Quick, light and tunable data processing for GOSAT target observations using RemoTeC retrieval algorithm

Abstract:

In addition to nominal grid observation, TANSO-FTS onboard GOSAT has functionality of target observations using agile pointing mechanism, which are scheduled and modified daily bases. At the beginning, target observations were limited to calibration and validation sites but now for large emission sources such as mega cities, oil field, landfill, and volcano. Towards emission inventory, combination of satellite, airplane and ground-based observations are essential. Generally large emission sources have thick aerosol and their topographies are not flat. For custom target observations of limited data set during the observation campaigns, satellite data processing must be quick and light. We had used RemoTeC algorithm for Level 1 algorithm for end (interferogram) to end (X_CO2 and X_CH4) test purposed. Now we apply its tunable functions to target observations. RemoTeC uses metrology surface pressure as an input in order to maximize retrievable aerosol parameters. In the operational RemoTeC retrieval, the ERA-Interim dataset, which is a global atmospheric reanalysis data with 80km spatial resolution and with 60 vertical levels from surface up to 0.1hPa provided by ECMWF, is applied as meteorology dataset. These dataset is published with charge free, and will be available about 3 months later from observation time. To quick validation between GOSAT observation and others within short time delay, NCEP FNL Operational Model Global Tropospheric Analyses dataset is newly applied. The special resolution and vertical levels of NCEP FNL data are 1degree x 1degree and 26, respectively. The spatial and vertical resolutions are coarser than that of ECWMF products, but the time delay from observation period is less. To characterize the biases of X_CO2 and X_CH4, and the accuracy of retrieval with NCEP meteorology dataset, over Lamont data are compared during June 2009 to March 2011. The results present clearly that the retrieved X_CO2 with NCEP dataset have around -0.37 ppm systematic bias against that with ECMWF dataset. In contrast, the agreement between NCEP based X_CH4 and ECMWF based X_CH4 is good as well as CHI2. Within a few days from observation time period, these retrieved data provide us X_CO2 and X_CH4 products with scatter of 2.09 ppm and 14.1 ppb, respectively.