Ground surface temperature (GST) measurements are scarce in the Tibetan plateau (TP), whereas the satellite observations and numerical weather model outputs are good alternatives to fill the spatial gaps among ground stations. However, the evaluation of different temperature products is challenging due to the distinct temporal and spatial dimensions in their acquisition methods. This paper develops an evaluation framework for comparing the performances of various temperature data, including the Advanced Along-Track Scanning Radiometer (AATSR) satellite land surface temperature (LST) data, the high Asia refined (HAR) analysis numerical outputs, and the GST.
In the proposed framework, we introduce a diurnal temperature cycle model and an aggregated weighted method to solve the temporal and spatial mismatch problem between different data sets. The results over TP show that the evaluation framework solves the temporal and spatial matching among different data sets. AATSR LST and HAR outputs are consistent regardless of the heterogeneous and weather conditions at Linzhi site indicating that the fully homogeneous land surface conditions are not the only way for the satellite/simulation validations.
Our results suggest that the proposed framework of time normalization and spatial aggregation method is appropriate for evaluating satellite thermal infrared retrieved data sets and numerical simulations even when the proper ground measurements are insufficient. Since it performs well in the high elevation and complex land surface-conditioned TP region, it will be easily adopted in the other regions with a variety of data sets. (Read More)