5 The energy input from the downward jets

The observational and theoretical results do not reveal a clear picture of the source of the solar oscillations. The asymmetries of the p-mode spectral lines have been interpreted as the signature of a spatially confined source, just below the surface of the p-mode cavity (Duvall et al. 1993). Recently, a model, not based on an explicit "depth dependent'' source, explains the same effect on the profiles (Magrí et al. 2001).

The continuous, randomly distributed downflow caused by rapid cooling at the surface is thought to be the trigger of solar oscillations (Goldreich et al. 1994). These downflows have been observed and correlated with seismic events (Goode et al. 1998) and have been used to explain the global behaviour of the solar background in the intensity-velocity phase difference (Skartlien & Rast 2000). These convective downflows have spatial scales of a few arcseconds and a duration of the order of few minutes.

From the analysis of several years of power spectra, the energy and energy input of the p-modes have been studied through the solar cycle. The most recent results give for the total energy and for the total energy input $4.2 \times10^{34}$ erg and $1.9 \times10^{29}$ ergs^-1 respectively (Komm et al. 2000). The former varies with magnetic activity (confirming their absorption) while the latter is compatible with zero changes (Komm et al. 2000; Chaplin et al. 2000). The data analysis of the ongoing maximum will confirm or reject this last conclusion.

In the following, we discuss the downflows associated with the flare occurrence to be a possible source of the solar oscillations.

This hypotheses is suggested by the observational evidence of the impact into the photosphere of the downward plume caused by a big flare (Kosovichev & Zarchova 1998) and of the recent correlations between the H$\alpha$ bright points and the locations of the solar background described in this paper.

The UV brightening, the X-ray blue asymmetry in the coronal Ca lines and the H$\alpha$ red asymmetry in the bright kernels are associated with the occurrence of flares. The physical mechanisms advanced to explain these phenomena depend on the observed velocities, momentum, energy etc. Let us select the explosive chromospheric evaporation model (Fisher et al. 1985a) in order to compute the downward energy flow. Why this process? One reason is that the observed main characteristics of these events seem to reproduce those we want in the local behaviour of the I-V phase: they have H$\alpha$ counterpart, usually reported in the bright points at spatial scales of some arcsec and one minute timescales (Canfield et al. 1987; Canfield & Metcalf 1987) and they are usually located at the border of the supergranular cells (Henoux 2001).

The leit-motif of this discussion is: several flares are observed, many are related to the sudden creation of a high-pressure region at the footpoint of a coronal loop causing balanced downward-upward plasma jets (Canfield et al. 1987); the downward jets are those observed in the H$\alpha$ line; the upward are observed in UV and X-rays and are invoked to explain a portion of the coronal heating. We estimate, only from observational evidence, the energy input to the p-modes as computed from the kinetic energy injected by the downward counterparts of the UV jets at scales of approximately 2 arcsec.

There is an energy flux threshold for explosive evaporation: it occurs when the upper chromosphere is unable to radiate the flare energy deposited there. If it occurs, the plasma driven downward (with a maximum velocity of 100 m/s) is cool and dense (Fisher et al. 1985b). For this mechanism, the downward and upward momenta are balanced (and should be seen in the H$\alpha$ and UV as impulsive events). We estimate the downward energy input using only some conservative values obtained from the observations, that is, we do not include the contribution below the limit of detection of the actual instrumentation.

Following Canfield et al. (1987), the momentum of the downward plume is $2 \times10^{21}$ g cms^-1 and the average velocity 30 Kms^-1, that is, a kinetic energy per event equal to $3 \times10^{27}$ erg. In order to estimate the total injected energy input, the global Sun birthrate is needed. From the related UV measurements, the global Sun birthrate of the impulsive brightenings strongly depends on the threshold used to select the events. Using the recent results from EIT on board SOHO (Berghmans et al. 1998), it can vary from 10 to 40 s^-1. In summary, the energy input spans from $\simeq$ $3.3\times10^{28}$ ergs^-1 to $\simeq$ $1.3\times10^{29}$ ergs^-1. This range is consistent with the value measured by Komm et al. (2000).
We have assumed that when the H$\alpha$ and UV jets are observed at the arcsec scales, the energy deposition rate in the chromosphere has been enough to inject the plasma in the lower layers of the solar atmosphere (referred as an "electron beam flare'', Mullan 1989). Did those downward jets reach the photosphere? Are they reflected by the magnetic topology or squashed into the higher density layers? The total energy input can be multiplied by a penetration coefficient to take into account these possibilities. Nevertheless, the results obtained by the Kosovichev & Zarchova show the seismic event to mainly preserve in the photosphere the spatial and temporal characteristics observed in the higher layers. Moreover, signatures at the sodium formation layers have been reported (Warmuth et al. 2000) and are commonly revealed (Moretti et al., in preparation). This suggests the penetration of such jets to be high.

Larger flares have longer temporal and larger spatial scales and should not satisfy the threshold imposed by the energy deposition rate to initiate the explosions. Hence, these events should not be included in any energy balance. Furthermore, the sole energy input estimate is not a selective constraint for the candidates of the source: the involved energy is so small that many processes can easily match the values.

In summary, if we believe that the downward plasma jets produced by the explosive chromospheric evaporation release their energy "somewhere'' in the photosphere, the observed energy input for the p-modes can be explained. Moreover, these plumes show the same spatial and temporal characteristics of the events required to explain the solar background as the signature of the source of the oscillations. Last, but not least, the results from a local analysis, even with a low spatial resolution, suggest that the location of the solar background mimics the same events.

We are not claiming the downflows observed in the photosphere to be the "surviving" tails of the jets produced above the surface, but rather, suggest that this plasma rain is a forcing mechanism for the convective instabilities causing the downflows. Higher resolution, multi-wavelength data are needed to confirm a physical connection between the downward plasma injected from the explosive chromospheric events and the photospheric motions underlying the negative intensity-velocity phase difference regime.

Acknowledgements

P.F.M. thanks Federica Brandizzi.
We also thank V. Andretta, E. Busá, B. Cullmann, J. Ireland, S. M. Jefferies, C. Lindsey, Th. Pettauer and Th. Straus.