rugged topography. Such uncertainties need to be considered when the relationships developed based on station data are applied to remote-sensing observations and model-generated data to separate rain from snowfall. The sensitivity of hydrologic model outputs to uncertainty in precipitation phase delineation was also assessed over two major basins in California by modifying default near-surface temperatures used in the Variable Infiltration Capacity (VIC) model. It was found that regional and scaling uncertainties in determining temperature thresholds can largely influence the accuracy of simulated downstream runoff and snow water equivalent (SWE) (e.g. up to 40% change in SWE for 2 degrees C shift in temperature threshold). Therefore, to reduce simulation uncertainties, it is important to improve rain-snow partitioning methods, consider regional variabilities in determining temperature thresholds, and perform the analysis at the highest possible resolutions to mitigate scale-related uncertainties.
Variational particle smoothers and their localizationGiven the success of 4D-variational methods (4D-Var) in numerical weather prediction, and recent efforts to merge ensemble Kalman filters with 4D-Var, we revisit how one can use importance sampling and particle filtering ideas within a 4D-Var framework. This leads us to variational particle smoothers (varPS) and we study how weight-localization can prevent the collapse of varPS in high-dimensional problems. We also discuss the relevance of (localized) weights in near-Gaussian problems. We test our ideas on the Lorenz'96 model of dimensions n = 40, n = 400, and n = 2, 000. In our numerical experiments the localized varPS does not collapse and yields results comparable to ensemble formulations of 4D-Var, while tuned EnKFs and the local particle filter lead to larger estimation errors. Additional numerical experiments suggest that using localized weights may not yield significant advantages over unweighted or linearized solutions in near-Gaussian problems.