Determination of atmospheric temperature, water vapor, and heating rates from mid- and far- infrared hyperspectral measurementsPresentation OutlineThe Far-Infrared FrontierFIRST: Far Infrared Spectroscopy of the TroposphereRetrieval Sensitivity Test Flow ChartClear-Sky Retrieval TestClear-Sky Heating RatesExtrapolating Far-IR with CloudsTest Flight on September 18, 2006: Ft, Sumner NMCloudSat/CALIPSO signalsFIRST and AIRS Cloud SignaturesClimate Model ConsiderationsConclusionsAcknowledgementsCloud Radiative Effect (CRE)Towards CLARREODetermination of atmospheric temperature, water vapor, and heating rates from mid- and far- infrared hyperspectral measurementsAGU Fall Meeting, Wednesday, December 12, 2007GC34A-02D.R. Feldman (Caltech); K.N. Liou (UCLA); Y.L. Yung (Caltech); D. G. Johnson (LaRC); M. L. Mlynczak (LaRC)Presentation Outline•Motivation for studying the far-infrared•FIRST instrument description•Sensitivity tests of mid-IR vs far-IR capabilities–Clear-sky–Cloudy-sky•Multi-instrument data comparison•Climate model considerations•ConclusionsOutlineOutline2The Far-Infrared Frontier3•Current EOS A-Train measure 3.4 to 15 μm, don’t measure 15-100 μm•IRIS-D measured to 25 μm in 1970•Far-IR, through H2O rotational band, affects OLR, tropospheric cooling rates•Far-IR processes inferred from other spectral regions• Mid-IR, Microwave, Vis/NIR•Interaction between UT H2O and cirrus clouds requires knowledge of both• Currently inferred from measurements in other spectral regionsFigures derived from Mlynczak et al, SPIE, 2002MotivationMotivationNo spectral measurementsto the right of lineFIRST: Far Infrared Spectroscopy of the Troposphere•FTS w/ 0.6 cm-1 unapodized resolution, ±0.8 cm scan length•Multilayer beamsplitter–Germanium on polypropylene–Good performance over broad spectral ranges in the far-infrared•5-200 μm (50 – 2000 cm-1) spectral range•NeDT goal ~0.2 K (10-60 μm), ~0.5 K (60-100 μm)•10 km IFOV, 10 multiplexed detectors•Cooling•Spectrometer LN2 cooled•Detectors liquid He cooled•Scan time: 1.4-8.5 sec•Balloon-borne & ground-based observationsFIRST instrumentFIRST instrument4FIRSTAIRS AIRSRetrieval Sensitivity TestFlow Chart5Model Atmosphere A priori Atmospheric State)Random PerturbationsSynthetic MeasurementRTM + NoiseA priori spectrumRTMRetrieval algorithmA prioriuncertaintyAnalyze retrieved state, spectra, and associated statisticsSensitivity testsSensitivity testsT(z)H2O(z)O3(z)CWC(z)CER(z)Clear-Sky Retrieval Test6Sensitivity testsSensitivity tests• AIRS and FIRST T(z) retrievals comparable.• FIRST better than AIRS in H2O(z) retrievals 200-300 mbar.• Residual signal in far IR seen 100-200 cm-1 → low NeDT criticalClear-Sky Heating Rates•Spectra provide information about fluxes/heating rates•Error propagation (Taylor et al, 1994; Feldman et al, In Review) can be used •Heating rate error for scenes with clouds generally higher due to lack of vertical cloud information7Heating RatesHeating RatesTropical Conditions Sub-Artic Winter ConditionsExtrapolating Far-IR with Clouds•Retrieval of T(z), H2O(z), CWC(z), CER(z) difficult with AIRS spectra•Use AIRS channels to extrapolate far-IR channels?–Depends on cloud conditions, T(Z), H2O(z)–Low BT channels from 6.3 μm band ≈ low BT channels in far-IR–High BT channels scale from mid- to far-IR–For tropics, channels with BT 250-270 K (emitting ~ 5-8 km) are complicated8CloudsCloudsTest Flight on September 18, 2006:Ft, Sumner NMAQUA MODIS L1B RGB Image9AIRS FootprintsFIRST BalloonCloudSat/CALIPSO TrackTest flightTest flightCloudSat/CALIPSO signals10Test flightTest flight• CloudSat and CALIPSO near collocation• No signal from CloudSat • CALIPSO signal consistent with FIRST residual• Instrument collocation• FIRST balloon-borne spectra• AIRS• MODIS • Residuals are consistent with clouds ~ 5 km, De ~ 60 μm11FIRST and AIRS Cloud SignaturesCloudDetected !Test flightTest flightClimate Model Considerations12•Climate models produce fields that specify mid- & far-IR spectra.•Multi-moment statistical comparisons of measured spectra and modeled spectra avoid subtle biases from data processing.–Spectral and atmospheric state spaces should be considered jointly.•Far-IR climate model analysis requires more far-IR data–Far-IR extrapolation should retain physical basis and be verified with measurements.–Agreement with CERES is only partial verification and presents a non-unique checksum•Future work to assess how spectra impart information towards climate model processes.Model evaluationModel evaluationConclusions•AIRS measures mid-IR, but far-IR is not covered A-Train spectrometers.•FIRST provides thorough description of far-IR but limited spectra are available.•FIRST clear-sky T retrievals comparable, improved UT H2O retrieval relative to AIRS–Implied cooling rate information difference is small .•Extrapolating far-IR channels good for Tb ~ 220 K, ok for Tb ~ 300 K, difficult for Tb ~250-270 K.•Multi-instrument analysis with A-Train facilitates comprehensive understanding of FIRST test flight spectra.•AIRS mid-IR spectra can validate climate models, but far-IR should not be neglected.13ConclusionsConclusionsAcknowledgements•NASA Earth Systems Science Fellowship, grant number NNG05GP90H.•Yuk Yung Radiation Group: Jack Margolis, Vijay Natraj, King-Fai Li, & Kuai Le•George Aumann and Duane Waliser from JPL•Xianglei Huang from U. Michigan and Yi Huang from Princeton•AIRS, CloudSat, and CALIPSO Data Processing Teams14Thank you for your timeThank you for your timeCloud Radiative Effect (CRE)15•CRE = TOA clear broadband flux – TOA broadband flux •CERES provides collocated measurements of CRE from broadband radiometers–Most CERES products contain multiple data-streams•AIRS L3 CRE lower than CERES CRE–Other A-Train sets (CloudSat/CALIPSO) can arbitrate differenceCloudsCloudsTowards CLARREO16•NRC Decadal Survey recommended CLARREO for–Radiance calibration–Climate monitoring•CLARREO specified to cover 200 – 2000 cm-1 with < 2 cm-1 resolution–NIST traceability requirement•Prototyped far-IR instruments provide a science and engineering test-bed for next generation of satellite instruments•Further orbital simulations required to test how mid-IR state space uncertainties appear as far-IR spectral residuals•More integrated A-train analyses