0806.1547/ms.tex
1:  %######################################################################
2: \documentstyle[12pt,aaspp4]{article}
3: 
4: \lefthead{Magnetic Dissipation and Electric Currents in the Photosphere}
5: \righthead{Abramenko}
6: 
7: 
8: \begin{document}
9: 
10: \title{Spectrum of Magnetic Dissipation and Horizontal Electric Currents in
11: the Solar Photosphere}
12: 
13: \author{Valentyna Abramenko}
14: \affil{Big Bear Solar Observatory, 40386 N. Shore Lane, Big Bear City, CA
15: 92314}
16: 
17: \begin{abstract}
18: 
19: A proxy for horizontal electric currents in the solar photosphere was suggested. 
20:  For a set of evolving active regions (ARs) observed with {\it Solar and Heliospheric 
21: Observatory (SOHO)} Michelson Doppler Imager (MDI) in the 
22: high resolution mode, the dissipation spectrum, $k^2E(k)$, and the spatial 
23: structure of dissipation, i.e., the Stokes dissipation function 
24: $\varepsilon(x,y)$, were calculated from  the observed $B_z$ component 
25: of the magnetic field. These functions allowed us to calculate (a part of) the 
26: horizontal electric current density in the photosphere.
27: It was shown that as an active region emerges,  
28: large-scale horizontal electric currents are gradually generated and
29: determine a bulk of dissipation. When an active region decays, the 
30: large-scale horizontal currents decay faster than the small-scale ones.
31: The density of horizontal currents in active regions is in the range 
32: of $\langle j_h \rangle \sim (0.008 - 0.028)$ A/m$^2$, that
33: is compatible with the density of vertical currents in active regions.
34: We suggest two possible mechanisms for generation of such horizontal currents 
35: in the photosphere. One of them is the drift motions of charged particles in 
36: the medium of varying plasma pressure gradient in a horizontal plane at the periphery 
37: of a sunspot. Such a drift can produce quasi-circular closed horizontal currents around  
38: sunspots. Another possibility could be an existence of  
39: horizontal axial current inside a highly twisted horizontal magnetic structure 
40: laying in the photosphere along the magnetic neutral line.
41: The horizontal currents may contribute significantly to the dynamics of the 
42: photosphere/corona coupling, as well as the estimation of non-potentiality of ARs.
43: 
44: 
45: 
46: 
47: \end{abstract}
48: 
49: \keywords{Sun: magnetic field; photosphere}
50: 
51: 
52: \section { Introduction}
53: 
54: The ultimate source of energy for the majority of non-stationary processes
55: in the solar atmosphere is believed to be a dissipation of the free magnetic
56: energy. This energy, accumulated via plasma motions, manifests as
57: non-potentiality of the magnetic field or, in other words, as electric currents
58: that permeate through all levels of the solar atmosphere. Electric currents,
59: which can be compared to vortex filaments in a turbulent field, tend to be
60: formed wherever a magnetic field has created a thin layer with both shearing
61: and stretching. Therefore analysis of the electric current evolution seems to
62: be a promising approach to better understand processes of generation
63: and dissipation of the magnetic field as well as accumulation of the free
64: magnetic energy to be released in flares and CMEs.
65: 
66: 
67: In spite of the obvious usefulness of information that electric currents could
68: provide, it is not trivial to reliably derive them from observed data. 
69: Rare and mostly state-of-art measurements
70: of the vector magnetic field in the photosphere can provide the only vertical
71: component of the current. To derive the horizontal component of the current,
72: vector magnetic field measurements at several heights in the photosphere are
73: required, which are not yet available. 
74: 
75: 
76: Here we present a new approach to indirectly probe horizontal electric currents.
77: The approach is based on two well known facts, namely, that i) electric current
78: is a product of the magnetic energy dissipation (Moffatt 1978; Parker
79: 1979; Priest 1982) and ii) the magnetic field in the solar photosphere is in a
80: turbulent state (Parker 1979, Petrovay \& Szakaly 1993). This allows us to
81: consider the vertical component of the magnetic field, $B_z$, as a passive
82: scalar in a turbulent flow and thus to analyze its dissipation in the framework
83: of the turbulence theory. Spatial and spectral distributions of the dissipation
84: can then be derived. The former can give us an image of the dissipation over the
85: active region's area, while the later indicates at which spatial scales the
86: dissipation dominates. Physical inferences from such consideration can be made
87: when using the fact that the dissipation of the vertical magnetic field
88: component is intrinsically related to horizontal electric currents.
89: 
90: In Section 2 of this paper we present two approaches (spatial and
91: spectral) to derive the dissipation of the $B_z$ component of the magnetic field
92: and corresponding electric current density. Section 3 is devoted to the 
93: application of the technique to emerging and decaying active regions.
94: Discussion is presented in the last section of the paper.
95: 
96: 
97: \section{\bf Techniques to derive the magnetic dissipation}
98: 
99: \subsection {Spatial structures of $B_z$ dissipation}
100: 
101: Because the $B_z$ component is transported in the photosphere in the same way as
102: a passive scalar does (Parker 1979, Petrovay \& Szakaly 1993), one may apply the
103: Stokes dissipation function (Monin \& Yaglom 1975) to obtain the spatial
104: structure of dissipation:
105: 
106: \begin{equation}
107: \varepsilon(x,y) = \frac{1}{2} \left ( \left ( \frac {{\it d}B_z}{{\it d}x}
108: \right ) ^2 + \left ( \frac {{\it d}B_z}{{\it d}y} \right ) ^2 \right ),
109: \label{eps}
110: \end{equation}
111: where $(x,y)$ is a current point on the horizontal plane and $z$ represents
112: direction normal to the solar surface.
113:  In general, the right hand part of Eq.
114: (\ref{eps}) requires the magnetic diffusion coefficient, $\eta$, which is
115: unknown. Assuming the quasi-uniform behavior of $\eta$ over the active region
116: area, we will analyze the spatial distribution of
117: dissipation as it is derived from Eq. (\ref{eps}) and measured in
118: $Gs^2/Mm^2$ units.
119: 
120: A relationship between the dissipation $\varepsilon$ and the squared electric
121: current, $\bf j^2$, can be easily derived from the general formula for the
122: squared curl of the magnetic field vector, ${\bf B} \equiv (B_x,B_y,B_z)$:
123: \begin{eqnarray}
124: \mu^2 {\bf j}^2 =(\nabla \times {\bf B})^2  = & \nonumber \\
125:  & \left ( \frac {{\it d}B_z}{{\it d}y} -  \frac {{\it d}B_y}{{\it d}z} \right
126: ) ^2 + \left ( \frac {{\it d}B_x}{{\it d}z} -  \frac {{\it d}B_z}{{\it d}x}
127: \right )
128: ^2 + \left ( \frac {{\it d}B_y}{{\it d}x} -  \frac {{\it d}B_x}{{\it d}y} \right
129: ) ^2 =   \label{j2} \\
130: & \varepsilon(x,y)  + \mu^2j_z^2 + f(B_x,B_y,{\it d/dz}). \nonumber
131: \end{eqnarray}
132: Thus, dissipation $\varepsilon(x,y)$ represents a part of the horizontal
133: electric current. The spatial structure of $\varepsilon(x,y)$ is shown in
134: Figures \ref{Fig0}, \ref{Fig2}, and \ref{Fig3}. Areas of strong dissipation,
135: and, therefore, strong horizontal currents, tend to be concentrated around
136: strong magnetic elements. Below we discuss the distribution of horizontal
137: currents with possible mechanisms for their formation.
138: 
139: 
140: %#####################################################################
141: \begin{figure}[!h] \centerline {\epsfxsize=4.0truein
142: \epsffile{Fig_0.eps}}
143: \caption{\sf MDI/HR magnetograms ({\it left}) and corresponding dissipation
144: structures, $\varepsilon(x,y)$, calculated for NOAA AR 0365 observed on 2003
145: May 26/01:01 UT. North is to the top, west is to the right. The image size is $
146: 156 \times 126$ arcsec. The magnetogram is scaled from -500 G to 500G. The
147: dissipation map is scaled from 0 to $ 10^5$ G$^2$/Mm$^2$.}
148: \label{Fig0} 
149: \end{figure}
150: %#####################################################################
151: 
152: Averaging $\varepsilon(x,y)$ over the magnetogram area, we obtain the mean
153: value of dissipation $\langle \varepsilon \rangle$ associated with $B_z$ component. 
154: 
155: Having the value of $\langle \varepsilon \rangle$ for an active region, we can estimate (a
156: part of) the horizontal electric current by using Eq. (\ref{j2}):
157: \begin{equation}
158: \langle {\bf j}_h^2\rangle =\frac{1}{\mu^2}\langle \varepsilon \rangle.
159: \label{jtr}
160: \end{equation}
161: Here $\mu=\mu_0=4\pi 10^{-7}$ H/m is the vacuum permeability.
162: 
163: 
164: \subsection {Spectrum of dissipation of $B_z$}
165: 
166: For any measured quantity, which characterizes the
167: photospheric magnetized plasma, one can calculate the spectrum of energy, i.e.,
168: a function that shows how the energy (per unit volume) is distributed in the
169: wavenumber space, $k=2\pi/s$, where $s$ is a spatial scale. The energy spectrum,
170: $E(k)$ (often referred also as a power spectrum), is essentially a squared
171: Fourier transform of a measured array (here we consider a one-dimensional
172: spectrum of a two-dimensional array, see Abramenko (2005) for details of calculation).
173: For a turbulent medium, the energy spectrum displays a maximum at large scales 
174: (see Figure \ref{Fig1}, blue curve) followed by a power-law decay at smaller 
175: scales. The wavenumber where the maximum is reached, marked as $k_e$ ($e$ stands 
176: for {\it energy}), refers to the scales which bear the maximum energy and where 
177: the energy input occurs (in our case, this is the largest magnetic elements on 
178: a magnetogram). 
179: 
180: The spectrum of {\it dissipation} can be derived from the energy
181: spectrum (Monin \& Yaglom 1975) as 
182: \begin{equation}
183: W(k) = k^2E(k). 
184: \label{W}
185: \end{equation} 
186: Usually, for fully developed turbulence at high magnetic Reynolds number (the
187: ratio of the inertial force to the viscous force), the magnetic dissipation 
188: becomes significant at high wavenumbers only. As a result, the maximum of the 
189: dissipation spectrum, $k_d$ (see Figure \ref{Fig1}), is located at higher wavenumbers
190: than the maximum, $k_e$, of the energy spectrum. The energy interval (which corresponds
191: to the bulk of energy) and the dissipation interval (which corresponds
192: to the bulk of dissipation) are separated in the wavenumber space: the
193: energy is stored mostly at large spatial scales and dissipated at much smaller scales.
194: The energy cascade is formed from large to small scales, which manifests the
195: presence of the fully developed turbulent state. Turbulent magnetic diffusion
196: seems to be a key mechanism for dissipation in this case.
197: 
198: %#####################################################################
199: \begin{figure}[!h] \centerline {\epsfxsize=5.0truein
200: \epsffile{Fig_1.eps}}
201: \caption{\sf MDI/HR magnetograms ({\it left}) and corresponding energy spectrum,
202: $E(k)$ ({\it blue curve and right axis}) and dissipation spectrum,
203: $W(k)=k^2E(k)$ ({\it red curve, left axis}), calculated for a coronal hole area
204: recorded on 2001 June 5/13:07 UT. Noise influence on the spectra was eliminated.
205: North is to the top, west is to the right. The image size is $ 266 \times 202$
206: arcsec. The magnetograms are scaled from -150 G to 150G. The arrows $k_e$  and
207: $k_d$ mark the maxima of the $E(k)$ and $W(k)$ spectra, respectively.  The
208: locations of the maxima were derived from 5 point box car averaging ({\it thin
209: curves}). Peaks of $E(k)$ and $W(k)$ are distinctly separated in the wavenumber
210: space. }
211: \label{Fig1} 
212: \end{figure}
213: %#####################################################################
214: 
215: 
216: 
217: Sometimes the maximum of dissipation, $k_d$, is located close to $k_e$, so that
218: the energy and dissipation intervals are overlapped. In such a case, the
219: dissipation regime differs from the one described above. Namely, the bulk of 
220: dissipation occurs at large scales, and both the ohmic dissipation of large-scale 
221: currents and the turbulent diffusion at smaller scales contribute into the 
222: dissipation process. Thus, the dissipation spectrum seems to be useful to diagnose 
223: the turbulence and the dissipation regime.
224: 
225: Another advantage of the dissipation spectrum arises from its straightforward
226: relation with the mean squared curl of a field $\bf B$ (Monin \& Yaglom
227: 1975, Biskamp 1993), in other words, with the electric currents: 
228: 
229: \begin{equation}
230: \langle {(\nabla \times {\bf B} )}^2 \rangle = 2 {\int_0^{\infty}}  k^2 E(k)
231: dk.
232: \label{rot}
233: \end{equation} 
234: This quantity gives us the mean value of dissipation per unit volume per unit
235: time. 
236: We have in our possession the $B_z$ components of the magnetic field only. When
237: we apply Eq. (\ref{rot}) for $B_z$ by substituting the energy spectrum of $B_z$
238: as $E(k)$, we acquire the part of dissipation associated with the $B_z$ component,
239: i.e., the mean value of dissipation per unit area per unit time measured in
240: units of G$^2$/Mm$^2$. We denoted this value as $\langle W \rangle$:
241: \begin{equation}
242: \langle W \rangle = 2{\int_0^{\infty}}  k^2 E(k) dk.
243: \label{Wsr}
244: \end{equation} 
245: 
246: One can see that $\langle W \rangle$ and $\langle \varepsilon \rangle$
247: determine the same physical quantity derived, however, from different
248: approaches: from spatial and spectral considerations. This circumstance will
249: allow us to provide two estimations for the mean magnetic energy
250: dissipation. Besides, the two approaches allow us to study different aspects
251: of the horizontal electric currents. The current density can be derived as
252: \begin{equation}
253: \langle {\bf j}_h^2\rangle =\frac{1}{\mu^2}\langle W \rangle.
254: \label{jtrW}
255: \end{equation}
256: 
257: 
258: 
259: 
260: 
261: 
262: \section{Application to active regions}
263: 
264: 
265: We utilized SOHO/MDI (Scherrer at al. 1995) magnetograms recorded in the high
266: resolution mode. We analyzed areas located at the center of the solar disk (no
267: farther than 15 degree away from the central meridian), so that the projection effect was
268: negligible and the direction of $z$ coincides quite well with the line-of-sight
269: direction.
270: 
271: In this study, we focused on the problem of how the dissipation spectrum and
272: horizontal currents evolve when an active region emerges and decays. 
273: An early emergence of an active region can be detected during a 2-day time interval 
274: when the magnetic structure is in the close vicinity of the central meridian. 
275: We selected three such cases: NOAA ARs 0314, 9574, 0798 (magnetograms and 
276: dissipation maps for the last one are shown in Figure \ref{Fig2}). 
277: However, to detect significant changes at the decay phase, the 
278: 2-day interval is rather short. We decided to analyze the decay of a magnetic structure 
279: by using observations at two consecutive rotations. For three strong, 
280: well-developed active regions: NOAA ARs 9682, 9169, 9165, we determined
281: their remains at the next rotation, namely, NOAA ARs 9712, 9199, 9189, 
282: respectively. Magnetograms and dissipation maps for the first pair, NOAA ARs 9682/9712, 
283: are presented in Figure \ref{Fig3}.
284: As a reference, we also analyzed a coronal hole area shown
285: in Figure \ref{Fig1}. 
286: 
287: The data for all analyzed magnetograms are compiled in Table 1. 
288: The first row refers to a coronal hole magnetogram;
289: three blocks below correspond to three emerging active regions;
290: three pairs below show the data for decaying complexes. The third (fourth) column, $r_e$ ($r_d$), 
291: shows the scale where the maximum of the energy (dissipation) spectrum was located, 
292: the fifth column shows the distance between the dissipation and energy maxima in 
293: the wavenumber space, ($k_d - k_e$). The next three columns show
294: the three estimations of the mean value of the magnetic energy dissipation. 
295: The sixth column, $\langle \varepsilon \rangle$, 
296: is derived by averaging results from Eq. (\ref{eps}) over a magnetogram. 
297: The seventh column represents the values $W_n$ calculated from Eq. (\ref{Wsr}) 
298: when noise was not eliminated in calculation of $E(k)$ (for the noise 
299: elimination details see Abramenko et al. 2001). 
300: The eighth column shows the values $W_y$ calculated 
301: from the same Eq. (\ref{Wsr}), however, noise was preliminary eliminated. Thus, 
302: for each magnetogram, we obtained values of 
303: $\langle \varepsilon \rangle$, $W_n$, and $W_y$, which allowed us to use 
304: Eqs. (\ref{jtr}, \ref{jtrW}) to derive three estimations of the squared horizontal 
305: electric current density, $\langle {\bf j}_h ^2\rangle$. The average from the 
306: three estimations $(\langle {\bf j}_h^2\rangle)^{1/2}$ is presented in the last column 
307: of Table 1. 
308: %The data of Table 1 are graphically represented in Figures (\ref{Fig4} - \ref{Fig5}).
309: 
310: \begin{table}[!ht]
311: \caption{\sf List of studied active regions and calculated parameters}
312: \footnotesize
313: \begin{center}
314: \begin{tabular}{llcrccccc}
315: \hline
316: NOAA    & Data            & $r_e$  &$r_d$  &$k_d - k_e$  &$\langle \varepsilon \rangle$&$\langle W_n \rangle$ &$\langle W_y \rangle$   &$(\langle {\bf j}_h^2\rangle)^{1/2}$\\
317:  AR     &                 &  Mm    &  Mm   &   Mm$^{-1}$ &      G$^2$/Mm$^2$          &    G$^2$/Mm$^2$     &       G$^2$/Mm$^2$    &   Amper/m$^2$            \\
318: \hline
319: CH      & 2001 Jun05/13:07 & 14.129 & 2.018 &   2.669     &      3106  	               &       3806           &          1581          &   0.0059 $\pm$ 0.0013    \\
320: \hline
321: 0314 a  & 2003 Mar13/22:19 & 18.988 & 6.995 &   0.567     &      6929  	               &       6872           &          4500          &   0.0087 $\pm$ 0.0011    \\
322: 0314 b  & 2003 Mar14/11:44 & 44.310 & 10.223&   0.473     &     20471  	               &      19479           &         16878          &   0.0155 $\pm$ 0.0008    \\
323: 0314 c  & 2003 Mar15/11:10 & 44.310 & 14.766&   0.284     &     35149  	               &      33228           &         30396          &   0.0204 $\pm$ 0.0007    \\
324: \hline			    
325: 9574 a  & 2001 Aug10/05:20 & 31.653 &  4.996&   1.059     &     12726                   &      12708           &         10333          &   0.0122 $\pm$ 0.0007    \\
326: 9574 b  & 2001 Aug10/17:11 & 31.653 & 10.547&   0.397     &     39459  	               &      38421           &         37117          &   0.0220 $\pm$ 0.0003    \\
327: 9574 c  & 2001 Aug11/08:43 & 31.653 & 31.653&   0.000     &     58558  	               &      60528           &         59573          &   0.0275 $\pm$ 0.0003    \\
328: \hline			    
329: 0798 a  & 2005 Aug18/08:27 & 47.136 &  4.873&   1.156     &      6564  	               &       6570           &          4067          &   0.0085 $\pm$ 0.0011    \\
330: 0798 b  & 2005 Aug18/16:09 & 47.136 &  6.729&   0.800     &      9227  	               &       8833           &          6290          &   0.0101 $\pm$ 0.0010    \\ 
331: 0798 c  & 2005 Aug19/08:38 & 47.136 & 28.264&   0.089     &     13839  	               &      12736           &         10119          &   0.0124 $\pm$ 0.0010    \\
332: \hline
333: 
334: 9682    & 2001 Oct30/17:02 & 63.853 &  7.093&   0.788     &     40802                   &      40082           &         36922          &   0.0223 $\pm$ 0.0006    \\
335: 9712    & 2001 Nov26/15:46 & 63.853 &  2.487&   2.428     &     33839                   &      35281           &         32545          &   0.0207 $\pm$ 0.0004    \\
336: \hline
337: 9169    & 2000 Sep23/19:00 & 86.307 &  3.547&   1.698     &     59743                   &      59257           &         55694          &   0.0271 $\pm$ 0.0006    \\
338: 9199    & 2000 Oct20/20:53 & 86.307 &  2.785&   2.184     &     25770                   &      26340           &         23946          &   0.0179 $\pm$ 0.0004    \\
339: \hline			    
340: 9165    & 2000 Sep15/02:00 & 44.848 &  3.281&   1.775     &     44853                   &      43850           &         40661          &   0.0233 $\pm$ 0.0006    \\
341: 9189    & 2000 Oct12/10:03 & 26.908 &  2.862&   1.962     &     15050                   &      15671           &         14051          &   0.0137 $\pm$ 0.0004    \\
342: 			    
343: 
344: \hline
345: \end{tabular}
346: \end{center}
347: 
348: \end{table}
349: 
350: %#####################################################################
351: \begin{figure}[!h] \centerline {\epsfxsize=6.5truein
352: \epsffile{Fig_2.eps}}
353: \caption{\sf MDI/HR magnetograms ({\it top row}) and corresponding dissipation
354: structures, $\varepsilon(x,y)$, calculated for emerging NOAA AR 0798 (E8
355: -W15, S10). North is to the top, west is to the right. The image size is $ 204
356: \times 126$ arcsec. The magnetograms are scaled from -500 G to 500G. The
357: dissipation maps are scaled from 0 to $6 \times 10^4$ G$^2$/Mm$^2$.}
358: \label{Fig2} 
359: \end{figure}
360: %#####################################################################
361: 
362: %#####################################################################
363: \begin{figure}[!h] \centerline {\epsfxsize=5.5truein
364: \epsffile{Fig_3.eps}}
365: \caption{\sf MDI/HR magnetograms ({\it top row}) and corresponding dissipation
366: structures, $\varepsilon(x,y)$, calculated for a decaying complex NOAA AR 9682/9712. 
367: North is to the top, west is to the right. The image size is $ 264
368: \times 174$ arcsec. The magnetograms are scaled from -500 G to 500G. The
369: dissipation maps are scaled from 0 to $9 \times 10^4$ G$^2$/Mm$^2$.}
370: \label{Fig3} 
371: \end{figure}
372: %#####################################################################
373: 
374: %#####################################################################
375: \begin{figure}[!h] \centerline {\epsfxsize=5.5truein
376: \epsffile{Fig_4.eps}}
377: \caption{\sf Energy spectra, $E(k)$ ({\it blue lines}), and dissipation spectra,
378: $k^2E(k)$ ({\it double red lines}), for three emerging active regions. For each AR, 
379: frames ({\it a, b, c}) correspond to a row with the same mark in Table 1. Vertical 
380: blue (red) bars mark the maximum of the energy (dissipation) spectrum. Blue bars
381:  correspond to $k_e$ and red bars correspond to $k_d$. As the active region 
382: emerges, $k_d$ shifts toward the smaller wavenumbers.}
383: \label{Fig4} 
384: \end{figure}
385: %#####################################################################
386: 
387: %#####################################################################
388: \begin{figure}[!h] \centerline {\epsfxsize=5.5truein
389: \epsffile{Fig_5.eps}}
390: \caption{\sf Energy spectra, $E(k)$, and dissipation spectra,
391: $k^2E(k)$, for three decaying magnetic complexes (three rows). 
392: The right frame in each row 
393: shows a superposition of the dissipation spectra at the well-developed 
394: ({\it double red line}) and decaying ({\it solid green line}) state 
395: of a magnetic complex. As the magnetic complex decays, $k_d$ shifts 
396: toward the larger wavenumbers.
397: Other notations are the same as in Figure \ref{Fig4}.}
398: \label{Fig5} 
399: \end{figure}
400: %#####################################################################
401: 
402: The data from Table 1 show that the coronal hole displays the smallest
403: energy input scales, $r_e$,  as compared to active regions. This can be 
404: attributed to smaller typical size of magnetic elements inside coronal holes. 
405: A non-trivial inference from the CH data is that 
406: the energy dissipation scale, $r_d$, is smaller and the distance ($k_d - k_e$) 
407:  is larger than that for active regions. This implies 
408: that in a CH we observe the state of more developed turbulence as compared to ARs. 
409: Thus, an assumption of fully developed turbulence seems to be valid for CHs.
410: The current density is smallest for the CH.
411: 
412: Three emerging ARs show a well pronounced trend as the emergence proceeds. 
413: Namely, at the very beginning of emergence, we observe the largest distance 
414: ($k_d - k_e$), which resembles the situation in the CH. Later during the emergence 
415: this distance decreases by means of shifting of $k_d$ toward the small wavenumbers.
416: The intervals of the energy and dissipation become more and more overlapped 
417: (see Figure \ref{Fig4}, frames from {\it a} to {\it c}) and the current density 
418: increases (the last column in Table 1). The observed fact, that during the emergence 
419: the bulk of dissipation shifts toward larger scales, implies the formation 
420: of large-scale strong currents. In the map of spatial distribution of dissipation 
421: (see Figure \ref{Fig2}) we see formation of bright contours (frequently closed) 
422: surrounding the magnetic flux tubes. Their size increases as the AR emerges.
423: 
424: Three decaying magnetic complexes showed an opposite behavior, see Figure \ref{Fig5}.
425: As the magnetic structure decays, the maximum of dissipation shifts toward the 
426: smaller scales, large-scale currents decay, the current density decreases and 
427: the entire picture begins to resemble what we saw for a CH. Disappearance of large-scale 
428: dissipation features is distinctively visible in Figure \ref{Fig3} (compare 
429: left and right frames in the bottom row.) 
430: 
431: 
432: The data obtained allow us to estimate the upper boundary of the plasma conductivity,
433:  $\sigma=\tau/l^2$. Here $l$ is a typical scale of decayed current and $\tau$ 
434: is a typical dissipation time. The data for NOAA ARs 9682/9712 pair show that 
435: currents of scales about 7 Mm  dissipated during $\tau <$ 27 days. This gives us 
436: $\sigma < 5\times 10^{-8}$ s/m$^2$, which is four orders of magnitude lower than 
437: the classical ohmic conductivity for the photosphere.
438: The magnetic diffusion coefficient in the photospheric plasma of active regions: 
439: $\eta=l^2/\tau$ (Priest 1982, here $l$ is a typical scale of decayed magnetic element and $\tau$ 
440: is a typical decay time). For magnetic elements of 7 Mm  dissipated during 
441: $\tau <$ 27 days we derive $\eta > 2\times 10^7$ m$^2$/s, which is four 
442: orders of magnitude higher than the classical diffusion coefficient for the photosphere.
443: 
444: 
445: \section { Discussion}
446: 
447: The study presented here suggests that in an active region there exist large-scale 
448: horizontal electric currents, which appear and grow during 1-2 days as the AR
449: emerges and slowly decay, during a couple of weeks, when the AR decays. 
450: A bulk of magnetic dissipation of the $B_z$ component is associated with these 
451: large-scale currents rather than with small-scale currents. 
452: 
453: This inference recalls the publications where the 
454: photospheric and chromospheric magnetic fields were compared  (Abramenko et al. (1992), 
455: Choudhary et al. (2002), Leka \& Metcalf (2003), Balasubramaniam et al. (2004);
456:  see also a resent review by Nagaraju et al. (2008) and references in).
457: Their results  
458: suggest that the stronger the line-of-sight magnetic field, the faster it 
459: decreases with height. Nagaraju et al. (2008) argued that for the photospheric 
460: field stronger than 700 G, the chromospheric field values "are much weaker 
461: than what one gets from the linear relationship [between the photospheric and 
462: chromospheric fields] and also from those expected from the extrapolation 
463: of the photospheric magnetic field" in the potential approximation. 
464: Leka \& Metcalf (2003) observed that the potential field of NOAA 8299 extrapolated from the 
465: photosphere to reasonable heights (lower than 2 Mm) statistically exceeds the observed 
466:  chromospheric field. 
467: In 1992 Abramenko and co-authors reported that for NOAA 6280 the chromospheric field 
468: measured in $H_{\beta}$ spectral line is lower than the extrapolated potential 
469: field. They also found that the vertical gradient of the line-of-sight magnetic fields 
470: is higher for strong fields. 
471: Note, that the above results were obtained from different types of measurements, 
472: different spectral lines, different instruments, etc., and still they are 
473: consistent with each other.
474: Several reasons were suggested to explain the effect (see Nagaraju et al. (2008) for review).
475: However, only one of them addresses the effect of horizontal electric current formation.
476: Namely, in Abramenko et al. (1992) the conclusion was made that the possible reason 
477: for such a systematical deviation between the observed and potential fields
478: is the presence of horizontal electric currents enclosing sunspots. The magnetic fields of 
479: such currents are predominantly vertical and directed opposite to the original 
480: field of a sunspot, so that superposition of the two fields results in the 
481: fast weakening with height of the observed field.
482: 
483: We suggest that one of the possible mechanisms for formation of such horizontal currents
484: could be the drift motions of charged particles in the medium of varying plasma 
485: pressure gradient in a horizontal plane. Presence of such a medium is unavoidable 
486: at the periphery of a sunspot. Such a drift can produce quasi-circular closed 
487: horizontal currents around sunspots. This suggestion is in part confirmed by 
488: the imaging of the horizontal currents, see Figures \ref{Fig0}, \ref{Fig2}, 
489: \ref{Fig3}: in all of them, quasi-circular closed bright structures around 
490: sunspots are noticeable. 
491: 
492: Another mechanism for the horizontal currents formation can be suggested. 
493: Often inside strong active  regions with stressed and sheared neutral line 
494: one observe strongly elongated magnetic features of both polarities along 
495: the neutral line, see, for example, Figures \ref{Fig0}, \ref{Fig3}. The most 
496: obvious way to explain such a phenomenon is to suggest a presence of highly 
497: twisted horizontal magnetic structure laying in the photosphere. A strong 
498: electric current along the axis of a magnetic helical tube is unavoidably present. 
499: A straight long electric current $I$ generates an azimuthal magnetic field 
500: $B$ at a distance $r$ from the cur rent axis: $B=2I/cr$, where $B$ is in Gauss, 
501: $I$ in the CGSE system of units, and $c$ is the light velocity. The data for 
502: NOAA AR 9682 (see Figure \ref{Fig3}) shows that an azimuthal magnetic field 
503: of about 500 G is generated at a distance of approximately 3 Mm from the axis 
504: of the horizontal current channel laying along the neutral line. This gives us 
505: an estimation of the current in the channel as $0.75 \cdot 10^{12}$ A and the 
506: density of the current as $0.026$ A/m$^2$, which is compatible with the 
507: magnitude given in Table 1 for this active region.
508: 
509: Emergence of a horizontal helical flux rope along a neutral line was recently 
510: reported by Okamoto et al. (2008) from vector magnetic field observations with 
511: the Solar Optical Telescope (SOT) on board the {\it Hinode} satellite.
512: 
513: 
514: Taking into account all the above mentioned results, we suggest that strong 
515: {\it horizontal} electric currents are present in the solar atmosphere, and they 
516: occupy a large interval of heights: at least, from the photosphere to 
517:  chromosphere. Their density, $\langle j_h \rangle \sim (0.008 - 0.028) $ A/m$^2$, 
518: is quit compatible with the density of vertical currents in the photosphere, 
519: see, e.g., Abramenko et al. (1991): $\langle j_z \rangle \sim 0.06 $ A/m$^2$; 
520: Wheatland  (2000): (0.009 - 0.015) A/m$^2$. The horizontal currents may contribute 
521: significantly into the dynamics of the photosphere/corona coupling, as well 
522: as into estimation of non-potentiality of ARs. They have to be closely connected 
523: with transverse plasma motions, with creation of shear motions. From this standpoint, 
524: the recently suggested new approach to study plasma velocities (Nagaraju et al. 2008) 
525: seems to be very promising.
526: 
527: 
528: One more result of the above study deserves to be briefly discussed. 
529: The shifting of the maximum of the dissipation spectrum as an active region emerges 
530: and decays indicates that the turbulence regime varies with the evolution of the AR.
531: Namely, as an active region emerges, the state of fully developed turbulence 
532: becomes gradually replaced by a state of under-developed turbulence.
533: The state of under-developed turbulence in mature active regions 
534: still bears a signature of a turbulent cascade, that is confirmed by 
535: {\it gradual} decrease of energy along the energy spectrum.
536: High value of the magnetic diffusion coefficient obtained 
537: here implies that the turbulent diffusion is significant 
538: in well-developed active regions and is responsible to the energy cascading. 
539: However, the presence of a bulk of dissipation at large scales suggests that 
540: the ohmic dissipation of large-scale currents might compete with the turbulent 
541: diffusion in formation of a dissipative regime in the photosphere of 
542: well-developed active regions. As the active region decays, the large-scale 
543: currents decay faster, as compared to the small-scale currents, and the 
544: fully-developed turbulence tends to be restored. 
545: 
546: 
547: 
548: Author is thankful to Prof. Lennard Fisk for a suggestion to use the dissipation 
549: spectrum as a proxy for the mean squared curl, to Drs. Vasyl Yurchyshyn, 
550: Gregory Fleshman, Haimin Wang for helpful discussions, to Erika Norro and Aaron Coulter 
551: for their help with language issues. 
552: This work was supported, in part, by NASA NNX07AT16G grant and 
553: NSF grant ATM-0716512.
554: 
555: 
556: 
557: 
558: 
559: \begin{thebibliography}{}
560: \bibitem{}
561: Abramenko, V. 2005, ApJ, 629, 1141
562: 
563: \bibitem{}
564: Abramenko, V. I., Gopasiuk, S. I., Ogir', M. B. 1991, Sol. Phys., 134, 287
565: 
566: \bibitem{}
567: Abramenko, V. I., Gopasyuk, S. I., Ogir', M. B., Yurchishin, V. B. 1992, Kinematics Phys. Celest. Bodies, 8, No. 5, 43
568: 
569: 
570: \bibitem{}
571: Abramenko, V., Yurchyshyn, V., Wang, H. \& Goode, P. R. 2001, Sol. Phys., 201, 225
572: 
573: \bibitem{}
574: Balasubramaniam, K. S.; Christopoulou, E. B.; Uitenbroek, H. 2004, ApJ, 606, 1233
575: 
576: \bibitem{}
577: Biskamp, D. 1993, Nonlinear Magnetohydrodynamics (Cambridge University Press, New York)
578: 
579: \bibitem{}
580: Choudhary, D. P., Suematsu, Y., Ichimoto, K. 2002, Sol. Phys. 209,349
581: 
582: \bibitem{}
583: Leka, K.D.\& Metcalf, T.R. 2003, Sol. Phys. 212, 361
584: 
585: \bibitem{}
586: Nagaraju, K., Sankarasubramanian, K. \& Rangarajan, K. E. 2008, ApJ, 678, 531
587: 
588: \bibitem{}
589: Moffatt, H.K. 1978, Magnetic Field Generation in Electrically Conducting Fluids 
590: (Cambridge University Press, Cambridge, London - New York - Melburn)
591: 
592: \bibitem{}
593: Monin, A.S.\& Yaglom, A.M. 1975, Statistical Fluid Mechanics, Vol. 2,
594: ed. J.Lumley (MIT Press, Cambridge, MA)
595: 
596: \bibitem{}
597: Okamoto, T. J., Tsuneta, S., Lites, B. W., and 12 co-authors 2008, ApJ, 673, L215
598: 
599: 
600: \bibitem{}
601: Parker, E.N. 1979, Cosmical magnetic fields, their origin and their activity (Oxford
602: University Press)
603: 
604: \bibitem{}
605: Petrovay, K. \& Szakaly, G. 1993, Astron. Astrophys., 274, 543
606: 
607: \bibitem{}
608: Priest, E. 1982, Solar Magnetohydrodynamics (Reidel) 
609: 
610: \bibitem{}
611: Scherrer, P.H., Bogart, R.S., Bush, R.I., Hoeksema, J.T., Kosovichev, A.G.,
612: Schou, J., Rosenberg, W., Springer, L., Tarbell, T.D., Title, A., Wolfson, C.J.,
613: Zayer, I. and the MDI engineering team, 1995, Sol. Phys., 162,129
614: 
615: \bibitem{}
616: Wheatland, M.S. 2000, ApJ, 532, 616
617: 
618: \end{thebibliography}{}
619: 
620: \end{document}
621: 
622: 
623: 
624: 
625: