Find in Library

Air temperature at approximately 2 m above the ground (<inline-formula><tex-math notation="LaTeX">$T_{a}$</tex-math></inline-formula>) is one of the most important environmental and biophysical parameters to study various earth surface processes. <inline-formula><tex-math notation="LaTeX">$T_{a}$</tex-math></inline-formula> measured from meteorological stations is inadequate to study its spatio-temporal patterns since the stations are unevenly and sparsely distributed. Satellite-derived land surface temperature (LST) provides global coverage, and is generally utilized to estimate <inline-formula><tex-math notation="LaTeX">$T_{a}$</tex-math></inline-formula> due to the close relationship between LST and <inline-formula><tex-math notation="LaTeX">$T_{a}$</tex-math></inline-formula>. However, LST products are sensitive to cloud contamination, resulting in missing values in LST and leading to the estimated <inline-formula><tex-math notation="LaTeX">$T_{a}$</tex-math></inline-formula> being spatially incomplete. To solve the missing data problem, we propose a deep learning method to estimate spatially seamless <inline-formula><tex-math notation="LaTeX">$T_{a}$</tex-math></inline-formula> from LST that contains missing values. Experimental results on 5-year data of mainland China illustrate that the image-to-image training strategy alleviates the missing data problem and fills the gaps in LST implicitly. Plus, the strong linear relationships between observed daily mean <inline-formula><tex-math notation="LaTeX">$T_{a}$</tex-math></inline-formula> (<inline-formula><tex-math notation="LaTeX">$T_{\rm{mean}}$</tex-math></inline-formula>), daily minimum <inline-formula><tex-math notation="LaTeX">$T_{a}$</tex-math></inline-formula> (<inline-formula><tex-math notation="LaTeX">$T_{\min}$</tex-math></inline-formula>), and daily maximum <inline-formula><tex-math notation="LaTeX">$T_{a}$</tex-math></inline-formula> (<inline-formula><tex-math notation="LaTeX">$T_{\max}$</tex-math></inline-formula>) make the estimation of <inline-formula><tex-math notation="LaTeX">$T_{\rm{mean}}$</tex-math></inline-formula>, <inline-formula><tex-math notation="LaTeX">$T_{\min}$</tex-math></inline-formula>, and <inline-formula><tex-math notation="LaTeX">$T_{\max}$</tex-math></inline-formula> simultaneously possible. For mainland China, the proposed method achieves results with <inline-formula><tex-math notation="LaTeX">$R^{2}$</tex-math></inline-formula> of 0.962, 0.953, 0.944, mean absolute error (MAE) of 1.793 <inline-formula><tex-math notation="LaTeX">$^{\circ }$</tex-math></inline-formula>C, 2.143 <inline-formula><tex-math notation="LaTeX">$^{\circ }$</tex-math></inline-formula>C, and 2.125 <inline-formula><tex-math notation="LaTeX">$^{\circ }$</tex-math></inline-formula>C, and root-mean-square error (RMSE) of 2.376 <inline-formula><tex-math notation="LaTeX">$^{\circ }$</tex-math></inline-formula>C, 2.808 <inline-formula><tex-math notation="LaTeX">$^{\circ }$</tex-math></inline-formula>C, and 2.823 <inline-formula><tex-math notation="LaTeX">$^{\circ }$</tex-math></inline-formula>C for <inline-formula><tex-math notation="LaTeX">$T_{\rm{\rm{mean}}}$</tex-math></inline-formula>, <inline-formula><tex-math notation="LaTeX">$T_{\min}$</tex-math></inline-formula>, and <inline-formula><tex-math notation="LaTeX">$T_{\max}$</tex-math></inline-formula>, respectively. Our study provides a new paradigm for estimating spatially seamless ground-level parameters from satellite products. Code and more results are available at <uri>https://github.com/cvvsu/LSTa</uri>.