Steve Miller

Towards NPOESS and GOES-R: Some Advanced MODIS Applications: 

Towards NPOESS and GOES-R: Some Advanced MODIS Applications Steven D. Miller, and the Satellite Meteorological Applications Team Naval Research Laboratory 14th International TeraScan Conference Nanjing, China 6 June, 2004

Slide2: 

With an ever growing constellation of satellite/sensor assets at our disposal, NRL Monterey resides at the nexus of observations, modeling, and data assimilation

Our Targeted Audience: 

Making realtime products that characterize the current and near-future environment Our Targeted Audience The User: The Challenge: The Need: The Solution: Weather analysts and forecasters Simple, comprehensive, and easy-to-interpret graphical products for rapid analysis and extraction of salient information Post-process multi-spectral and/or model fusion data in an attempt to reveal the underlying physics and present the findings in terms of false-color, value-added satellite imagery products

MODIS Near Real-Time Data: 

36 narrowband spectral channels 0.4-14 m. 1km, 500m, 250m (649 & 853 nm). 2330 km swath (cross track, whisk broom). Terra (1030, local time descending node) and Aqua (1330, local time ascending node). MODIS Near Real-Time Data EOS Terra EOS Aqua

Sample Applications Developed by NRL (Monterey California USA): 

Sample Applications Developed by NRL (Monterey California USA)

Daytime Dust Detection: 

Daytime Dust Detection GOAL: Exploit physics to enhance regions of significant dust Coloration properties Strong temperature contrasts (esp. over deserts, TSKIN > 130° F) IR split window transparency differences reveal dust signature  Combine (1-3) to form a “false-color” (RGB) enhancement that depicts dust in strong contrast to background Miller, S. D., 2003

12/12/2003 Dust Outbreak: 

12/12/2003 Dust Outbreak

Convective Cloud Analysis: 

Convective Cloud Analysis GOAL: Toward improved identification and characterization of deep convective clouds Bi-spectral to deep convection detection* Cloud top heights from IR temperature and model (COAMPS™ or NOGAPS) profile Heights converted to altimeter readings * e.g., Schmetz et al. 1997

Aircraft Contrail Enhancements: 

Aircraft Contrail Enhancements GOAL: Enable detection and tracking of flight patterns Multi-spectral (8.5,11.0,12.0 m) technique enables phase decoupling  Split-window (11.0-12.0 m) transmission contrasting conducive to contrail enhancement. Modified microphysics allowing for detection of contrails embedded within cirrus layers  Contrails appear as enhanced linear features

Conventional Imagery vs. Multi-Spectral Innovations to Cloud/Snow Detection: 

Conventional Nighttime Infrared Imagery Over the Northern Arabian Sea Conventional Daytime Visible Imagery Over Afghanistan Low Clouds? Snow Cover? Conventional Imagery vs. Multi-Spectral Innovations to Cloud/Snow Detection (Night) (Day)

The NRL/NPOESS “CONUS” Web Page: Public Access to Satellite Products: 

The NRL/NPOESS “CONUS” Web Page: Public Access to Satellite Products

Fire and Burn Scars: 

Fire and Burn Scars

Snow & Contrails: 

Snow & Contrails

Significant Dust: 

Significant Dust TEXAS NEW MEXICO MEXICO

Lightning Detection: 

Application of new model/satellite data fusion algorithms and novel data sources (lightning) result in improved deep convection analysis for aviation support. GOES_12 IR Imagery Lightning Data Overlay KEY: (+/-)=(pos/neg) Red < 30 min old Blue =30-60 min old Ground Strikes: Green < 30 min old Cyan 30-60 min old Lightning Detection Convective diagnostic used for isolating the deepest cells Real-time lightning data from Vaisala, c/o J. Cook, M. Frost, and L. Phegley (NRL) NRL’s GOES/NOGAPS-Derived Convective Cloud Top Product (in k-ft)

Toward Next-Generation Operational Observing Systems…: 

Toward Next-Generation Operational Observing Systems…

The Value of True Color…: 

The Value of True Color…

GOES-R ABI Proposed Bands: 

Current GOES Imagers MSG/Sounder MODIS/MTG/etc Slide Courtesy of T. J. Schmit (NOAA/NESDIS/ORA) GOES-R ABI Proposed Bands Green Channel Omitted from GOES-R ABI notional baseline!! * * * * * * * * * * * * 3.7 SIR Channel Omitted from GOES-R ABI notional baseline

Leveraging MODIS to Approximate the Missing Green Band: 

Examples of un-approximated (left) and synthetic (right) natural color imagery products over Afghanistan (upper) and equatorial Africa (lower). The GOES-R Advanced Baseline Imager does not include the 555 nm (green) channel  the highly visually intuitive “natural color” imaging products popularized by MODIS, SeaWiFS, etc. will not be readily available. Simple approximations involving combinations of available channels are inadequate to characterize the highly nonlinear problem. One solution: Producing 3-D look-up tables based on MODIS data, relating 469, 645, and 865 nm information to 555 nm. Interpolations made in reflectance-space to “fill in the gaps” Current Challenge: Coastal zones misrepresented in ensemble statistics. Terra-Modis natural color image of Qatar and surrounding waters. Note shallow water effects east of Qatar resulting in green coloration R/G/NIR lookup table applied to synthesize 0.555 micron channel. Many shallow water features are lost (non-unique R/G/NIR) Absolute difference between true and approximated, revealing only minor departures (<3%) in the “problem areas” Relative difference between true and approximated, revealing significant departures (>30%) in the “problem areas” QATAR Leveraging MODIS to Approximate the Missing Green Band

Conclusion: 

Conclusion MODIS enables simulation of future operational sensors, such as NPOESS/VIIRS and GOES-R Series ABI Omission of 0.555 micron channel on ABI regarded as unfortunate in light of the positive feedback received from our operational users on natural color products, although LUT mitigates problem to great extent (with exception of littoral).