It could be the altitude: enviacc.modtran.sensor_altitude = 35786.0000 In FLAASH, this is defined in km so this would be 35786km. Did you enter this in meters?

Thank you for your help! I’m sure the sensor altitude is 35786km because the GOCI sensor is equipped in the geostationary orbit satellite. Here is the wiki about GOCI https://en.wikipedia.org/wiki/Geostationary_Ocean_Color_Imager Is there any limitation about the sensor altitude in FLAASH?

No, I do not believe it is a limit in altitude and it does look correct. This error sounds like it is unable to read something from a file. How does your spectral response (filter) function look? Can you also post the error from the journal file to see the last few processes that took place before the error?

I am sure the spectral response function is the same as the image in this link :http://222.236.46.36/page...201301/20130104.html And I convert the ASCII file to the .sli file which can be opened by the spectral library viewer in ENVI. Here is the journal.txt: Warning: first requested MODTRAN run calls for existing statistics; changing to no-statistics *****************FLAASH-SLANT**************** Spectral Sciences Atmospheric Correction Code 1998-2008 Version 01302013 ** See Instructions for Use ** ************* MAIN MENU *************** Commands: m = set up MODTRAN input file r = Run MODTRAN h = Hyperspectral data analysis a = About FLAASH (Instructions) c = Convert existing MODTRAN tape7 w = Read Warfighter run file, create initial tape5 q = Quit script command h Path for output files (except output reflectance cube) is D:\GOCI\ Enter u = Use this path or a = select Another path : script command u ************ DATA ANALYSIS MENU ************* Setup Commands: S = I/O Setup, interactive s = I/O Setup via serial queries d = new radiance Data cube a = Adjacency algorithm setup (create averaged image) q = Quit to main menu Processing Commands: i = simulate radiance Image from reflectance M = prepare MODTRAN run R = run MODTRAN WC = Channel Wavelength Correction Utility Commands: ? = menu help p = Plot pixel spectra e = display Existing reflectance file script command d Reading header (if any) from image file D:\GOCI\taihu_rad Enter u = Use this file, a = select Another file (dialog), or d = select another file (Direct filename entry) : script command u Select method for representing instrument spectral filter function. Type 1 for calculated function, 2 for a function defined by a file, 3 for external channel definitions passed in. ...type 99 to accept default from call. script command 99 ************ DATA ANALYSIS MENU ************* Setup Commands: S = I/O Setup, interactive s = I/O Setup via serial queries d = new radiance Data cube l = Load and display radiance data cube a = Adjacency algorithm setup (create averaged image) q = Quit to main menu Processing Commands: i = simulate radiance Image from reflectance v = aerosol Visibility determination k = Kaufman (ratio) visibility retrieval M = prepare MODTRAN run R = run MODTRAN WC = Channel Wavelength Correction Utility Commands: ? = menu help p = Plot pixel spectra e = display Existing reflectance file script command l Channel definitions: type 1 to use passed-in default, 2 to select manually between automatic and external-file channel definitions. script command 1 Warning: KT calculation not defined in channel selection file D:\GOCI\channel_def.txt Warning: vegetation parameters not found in channel selection file D:\GOCI\channel_def.txt Loading spectral data... 9045 blank pixels out of 18354 were found. *Not displaying RGB radiance image, cirrus cloud image and last-line spectra ************ DATA ANALYSIS MENU ************* Setup Commands: S = I/O Setup, interactive s = I/O Setup via serial queries d = new radiance Data cube l = Load and display radiance data cube a = Adjacency algorithm setup (create averaged image) q = Quit to main menu Processing Commands: i = simulate radiance Image from reflectance v = aerosol Visibility determination k = Kaufman (ratio) visibility retrieval c = Column water retrieval MSc = User-specified Column water M = prepare MODTRAN run R = run MODTRAN WC = Channel Wavelength Correction Utility Commands: ? = menu help p = Plot pixel spectra e = display Existing reflectance file script command M Select option: g = standard GUI interface v = visibility setup script command call Estimated MODTRAN run time: 0.035526826 min ACC error... IDL error: -361 Unable to apply format code F to input: "**********". in context: acc_called_t5set:read tape5 file back in to check format. calls: ACC_CALLED_T5SET ACC_HYPER2 ACC_ACC ENVI_ACC_EVENT ENVI_FLAASH_PROCESS_MANAGER IDLRTMAIN $MAIN$

I agree that the spectral filter function looks reasonable. I asked one of our engineers and they agreed that the error is typically caused by reading an unexpected ascii string. The only thing that might be unexpected that I can see is the sensor elevation. Unfortunately, I do not have access to any GOCI data to test myself. If you set the sensor to something like 700km, do you get the same error?

No,this time it works well when I set the sensor altitude to 700km though the spectrum curve is absolutely wrong.The reflectance is too high in the blue(412nm and 443nm).

The developers thought that using the lower altitude should not make much difference since there is little atmosphere beyond 100km. You can also turn off adjacency correction as it is negligible with large pixel sizes. I was able to find a GOCI dataset and process it in FLAASH. I did not think the spectra looked unusual: How did you convert to radiance?