Passive and Active Microwave Sensors for Precipitation Research

Transcription

1 Passive and Active Microwave Sensors for Precipitation Research Joe Turk Jet Propulsion Laboratory California Institute of Technology Pasadena, CA

2 Characteristics of Precipitation Wide variability in space, time, and intensity Discontinuous; wide range of extremes; intermittent Controlled across scales from microscale (microphysics), mesoscale (sea breezes), synoptic (fronts) Spaceborne measurement of precipitation: Constrained by observing system limitations (revisit, resolution, frequency, swath, noise) More challenging over land than over ocean Used by an increasingly broad and diverse user community Not straightforward to validate

3 Topics that I will discuss Satellite Platforms with Precipitation-Relevant Capabilities General issues with current datasets and products Focus on GPM and what it will do differently from TRMM A few unique proposed mission concepts

4 Local Time of Satellite Observations TRMM TMI+PR 28-day repeat (equator)? TRMM launch Nov 1997 F19 launch: As soon as Oct 2013 F14 only direct-broadcast since 24 Aug 2008 F13 only direct-broadcast since 9 Nov 2009 F15 RADCAL beacon activated 14 Aug 2006 FY-3A launch: May 2008 Megha-Tropiques launch: Oct 2012

5

6 Why is Observation Time So Important? Technical Aspects Solar panels in the dark Solar intrusions into detectors or calibration references Example from EOS-Terra morning overpass on January 1, Solar zenith is indicated by the white shades and the red stripe indicates the daynight terminator. Note how Terra passes over at nearly the same local time each orbit. Terminator Orbit First-light (spy satellites) Limit Sun s influence on passive measurements Scene Conditions Precipitation changes very rapidly and often has a predominant daily cycle

7 The Satellite Beamfilling Problem We don t know the spatial pattern of the underlying rainfall at the time that the satellite flies over Satellite movement Therefore, when one interprets the satellite signal (radiances), there will be a systematic underestimate of the rainfall (e.g, 10 mm/hour) But it s only raining in this fraction of the sensor s field of view (e.g., 25 mm/hour) Satellite sensor receives a signal for all Earth scenes that fall within this cone ( field of view ) Earth s surface 50-km (not drawn to scale)

8 Two Types of Satellite Measurements Sun T=6000 Kelvin Satellite Solar Reflection Sunlight that is reflected off the top of the cloud and then into the satellite field of view. Daytime-only. Does not directly measure rainfall, only cloud top properties Cloud T=300 K Earth s surface (70% ocean, 30% land) Thermal Emission Thermal emission that originates near the top of the cloud (infrared to near infrared) or within the cloud (microwave) and into the satellite field of view. Day and night. (not drawn to scale)

9 Electromagnetic Spectrum Utilized For Satellite-Based Remote Sensing of Precipitation Maximum emission from the sun (T=6000 K) is from microns wavelength (Visible spectrum) Maximum thermal emission from the Earth (T=300 K) is in the infrared (IR) wavelengths near 11 microns Passive microwave (PMW) emission is much weaker, but able to penetrate clouds and more directly sense rainfall. 10, 19, 37, 85 GHz are commonly used window channels

10 Different Satellite Instrument Scanning Patterns Across-Track Scanning Conically Scanning AMSU Sensors on NOAA/METOP TRMM Precipitation Radar Pixels expand across track Viewing angle changes across scan SSMI Sensor on DMSP AMSR-E Sensor on EOS-Aqua Pixels maintain size across track Viewing angle constant across scan click on image to animate

11 Concept of Atmospheric Transmittance Any signal received by a satellite sensor is affected by atmospheric constituents such as water vapor and oxygen molecules However there are windows in the spectra where the atmosphere is more transparent and hence the signal received by the satellite better represents the feature that you want to observe (clouds, rainfall, terrain, sea surface) TRANSPARENT SEMI-TRANSPARENT OPAQUE

12 8 May Z Aqua Overpass Clear-Air Flight Date During MC3E 85H Precipitation across south-central Minnesota Open water lakes Partially frozen lakes River channels Antecedent precipitation effects Lower and Upper Mississippi floods Upper Midwest floods (Red River) Mississippi alluvial plains Small inland lakes and reservoirs

13 8 May Z Aqua Overpass Clear-Air Flight Date During MC3E 37H Precipitation across south-central Minnesota Open water lakes Partially frozen lakes River channels Antecedent precipitation effects Lower and Upper Mississippi floods Upper Midwest floods (Red River) Mississippi alluvial plains Small inland lakes and reservoirs

14 8 May Z Aqua Overpass Clear-Air Flight Date During MC3E 19H Precipitation across south-central Minnesota Open water lakes Partially frozen lakes River channels Antecedent precipitation effects Lower and Upper Mississippi floods Upper Midwest floods (Red River) Mississippi alluvial plains Small inland lakes and reservoirs

15 8 May Z Aqua Overpass Clear-Air Flight Date During MC3E 10H Precipitation across south-central Minnesota Open water lakes Partially frozen lakes River channels Antecedent precipitation effects Lower and Upper Mississippi floods Upper Midwest floods (Red River) Mississippi alluvial plains Small inland lakes and reservoirs

16 MC3E AMSR-E Afternoon Overpasses 10H 19H 37H 89H 8 May May May May May 2011

17 SSMIS Scan Geometry Main Reflector Cold Calibration Reflector Warm Load Feedhorns 1707-km SSMIS 1400-km SSMI 180 samples/scan (91, 150, 183 GHz) 90 samples/scan (19-37 GHz) 60 samples/scan (lower sounding) 30 samples/scan (upper sounding) 17

18 SSMIS Orbit Times Satellites fly in a near terminator orbit to capture first-light F19 launch as soon as October 2013 Future of this satellite system (after F-20) unclear at this time

19 Global Precipitation Measurement (GPM) core satellite Orbit will provide sampling of considerably more high latitude scenes Use combined GMI+DPR retrievals over narrow inner swath for radiometer only retrievals

20 TRMM and GPM TRMM GPM Orbit 38-deg NSS 1, 405-km 65-deg NSS, 410-km Launch H-2 (JAXA) HY-2 (JAXA) Radar Single frequency, ±17 o scan Dual frequency/interlaced Radiometer TMI (SSMI + 10 GHz) GMI (TMI + 157/183 GHz) Revisit Data System Sufficient sampling to study tropical climate TSDIS (now PPS) Realtime was afterthought and best effort Aggregate 3-hr revisit with partner satellites PPS Realtime essential role 1 Non Sun Synchronous

21 GPM Microwave Radiometer (GMI) atmospheric transmittance TRMM (9 channels, GHz) GPM (11 channels, GHz) window channels (10.7H/V, H/V, 37.1H/V, 89.0H/V, 166H GHz) sounding channels (23.8, 190 GHz)

22 Main Characteristics of DPR Item KuPR KaPR TRMM PR Antenna Type Active Phased Array (128) Active Phased Array (128) Active Phased Array (128) Frequency & GHz & GHz & GHz Swath Width 245 km 120 km 215 km Horizontal Reso 5 km (at nadir) 5 km (at nadir) 4.3 km (at nadir) Tx Pulse Width 1.6 us (x2) 1.6/3.2 us (x2) 1.6 us (x2) Range Reso 250 m (1.67 s) 250 m/500 m (1.67/3.34 s) 250m Observation Range 18 km to -5 km (mirror image around nadir) 18 km to -3 km (mirror image around nadir) 15km to -5km (mirror image at nadir) PRF VPRF (4206 Hz±170 Hz) VPRF (4275 Hz±100 Hz) Fixed PRF (2776Hz) Sampling Num 104~ ~ Tx Peak Power > 1013 W > 146 W > 500 W Min Detect Ze (Rainfall Rate) < 18 dbz ( < 0.5 mm/hr ) < 12 dbz (500m res) ( < 0.2 mm/hr ) Courtesy T. Iguchi, NICT < 18 dbz ( < 0.7 mm/hr ) Measure Accuracy within ±1 db within ±1 db within ±1 db Data Rate < 112 Kbps < 78 Kbps < 93.5 Kbps Mass < 365 kg < 300 kg < 465 kg Power Consumption < 383 W < 297 W < 250 W Size m m m * Minimum detectable rainfall rate is defined by Ze=200 R 1.6 (TRMM/PR: Ze=372.4 R 1.54 )

23 Dual Frequency Precipitation Radar (DPR) 14GHz radar beam 35GHz radar beam Detection limit in 35GHz channel Detection limit in 14GHz channel Higher sensitivity at higher frequency Ice Snow Height 14GHz Discrimination between snow and rain by attenuation difference Melting layer Rain 35GHz Accurate rain estimation based on attenuation difference Radar Reflectivity Factor Roles of DPR Accurate 3D measurements of precipitation as TRMM, but with better sensitivity Improvement of estimation accuracy Identification of hydrometer type, phase state Improvement of MWR algorithms Simultaneous measurements with GPM Microwave Imager (GMI) Courtesy T. Iguchi, NICT

24 Concept of the DPR antenna scan Courtesy T. Iguchi, NICT KuPR footprint : Δz = 250 m KaPR footprint (Matched-beam with KuPR) : Δz = 250 m KaPR footprint (High-sensitivity beam) : Δz = 500 m KaPR: 120 km (24 beams) KuPR: 245 km (49 beams) In the interlacing scan area ( ), the KaPR can measure snow and light rain in a high-sensitivity mode with a double pulse width. The synchronized matched beam ( ) is necessary for the dualfrequency algorithm.

25 GPM Two-Frequency Radar

26 Retrieval of Near-Surface Rainrate Examine Specific Attenuation Adjacent gates are analyzed and the Z is analyzed Specific Attenuation (db km -1 ) A = (Z 1 Z 2 ) / z then A= ar b where a,b are constants that relate A to reflectivity Z z= 250m for PR/DPR Z at gate 1 Z at gate 2 etc etc Errors gradually build up in topdown and bottom-up approaches Natural variability of the raindrop size distribution (DSD) and A-Z relations Single frequency radar does not directly provide information about hydrometeor phase Ka-band (35 GHz) and higher: More impact from clouds/vapor, multiple scattering

27 Retrieval of Near-Surface Rainrate Examine Specific Attenuation Adjacent gates are analyzed and the Z is analyzed Specific Attenuation (db km -1 ) A = (Z 1 Z 2 ) / z then A= ar b where a,b are constants that relate A to reflectivity Z z= 250m for PR/DPR Z at gate 1 Z at gate 2 etc etc Use Path Integrated Attenuation (PIA) As a Constraint Knowledge of clear-sky PIA is used as a constraint to partition per-gate attenuation and derive rainrate Nearby clear-sky scene is used as a reference Clear-sky scene 30 db is measured from the gate that hits the surface Rainy scene 20 db measured= 10 db PIA (2-way) Strongly reflecting ocean surface (easier) or Highly variable land surface (difficult)

28 TRMM TMI/PR 15 Sep UTC Over-Ocean TMI can t delineate fine-scale structure PR-estimated precip is displaced from the TMIestimated precip due to parallax satellite motion

29 TRMM TMI/PR 17 Sep UTC Over-Land TMI can t capture the heavy isolated rain events (the tail of the rainfall histogram, i.e. the few big events ) that the PR picks up satellite motion

30

31 Tropical Rainfall convective clouds with ice region above rain, warm ocean background Higher Latitude Drizzle Lower altitude clouds, mainly non-convective, cold ocean background Using Microwave Imaging and Sounding Channels GHz imaging channels are the most useful 183 GHz sounding channels less useful water vapor and temperature variations less significant Significant surface contribution between GHz Surface is usually opaque at 150 and 183 GHz signals due to snow and drizzle 150 ±1 ±3 ± shading denotes water vapor thermal emission 19 GHz 19 GHz 37 85

32 Hurricane Katrina 29 August UTC SSMIS 19V GHz ~ 50-km resolution Over Water V > H Over Land V ~ H Hurricane eye 280K Cold ocean due to lower surface emissivity Gulf 220K Great Lakes 200K Coarse sensor resolution effects near coastlines Warm land due to high surface emissivity 280K

33 Hurricane Katrina 29 August UTC (Morning Overpass) Lakes and Irrigated Lands Apparent SSMIS 91H GHz (Over Water V > H) (Over Land V ~ H) Surface Depolarizes for Increasingly Moist Atmospheres Cloud-Free Land-Ocean TB Difference ~ 60K Great Lakes ~ 20K Gulf of Mexico Heaviest Convection in southern Mississippi

34 Hurricane Katrina 29 August UTC GOES-12 IR image SSMIS 91V GHz ~15-km resolution Cloud-Free LandOcean TB Difference ~ 15K Gulf of Mexico ~ 30K Great Lakes Land/water differences are almost completely masked by atmospheric emission Surface precipitation shows up at 91 GHz

35 Hurricane Katrina 29 August UTC (Morning Overpass) SSMIS 150 GHz Cloud-Free Land- Ocean T B Difference Disappears over Gulf of Mexico ~ 10K over Great Lakes Masks Nearly All But the Most Convective Precipitation

36 Hurricane Katrina 29 August UTC (Morning Overpass) SSMIS 183.3±7 GHz (Outer Skirt of Strong Water Absorption) Lower Level Moisture (700 hpa) Patterns Revealed (Similar to a 6.2 µm MSG Water Vapor Image) Strongest Regions of Convection Precip Embedded in Upper-Level WV

37 Hurricane Katrina 29 August UTC (Morning Overpass) SSMIS 183.3±3 GHz (Mid-Skirt of Strong Water Absorption) Mid-Level Moisture (500 hpa) Patterns Revealed Strongest Regions of Convection Contrast but are Weaker

38 Hurricane Katrina 29 August UTC (Morning Overpass) SSMIS 183.3±1 GHz (Inner-Skirt of Strong Water Absorption) Upper Level Moisture (350 hpa) Patterns Revealed Strongest Regions of Convection Contrast Primarily Upper-Level WV with the very strongest convection

39 Increasing Refresh and Coverage with Multi-Dataset Merging Techniques Multiple LEO (Microwave) Satellite Merging DMSP orbits Aqua (AMSR) TRMM (TMI+PR) Characteristics Only a few obs per point per day Intermittently spaced in time Inter-sensor differences (resolution, calibration, algorithm) Open issues: high latitudes, snow, cold/variable surfaces, drizzle 3-hour 6-hour 12-hour 24-hour etc.

40 Multi-Sensor & Techniques = DMSP orbits Aqua (AMSR) TRMM (TMI+PR) geostationary Across-track scanning microwave sounders (15-50 km) Conically scanning microwave imagers (5-70 km) Space Radars (5 km) TRMM 2A12 Instantaneous swath-level precipitation product (L2) CMORPH TRMM 3B42 Blend/tracking with geostationary VIS/IR (25 km, 3-hourly) High resolution precipitation products HRPP GPCP CMAP HOAPS etc Raingauge analyses (1-deg, daily) (2.5-deg, monthly

41 Merging Procedures Geostationary imagers 4-km IR Microwave imagers km resolution min refresh intermittent refresh time 0000 UTC 6-hour time window 0600 UTC

42 IR and microwave rain estimation Infrared Microwave GHz

43 Additional precipitation estimates for climate monitoring Product Source Resolution Data record GPCP NOAA 2.5º, pentad & monthly, global CMAP NOAA 2.5º, pentad & monthly, global 1979-present 1979-present TRMM 3B42 NASA 0.25º, 3 hr, 50ºS-50ºN 1998-present

44 Near-real time ~global precipitation Product Source Resolution TMPA NASA 0.25º, 3 hr NRLblend NRL 0.10º, 1 hr PERSIANN UC Irvine 0.25º, 1 hr CMORPH NOAA 8 km, 30 min HYDRO NOAA 5 km, 1 hr MIRS NOAA 0.25º, daily GSMaP JAXA 0.10º, 1 hr

45 Tools GSMaP nearreal time data viewer

46 Geostationary Earth-Orbiting Satellites Need at least 5 to image the entire Earth at any one time Don t see above (below) 70N (70S) degrees latitude Very rapid time updating: New images transmitted as fast as every 15 minutes Subpoint pixel resolution near 1-km during day, 3-km at night Composite IR from GOES-11, GOES-12, Meteo-8, Meteo-5, GMS-6 between +/-60 degrees