Publications

Shi, CY; Wang, TX; Wang, SY; Jia, AL; Zheng, XP; Leng, WC; Du, YH (2024). MDINEOF: A scheme to recover land surface temperatures under cloudy-sky conditions by incorporating radiation fluxes. REMOTE SENSING OF ENVIRONMENT, 309, 114208.

Abstract
Land Surface Temperature (LST) is widely used as a crucial parameter to monitor the energy exchange and cycle processes over the globe. Constructing seamless LST datasets with high accuracy is crucial to study its influence on climate. However, existing remote sensing methods for estimating LST are inevitably limited by cloud contaminations, leading to spatiotemporal discontinuity of the derived LST products. Thus, many reconstruction methods have been developed to recover the missing LST values. Unfortunately, there is still a great room for improvement in accuracy or practicality of the existing reconstruction studies. Moreover, few interpolationbased methods effectively take radiation fluxes into account, even though numerous studies have shown that LST is closely related to radiation variations. To this end, a new scheme named Multivariate Data Interpolating Empirical Orthogonal Functions (MDINEOF) is proposed as a simple yet effective method to accurately reconstruct the cloudy-sky LST by incorporating appropriate radiation components. The proposed method achieves satisfactory performance in terms of both simulated and remotely sensed data, with the root mean square error (RMSE) reduced from 6.86 K to 3.28 K compared to the original DINEOF algorithm in Moderate Resolution Imaging Spectroradiometer (MODIS) data. This study proposes a practical solution to derive all-sky LST dataset over the globe and provides a perspective in the reconstruction fields of LST or other Essential climate variables (ECVs).

DOI:
10.1016/j.rse.2024.114208

ISSN:
1879-0704