Applications of Advanced Imagers James F.W. Purdom CIRA Colorado State University Fort Collins, CO Applications of Advanced Imagers in synergy with other space based observing capabilities (It’s a system of space based observations, but we will focus on imagers) Environmental Satellite Sensor Synergy Goal and Challenge: Dynamic Tasking and Adaptive Sensing Observing system characteristics – Polar is global and fixed (well determined orbit and sensor operational capability) – Geostationary is quasi-hemispheric and adaptive (point and shoot) Selected sensors will operate over very similar spectral regions (visible to infrared) The spatial, spectral and radiometric resolutions of future geostationary satellite systems’ sensors will in some cases approach and in other cases surpass those of the polar orbiting satellite systems’ sensors. Example of Synergy in GOES-R Era Goal and Challenge: Dynamic Tasking and Adaptive Sensing Intra-Satellite (USA Operational) – GOES: ABI and HES (adaptive) – NPOESS: VIIRS and CrIS (fixed) Intra-System (USA Geo + operational polar) – GOES-E and GOES-W (or other possibilities) – NPOESS AM and PM, METOP, FY-3, METEOR-3 Inter-System – GOES-R and NPOESS (ABI, VIIRS, HES and CrIS) – GOES-R, NPOESS and other operational and research satellites Intra-Satellite (USA Operational) GOES: ABI and HES (adaptive) Evolving instability field from HES triggers super rapid scan over severe storm area Example: Use ABI Data to Task HES VNIR GOES-8 loop from 1615 to 2345: this loop illustrates the changes that occur in the cloud field after the MODIS pass and the need to dynamically task HES. Despite increasing cloud cover, the Florida Bay and Northern Keys could be successfully imaged over several hours which will allow for observations of ocean color as well as changes due to tidal effects. Florida Bay Northern Keys Intra-System (USA Geo + operational polar) GOES-E and GOES-W (or other possibilities) Intra-System (USA Geo + operational polar) NPOESS AM and PM, METOP, FY-3, METEOR-3 Inter-System (USA Geo + operational polar) NPOESS AM and PM, METOP, FY-3, METEOR-3, etc Intra-System (USA Geo + operational polar) NPOESS AM and PM, METOP, FY-3, METEOR-3 EO-1 Landsat-7 SAC-C Terra “True color” movie made from multispectral images from the A.M. Constellation While this movie focuses on storm development, one can imagine how high spatial and spectral resolution imagery from Terra and EO-1, separated in time by 39 minutes could be used to investigate the development and evolution of ocean and coastal zone phenomena Inter-System GOES-R, NPOESS and other satellites • Can you imagine the impact of Radio Occultation when combined with the frequency of GOES-R for applications such as nowcasting severe weather ? • Collard and Healy, QJRMS, 129, p.2741, 2003 Inter-System GOES-R, NPOESS and other satellites • Can you imagine the impact of Radio Occultation when combined with the frequency of GOES-R for applications such as nowcasting or hurricane analysis ? • Collard and Healy, QJRMS, 129, p.2741, 2003 Inter-System GOES-R, NPOESS and other satellites GOES-R total HES can serve as stable reference basis for other satellites (operational polar and other LEO) – Contiguous and high resolution spectral measurements required for inter-calibration Spectral flexibility (adaptability) allows for spectral matching with other systems’ instrumentation GOES-R total HES is a baseline Already looking at AIRS and LEO channel calibration GOES-R ability to track diurnal cycle – Simultaneity with LEO constellation – Contribution to GPM (Global Precipitation Mission) High spectral resolution radiance matching for scenes over long time periods by one fixed system at the same viewing angle and from the same altitude (Goody concept) – Detect and monitor long term changes (trends) in water vapor and other gasses The spatial and temporal domains of the phenomena being observed drive the satellite systems’ spectral needs as a function of space, time, and signal to noise. Synergy in its infancy – hurricane analysis : geostationary, polar and other Synergy in its infancy – hurricane analysis: geostationary, polar and other Synergy in its infancy – hurricane analysis : geostationary, polar and other Synergy in its infancy – hurricane analysis : geostationary, polar and other Synergy in its infancy – hurricane analysis : geostationary, polar and other Synergy in its infancy – hurricane analysis : geostationary, polar and other Synergy in its infancy – hurricane analysis : geostationary, polar and other Synergy in its infancy – hurricane analysis : geostationary, polar and other ENTERING AN AGE OF MULTI-PLATFORM MULTI-SENSOR PRODUCTS But we’re not there yet To assure full utilization we need to capitalize on the adaptive observing nature of the geostationary system in synergy with the fixed polar system Period Climate Long term Seasonal Interan’l Spatial Temporal Spectral Radiometric Issue L Weekly* H vH Climate Year Months M/L Days to Weekly* M M Climate NWP 1-14 day M/L (obs) M (surface) 4 hr Sounder: H Image: M H Weather, Transportation Regional NWP 6-24 hr M (obs) H (surface) 1 hr Sounder: H Image: M H Weather, Trans. Nowcast Severe Wx 0-6 hours H Sounder: 0.5 Sounder: H hr Image: M Image: 1min H Weather, Trans. Coastal Littoral Hr to daily H to vH 1-3 hr H – VNIR vH Littoral, Weather, Trans. Rapid Action Disaster Immediate vH 1 min H – VNIR Image: M H Homeland Security + L~100km, M~10km, H~1 km, vH~100m *cumulative from more freq obs system where instrument stability over very long time period is important The spatial and temporal domains of the phenomena being observed drive the satellite systems’ spectral needs as a function of space, time, and signal to noise. Spectral Awareness In Terms of Space, Time, Signal to Noise and Scene Characteristics Today we’re digital - Basic Premise Spectral Awareness In Terms of Space, Time, S/N and Scene Properties AVIRIS spectral movie from 0.4 to 2.4 microns Fire scene Spectral Awareness In Terms of Space, Time, S/N and Scene Properties AVIRIS spectral movie from 0.4 to 2.4 microns Fire scene AVIRIS Spectral Information from the Scene Depicting Cloud, Smoke and Active Burn Areas AVIRIS Image - Linden CA 20-Aug-1992 224 Spectral Bands: 0.4 - 2.5 mm Pixel: 20m x 20m Scene: 10km x 10km Spectral Signatures of Selected Pixels 1 0 . 0 C l o u d F i r e H o t A r e a G r a s s L a k e B a r e S o i l S m o k e ( s m . p a r t . ) S m o k e ( l g . p a r t . ) S h a d o w Smoke - large part. Shadow 1 . 0 Fire Hot Area Soil Cloud AparentRflcane Grass Lake Smoke small part. 0 . 1 4 0 0 7 0 0 1 0 0 01 3 0 01 6 0 01 9 0 02 2 0 02 5 0 0 W a v e l e n g t h ( n m ) The need to filter atmospheric effects: note water vapor change every 15 minutes HES VIS/NIR at high resolutions will be able to monitor pre-cumulus cloud moisture and moisture convergence – this will be enhanced by HES-IR Note water vapor change every 15 minutes HES VIS/NIR at high resolutions will be able to monitor pre-cumulus cloud moisture and moisture convergence – this will be enhanced by HES-IR VIIRS, MODIS, FY-1C, AVHRR Reality: CO2 O2 O3 H2O O2 H2O H2O H2O O2 H2O H2O CO2 Spectral Awareness In Terms of Space, Time, S/N and Scene Properties VIIRS MODIS Channel Positions of Various Ocean-Color Sensors, 1978-2000* (380 – 950 nm) For a multi spectral sensor Many spectral bands are identified for various applications Selection of band location and width are also important * IOCCG Report #1 GOES Sounder spectral coverage - Example of spectral variability Associated with channel radiometry High Spectral Resolution (AIRS) resolves H2O spectral Features (right) GOES-I/M era sounder H20 Channels (above) With satellite remote sensing, there are four basic questions that need to be addressed They all deal with resolution: – temporal (how often) – spatial (what size) – spectral (what wavelengths and their width) – radiometric (signal-tonoise) With satellite remote sensing, there are four basic questions that need to be addressed Spectral Awareness Spectral Awareness Scattering from water versus ice particles at 3.9 microns Response of 3.9 vs. 10.7 microns to Temperature variability in a FOV Visible loop (left) and 3.9 micron reflective component loop (right) from GOES-West (aspect ratio not 1:1) Clouds separate into classes when multispectral radiance information is viewed Hi cld Mid cld vis 1.6 um Lo cld Snow Clear LSD 1.6 um 8.6-11 um 11 um 11 um Cloud Composition Contrails Image Over Kansas - 21 April 1996 Ice Cloud Infrared Temperature Difference - 8.6 mm (Band 29) - 11.0 mm (Band 31) Contrails Water Cloud Infrared Temperature Difference - 11.0 mm (Band 31) - 12.0 mm (Band 32) Focus Maximizing utilization from the satellite systems available in the 2010 era – The amount of data and potential products will be so massive that it can swamp the ordinary user – Growth in applications and sophistication What we can do today to improve current utilization – User requirements versus system capability – Educating the user – Keeping science and vision in play The spatial and temporal domains of the phenomena being observed drive the satellite systems’ spectral needs as a function of space, time, and signal to noise. GOES-R: Unique in spectral, spatial and temporal domains GOES-R sensors spatial and spectral resolutions approach and in some instances surpass those of its operational polar orbiting counterparts Spatial, Spectral, Temporal Improvements: Predicting when and where severe weather will occur (better location accuracy) – Severe storms – Hurricanes GOES I-P Increased lead-times for severe weather warning helping meet NWS goals Track path of severe weather more accurately Greatly enhanced hurricane monitoring GOES-R Temporal Comparison of animation sequences of severe thunderstorm over western Kansas. Movies at 30, 15, 5 and 1 minute intervals. While 5 minute interval imaging is routine for GOES-R, special imaging like this is possible at 1 minute intervals or less at 4(ABI) to 30 (HES-VNIR) times better spatial resolution than today. Spatial 1 Km to 250 m Hurricane Erin 09/09/01 ~1530 Z •Note the detail in the eye wall (you can see up its side), improving the resolution of visible imagery provides enhanced ability to analyze a cloud field GOES-8: ~1 km Spectral - ABI Simulated GOES-R Imagery Intra-satellite Intra-satellite Spectral Intra-satellite Observe Phenomena With Greater Information Content Simulated GOES-R Multi-channel Product Intra-satellite Water And Wet Ground Middle Level Moisture Boundary Between Dry And Unstable Air Cirrus Cloud Severe Thunderstorms New products based on mathematical analysis of multi-channel images – every 5 minutes or less! Different Combinations Of Mathematically Derived Images Highlight Different Features: Adaptive Observing Leading To Adaptive Analysis The phenomena being analyzed defines the spectral, spatial and temporal requirements of the satellite observing system. For satellite applications that employ animation that means that different applications and procedures depending on scale Roles of various sensors (Polar are basically 4 hr repeat; since Geo may do adaptive observing, delta time may be varied to optimize strategy) Cloud Base, Top and Layer related products – Daytime cloud base & top stereo ABI and/or HES-VNIR – VIIRS with GOES imagers every 4 hrs shadow from ABI or HES-VNIR, VIIRS at 4 hrs HES-VNIR reflectance and HES-IR radiances – Day or night cloud base Derived HES-IR, CrIS, IASI augmented by GPS for stability, ABI surface temperature – daytime HES-VNIR for moisture in cloudy regions over land) – Day or night cloud top Multispectral slicing methods – HES-IR and ABI every 5 to 60 minutes global – CrIS and VIIRS every 4 hours global – Cloud layer high resolution imager within sounder using n* type methods (requires model baseline), cloud phase with ABI to delineate ice from water cloud, stereo improvement and limited layer thickness using ABI and HES-VNIR Adaptive observational needs of NWP and nowcasting will help define sensor activity and application in GOES-R era Atmospheric Motion Products Cloud and plume motion vectors (Cloud height from prior chart) – ABI 5-min interval for hemispheric cmv’s Rapid scan for special applications – HES-VNIR (10 to 36x improvement in spatial resolution over ABI) < 300 m and 1 min intervals for severe weather and hazard applications Moisture motion vectors – VIIRS in polar regions - need interim polar sounder & imager gap filler – Routine ABI @ 5-min intervals Hemispheric and local scale – Routine HES-IR @ 30-60 min intervals Hemispheric and local scale – Improved vertical definition for certain applications with HESVNIR and ABI – HES-VNIR @ 300 m and ~10 min intervals moisture motion in convective boundary layer in synergy with HES-IR Ocean surface winds – CMIS @ 4 hr intervals Viewing Perspective, dt and l, determine what we see Differences in scattering as a function of sun-scatterer-detector geometry allow for a variety of atmospheric, land, costal zone and ocean applications (think of MISR) Stereo cloud height determinations: accuracy is in large part a function of spatial resolution (shadows can also provide exceptionally accurate cloud height depending on time of day and viewing geometries) Exceptional CMV’s (u, u', v, v', w') in complex situations: potential for nearly 50 times higher resolution than today (150m vs 1000m) and over 10 times higher than GOES-R’s ABI (150m vs 500m) Pre-cumulus moisture field and its changes in time Adaptive observational needs of NWP and nowcasting will help define sensor activity and application in GOES-R era Atmospheric Profile Products Resolutions of various sensors CrIS/IASI 4 hr T and H2O profiles, globally, 14 km clear and above cloud top GPS/OS – very accurate profiles above tropopause around 100 km spatial ATMS/AMSU 4 hr T and H2O, globally, ~30 km through cloud over ocean HES-IR 15 min to hourly T and H2O profiles, clear and above cloud HES-VNIR mesoscale TCM (total column moisture), 300m best used in conjunction w/ HES-IR and ABI (surface temperature) Sounding applications should be in synergy Global scale analysis and modeling – 4 hr radiances and GPS OS Regional scale modeling – 1-4 hr sounding/radiances and GPS OS Local scale/ mesoscale for severe storm prediction – <hourly radiances, sounding, with GPS OS Surface parameters from sounder (in conjunction w/ VIIRS and ABI) – SST 4 hr over open ocean and hourly to 5 minutes* over coastal waters – Surface Temperature and moisture hourly to 5 minutes* – (*cloud cover, temporal and spatial requirement play crucial role) Increased Sounder Coverage 60 Minutes GOES I & N GOES R HES Improved Clear Radiance Data from Combining Imager and Sounder Observations RMS radiance differences between true clear radiance and cloud re-moved clear column Note improvement in sounder clear column radiance estimation when higher resolution imager data is used in synergy with sounder data (Smith et al. 2003) Inter-System GOES-R, NPOESS and other satellites • A major challenge that will be faced when trying to achieve such high resolution will be combining information from sensors with highly differing horizontal resolutions. For example IR sounders in the 2010 era will have horizontal resolutions around 10 km, while GPS at the surface represents a traverse across around 200 km. • Another challenge will be NWP models ability to handle clouds and moisture so that the entire radiance information is utilized. • Collard and Healy, QJRMS, 129, p.2741, 2003 Nowcasting requires detailed information on mesoscale thermodynamic structure of atmosphere, cloud type and vertical wind shear Convection and Severe Weather Important for Nowcasting Convection and Severe Weather • Vertical wind shear • Evolving instability field • Strength of storm produced cold pool • Updraft strength • Anvil characteristics Development Temperature structure • Storm-environment interaction • Cloud top rotation • Storm damage Vertical shear – ABI (cloud motion) – HES-IR (moisture motion) – HES-VNIR (cloud and moisture motion) Evolving instability field – ABI (surface heating) – HES-IR (instability and surface heating) – HES-VNIR (detailed moisture field) Cold pool production – HES-IR Updraft strength – ABI (IR top temperature) – ABI and HES-VNIR (overshooting top height) – Above with HES-IR (updraft efficiency) Anvil characteristics & storm environment interaction – ABI (growth and detailed upper level atmospheric motion and water vapor behavior) – HES-IR and VNIR (as ABI but with better spectral definition) Rotating overshooting top – ABI and HES-VNIR Storm damage – HES-VNIR The development and evolution of deep convection GOES-R: Unique in spectra, space and time The spatial and temporal domains of the phenomena drive the spectral needs as a function of space, time, and signal to noise. Nowcasting severe convection requires frequent imaging and sounding that can only be provided by geostationary satellites. GOES Lightning Mapper New NOAA instrument – Severe storm warning times – Lightning danger alerts – Disaster team response – Nitrogen production Hemispheric or CONUS Coverage Parameters – 10 km spatial resolution (1 km goal) – 1 km vertical resolution Increased coverage over oceans and lands – Currently No Ocean Brightness Temperature (K) Detection of Temperature Inversions Possible with Interferometer Texas Spikes down cooling with height (No inversion) Ontario Spikes up warming with height (low-level inversion) Wavenumber (cm-1) Detection of inversions is critical for severe weather forecasting. Combined with improved low-level moisture depiction, key ingredients for night-time severe storm development can be monitored. Ocean Color Products Coastal: At-sensor radiances for ocean color products including chlorophyll, CDOM, suspended matter, and bottom properties HES-VNIR optimal for multiple cloud free views /day @ <300 m – Hourly best for modeling coastal ocean dynamics and currents by feature tracking – Tasked by ABI defined “cloud free” (%TBD) FOV every 5 min VIIRS Not optimal, spatial, spectral or temporal for coastal zone – 4 hourly fixed views @ 1000 m not adequate to resolve tidal related features – May not be cloud free FOV – Complex features and atmospheric correction best resolved by HES (different satellites-solar-atmosphere scattering angles) Selected commercial high-spatial resolution data may be available Open Ocean (chlorophyll major constituent) VIIRS 4 hourly fixed views @ 1000 m appears adequate HES-VNIR may be tasked for selected cloud free views @ <300 m – Tasked by ABI defined cloudy FOVs Selected international data may be available Example: Use ABI Data to Task HES VNIR GOES-8 loop from 1615 to 2345: this loop illustrates the changes that occur in the cloud field after the MODIS pass and the need to dynamically task HES. Despite increasing cloud cover, the Florida Bay and Northern Keys could be successfully imaged over several hours which will allow for observations of ocean color as well as changes due to tidal effects. Florida Bay Northern Keys Then along came Floyd It will be important to monitor such disasters hourly at very high resolution as will be available from HES’VNIR capability Ocean color showing result of flooding interacting with pig farms. You want to be able to make daily cloud free images of this consequence of a natural disaster immediately and blend with SST, ocean currents and other information. Climate products require long term, stable and accurate sensor measurements GOES-R total HES can serve as stable reference basis for other satellites (operational polar and other LEO) – Contiguous and high resolution spectral measurements required for inter-calibration Spectral flexibility (adaptability) allows for spectral matching with other systems’ instrumentation GOES-R total HES is a baseline GOES-R ability to track diurnal cycle – Simultaneity with LEO constellation – Contribution to GPM (Global Precipitation Mission) High spectral resolution radiance matching for scenes over long time periods by one fixed system at the same viewing angle and from the same altitude (Goody concept) – Detect and monitor long term changes (trends) in water vapor and other gasses GPS - 1st order measurement Climate products require long term, stable and accurate sensor measurements GOES-R ability to track diurnal cycle spectrally with both ABI and HES! Hourly Vis Average Animation for 31 Days in January 2001 over Australia Seasonal to Interannual Range - are we missing something? Hourly Vis Average Animation for 31 Days in May 2001 over Western Pacific Seasonal to Inter-annual Range: multi information Daily cloud free SST anomaly from cloud cleared GOES IR data (1998) showing return current with la Nina Hourly Average Animation for 31 Days in May 2001 over Western Pacific Seasonal to Inter-annual Range: multi information Daily cloud free SST anomaly from cloud cleared GOES IR data (1998) showing return current with la Nina Aerosol Products Primary VIIRS/APS global product every 4 hours Marine environment – ABI adds more frequent observations for variation as function time – HES VNIR provides more frequent and @ <300 m resolution (less cloud effects) – Moisture effect on aerosols (particle size) – Water leaving radiance for coastal water algorithms – CrIS/ 4 hr and HES-IR/ 1 hr provides better definition of moisture profile at 4-10-12 km scale Over land – ABI adds more frequent observations for variation as function time – HES VNIR provides more frequent and @ 300 m resolution (less cloud effects) – Moisture effect on aerosols (particle size) – CrIS/ 4 hr and HES-IR/ 1 hr provides better definition of moisture profile at 4-10-12 km scale Aerosol Products – Dust storm during daylight Filling the gaps between 4 hourly APS Aerosol Products – Dust storm day and night Filling the gaps between 4 hourly APS Rapid Response!! Catastrophic Products Catastrophic (on demand): utilize baseline information – every 5 min update required Accurate plume location and tracking – ABI @ 1 min – HES VNIR @ 1-5 min w/ <300 m resolution (less cloud effects) Dual satellite HES VNIR better plume depth HES-IR 15 min characterization of moisture profile and trace gases Damage area identification and possible assessment Fire and Plumes, as below, can be rapidly detected and assessed So can other type plumes •VIIRS every 4 hrs at various resolutions •ABI every 1 to 5 minutes at various resolutions •HES often, depending on aerial extent, at 300 meters or less This damage was due to a tornado, it could have occurred over a similar or larger area due to explosions from various causes. Do you want to wait for conventional monitoring methods to begin damage assessment? With HES you can view immediately with exceptionally high spatial, spectral and temporal resolutions. LaPlata tornado damage path at 120 m resolution LaPlata tornado damage path at 240 m resolution LaPlata Tornado Damage – 1 from EO-1 30 m 60 m 120 m 240 m The Satellite System of the GOES-R Era Will Lead To Improvements . . Improved prediction accuracy from improved observations – Observe phenomena with greater clarity – Observe phenomena with greater information content – Observe phenomena with greater frequency Observe the previously unobserved Particularly when: we capitalize on the adaptive observing nature of the geostationary system in synergy with the fixed polar system