Units¶
| Instrument | Data | Response function | Loci (Data/response) | DEM output units |
|---|---|---|---|---|
| AIA | DN s$^{-1}$ px$^{-1}$ | DN s$^{-1}$ px$^{-1}$ cm$^5$ K$^{-1}$ | cm$^{-5}$ K$^{-1}$ | cm$^{-5}$ K$^{-1}$ |
| XRT | DN s$^{-1}$ px$^{-1}$ | DN s$^{-1}$ px$^{-1}$ cm$^5$ K$^{-1}$ | cm$^{-5}$ K$^{-1}$ | cm$^{-5}$ K$^{-1}$ |
| NuSTAR | counts s$^{-1}$ | counts s$^{-1}$ cm$^3$ K$^{-1}$ | cm$^{-3}$ K$^{-1}$ | cm$^{-5}$ K$^{-1}$ |
| NuSTAR response*area | counts s$^{-1}$ | counts s$^{-1}$ cm$^5$ K$^{-1}$ | cm$^{-5}$ K$^{-1}$ | cm$^{-5}$ K$^{-1}$ |
AIA processed in IDL with aia_prep, /normalize
Area in cm$^2$ determined from AIA images
XRT processed in IDL with xrt_prep, then adjusted for exposure here
XRT was summed on board by 16 pixels! So to get the true per-pixel values, divide the XRT-prepped data by 16
NuSTAR data 20s cadence images, adjusted such that pixel values are counts/s
Response Functions¶
trmatrix2=np.zeros((len(aia_tresp_logt),9))
trmatrix2[:,:6]=aia_trmatrix
trmatrix2[:,6]=xtrint
trmatrix2[:,7]=nstrint_lo
trmatrix2[:,8]=nstrint_hi
og_trmatrix=trmatrix2
pickle.dump(og_trmatrix,open('aia_xrt_nustar_trmatrix_091220.p','wb'))
trmatrix2=np.zeros((len(aia_tresp_logt),9))
trmatrix2[:,:6]=aia_trmatrix
trmatrix2[:,6]=xtrint
trmatrix2[:,7]=nstrint_lo*area #run cells in Data section first
trmatrix2[:,8]=nstrint_hi*area
full_trmatrix=trmatrix2
Load data¶
XRT¶
select 10"x10" submap. Because of pixel size and source location, this ends up as >10" x ~ 10" or 3x2 px
AIA¶
tidx=single_time_indices(dfaia,obstime)
tmaps=[dfaia.maps[idx] for idx in tidx]
type(tmaps[0]),tmaps[0].meta['date-obs']
(sunpy.map.sources.sdo.AIAMap, '2020-09-12T20:40:23.12')
all_six_AIA(tmaps)
cmask=make_contour_mask(211,submap=tmaps[4],contour=[80],plot=True)
Could not make difference image, using peak flare image only 1
npix=np.product(cmask.shape)-np.sum(cmask)
npix
9
ROI based on AIA contour
area=npix*(arcsec_to_cm(tmaps[0].meta['cdelt2']*u.arcsec))**2
print(area)
1.724874971917349e+16 cm2
# Correct the AIA data for the degradation
cor_data=datavec/degs
print(datavec)
print(cor_data)
# Get into DN/s/px for the DEM stuff
adn_in=cor_data/durs
print('DN/s/px: ',adn_in,'\n')
[8.358437, 73.33333333333333, 989.6666666666666, 1104.0, 616.5555555555555, 14.11111111111111] [ 9.25448016 144.58598185 1337.50868497 2239.15511161 1532.82583401 81.93785425] DN/s/px: [ 3.18895019 49.83413958 668.69549728 1119.78135601 528.34382926 28.24647479]
NuSTAR¶
lowE,highE=nustar_dem_prep(bottom_left,top_right,how='sum') #contour option not activated yet..
lowE,highE #total counts/s in each channel
(0.026794877, 0.0027598343)
put all the data together, calculate error
dn_in=np.hstack([adn_in,xdnspx,lowE,highE])
dn_in
array([3.18895019e+00, 4.98341396e+01, 6.68695497e+02, 1.11978136e+03,
5.28343829e+02, 2.82464748e+01, 3.60276434e+00, 2.67948769e-02,
2.75983429e-03])
aia_err=generate_errors(0,0,6,dn_in[:6])
aia_err
array([ 2.7989203 , 9.98465407, 72.27363412, 116.69425551,
57.07644205, 5.06830841])
#assume 1% error for the rest:
edn_in=np.hstack([aia_err,0.1*np.copy(dn_in)[6:]])
edn_in
array([2.79892030e+00, 9.98465407e+00, 7.22736341e+01, 1.16694256e+02,
5.70764421e+01, 5.06830841e+00, 3.60276434e-01, 2.67948769e-03,
2.75983429e-04])
DEM calculations¶
# What temperature binning do we want for the DEM ?
temps=np.logspace(5.6,6.8,num=42)
dtemps=([temps[i+1]-temps[i] for i in np.arange(0,len(temps)-1)])
mlogt=([np.mean([(np.log10(temps[i])),np.log10((temps[i+1]))]) \
for i in np.arange(0,len(temps)-1)])
Compare everything¶
- AIA only, selfnorm
- AIA + XRT selfnorm
- AIA + XRT - EMloci min
- AIA + 2*XRT - EMloci min
- AIA + XRT + NuSTAR - selfnorm
- AIA + XRT + NuSTAR - EMloci min
- AIA + 2*XRT + NuSTAR - EMloci min
- AIA + NuSTAR - selfnorm
- AIA + NuSTAR - EMloci min
DEM Comparison¶
Parameter grid search¶
aia_contours=[70,75,80,85] #also no large correlation
xrt_submap_radii=[5,10,15] #nustar radii no effect really
xrt_facs=[.5,1,1.5,2,2.5,3] #multiplicative factor for data
nustar_facs=[.01,.1,1,2,5,10,15,25,50]
nustar_areas=['AIA','XRT'] #which instrument does NuSTAR take the area measurement from
initial_weightvals=['guess','loci_min',6,6.5]
results.describe()
| index | aia_area | aia_contour | aia_mask_channel | aia_ratio | chisq | nustar_fac | nustar_submap_radius | xray_ratio | xrt_area | xrt_fac | xrt_submap_radius | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| count | 5184.0 | 5.184000e+03 | 5184.000000 | 5184.0 | 5184.000000 | 5184.000000 | 5184.000000 | 5184.0 | 5184.000000 | 5.184000e+03 | 5184.000000 | 5184.000000 |
| mean | 0.0 | 2.491486e+16 | 77.500000 | 211.0 | 0.157363 | 37.708994 | 12.012222 | 55.0 | 0.764704 | 2.613507e+18 | 1.750000 | 10.000000 |
| std | 0.0 | 1.005133e+16 | 5.590709 | 0.0 | 0.129971 | 3.660444 | 15.569877 | 0.0 | 0.204740 | 1.857341e+18 | 0.853995 | 4.082877 |
| min | 0.0 | 1.341569e+16 | 70.000000 | 211.0 | 0.030099 | 16.182034 | 0.010000 | 55.0 | 0.153513 | 5.407255e+17 | 0.500000 | 5.000000 |
| 25% | 0.0 | 1.629049e+16 | 73.750000 | 211.0 | 0.064819 | 36.945918 | 1.000000 | 55.0 | 0.680621 | 5.407255e+17 | 1.000000 | 5.000000 |
| 50% | 0.0 | 2.395660e+16 | 77.500000 | 211.0 | 0.106633 | 38.366676 | 5.000000 | 55.0 | 0.804067 | 2.253023e+18 | 1.750000 | 10.000000 |
| 75% | 0.0 | 3.258097e+16 | 81.250000 | 211.0 | 0.184983 | 39.651748 | 15.000000 | 55.0 | 0.929853 | 5.046771e+18 | 2.500000 | 15.000000 |
| max | 0.0 | 3.833055e+16 | 85.000000 | 211.0 | 0.579205 | 43.810567 | 50.000000 | 55.0 | 1.010971 | 5.046771e+18 | 3.000000 | 15.000000 |
good_results=results.where(np.log10(results.ratio_ratio)>np.log10(.85))
good_results=good_results.where(np.log10(good_results.ratio_ratio)<np.log10(1.15))
good_results['ratio_sum']=good_results.aia_ratio+good_results.xray_ratio
good_results.sort_values(['ratio_sum'],ascending=False,inplace=True)
good_results[['aia_ratio','xray_ratio','ratio_ratio','ratio_sum']].head(10)
| aia_ratio | xray_ratio | ratio_ratio | ratio_sum | |
|---|---|---|---|---|
| 1562 | 0.543496 | 0.606009 | 0.896845 | 1.149505 |
| 1560 | 0.543496 | 0.606009 | 0.896845 | 1.149505 |
| 2920 | 0.552761 | 0.590615 | 0.935908 | 1.143376 |
| 2922 | 0.552761 | 0.590615 | 0.935908 | 1.143376 |
| 4138 | 0.520980 | 0.607123 | 0.858113 | 1.128103 |
| 4136 | 0.520980 | 0.607123 | 0.858113 | 1.128103 |
| 344 | 0.562062 | 0.558999 | 1.005480 | 1.121061 |
| 346 | 0.562062 | 0.558999 | 1.005480 | 1.121061 |
| 266 | 0.531715 | 0.588653 | 0.903274 | 1.120369 |
| 264 | 0.531715 | 0.588653 | 0.903274 | 1.120369 |
plot_results=good_results.drop(columns=['aia_mask_channel','initial_weights','nustar_area','dem','dn_in','dn_reg','edem','elogt','mlogt','nustar_submap_radius','selfnorm','ratio_ratio','ratio_sum']).head(20)
best=results.loc[good_results.head().index[0]] #first one...
best
index 0 aia_area 3.06644e+16 aia_contour 75 aia_mask_channel 211 aia_ratio 0.543496 chisq 25.4777 dem [[15043975307.596079, 341417690159.0855, 62923... dn_in [[8.931356792940631, 141.21778568461554, 1307.... dn_reg [[4.436821585832275, 17.08305879154923, 1335.5... edem [[2170432409.9255633, 49491910394.78156, 90879... elogt [[0.03899325338481579, 0.024370783365509772, 0... initial_weights 6 mlogt [[5.614634146341463, 5.64390243902439, 5.67317... nustar_area AIA nustar_fac 15 nustar_submap_radius 55 selfnorm False xray_ratio 0.606009 xrt_area 5.40725e+17 xrt_fac 2 xrt_submap_radius 5 ratio_ratio 0.896845 Name: 1562, dtype: object
Correrlation heatmap
