Search for: All records

Subseasonal climate forecasting is the task of predicting climate variables, such as temperature and precipitation, in a two-week to two-month time horizon. The primary predictors for such prediction problem are spatio-temporal satellite and ground measurements of a variety of climate variables in the atmosphere, ocean, and land, which however have rather limited predictive signal at the subseasonal time horizon. We propose a carefully constructed spatial hierarchical Bayesian regression model that makes use of the inherent spatial structure of the subseasonal climate prediction task. We use our Bayesian model to then derive decision-theoretically optimal point estimates with respect to various performance measures of interest to climate science. As we show, our approach handily improves on various off-the-shelf ML baselines. Since our method is based on a Bayesian frame- work, we are also able to quantify the uncertainty in our predictions, which is particularly crucial for difficult tasks such as the subseasonal prediction, where we expect any model to have considerable uncertainty at different test locations under differ- ent scenarios. 

In light of the 2018 special report on climate change compiled by the United Nations, there is a renewed urgency to the rapid adoption of renewable energy technologies. A key roadblock to the large-scale/commercial conversion of tidal energy is the question concerning the operational efficiency of existing technologies in the non-homogeneous, turbulent and corrosive marine environment. A thorough understanding of the aforementioned aspects of full-scale deployment is vital in developing robust and cost-effective turbine designs and farm layouts. The current experimental work at Lehigh University aims to better the understanding of turbine performance and near-wake statistics in homogeneous and non-homogeneous turbulent flows, similar to actual marine conditions. A 1:20 laboratory scale tidal turbine model with a rotor diameter of 0.28m is used in the experiments and an active grid type turbulence generator, designed in-house, is employed to generate both homogeneous and non-homogeneous turbulent inflow conditions. To the knowledge of the authors, this is the first experimental study to explore the effects of non-homogeneous inflow turbulence on tidal turbines. From the data collected it was observed that the non-homogeneous inflow condition led to a considerable drop (15-20%) in the measured thrust coefficient. They also resulted in larger torque and thrust fluctuations on the rotor (~40% under the tested conditions). The effect of inflow non-homogeneity was evident in the asymmetric near-wake characteristics as well. Turbulence intensity and Reynolds stresses measured in the wake of the rotor were found to adapt quicker to inflow non-homogeneity than the wake velocity deficit and integral length scales.