AOMSUC-7@Incheon Rapidly Developing Cumulus Area RDCA detection using Himawari-8 data Hiroshi SUZUE and Yasuhiko SUMIDA Meteorological Satellite Center Japan Meteorological Agency
Contents Ø Outline of Himawari-8/9 ü Improved Resolutions ü Advantages of High Observation Frequency Ø Detection of Rapidly Developing Cumulus Area ü Algorithm ü Validation ü Case Study Ø Future Plan Ø Summary 2
Himawari-8 began operation at 02:00 UTC on 7th July 2015. 3
Outline of Himawari-8/9 Advanced Himawari Imager (AHI) solar panel communication antennas Geostationary position Attitude control Communication Around 140.7 E 3-axis attitude-controlled geostationary satellite 1) Raw observation data transmission Ka-band, 18.1-18.4 GHz (downlink) 2) DCS International channel 402.0-402.1 MHz (uplink) Domestic channel 402.1-402.4 MHz (uplink) Transmission to ground segments Ka-band, 18.1-18.4 GHz (downlink) 3) Telemetry and command Ku-band, 12.2-12.75 GHz (downlink) 13.75-14.5 GHz (uplink) Himawari-8 began operation on 7 July 2015, replacing the previous MTSAT-2 operational satellite 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 MTSAT-1R MTSAT-2 operation standby standby operation standby Himawari-8 Himawari-9 a package purchase manufacture manufacture launch launch operation standby standby operation standby 4
MSC/JMA Improved Resolutions Spatial At sub-satellite point VIS 1 km IR 4 km VIS 0.5/1 km IR 2 km MTSAT-1R/2 Himawari-8/9 Spectral VIS 1 band 3 bands VIS R G B Temporal NIR 3 bands Observation Frequency IR 4 bands IR 10 bands full-disk obs. MTSAT-1R/2 Himawari-8/9 5 bands 16 bands MTSAT-1R/2 Himawari-8/9 5
Observation modes and intervals Visible band True Color RGB July 9-10, 2015 Japan & Vicinity Obs. 2.5 min. Targeted Area obs. 2.5 min. Full Disk Obs. 10 min. Visible band 6
Developing Cumulus and Radar Echo height Developing of Cumulus (model) heavy rain area of Met. Radar 0 min 10 min 15 min 20 min 25 min 30 min time Chisholm, A. J. and Renick, J. H. (1972) [traced and added] 3min 3min If we can detect cumulus that is growing rapidly, we get to know thunderstorm coming earlier than the radar! Prepare for thunderstorm! 7
MSC/JMA RDCA Product Convective Cloud Information Cumulonimbus Rapidly Developing Cumulus Mid/Low cloud unknown 8
l Rapidly Developing Cumulus Area (RDCA) ü Developing cumulous ü Current/Future disturbance is expected RDCA Product Rapidly Developing Cumulus Area Cumulonimbus Area Mid/Low cloud unknown Area l Cumulonimbus Area ü A round top, except for anvil cirrus ü Strong upward flow is expected? l Mid/Low Cloud Unknown Area ü Anvil cirrus ü Anvil cirrus hides clouds below 9
Concept of RDCA Detection Cloud Height After 5 min. Cloud top adjacent Developing cumulus Cloud top is higher Brightness temperature is getting low. Roughness of cloud top increases Contrast between light and dark is getting clear. e.g. Difference of reflective intensity is increasing in visible image. Cloud microphysical parameters change Ice particles are produced near cloud top 10
RDCA : Decision Process Logistic Regression Model 1 p = ì æ 1+ expí- ça0 + î è å i a i x i öü ý øþ Three class parameters; :<250K, :250~273.15K, :>273.15K Probability (forecast) Detection parameters Actual Probability lightning obs. [num/area ] (13. Jul. 2011 ) The correlation between lightning and regression p developing Coefficients a i are determined by the logistic regression analysis when lightning occurs within 1 hour after observed variable x i. => High P area is decided as RDCA Predicted Probability p 11
Ø RDCA detection parameters Only day time New Only day time New No. Detection Parameter Main Objective 1 B03(0.64um):Max-Ave.* 2 Cloud Top Roughness Detection 2 B03:Standad Deviation* 2 Cloud Top Roughness Detection 3 B13(10.4um):Min.-Ave. Cloud Top Roughness Detection 4 B13:Standard Deviation Cloud Top Roughness Detection 5 B16(13.3um)-B13 Ice Cloud Detection 6 B08(6.2um)-B13 Water Vapor Detection above Cloud Top 9 B15(12.4um)-B13 Ice Cloud Detection Water Vapor Detection above 10 B10(7.3um)-B08 Cloud Top 11 B11(8.6um)-B13 Ice Cloud Detection Temporal Variation of Presumption of Developing 21 B03 Average Value* 2 Level of Cloud Temporal Variation of Presumption of Developing 23 B13 Average Value Level of Cloud Temporal Variation of Developing Ice Particle 24 B11-B13 Average Value* Detection Temporal Variation of Developing Ice Particle 25 B15-B13 Average Value* Detection One Scene Parameters Time Change Parameters 12
Validation of RDCA Day time (00-09UTC) / Night time (12-21UTC) Truth:lightning (CC and CG) detection Validation Period 1 27 31July (5 days) 2 28 Aug. 1 Sep. (5 days) 3 24 28 Sep. (5 days) Total: 15 days + 0.1-0.1 Hit Range Hit 0 +1 h time Lightning detection number 1 2 3 Mutch Up Validation 13
Validation of RDCA Old Algorithm Year Time Zone Satellite Period POD [%] FAR [%] 2012 Daytime MTSAT-1R Jun. Sep. 44.03 68.81 2013 Daytime MTSAT-1R Jun. Sep. 39.62 75.60 2014 Daytime MTSAT-1R Jun. Sep. 51.48 70.70 2015 Daytime Himawari-8 Jun. Sep. 79.31 90.60 Summer in 2015 Method Time Zone Satellite Period POD [%] FAR [%] Old Daytime Himawari-8 15 days 71.11 85.37 New Daytime Himawari-8 15 days 63.68 57.79 New Nighttime Himawari-8 15 days 57.83 63.26 14
MSC/JMA Case Study #1 Ø Early detection of convective cloud with lightning RDCA product can detect developing cumulus earlier than a radar echo. 02 30 UTC RDCA Detection 02 50 UTC Radar Echo Detection 03 20 UTC Lightning Detection On 4 Aug. 2015 15
MSC/JMA Case Study #2 Ø False detection due to passing upper clouds Brightness temperature seems to decrease rapidly because upper clouds pass over lower clouds 00 00 UTC 00 30 UTC 01 00 UTC 01 30 UTC On 6 June 201616
MSC/JMA Future Plan Ø Domain extension of the RDCA product using Himawari-8/AHI Full Disk observation data for safety and air traffic control over Asia and Western Pacific regions Sample Animation Sample of extended domain RDCA product 17
MSC/JMA Future Plan Ø Improvement of the RDCA detection algorithm (e.g. cloud tracking) Sample Animation Sample of cloud object tracking 18
Summary Ø Improved Observation Function by Himawari-8/AHI ü High-resolution and high-frequency observation using multiple bands enables to capture server weather phenomena Ø Detection of Rapidly Developing Cumulus Area ü Statistical method is used for rapidly developing cumulus detection ü RDCA product has been operational all day using multiple observation bands data ü POD and FAR are about 60 % Ø Future Plan ü Domain extension of the RDCA product ü Improvement of the RDCA detection algorithm 19
Thank you for your kind attention JMA mascot character Harerun 20