AAAI 2020 Spring Symposium on Combining Artificial Intelligence and Machine Learning with Physical Sciences
LSTMs for Inferring Planetary Boundary Layer Height (PBLH)
March 22, 2021
In this paper, we describe new work which is part of a larger study to understand how machine learning could be used to augment existing methods for calculating and estimating the Planetary Boundary Layer Height (PBLH). We describe how a Long Short-Term Memory (LSTM) Network could be used to learn PBLH changes over time for different geographical locations across the United States, used in conjunction with the WRF-Chem model. If the machine learning method could achieve accuracy levels similar to the model-based calculations, then it is feasible for the deep learning model to be used as an embedded method for the WRF-Chem model. The paper shows promising results that warrant more exploration. We describe results for two experiments in particular. The first experiment used 20 geographical locations for a two-month period of hourly WRF-Chem calculated PBLH. In this experiment, we evaluated how well the LSTM could learn PBLH by using limited data across a set of nearby locations. This model achieved RMSE of .11 on predicted PBLH. The second experiment used one year of hourly PBLH calculations from the WRF-Chem model to evaluate the LSTM prediction for a selection of three locations with separate LSTM models, achieving RMSE scores of 0.04, 0.01 and 0.05, respectively. We describe these results and the future plans for this work.
Downloads: 54 downloads