Deep learning-based air temperature mapping by fusing remote sensing, station, simulation and socioeconomic data

TL;DR

This study employs a deep belief network to accurately map daily maximum air temperature across China by integrating remote sensing, station, simulation, and socioeconomic data, outperforming traditional methods.

Contribution

It introduces a multi-source data fusion approach using deep learning for detailed and accurate air temperature mapping, a novel application in this context.

Findings

DBN achieves RMSE of 1.996°C and R of 0.986 at national scale.

DBN outperforms MLR, BPNN, and RF in MAE reduction.

Spatial and temporal analysis confirms DBN's effectiveness in Ta estimation.

Abstract

Air temperature (Ta) is an essential climatological component that controls and influences various earth surface processes. In this study, we make the first attempt to employ deep learning for Ta mapping mainly based on space remote sensing and ground station observations. Considering that Ta varies greatly in space and time and is sensitive to many factors, assimilation data and socioeconomic data are also included for a multi-source data fusion based estimation. Specifically, a 5-layers structured deep belief network (DBN) is employed to better capture the complicated and non-linear relationships between Ta and different predictor variables. Layer-wise pre-training process for essential features extraction and fine-tuning process for weight parameters optimization ensure the robust prediction of Ta spatio-temporal distribution. The DBN model was implemented for 0.01{\deg} daily maximum Ta mapping across China. The ten-fold cross-validation results indicate that the DBN model achieves promising results with the RMSE of 1.996{\deg}C, MAE of 1.539{\deg}C, and R of 0.986 at the national scale. Compared with multiple linear regression (MLR), back-propagation neural network (BPNN) and random forest (RF) method, the DBN model reduces the MAE values by 1.340{\deg}C, 0.387{\deg}C and 0.222{\deg}C, respectively. Further analysis on spatial distribution and temporal tendency of prediction errors both validate the great potentials of DBN in Ta estimation.