Long-Term Monitoring of Mountain Meteorological Variables using Multiple Satellite Remote Sensing Data Fusion over the Korean Peninsula

Mountain meteorology monitoring such as air temperature and humidity is important for the forest disaster and forest ecosystem management. The satellite remotely sensed data provide an opportunity for retrieving meteorological variables with high spatio-temporal resolutions in the mountainous region. However, frequent cloud cover leads to the substantial retrieval error and signal loss of remote sensing, limiting the capability of continuous monitoring. In this study, we propose the estimates method for the mountain meteorology from 2003 to 2015 using multiple satellite data combination approach. A simple pixel-wise empirical regression method between the Aqua Moderate Resolution Imaging Spectroradiometer (MODIS) atmospheric products and the Advanced Microwave Scanning Radiometer-EOS (AMSR-E) and GCOM-W1 AMSR2 brightness temperature at 37 GHz frequency was applied to estimate surface meteorology for all sky conditions. The ambient temperature lapse rate was applied to improve the retrieval accuracy from MODIS atmospheric products in complex terrain. Estimated results were evaluated using the 150 Automated Mountain Meteorology Stations (AMOS) records of the Korea Forest Service (KFS), which is observing the weather situation with 1 minute interval over the complex mountain region. Air and dew temperature (T_air and T_dew) retrievals from satellite dataset showed good agreements (RMSE < 4K) with measurements for clear sky conditions. Furthermore, the estimates accuracy was improved approximated 5% by applying the ambient temperature lapse rate. Substantial retrievals of meteorology were estimated for cloudy sky conditions (n=2,657) by using the proposed pixel-wise regression method. The relative humidity calculated by using the T_air and T_dew retrievals showed a favorable agreement in comparison with AMOS measurements. The results presented in this study indicate that the multiple satellite data fusion can produce the surface meteorological variables with reasonable accuracy for all sky conditions over complex terrains such as the Korean Peninsula.