Spaceweather HMI Active Region Patch (SHARP)

Motivation

[1] It is believed that parameterizations of the photospheric magnetic field can be correlated with solar activity (Falconer et al., 2002 ; Leka and Barnes, 2003a and 2003b; Schrijver 2007; Mason and Hoeksema, 2010).

[2] Collecting, storing, tracking, and analyzing data in the number of full-disk vector magnetograms necessary to follow an active region as it crosses the disk is difficult.

Data [1]: What is a SHARP?

SHARP stands for Spaceweather HMI Active Region Patch. A SHARP is a DRMS series that contains

SHARPs are calculated every 12 minutes. Often, there is more than one active region on the solar disk at any given time. Thus, SHARPs are indexed by two prime keys: time, T_REC, and HMI Active Region Patch Number, HARPNUM. There are four SHARP DRMS series:

[1] hmi.sharp_720s: definitive data with segments in CCD coordinates wherein the vector B has been decomposed into azimuth, inclination, and field,
[2] hmi.sharp_cea_720s: definitive data wherein the vector B has been remapped to a Lambert Cylindrical Equal-Area projection and decomposed into B_radial, B_phi, and B_theta,
[3] hmi.sharp_720s_nrt: near real-time data in CCD coordinates wherein the vector B has been decomposed into azimuth, inclination, and field,
[4] hmi.sharp_cea_720s_nrt: near real-time data wherein the vector B has been remapped to a Lambert Cylindrical Equal-Area projection and decomposed into B_radial, B_phi, and B_theta.

This page describes the methodology for creating SHARPs and uses an example results for NOAA Active Region 11429, which produced an X5.4-class flare on 7 March 2012.

The HARPNUM of this region is 1449.

Disclaimer : These data are not to be used uncritically

The SHARP data series contain the HMI team's current best effort attempt to determine the vector magnetic field and uncertainties and the data provide a great deal of information. However, the instrument has known limitations due to its spatial and spectral resolution. There are a variety of known and unknown systematic errors and limitations induced by the orbital and other factors. The inversion and disambiguation methods are also known to be far less than perfect. So, please use these data carefully and examine them critically.

See Vector Magnetic Field for more documentation.

Please contact any member of the Vector Field Team with questions or concerns. We welcome your comments and suggestions.

Methodology [1]: Automatically Identifying Active Region Patches

The HMI pipeline analysis code automatically detects active regions in photospheric line-of-sight magnetograms and photograms. This code (i) identifies a patch on the full-disk imagery that encompasses each active region and (ii) tracks every active region as it rotates across the face of the solar disk. One active region crossing the face of the solar disk is referred to as an HMI Active Region Patch (HARP) and is assigned a HARP number, similar in concept to a NOAA Active Region number. (We depart from the NOAA Active Region number because HARPs are detected before NOAA Active Region numbers are assigned and because we identify many smaller magnetic regions at HARPs. However, we include the NOAA Active Region number post-facto as FITS keyword NOAA_AR).

There are two kinds of HARPS: (i) near-real time HARPS and (ii) definitive HARPs. Near-real time HARPs are calculated as soon as possible; in this case, the heliographic bounding box for the active region can change from one instant in time to another. Definitive HARPs are calculated only after an active region has crossed the face of the disk or 5 days after an active region disappears (whichever comes first); in this case, the heliographic coordinates enclosed by the bounding box for the active region are constant over time. HARPs provide all of the required geometric and heliographic information needed to track active patches in HMI and other solar data sets. Note that HARPNUMs in the nrt and defnitive data series will be different.

The movie below (courtesy Mike Turmon) shows near-real time HARPs crossing the face of the solar disk in March 2012 (NOAA AR 11429 corresponds to HARP Number 1449):

Definitive HARP data are available at hmi.Mharp_720s; near real time (nrt) HARP data are available at hmi.Mharp_720s_nrt.

Methodology [2]: Calculating the Vector Field and Spaceweather Quantities

The HARP analysis code defines a rectangular (in CCD coordinates) bounding box, which contains a smooth bounding curve. The bounding box, which is centered at the flux-weighted centroid, is defined as a HARP; the area within the bounding curve is defined as an active region. This information is available within the BITMAP segment of the SHARP data series.

The HMI Stokes I, Q, U, and V data are inverted within the HARP bounding box by using the Very Fast Inversion of the Stokes Vector (VFISV) code, which assumes a Milne-Eddington model of the solar atmosphere, to yield the vector magnetogram and other plasma parameter data Borrero et al, 2011).

The azimuthal component of the vector magnetic field is disambiguated using a minimum energy method to resolve the 180° ambiguity (Metcalf, 1994; Leka et al., 2009). Input parameters to the disambiguation module determine the rate of minimization and number of reconfigurations attempted per minimization step. These parameters are necessarily different for the nrt and definitive data as the nrt data must be produced quickly. A non-linear mask determines the noise thresholds as a function of solar radius. (Pixels above this mask threshold can be identified in the CONF_DISAMBIG segment of the SHARP series). Pixels below the threshold are smoothed during the annealing; pixels above the threshold are not.

Spaceweather quantities (listed in Results[1], below) are calculated every 12 minuets on the inverted and disambiguated data wherein the vector B has been remapped to a Lambert Cylindrical Equal-Area projection and decomposed into Bx, By, and Bz. Some of the spaceweather quantities are calculated directly on the data (such as z-component of the current density; for quantities that require a computational derivative, we use a finite-difference method with a 9-point stencil) and others require a potential field model (such as mean photospheric excess magnetic energy density). The potential field is calculated using Green’s functions and a monopole depth of 0.00001 pixels. Only data that are the logical AND of (i) pixels within the smooth bounding curve in BITMAP and (ii) pixels above the noise threshold (indicated by the coded values greater than 60 in CONF_DISAMBIG) contribute to the spaceweather indices.

Below is an example for nrt SHARP 606 at 2012.10.28_18:24:00_TAI; see hmi.sharp_720s_nrt[606][2012.10.28_18:24:00_TAI].

Results [1]: Spaceweather FITS Header Keywords

The SHARP analysis code calculates the following spaceweather quantities for the region (keyword, quantity, units):

USFLUX Total unsigned flux in Maxwells
MEANGAM Mean inclination angle, gamma, in degrees
MEANGBT Mean value of the total field gradient, in Gauss/Mm
MEANGBZ Mean value of the vertical field gradient, in Gauss/Mm
MEANGBH Mean value of the horizontal field gradient, in Gauss/Mm
MEANJZD Mean vertical current density, in mA/m2
TOTUSJZ Total unsigned vertical current, in Amperes
MEANALP Mean twist parameter, alpha, in 1/Mm
MEANJZH Mean current helicity in G2/m
TOTUSJH Total unsigned current helicity in G2/m
ABSNJZH Absolute value of the net current helicity in G2/m
SAVNCPP Sum of the Absolute Value of the Net Currents Per Polarity in Amperes
MEANPOT Mean photospheric excess magnetic energy density in ergs per cubic centimeter
TOTPOT Total photospheric magnetic energy density in ergs per cubic centimeter
MEANSHR Mean shear angle (measured using Btotal) in degrees

Plots of several quantities computed for HARPNUM 1449 = NOAA AR 11429, appear at the bottom of this web page. One can plot any keyword value directly in JSOC lookdata under the graph tab.

Data [1]: SHARP Data Segment Descriptions

Below are descriptions for each segment (segment name, units, description) in order of appearance in Lookdata.
The remapped CEA (cylindrical equal area) segments are slightly different (as noted) and only a subset of quantities are present.

Observables and HARP Information

MAGNETOGRAM, [Mx/cm2]
The magnetogram segment contains HARP-sized line-of-sight magnetogram data from the DRMS series hmi.M_720s.
CEA version: the field is remapped, but not projected, i.e. it is still the line-of-sight component relative to HMI.

BITMAP, [dimensionless]
The bitmap segment identifies the pixels located within the smooth bounding curve by labeling each pixel with the following:
1   : weak field outside smooth bounding curve
2   : strong field outside smooth bounding curve
33 : weak field inside smooth bounding curve
34 : strong field inside smooth bounding curve
CEA version: same.

DOPPLERGRAM, [m/s]
The dopplergram segment contains HARP-sized dopplergram data from the DRMS series hmi.V_720s.
CEA version: the Doppler velocity is remapped, but not projected, i.e. it is still the line-of-sight component relative to HMI.

CONTINUUM, [DN/s]
The dopplergram segment contains HARP-sized dopplergram data from the DRMS series hmi.Ic_720s.
CEA Version: same.

Inversion Module Outputs

In order to solve the inverse problem of inferring a vector magnetic field from polarization profiles, the Very Fast Inversion of the Stokes Vector (VFISV) module solves a set of differential equations that fit the following parameters:

INCLINATION, [degrees]
The inclination segment contains the magnetic field inclination with respect to the line-of-sight field.
CEA version: not included, See below

AZIMUTH, [degrees]
The azimuth segment contains the magnetic field azimuth. Zero corresponds to the direction of a column of pixels on HMI’s CCD; values increase counter-clockwise.
CEA version: not included, See below

FIELD, [Mx/cm2]
The field segment contains the magnetic field strength. Currently, the filling factor is equal to unity, so this quantity is representative of the magnetic flux density. Values of 120 Mx/cm2 or less are generally considered to be noise.
CEA version: not included, See below

VLOS_MAG, [cm/s]
The vlos_mag segment contains the velocity of the plasma along the line-of-sight. Positive means redshift. [Note: The dopplergram data are in m/s, whereas these data are in cm/s.]
CEA version: not included

DOP_WIDTH, [mÅ]
The dop_width segment contains the doppler width of the spectral line, which is assumed to be a Gaussian.
CEA version: not included

ETA_0, [dimensionless]
The eta_0 segment contains the center-to-continuum absorption coefficient. This is a measure of the difference between the absorption from the center and the continuum of the spectral line. (Differences arise because the center of the spectral line is formed at a different height in the atmosphere than the continuum of the spectral line. These quantities are not constant across the solar disk due to (i) thermodynamic variations and (ii) gradients in height.)
CEA version: not included

DAMPING, [mÅ]
The damping segment contains the electron dipole oscillation approximated as a simple harmonic oscillator. This quantity is affected by the characteristics of the electromagnetic field. In the current version of the code, this parameter is constant and set to unity.
CEA version: not included

SRC_CONTINUUM, [DN/s]
The src_continuum segment contains the source function at the base of the photosphere. In the Milne-Eddington approximation, the source function varies linearly with optical depth.
CEA version: not included

SRC_GRAD, [DN/s]
The src_grad segment contains gradient of the source function with optical depth. By definition, src_continuum + src_grad = observed continuum intensity.
CEA version: not included

ALPHA_MAG, [dimensionless]
The segment alpha_mag is defined as the portion of the resolution element that is filled with magnetized plasma. However, in the current version of the code, this parameter is constant and set to unity.
CEA version: not included

CHISQ, [dimensionless]
The segment chisq is a measure of how well the profiles are fit in the least squares iteration. Chisq is not normalized.
CEA version: not included

CONV_FLAG, [dimensionless]
This is currently not calcuated.
CEA version: not included

INFO_MAP, [dimensionless]
The info_map segment identifies the quality index of the inversion output. The index at each pixel is defined as follows:
0x00000000  : Good pixel, with no severe issue.
0x0[0-6]00000 : Same as the convergence index in CONV_FLAG
0x08000000  : Bad pixel, defined as the same criteria as #5 of CONFID_MAP.
0x10000000  : Low Q or U signal. sqrt((Q_0 + ··· + Q_5)^2 + (U_0 + ··· + U_5)^2) was smaller than 0.206 sqrt(I_0 + ··· + I_5 )
0x20000000  : Low V signal; |V_0| + |V_1| + ··· + |V_5| was smaller than 0.206 sqrt(I_0 + ··· + I_5 )
0x40000000  : Low B_los value. |Blos| from magnetogram algorithm was smaller than 6.2 Gauss.
0x80000000  : Missing data
CEA version: not included

CONFID_MAP, [dimensionless]
The confid_map segment identifies the confidence index of the inversion output. The index value at each pixel will take the integer value from 0 (best) to 6(worst), defined as the highest item number satisfying the following conditions:
0: No issue found in the input Stokes
1: Signals for the transverse field component in the input Stokes parameters (Q & U) were weak.
2: Signal for the line-of-sight field component in the input Stokes parameters (V) was weak.
3: Magnetic field signals of both LoS and transverse component were weak.
4: The ME-VFISV iteration did not converge within the iteration maximum.
In the test data release, we set very small , thus, the confidence ε index at most pixels may be 4.
5: If the difference between the absolute value of the line-of-sight field strength derived from magnetogram algorithm and the absolute value of the LoS component from VFISV inversion |B cos(inclination)| is greater than 500 Gauss, we regard the inversion did not well solve the problem.
6: One (or more) of the 24 input Stokes arrays had NaN value.
CEA version: same for nearest un-remapped pixel

Errors

The following segments contain standard deviations and correlation coeffients that can be used to determine the statistical errors of the vector magnetic field. As described in Section 4 of the AR 11158 Vector Magnetic Field, the calculated variances and covariances are only reliable if the VFISV solution is close to an absolute minimum.
CEA version: not included

INCLINATION_ERR, [degrees]
AZIMUTH_ERR, [degrees]
FIELD_ERR, [Mx/cm2]
VLOS_ERR, [cm/s]
ALPHA_ERR, [dimensionless]
FIELD_INCLINATION_ERR, [dimensionless]
FIELD_AZ_ERR, [dimensionless]
FIELD_ALPHA_ERR, [dimensionless]
INCLIN_AZIMUTH_ERR, [dimensionless]
FIELD_ALPHA_ERR, [dimensionless]
INCLINATION_ALPHA_ERR, [dimensionless]
AZIMUTH_ALPHA_ERR, [dimensionless]

Disambiguation Module Outputs

The following segments are outputs of the disambiguation module:

DISAMBIG, [dimensionless]
The disambig segment identifies the type of disambiguation solution for each pixel by setting the following bits (if multiple bits are set, this indicates that the various solutions agree). Currently the SHARP code uses the minimum energy solution to determine the disambiguation.
bit 1 (lowest bit): Returned from the minimum energy algorithm. For weak field region pixels (below estimated noise), additional smoothing is applied (marked 60 in the conf_disambig segment). For the strong field pixels (above estimated noise), the minimum energy solution is directly reported (marked 90 in the conf_disambig segment).
bit 2 (middle bit): A random solution is adopted for weak field pixels. The minimum energy solution is adopted for strong field pixels (identical to bit 1).
bit 3 (higher bit): The radial acute solution is adopted for weak field pixels. The minimum energy solution is adopted for strong field pixels (identical to bit 1).
CEA version: not included

CONF_DISAMBIG, [dimensionless]
The conf_disambig segment identifies the final disambiguation solution for each pixel with a value which maps to a confidence level in the result (roughly a probability):
50 : weak field method applied (random/radial/potential)
60 : smoothing applied
90 : annealed
0   : not disambiguated
CEA version: same for nearest un-remapped pixel

CEA VERSION - Unique Segments

The CEA magnetic field values are presented differently, as rectilinear components Br, Btheta, and Bphi at each grid point. Statistical errors are provided for each field component, but no cross-correlations are provided. The error in Br, Btheta, and Bphi at each projection grid are calculated from the variances of inverted magnetic field B_total, B_azimuth and B_inclination and the covariances between them. The near-neighbor method is used to get the values of the variances and covariances at that pixel. These values are then propagated to derive the errors for Br, Btheta, and Bphi. For further information on vector and map transformations, see: Vector Coordinates and Vector Transformation.

Bp, [Mx/cm2]
Phi (northward) component of the CEA vector magnetic field.


Bt, [Mx/cm2]
Theta (westward) component of the CEA vector magnetic field.


Br, [Mx/cm2]
Radial (out of photosphere) component of the CEA vector magnetic field.


Bp_err, [Mx/cm2]
Computed uncertainty (standard deviation) of the phi component of the CEA vector magnetic field.


Bt_err, [Mx/cm2]
Computed uncertainty (standard deviation) of the theta component of the CEA vector magnetic field.


Br_err, [Mx/cm2]
Computed uncertainty (standard deviation) of the radial component of the CEA vector magnetic field.


Plots of Space Weather Quantities for HARP 1449, NOAA 11429

Below are plots of nrt and definitive spaceweather quantities calculated on the HARP-sized vector magnetogram data that encapsulated NOAA Active Region 11429. The plots are from March 4 to 8, 2012. The GOES X-Ray flux for this event peaked at 0:40 UT on March 7, 2012, corresponding to an X5.4-class flare. The flare peak is indicated in the plots by a yellow vertical line. Wide scatter in the spaceweather parameters at about ~8 UT on March 7, 2012 is due to poor data quality following an eclipse. Further below are plots representing the percent difference between the nrt and definitive quantities. The active region is rotating toward disk center; therefore, the scatter in the percent difference plots decrease with time as expected.