46
MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell, R. Czerwinski, R. V. Leslie, M. Pieper, P. Rosenkranz*, J. Samra, D. H. Staelin*, C. Surrussavadee ‡, K. Wallenstein, & D. Zhang MIT Lincoln Laboratory * Research Laboratory of Electronics at MIT ‡ Prince of Songkla University IGARSS 2011: Vancouver, Canada 28 July 2011 This work was sponsored by the National Oceanic and Atmospheric Administration under contract FA8721-05-C-0002. Opinions, interpretations, conclusions, and recommendations are those of the author and are not necessarily endorsed by the United States Government.

MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

Embed Size (px)

Citation preview

Page 1: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIT Lincoln Laboratory

MIS IGARSS11-1RVL 9/15/2010

An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders

W. J. Blackwell, R. Czerwinski, R. V. Leslie, M. Pieper, P. Rosenkranz*, J. Samra, D. H. Staelin*, C. Surrussavadee ‡, K.

Wallenstein, & D. Zhang

MIT Lincoln Laboratory* Research Laboratory of Electronics at MIT

‡ Prince of Songkla University

IGARSS 2011: Vancouver, Canada

28 July 2011

This work was sponsored by the National Oceanic and Atmospheric Administration under contract FA8721-05-C-0002. Opinions, interpretations, conclusions, and recommendations are those of the author and are not necessarily endorsed by the United States Government.

Page 2: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-2RVL 9/15/2010

MIT Lincoln Laboratory

Outline

• Overview

• Physics

• Retrieval Approach– Neural Networks– Radiative Transfer

• Training Datasets

• Expected performance

• Summary

Page 3: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-3RVL 9/15/2010

MIT Lincoln Laboratory

Atmosphere EDR Suite

• Atmospheric Vertical Temperature Profile (AVTP) – Kelvin– Lower Atmospheric Sounding (Surface to 10 mb)– Upper Atmospheric Sounding (10 mb to ~0.01 mb)

• Atmospheric Vertical Moisture Profile (AVMP) – MMR g/kg• Atmospheric Pressure Profile (APP) – millibar• Total Water Content (TWC) - kg/m2 or mm in a 3-km vertical segment

• Total Integrated Water Vapor (TIWV) - kg/m2 or mm (a.k.a., precipitable water)

• Precipitation Rate/Type (PRT) – mm/hr and types: rain or ice• Cloud Liquid Water Content (CLWC) – kg/m2 or mm• Cloud Ice Water Path (CIWP) - kg/m2 or mm

Profile Subset

2-D Field Subset

Page 4: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-4RVL 9/15/2010

MIT Lincoln Laboratory

Algorithm Simulation Methodology

Page 5: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-5RVL 9/15/2010

MIT Lincoln Laboratory

MIS Atmospheric Algorithm Methodology

• Cloud/precipitation products derived from cloud-resolving NWP models combined with multi-stream scattering models

– Global NWP runs over ~5M pixels– Multi-phase microphysical modeling

• Profile products derived from global high-resolution analysis fields

– Performance validated over many years (millions of pixels) for similar AMSU/AIRS algorithm

– Framework allows for optimization of product spatial resolution

• Neural network estimators offer accuracy/robustness/speed– Very easy to code (large infrastructure currently available)– Very easy to upgrade (simply replace coefficient file)– Very low computational burden – can run on mobile terminals

Physical Models + Stochastic Processing

Page 6: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-6RVL 9/15/2010

MIT Lincoln Laboratory

PHYSICS AND

PHENOMENOLOGY

Page 7: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-7RVL 9/15/2010

MIT Lincoln Laboratory

Microwave Scattering and Absorption

Atmospheric Transmission

Hydrometeor Mie Scattering and Absorption

Liquid water

Ice

Frequency [GHz]

Frequency [GHz]

Frequency [GHz]

Page 8: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-8RVL 9/15/2010

MIT Lincoln Laboratory

Passive Microwave Sensing of Precipitation

35 km

45 k

m

Page 9: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-9RVL 9/15/2010

MIT Lincoln Laboratory

Overview of SSMIS Channel Setand Spatial Resolutions

V = vertical pol.H = horizontal pol.R = right-hand circ.* subset in precipitation algorithm

km

Page 10: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-10RVL 9/15/2010

MIT Lincoln Laboratory

SSMIS UAS Channel Characteristics

Page 11: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-11RVL 9/15/2010

MIT Lincoln Laboratory

Temperature and Water VaporWeighting Functions

Temperature Water Vapor

45° off-nadir angle

Page 12: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-12RVL 9/15/2010

MIT Lincoln Laboratory

Upper Air TemperatureWeighting Functions

26 uT 90 deg. (tropical)

65 uT 53 deg. (polar)

Page 13: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-13RVL 9/15/2010

MIT Lincoln Laboratory

MULTILAYER FEEDFORWARD

NEURAL NETWORKS

Page 14: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-14RVL 9/15/2010

MIT Lincoln Laboratory

Neural NetworksNonlinear, Parameterized Function Approximators

Page 15: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-15RVL 9/15/2010

MIT Lincoln Laboratory

Example: Temperature Profile RetrievalAdvantages Relative to Linear Regression (LLSE)

Page 16: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-16RVL 9/15/2010

MIT Lincoln Laboratory

Advantages Relative to Linear RegressionBetter Noise Immunity and Physical Representation

Noise contribution: Component of retrieval error due only to sensor noiseAtmosphere contribution: Retrieval error in the absence of sensor noise

Page 17: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-17RVL 9/15/2010

MIT Lincoln Laboratory

RADIATIVE TRANSFER AND

SIMULATION METHODOLOGY

Page 18: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-18RVL 9/15/2010

MIT Lincoln Laboratory

Algorithm Simulation Methodology

Page 19: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-19RVL 9/15/2010

MIT Lincoln Laboratory

Radiative Transfer / NWP Interface Issues

Each level requires hydrometeor densityper drop radius

MM5

Pre

ssur

e [m

b]

Mas

s D

ensi

ty [

g/m

3 ]

Radius [mm]

Mass Density [g/m3]

graupel

snow

rain

10 mb

Sekhon-Srivastava

Marshall-Palmer

Image courtesy of Colorado State University

SSMIS(NGES)

Page 20: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-20RVL 9/15/2010

MIT Lincoln Laboratory

PERFORMANCE

VERIFICATION DATASETS

Page 21: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-21RVL 9/15/2010

MIT Lincoln Laboratory

Geographical locations of the pixels in the MM5 and NOAA88b data sets

Page 22: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-22RVL 9/15/2010

MIT Lincoln Laboratory

Mean and Standard Deviation of NOAA/MM5 Data Sets

Temperature Water Vapor

Page 23: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-23RVL 9/15/2010

MIT Lincoln Laboratory

MM5 Cloudy Data Set

Page 24: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-24RVL 9/15/2010

MIT Lincoln Laboratory

PERFORMANCE

VERIFICATION

Page 25: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-25RVL 9/15/2010

MIT Lincoln Laboratory

Precipitation Rate Retrieval Performance

Page 26: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-26RVL 9/15/2010

MIT Lincoln Laboratory

Summary of Cloud Water/Ice Retrieval Performance

Page 27: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-27RVL 9/15/2010

MIT Lincoln Laboratory

AVTP Retrieval PerformanceCloudy (40 km)

MM5 not valid at these high altitudes

Page 28: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-28RVL 9/15/2010

MIT Lincoln Laboratory

Upper Air Sounding Performance

• SSMIS UAS channels (CH20-24)

• No Doppler effects

• IGRF-11 geomagnetic model

• Multi-layer Feedforward Neural Network

• NOAA88b dataset

• SSMIS Spec:– 7-1 mb: 5 K– 0.4 mb: 5.5 K– 0.2-0.03 mb: 8 K

Page 29: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-29RVL 9/15/2010

MIT Lincoln Laboratory

AVMP Retrieval PerformanceCloudy (40 km)

• SSMIS: Greater of 1.5 g/kg or 20%

• IORDII: • 10% objective• Greater of 0.2 g/kg or 20%

(surf. to 600 mb)

Page 30: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-30RVL 9/15/2010

MIT Lincoln Laboratory

Clear-Air Atmospheric Pressure Profile Performance (40 km)

APP derived using AVTP and AVMP retrievals and surface pressure (assumed perfect)Quality-controlled global radiosondes used for ground truth

Land Ocean

Page 31: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-31RVL 9/15/2010

MIT Lincoln Laboratory

Summary

• Comprehensive, end-to-end performance assessment capability in place for all products in the Atmosphere EDR Suite

– Minimal retrieval optimization performed at this point– Clear path to requirement compliance for all products

• Flexible, modular algorithm architecture easily accommodates changes to sensor characteristics and performance

Page 32: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-32RVL 9/15/2010

MIT Lincoln Laboratory

Backup Slides

Page 33: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-33RVL 9/15/2010

MIT Lincoln Laboratory

Simulated SSMIS Pass Over CONUS

• 50.3-GHz brightness temperature

• 40-km Spatial resolution

• 2/3 CONUS HRRR – 3 km

• CCA antenna pattern

Page 34: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-34RVL 9/15/2010

MIT Lincoln Laboratory

SSMIS and AMSU Precipitation Rate Retrievals

8

Page 35: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-35RVL 9/15/2010

MIT Lincoln Laboratory

Structure of the SSMIS Precipitation Algorithm

Pixel Longitude/Latitude Brightness Temperatures

Bias correction Interpolate to fine retrieval grid

Surface classification

PCA Transform

Channel Selection Channel Selection

Spatial Perturbations

Specialized Neural Network

Surface-Classification-Dependent Weighting

Retrieved Precipitation Parameters

Channel Selection

Page 36: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-36RVL 9/15/2010

MIT Lincoln Laboratory

Radiance Simulation Methodology

• CRM = MM5 1-km saved every 15 min• RTM = multiple-stream radiative transfer

solution (TBSCAT† or TBSOI*) • Simulated NAST-M radiances• Developed and adapted MIT software to

LLGrid parallel computing facility

MM5 grid levels

Cloud Resolving Model (CRM)

Radiative Transfer Model (RTM)

Simulated Radiances

SP

AT

IAL

FIL

TE

RIN

G

“Satellite Geometry”Toolbox (MATLAB)

* Successive Order of Interaction: Heidinger A. K., et al., J. Appl. Meteor. Climatol., 2006† TBSCAT: Rosenkranz, P. W., IEEE Trans. Geosci. Remote Sens. 2002

Page 37: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-37RVL 9/15/2010

MIT Lincoln Laboratory

Histogram of Surface Pressures for the Synoptic Radiosonde Data Set

Page 38: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-38RVL 9/15/2010

MIT Lincoln Laboratory

Geographical Locations of the Pixels in the Synoptic Radiosonde Data Sets

~200,000 quality-controlled radiosondes from 2009-2010 representing all seasons

Page 39: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-39RVL 9/15/2010

MIT Lincoln Laboratory

Precipitation Rate Performance

Page 40: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-40RVL 9/15/2010

MIT Lincoln Laboratory

Precipitation Type Retrieval

Page 41: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-41RVL 9/15/2010

MIT Lincoln Laboratory

Precipitation Rate Performance Stratified by Precipitation Type

Page 42: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-42RVL 9/15/2010

MIT Lincoln Laboratory

Cloud Water/Ice Retrieval Performance

Page 43: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-43RVL 9/15/2010

MIT Lincoln Laboratory

Total Integrated Water Vapor Performance (25 km)

Page 44: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-44RVL 9/15/2010

MIT Lincoln Laboratory

AVMP Retrieval PerformanceClear-air (40 km)

• Black = Ocean• Green = Land• Blue = Global

• SSMIS: Greater of 1.5 g/kg or 20%

• IORDII: • 10% objective• Greater of 0.2 g/kg or 20%

(surf. to 600 mb)

Page 45: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-45RVL 9/15/2010

MIT Lincoln Laboratory

Total Water Content Performance

Altitude Ocean Land Global Spec. (IORDII)

surface 1.20 2.00 1.44 2.0 kg/m2

5 km 0.80 1.40 1.10 2.0 kg/m2

7.5 km 0.45 0.48 0.46 2.0 kg/m2

10 km 0.10 0.12 0.11 2.0 kg/m2

• 3-km “slabs”• 25 km resolution• cloudy MM5 dataset

Page 46: MIT Lincoln Laboratory MIS IGARSS11-1 RVL 9/15/2010 An Atmospheric Algorithm Suite Based on Neural Networks for Microwave Imager/Sounders W. J. Blackwell,

MIS IGARSS11-46RVL 9/15/2010

MIT Lincoln Laboratory

Limitations and Degradation

• Precipitation– Effects all atmos. EDRs except PRT– Nominally, atmos. EDRs will be retrieved under 1 mm/hr– Difficult to quantify 1 mm/hr, will use status flags to classify the

precipitation (e.g., “no precip.”, “stratiform”, “light convective”)– Status flags must determine if a CFOV has even one precipitation-

impacted EFOV

• Land emissivity– Properly classifying land conditions (e.g., flooded or snow-covered) will

make stratifications (i.e., a condition specific NN) more difficult to implement

– Difficult to obtain a statistically-adequate sample set

• Land elevation– Difficult to obtain a statistically significant sample set to train on– Must evaluate whether training many altitude stratifications is worth the

effort and cost