SH31C-08: The Photospheric Poynting Flux and Coronal Heatin - PowerPoint Presentation

Slide1

SH31C-08: The Photospheric Poynting Flux and Coronal Heating

Some models of coronal heating suppose that convective motions at the photosphere shuffle the footpoints of coronal magnetic fields and thereby inject sufficient magnetic energy upward to account for observed coronal and chromospheric energy losses in active regions. Using high-resolution observations of plage magnetic fields made with the Solar Optical Telescope aboard the Hinode satellite, we investigate this idea by estimating the upward transport of magnetic energy --- the vertical Poynting flux, Sz --- across the photosphere in a plage region. To do so, we combine: (i) estimates of photospheric horizontal velocities, vh, determined by local correlation tracking applied to a sequence of line-of-sight magnetic field maps from the Narrowband Filter Imager, with (ii) a vector magnetic field measurement from the SpectroPolarimeter. Plage fields are ideal observational targets for estimating energy injection by convection, because they are: (i) strong enough to be measured with relatively small uncertainties; (ii) not so strong that convection is heavily suppressed (as within umbrae); and (iii) unipolar, so Sz in plage is not influenced by mixed-polarity processes (e.g., flux emergence) unrelated to heating in stable, active-region fields. In this plage region, we found that the average Sz varied in space, but was positive (upward) and sufficient to explain coronal heating, with values near (5 +/- 1)?107 erg /cm2 /s. We find the energy input per unit magnetic flux to be on the order of 105 erg /s /Mx. This is consistent, within order of magnitude, with luminosities per unit magnetic flux observed in soft X-ray emission. We also found upward fluxes to predominate in most other plage regions, indicating that upward energy flux is a generic property of plage fields, and suggesting this is a manifestation of the energy required for coronal heating.

Brian T. Welsch

Space Sciences Laboratory, University of California, BerkeleySlide2

If I’ve seemed a little fuzzy minded for the past few days, there’s a reason...

I apologize for not functioning at full capacity!Owen, age 4.5 months Slide3

What heats the chromosphere & corona to 10

4 & 106K? Dissipation of magnetic energy is powers heating.This energy probably enters the corona from the interior as waves or via convectively driven footpoint shuffling: Parker 1983 Marsh et al. 2009 wave heating (“AC”?)“DC” heatingSlide4

The Poynting flux of

magnetic energy across the photosphere into the corona is given by: dU/dt = ∫ dA (cEph x Bph)z /4π Bph is the photospheric magnetic field, and Eph is the photospheric electric field.Parker (and others) convincingly argued that Eph is ideal, i.e., Eph = -(vph x Bph)/c Q: If magnetic energy powers the heating, can we see any flux of magnetic energy across the photosphere?Slide5

dU/dt = ∫ dA Sz =∫ dA (Bph x [vph x Bph])z /4π =∫ dA (vz Bh2 – [vh⋅Bh]Bz)/4π In fact, both terms involve the transport of magnetized plasma into the corona – so emergence of fields!But they do differ: “emergence” increases total unsigned flux.

“emergence”

“shearing”

v

Assuming an ideal Ohm’s law,

c

E

ph

= -(

v

ph

x

B

ph

), the Poynting

flux can be written in terms of

photospheric velocities,

v

ph

:Slide6

Plage

magnetic fields are mostly-vertical fields in active regions, where “shearing” term should dominate. Emergence of new flux is not observed in plage.So Szplage≈ - f [vh⋅Bh]Bz/4π In this box, inclinations were ~20o from vertical.Estimating Szplage requires\estimating vh & Bh0.32’’ pix, so ~100 Mm

must include fill fraction!Slide7

7

We used Fourier local correlation tracking (FLCT) to get vh( x, y) by correlating LOS magnetograms’ subregions.

1) for ea. (x

i

,

y

i

)

above |

B

|

threshold

2) apply Gaussian

window

at (x

i

,

y

i

)

3) truncate and

cross-correlate

*

4)

Δ

x

(x

i

,

y

i

) is inter-

polated

max. of

correlation

funct

=

=

=

B

LOS

cadence = 2 min.

;



t = 8

min.; w

indow parameter

σ

=

4

pix.Slide8

Vector

B is measured by relatively slow rastering – 45 minutes to scan across the entire active region!So we co-registered the (tracked) BLOS with B from the vector magnetogram in our plage region.Aligned Blos & Bz Correlation ≈ 0.8Slide9

Flows & fields are combined to make Poynting flux maps.

Withbroe & Noyes (1977) estimated coronal & chromospheric energy demands to be 1 x 107 erg cm-2 s-1 and 2 x 107 erg cm-2 s-1, resp.There are regions of positive & negative Poynting flux, but mean is positive, about (5 ± 1) x 107 erg cm-2 s-1 – enough power to heat both! Slide10

Flows & fieldsSlide11

Poynting flux

mapSlide12

We checked our flow estimates with another LCT code:

Velocities agreed in many regions, disagreed in others. Correlation coefficients were ≈ 0.7The mean is again upward, but ~10% larger, at 5.5 x 107 erg cm-2 s-1 – so uncertainties on the order of ~10% (or more) from tracking.Slide13

Poynting flux is

upward in the one region we study in detail. Is the Poynting flux also upward elsewhere?We defined all pixels with filling-factor-weighted 100 Mx cm−2 <|Bz|< 1500 Mx cm−2 and inclination < 30o from vertical as “plage-like.”Slide14

To construct a flow map simultaneous with the

rastered B( x, y) map, we sampled near-simultaneous columns in velocity maps, vh( x, y, t). Again, both upward and downward Poynting fluxes, but mostly up!Blue is histogram of upwardPoynting fluxes.Red

is

histogram of

downward

Poynting fluxes

.

Poynting flux is

upward

in the one region we study in detail.

Is the Poynting flux also upward

elsewhere

?

(cont’d)Slide15

Physically, magnetic energy travels both up & down --

BUT the corona “extracts” a small but significant “heating tax!” LMSAL/TRACEDoes the energy propagate along B? Or is it (mostly) reflected – up & back down? Would this lead to waves, -- and non-thermal broadening in IRIS lines? Coronal dissipation & heating should drive intensity variations on moss above footpoints – correlated with Poynting flux?Slide16

Is this is a “DC” Poynting flux?

Autocorrelation of vx & vy in this plage region ==> lifetime ~500 sec.Alfvén transit times for coronal portion of loops are ~100 sec. Photosphere-to-corona transit times are ~60-200 sec. each way (van Ballegooijen et al. 2011) ==> Footpoint motions evolve on loops’ approximate global relaxation timescale – coronal B never relaxes!Slide17

The energy flux we find can be expressed as a luminosity per unit magnetic flux,

LPoynt ≈105 erg/s/Mx . Pevtsov et al. (2003)“Pevtsov’s law:” magnetic fields like to radiate in soft X-raysIn soft X-rays, luminosities per unit flux are Lx ≈103 erg/s/Mx; matches rule of thumb that Lx ≈ 0.01 Lheat, the total input from heating.Slide18

What type of flows produce this upward

energy flux?No clear twisting & braiding; correlation of Poynting flux with flow vorticity was weak. Mismatch between structure in vh & Bh produces bipolar Poynting fluxSlide19

Main Idea: studies of the Poynting flux can help understand energy input for coronal heating.

We see enough energy to heat both the corona & chromosphere --- about 50% more than Withbroe & Noyes (1977) said is needed. There must be some scales below which the flux decreases (so no ultraviolet catastrophe) – higher-res observations are needed.Additional studies with Hinode/SOT can address these questions. Comparisons with IRIS and AIA observations are planned.Poynting flux studies are a “Critical Science” item for DKIST!