Machine Learning Emulation of Urban Land Surface Processes

Can we improve the modeling of urban land surface processes with machine learning (ML)? A prior comparison models (ULSMs) found that no single model is 'best' at predicting all common fluxes. Here, develop an neural network (UNN) trained on mean predicted fluxes from 22 ULSMs one site. The UNN emulates output accurately. When compared to a reference ULSM (Town Energy Balance; TEB), has greater accuracy relative flux observations, less computational cost, and requires fewer input parameters. coupled Weather Research Forecasting (WRF) using TensorFlow bindings, WRF-UNN stable more accurate than WRF-TEB. Although application currently constrained by training data (1 site), show novel approach combining strengths several into ML.