Geolocation And Cross Calibration Of FY-3B/C Microwave Radiation Imager Data And Its Application In Liquid Water Path Retrieval

The Chinese FengYun-3B and FengYun-3C polar-orbiting meteorological satellites(FY-3B and FY-3C),operating in an afternoon and midmorning configured orbits,can provide global observations almost four times a day at any location.The MicroWave Radiation Imager(MWRI)onboard FY-3B/C satellites measures the radiance from channels at 10.65,18.7,23.8,36.5 and 89 GHz.The measurements and their retrieval products play significant roles in the monitoring and numerical forecasting of disastrous weathers,such as typhoon.The efficient application of MWRI relies on the geolocation accuracy,bias correction and cloud retrieval products.This study first establishes a geolocation error estimation and correction model for the FY-3B/C MWRI measurements to improve geolocation accuracy.Then,a double difference method is used to evaluate MWRI calibration bias based on the TRMM TMI measurements.After geolocation error correction and cross calibration,the study focuses on the retrieval of liquid water path(LWP)over oceans and its diurnal variations.A modified multi-channel LWP retrieval algorithm is developed for retrieving LWP of different magnitudes by using the lowest frequency channel(10.65 GHz)of MWRI.Characteristics of LWP in Super Typhoon"Neoguri" are given,and the diurnal variation of global LWP are consistent with results from TMI.The main results are summarized as follows:(1)To ensure the correctness of geolocation in MWRI data when retrieving LWP,a satellite attitude model is established,and then the attitude offset is utilized to adjust the satellite attitude to correct the geolocation error in the FY-3B/C MWRI data.The node differential method(NDM)is applied to evaluate geolocation errors of FY-3B MWRI measurements at 89 GHz from January to September in 2015.The in-track error of 3 km is found,while the cross-track error is less than 1 km.The coastline crossing method(CCM)is used to calculate the inflection points from coastal regions along northwest coast of Australia,Libya,and Arabian Peninsula under clear-sky conditions from April to July 2014.Statistical results of distance from inflection points to fine-resolution shorelines indicate geolocation error in FY-3C MWRI data at 89 GHz channel is 5-6 km in cross-track and in-track.(2)A double difference method is used in cross calibration based on the TMI measurements.The brightness temperatures with consistent bias from FY-3B/C MWRI after cross calibration are the foundation for the following LWP retrieval.Observations from MWRI and TMI on board TRMM satellite under clear-sky conditions over ocean are collocated to evaluate calibration biases between the two sensors.Results show that bias for FY-3C MWRI descending orbit at 10.65 GHz channel is approximately 5-7 K,which is larger than that at other channels with biases less than 4 K.The bias of both FY-3B and FY-3C MWRI observations between ascending and descending orbits are approximately 2 K.(3)After geolocation and cross calibration bias correction,the brightness temperatures of MWRI are applied to retrieve LWP over oceans.In this study,a modified multi-channel LWP retrieval algorithm is developed for retrieving LWP of different magnitudes by using the lowest frequency channel(10.65 GHz)of MWRI.The theoretical estimates of the LWP retrieval errors are between 0.06-0.11 mm for 10.65 and 18.7 GHz channels,and 0.02-0.04 mm for 36.5 and 89 GHz channels,respectively.In this study,observations from FY-3C MWRI on July 7,2014 are applied to characteristrize the LWP of Super Typhoon "Neoguri".It shows that the brightness temperature observations at 10.65 GHz can be utilized for better retrieving the LWP greater than 3 mm in the eye wall region of Super Typhoon Neoguri.The spiral structure of cloud within and around Typhoon Neoguri can be well captured by combining the LWP retrievals from different frequency channels.(4)The diurnal variation of LWP is studied using multi-channel retrieval algorithm from combined FY-3B and FY-3C MWRI observations.Firstly,TMI observation which can provide 24 hour average LWP,are used as reference to characterize the diurnal variation over Southeast Pacific and South Indian Ocean.Then,after geolocation and cross calibration bias correction,the MWRI observations are applied to retrieve the global annual mean LWP of four daily intervals.Results show LWP retrieved from FY-3B/C MWRI observations is consistent with the diurnal variation curve,with the maximum and minimum occurred in the early morning and afternoon,respectively.