7.3 Utilization of Himawari-8 in the Global Satellite Mapping of Precipitation (GSMaP) Nowcast version

Wednesday, 13 January 2016: 4:30 PM
Room 252/254 ( New Orleans Ernest N. Morial Convention Center)
Misako Kachi, JAXA, Tsukuba-shi, Ibaraki, Japan; and T. Kubota and R. Oki

The Global Satellite Mapping of Precipitation (GSMaP) product is a global rainfall map that is highly accurate, highly frequent, and high-spatial resolution through the development of rain rate retrieval algorithms based on reliable precipitation physical models by using a number of microwave radiometer (imagers, sounders, and imager/sounders) observation data, and comprehensive use of the Precipitation Radar (PR) on board the Tropical Rainfall Measuring Mission (TRMM) satellite. To produce a highly frequent global rainfall map, sampling errors caused by un-observed regions by low-orbital satellites is a problem. The GSMaP project has developed a method to interpolate observations between each microwave imager by utilizing information from the Infrared imagers on board the five geostationary satellites, which are JMA's MTSAT/Himawari satellite, two NOAA's GOES satellites, and EUMETSAT's Meteosat satellites, and achieved production of an hourly global rainfall map in 0.1-degree latitude/longitude grid.

JAXA has provided the near-real-time GSMaP (GSMaP_NRT) product four hours after observation through the “JAXA Global Rainfall Watch” website (http://sharaku.eorc.jaxa.jp/GSMaP/) since 2008 as a prototype of the Global Precipitation Measurement (GPM) mission. To assure near-real-time data availability, the GSMaP_NRT system simplified part of the algorithm and its processing procedure. Therefore, the GSMaP_NRT product gives higher priority to data latency than accuracy. The GSMaP product became one of JAXA products in the GPM mission. The GPM mission consists of TRMM-like satellite, called the GPM Core Observatory, jointly developed by U.S. and Japan and Constellation Satellites that carry microwave radiometers. The GPM mission is an international collaboration to achieve highly accurate and highly frequent global precipitation observations. The GPM Core Observatory was launched on February 28, 2014, while the Constellation Satellites are launched by each partner agency sometime around 2014-2018 to expand observation coverage and increase observation frequency. The GPM-GSMaP product has been released to public since September 2, 2014 via JAXA G-Portal (https://www.gportal.jaxa.jp/), as well as the JAXA Global Rainfall Watch web site.

Since its data release, GSMaP_NRT data has been used by various users for various purposes, such as rainfall monitoring, flood alert and warning, drought monitoring, crop yield forecast, and agricultural insurance. There are, however, several requirements from users for GSMaP improvements not only for accuracy but also specification. Among those requests for data specification, the most popular ones are shortening of data latency time and higher horizontal resolution. To achieve shortening of data latency time, JAXA has developed beta version of GSMaP nowcast version (GSMaP_NOW) product for observation area of the Himawari geostationary satellite since June 18, 2015, and it is being validated by JAXA and related agencies.

After the launch of Himwari-8 satellite, JAXA has received its data since March 2015. The Himawari-8 IR data has been used in the GSMaP_NOW system. The GSMaP_NOW product uses passive microwave radiometer data that is available only within one hour after observation. Addition of extrapolation by using cloud moving vector, which is calculated from Himawari-8/AHI, of one hour forward (toward future) enables us to produce “nowcasting” rainfall map over Asian regions. This “nowcasting” capability enables operational users to apply the GSMaP_NOW data to rainfall monitoring more rapidly, flood alert in smaller basins. We also plan to use the GSMaP_NOW data in the research of crop yield monitoring using satellite data and its early analysis.

Early evaluation result comparing with Japan Meteorological Agency Radar-AMeDAS rainfall analysis (gauge-calibrated ground-based radar map) in 0.25 degree and daily average indicates that the GSMaP_NOW accuracy is slightly lower than that of GSMaP_NRT. For the period from June 11 to July 3, 2015, RMSE and correlation coefficient of GSMaP_NOW are 0.56 mm/hr and 0.76 respectively, while those of GSMaP_NRT are 0.52 and 0.79. At present, The GSMaP_NOW product and its web site will be open to public in the summer of 2015.

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner