nrlmsis 2.0: new formulation, new data...occul‐ tation (o 2) map tables (~cira‐86) misc rocket...

15
NRLMSIS 2.0: New Formulation, New Data Acknowledgement: This work was supported by the Chief of Naval Research and by NASA Many, many thanks to the data providers! John Emmert, Doug Drob, David Siskind Space Science Division, Naval Research Lab Mike Picone Voluntary Emeritus Program, Naval Research Lab 1) NRLMSISE00 History and Formulation 2) NRLMSIS 2.0 Data and Formulation 3) NRLMSIS 2.0 Results

Upload: others

Post on 25-Jan-2021

2 views

Category:

Documents


0 download

TRANSCRIPT

  • NRLMSIS 2.0: New Formulation, New Data

    Acknowledgement: This work was supported by the Chief of Naval Research and by NASA

    Many, many thanks to the data providers!

    John Emmert, Doug Drob, David SiskindSpace Science Division, Naval Research Lab

    Mike PiconeVoluntary Emeritus Program, Naval Research Lab

    1) NRLMSISE‐00 History and Formulation

    2) NRLMSIS 2.0 Data and Formulation

    3) NRLMSIS 2.0 Results

  • NRLMSIS History• The Mass Spectrometer and Incoherent Scatter radar model (MSIS) was created in 

    1977 by Alan Hedin at Goddard Space Flight Center, based in large part on Atmospheric Explorer data.

    • It grew out of a 1974 statistical model of Ogo 6 mass spectrometer data.• MSIS originally represented the upper thermosphere. Upgrades followed:

    • 1983: Rocket data, extended to lower thermosphere• 1986: DE‐2 data, atomic nitrogen added, expanded formulation• 1990: Extended to ground

    • After Alan Hedin retired from NASA in 1995,  Mike Picone of NRL’s Space Science Division continued development of the model with Alan’s assistance.

    • The current version, NRLMSISE‐00, added mass density from satellite drag, O2 data from solar occultation, and a new “anomalous O” species above 500 km.

  • MSIS papers are top 3 cited papers in JGR Space Physics

  • MSIS Applications• Prediction of geophysical conditions at specific times• Distilled view of the historical observational record• Benchmark for assessing new measurement techniques and first‐principles models

    • Boundary and initial conditions for first‐principles models• Interpolation among sparse observations• Attribution of observed variations• First guess (Bayesian prior) for measurement retrievals• Background conditions for other models (e.g., wave propagation)

    Temperatures at 90 km over Mohe, China (54N, 122E) from MSIS (left) and a Meteor Radar (right). From Liu et al., JGR, 2017

  • NRLMSISE‐00 OperationGTD7 ‐ Gets Temperature and Density: GTD7(IYD,SEC,ALT,GLAT,GLONG,STL,F107A,F107,AP,MASS,D,T)

    Input:IYD YEAR AND DAY AS YYDDD (day of year from 1 to 365 or 366; year ignored)

    SEC UNIVERSAL TIME (s); should be consistent with GLONG and STL

    ALT GEODETIC ALTITUDE (km)

    GLAT GEODETIC LATITUDE (degrees)

    GLONG GEODETIC LONGITUDE(degrees)

    STL LOCAL APPARENT SOLAR TIME (hours)

    F107A 81 day AVERAGE OF F10.7 FLUX (centered on day DDD)

    F107 DAILY F10.7 FLUX FOR PREVIOUS DAY

    AP 7‐element array: Daily Ap, ap(t), ap(t‐3h), ap(t‐6h), ap(t‐9h), Ap(12‐33h), Ap(36‐57h)First element used  when SW(9) = 1, all elements used when SW(9) = ‐1

    MASS MASS NUMBER; 48 for all, 0, for temperature, 1 for H, 2 for He, 14 for N, etc. 

    Output:1 2 3 4 5 6 7 8 9

    D = Number Density (cm‐3) He O N2 O2 Ar

    (g/cm3)

    H N Hot O

    T = Temperature Tex T(z)

  • Asymptotic temperature profile defined by exospheric temperature Fully mixed hydrostatic equilibrium below ~100 km Diffusive equilibrium (~species hydrostatic equilibrium ) above ~200 km Approximate hydrostatic equilibrium between 100 and 200 km Temperature profile constructed so that the integral can be computed in closed‐form

    Number Density

    O2N2ON

    HeH

    Solar Min

    Solar Max

    Diff

    usiv

    e Eq

    uilib

    rium

    Fully Mixed

    Asymptotic Temperature

    00

    00

    1ln ln ln mgTn n dT k T

    00

    ,00

    1ln ln 1 ln ii im gTn n d

    T k T

    NRLMSISE‐00 Physical Constraints

    Ref. Ideal gas Hydrostatic

    Ref. Ideal gas Hydrostatic(w/thermal diff)

  • NRLMSISE‐00 Formulation Vertical temperature profile (17 parameters)

    • Cubic splines in 1/T below 120 km (14 parameters)• Bates profile above 120 km (Tex, T120, )

    Species density parameters (8 per species)• Reference density (1)• Mixing ratio relative to N2 (1)• Turbopause height (1)• Corrections for dynamic flow and chemistry  (5)

    Expansion of vertical parameters (up to ~140 per param.)• Associated Legendre fns in latitude (up to degree 6)• Polynomials in daily and 81‐day average solar activity

    (F10.7) up to order 2• Intra‐annual harmonics up  to semiannual• Local time harmonics up to terdiurnal (migrating tides)• Time history of geomagnetic activity (ap index) via heating function• Longitude harmonics up to order 1 • Universal Time harmonics up to order 1

    Total number of nonzero model parameters: ~1280 Limitation: Important nonmigrating tides not included.

    120 120

    120 120

    exp

    Shape factorex ex

    ex

    T T T

    T T T

    Em

    mer

    t, A

    dv. S

    pace

    Res

    ., 20

    15

  • NRLMSISE‐00 Data

    Satellite Drag: Orbit‐derived, accelerometer

    ()

    Alti

    tude

    (km

    )0

    100

    20

    0

    3

    00

    400

    500

    600

    1960 1970 1980 1990 2000

    Mass Spectrometer

    (T, ni)

    Incoherent Scatter Radar (T)

    ISR (T)

    Occul‐tation(O2)

    MAP tables (~CIRA‐86)

    Misc RocketISR (T)

    Reanalysis (T, P)

    Lidar (T)

    Infrared

     (T, n

    i)

    Ultraviolet (T, n

    i)

    Microwave (T)

    Occultatio

    n (T)

    New Data for NRLMSIS 2.0:Orbit, accelerometer, ISR, and:

  • New Data• NOAA meteorological reanalyses in the troposphere/stratosphere.

    • Extensive new temperature and composition data in the mesosphere and lower thermosphere (incl. TIMED/SABER, Aura/MLS, ACE, AIM/Sofie, Odin/OSIRIS, Lidar).

    • Extensive new orbit‐derived and accelerometer densities, UV remote sensing data, and ground‐based FPI temperaturesin the thermosphere.

    1) NRLMSISE‐00 History and Formulation2) NRLMSIS 2.0 Data and Formulation3) NRLMSIS 2.0 Results

  • NRLMSIS 2.0: Formulation Changes

    NRLMSISE‐00 NRLMSIS 2.0

    Temperature Profile Continuity C1 Continuous C2 Continuous

    Transition from fully mixed to diffusively separated

    Geometric average of mixed and diffusive profiles

    Height‐dependent effective mass

    O profile near and below peak

    Hyperbolic tangent (logistic) function

    Above 85 km: Modified Chapman function

    Below 85 km: Cubic splines decoupled from T

    Thermal Diffusion Applied to H, He, Ar

    Not used: Incompatible with static representation, 

    inadequately supported by data

    Fourier expansion fitting parameters Amplitude and Phase Sine and cosine coefficients

    Altitude representationGeopotential height differences  between 

    geometric reference heights

    Internally uses global geopotential height function

  • Transition from Fully Mixed to Diffusively SeparatedNRLMSISE‐00 NRLMSIS 2.0

    260

    220

    180

    140

    100

    60

    20

    1

    2828.95

    AA Ad m

    s

    n n n

    Am

    0

    00

    0

    ln ln ln1 R R

    effH

    MgT Rn n dT k T e

    Mixed nmDiffusive ndMerged

    log nN2

    Ref. Ideal gas Hydrostatic Topside Corr.(O2, He, H)

    Effective Mass (Meff) transitions from 28.9 amu to species massEffective Mass (Meff) transitions from 28.9 amu to species mass

    Merged profile: weighted averageof diffusive andmixed profiles

  • 12

    Problems with merged profile in NRLMSISE‐00 (fixed in 2.0)

    NRLMSISE‐00 Helium Effective Mass Profile

    Negative effective mass Density increases with height!

    Discontinuities in effective mass Density profile not C1 continuous.Discontinuities in effective mass Density profile not C1 continuous.

    Geo

    pote

    ntia

    l H

    eigh

    t (km

    )

  • NRLMSIS 2.0: Test fits of O profile (5N, Day of year 150)

    Daytime (14 LT) Nighttime (2 LT)

    Species Hydrostatic Equilibrium with Variable Mass

    Modified Chapman Layer

    Cubic B Splines (decoupled from T)

    0

    00

    0

    ln ln ln expeff CC

    M Tgn n d Ck T T H

    -- NRLMSISE-00 -- SABER Climatology-- Fit with MSIS above 120 km, SABER below 80 km

    Ref. Hydrostatic Ideal gas Chapman term

  • 14

    NRLMSIS 2.0 Fortran Code Goals• Fortran 90• Compatible with gfortran compiler• OpenMP compatible• Speed comparable to or faster than NRLMSISE‐00

    Remaining Tasks• Add geomagnetic activity• Finalize model parameters• Code and speed enhancements• Documentation

    Planned Enhancements for NRLMSIS 2.1• Use other solar activity inputs besides F10.7• Add nitric oxide (NO)• Improve geomagnetic activity dependence

  • 17

    NRLMSIS 2.0 Summary Arguments: Position, time, solar irradiance, geomagnetic activity Output: T(z), Tex (K); N2, O2, O, He, H, Ar (cm‐3);  (g cm‐3) Domain: Ground to exosphere 

    Formulation Changes Include: • Seamless transition from fully mixed to diffusive separation• New O profile, down to 50 km• C2 continuity in temperature profile

    New Data Include: • NOAA meteorological reanalysis data in the lower atmosphere.• Extensive new temperature and composition data in the mesosphere 

    and lower thermosphere.• Extensive new orbit‐derived an accelerometer densities, UV remote 

    sensing data, and ground‐based FPI temperatures in the thermosphere.

    Interested in using the new model (version 1.95)?Contact John Emmert ([email protected]) or Doug Drob ([email protected])