Machine Learning Emulation of Urban Land Surface Processes

TL;DR

This paper introduces a neural network model trained on multiple urban land surface models to emulate their outputs, resulting in improved accuracy, reduced computational cost, and enhanced stability when integrated with weather forecasting models.

Contribution

The paper presents a novel urban neural network trained on multiple models, outperforming traditional models in accuracy and efficiency, and demonstrates its integration with weather forecasting systems.

Findings

UNN accurately emulates mean ULSM outputs

UNN outperforms reference ULSM in flux prediction accuracy

WRF-UNN is more stable and accurate than WRF-TEB

Abstract

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