Coronal differential emission measure (DEM) maps Contact: Huw Morgan, hmorgan@aber.ac.uk This archive contains full-disk DEM data files at ~3-day time steps for the SDO era (2010-present). These are created using the 7 Extreme UltraViolet (EUV) channels of AIA/SDO. The core DEM method used is the Solar Iterative Temperature Emission Solver (SITES), developed in Aberystwyth and published in reference [1] below. The method is made more efficient by the Grid-SITES implementation, also developed in Aberystwyth and published in reference [2]. DEMs give, for each image pixel, an estimate of the emission as a function of temperature. The measured intensity in each channel at that pixel, as well as the temperature response of each channel, is used to invert the observation into the DEM. The DEM array is of dimensions [nx,ny,ntemp], where ntemp is the number of temperature bins across which the DEM is calculated. So for a given pixel ix, iy, we can extract the DEM by DEM[ix,iy,*]. The DEM is in units of cm^-5 K^-1. The temperature for these DEM arrays have regular increments in Log_10 T, as given by the logt variable (descriptions below). So take care in converting the DEM to a temperature-integrated emission. The directory structure is by year, e.g. ./2007/ . The filename format is dem_aia_yyyymmdd_hhmmss.dat For each date where a tomography map exists, there is an IDL save file (.dat) and two corresponding JPEG images (.jpg). IDL save files can easily be opened in Python, see https://docs.scipy.org/doc/scipy/reference/generated/scipy.io.readsav.html. Users should quickly check the JPEG images for any obvious errors. The JPEG image with the _dem.jpg extension is the DEM. The other JPEG image with the _fem.jpg extension is a fractional emission measure (FEM). This gives the fraction of emission, at a given temperature, compared to the total emission over all temperatures. We find this is a very simple, and more effective way, to distinguish between different types of solar regions. See reference [1] for more details. At the time of writing (2021/10) the archive extends from the start of the SDO era (2010/05) to ~end of 2018. We will endeavour to update as soon as possible. If you require DEMs for dates not stored in this archive, please get in touch. Note: - for this archive, we have used the AIA/SDO 'synoptic' data (http://jsoc2.stanford.edu/data/aia/synoptic) - for efficiency, we have rebinned the synoptic data to 512x512 - NOT described in the references, we have also applied deconvolution to the data. We hope to publish the deconvolution method soon - for convenience, we have included the corresponding HMI magnetogram image, interpolated onto the same pixel coordinates as the AIA. Note on DEM errors: - Rather than saving an error on each DEM (this takes a lot of space!), we provide the array DEMERR. This is a [ntemp,2] array, so 2 values for each temperature bin. To calculate the error for a DEM value, D, at a given pixel and temperature index itemp (e.g. itemp=10, the 10th temperature bin) you evaluate ERROR=DEMERR[itemp,0]+D*DEMERR[itemp,1] For example, to calculate the error on the whole image at temperature bin number 10, do ERR=demerr[10,0]+dem[*,*,10]*demerr[10,1] ERR will then be a [nx,ny] image, giving the error at each pixel for the DEM at that temperature. The SITES software is available in the SolarSoft library, see https://www.lmsal.com/solarsoft/ssw/packages/dem_sites/idl/ If you use these data, please cite this web archive and publications: [1] https://link.springer.com/article/10.1007/s11207-019-1525-4 [2] https://link.springer.com/article/10.1007/s11207-019-1526-3 The variables contained in each file are: DATE STRING = '2010/07/01 12:00:04' MISSION STRING = 'sdo' INSTR STRING = 'aia' TIMECOMBINE FLOAT = 0.00000 WL INT = Array[7] HDR STRUCT = -> Array[1] LOGT FLOAT = Array[43] HMI_HDR STRUCT = -> Array[1] HMI DOUBLE = Array[512, 512] FILES POINTER = Array[7] DEM DOUBLE = Array[512, 512, 43] DEMERR DOUBLE = Array[43, 2] IS_SYNOPTIC BYTE = 1 DESCRIPTION STRING = Array[14] CREDIT STRING = 'If you use these data for your own research, please cite: https://link.spr'... DATE_CREATED STRING = '2021/10/22 11:20:35' The description variable contains more detail on each variable: description=['mission = acronym of mission', $ 'instr = acronym of instrument', $ 'date = Date of observation', $ 'timecombine = number of seconds over which observations have been averaged e.g. timecombine=300s means available data from date to date+300s have been averaged. If zero only 1 observation used (for each channel)', $ 'is_synoptic = 0/1 (false/true) if synoptic data is used rather than full resolution', $ 'wl = the wavelength of the channels used for the DEM', $ 'hdr = a master data header, corresponding to one of the images used for the DEM, useful for extracting co-ordinates or other values', $ 'logt = log base 10 of the temperatures where DEM are evaluated', $ 'hmi_hdr = data header for the HMI image', $ 'hmi = HMI magnetogram, conveniently interpolated to the same pixel locations as the AIA DEM', $ 'files = array of pointers, an element for each channel, giving the filenames of the original AIA data fits files', $ 'dem = the main DEM array with dimensions [x,y,temperature]', $ 'demerr = a N_temp x 2 array. To calculate error on a DEM at temperature index i, do err=demerr[i,0]+dem[i]*demerr[i,1], where dem[i] is the DEM value at that pixel and temperature i. This is very accurate down to small DEM values.', $ 'date_created = date and time DEM file has been created']