Extended Cyclotron Resonant Heating of the Turbulent Solar Wind
Abstract
Circularly polarized, nearly parallel propagating waves are prevalent in the solar wind at ion-kinetic scales. At these scales, the spectrum of turbulent fluctuations in the solar wind steepens, often called the transition-range, before flattening at sub-ion scales. Circularly polarized waves have been proposed as a mechanism to couple electromagnetic fluctuations to ion gyromotion, enabling ion-scale dissipation that results in observed ion-scale steepening. Here, we study Parker Solar Probe observations of an extended stream of fast solar wind ranging from . We demonstrate that, throughout the stream, transition-range steepening at ion-scales is associated with the presence of significant left-handed ion-kinetic scale waves, which are thought to be ion-cyclotron waves. We implement quasilinear theory to compute the rate at which ions are heated via cyclotron resonance with the observed circularly polarized waves given the empirically measured proton velocity distribution functions. We apply the Von Kármán decay law to estimate the turbulent decay of the large-scale fluctuations, which is equal to the turbulent energy cascade rate. We find that the ion-cyclotron heating rates are correlated with, and amount to a significant fraction of, the turbulent energy cascade rate, implying that cyclotron heating is an important dissipation mechanism in the solar wind.
1 Introduction
Weakly collisional plasmas, which are common in astrophysical environments, are fundamentally governed by kinetic processes (Marsch, 2006). Our understanding of kinetic processes responsible for turbulent dissipation, heating, and energy transfer in collisionless environments is relatively incomplete, and necessary to explain phenomena such as solar wind acceleration and coronal heating (Parker, 1958; Richardson et al., 1995; Hellinger et al., 2013; Fox et al., 2016), and analogous astrophysical processes.
Recent work on the near-Sun solar wind has highlighted the significant presence of circularly polarized ion-scale waves (Bale et al., 2019; Bowen et al., 2020a), and their association with non-thermal features in particle distributions (Verniero et al., 2020; Klein et al., 2021; Verniero et al., 2022). These waves are characterized by their quasi-parallel propagation along the magnetic field (Jian et al., 2014; Boardsen et al., 2015; Bowen et al., 2020a; Liu et al., 2023). Electric field measurements suggest that these wave predominantly propagate outward from the sun (Bowen et al., 2020b), similar to the observed propagation direction of larger-scale Alfvénic turbulent fluctuations (Roberts et al., 1987; Tu & Marsch, 1995; Bavassano et al., 1998; McManus et al., 2020).
While ion-scale waves are often associated with processes related to kinetic plasma distributions, e.g. instabilities and resonant damping, (Gary, 1993; Isenberg & Lee, 1996; Hollweg & Markovskii, 2002; Marsch, 2006; Klein et al., 2018, 2021), they may also serve as a mechanism to transfer turbulent energy via cyclotron resonance to particle thermal motion (Hollweg & Johnson, 1988; Tu & Marsch, 1997; Cranmer, 2000; Hollweg & Isenberg, 2002; Cranmer, 2014). Signatures of quasilinear cyclotron resonance in the observed proton distribution functions have been suggested in various spacecraft observations (Marsch & Tu, 2001a; He et al., 2015; Verniero et al., 2022; Bowen et al., 2022). Furthermore, steepening of turbulent spectra at ion cyclotron resonant scales has been interpreted as a signature of cyclotron resonant damping (Denskat et al., 1983; Woodham et al., 2018; Lotz et al., 2023) and is correlated with the presence of circularly polarized signatures (Goldstein et al., 1994; Leamon et al., 1998; He et al., 2011; Lion et al., 2016; Zhao et al., 2021). These helical signatures are both correlated with proton temperature anisotropy as well as the turbulent amplitudes (Telloni et al., 2019), suggesting they may be associated with turbulent dissipation.
The quasilinear damping of the in situ population of ion cyclotron waves (ICW) has been measured with a heating rate accounting for 10-20% of the turbulent energy flux (Bowen et al., 2022). Further studies have shown the direct transfer of energy from waves to protons using wave-particle correlation methods (Vech et al., 2020; Luo et al., 2022). The presence of these waves has been observed to correlate both to turbulent features such as the large-scale cross helicity and sub-ion-scale intermittency (Bowen et al., 2023).
In this Letter, we study an extended stream of fast solar wind observed by PSP from to with persistent signatures of left-hand polarized waves. We show that left-handed polarization directly corresponds to turbulent steepening in the ion-kinetic scale transition range (Sahraoui et al., 2009; Kiyani et al., 2009; Bowen et al., 2020c; Duan et al., 2020, 2021), which has historically been interpreted as a signature of dissipation (Denskat et al., 1983; Goldstein et al., 1994; Leamon et al., 1998; Smith et al., 2006, 2012; Lion et al., 2016; Bowen et al., 2020c, 2023). Using the cold plasma dispersion to determine the internal energy and Poynting flux of the measured wave spectrum suggests that the energy to generate waves is stored within the turbulent fluctuations, indicating that the waves are a pathway to ion-scale turbulent dissipation. By applying drifting bi-Maxwellian fits to the distribution function, we estimate the empirical quasilinear heating rate of the local ion-scale waves (Kennel & Engelmann, 1966; Isenberg & Lee, 1996; Bowen et al., 2022). We similarly estimate local turbulent dissipation rates via turbulent amplitudes and the Von Kármán decay law to determine the decay of the largest, outer-scale turbulent fluctuations (Hossain et al., 1995; Wan et al., 2012; Bandyopadhyay et al., 2020; Wu et al., 2022). We demonstrate strong correlations between the quasilinear heating and turbulent dissipation rates. These strong correlations are evident in both global scaling as well as local fluctuations in the turbulent cascade and quasilinear heating rates. These results suggest that cyclotron heating plays an important role in extended solar wind heating. These results may have strong implications for the nature of ion-scale heating in the corona and other weakly-collisonal astrophysical plasmas.
2 Methods & Results
We focus on a stream of fast solar wind observed in the inner heliosphere by Parker Solar Probe (PSP) from 11/16/2021 to 11/20/2021. During this interval PSP was flying inwards towards the sun in near co-rotation with the solar surface, sampling a relatively singular source region over the range of 14-55 (Badman et al., 2023; Davis et al., 2023). We use measurements from the PSP Solar wind Electron Alpha and Proton experiment’s Solar Probe Analyzer (Livi et al., 2022, SPAN). The PSP FIELDS experiment (Bale et al., 2016) provides measurements of the magnetic field from DC to sub-ion kinetic scales using merged search coil (Jannet et al., 2021) and fluxgate magnetometer (SCaM) measurements (Bowen et al., 2020d).
We break the 4 day stream into a set of intervals of 128 s with 50% overlap to study the evolution of the turbulent spectra alongside ion-scale waves. SCaM data are sensitive well into the sub-ion kinetic scales (Dudok de Wit et al., 2022), but are only available for two axes; as a result, the SCaM data are only used to measure the shape of the spectra. The fluxgate magnetometer data provides three component magnetic field measurements that enable comprehensive study of the properties of observed waves. The fluxgate magnetometer and SCaM data are re-sampled to a uniform 146.4845 Sa/sec rate. Data were discarded when discontinuities or changes in instrumental modes led to artifacts. In total 6737 intervals of 128 s were analyzed.
Turbulent Spectra
We compute spectral densities , with units nT2/Hz, from the two-component SCaM data through ensemble averaging over 4096 point FFTs in each 128s interval. Each spectral density is interpolated onto 320 logarithmically spaced frequencies. Figure 1(a) shows spectral densities computed from the strem intervals, though for clarity only every 40th spectra is plotted. The colors corresponding to radial distance from the sun, which is between 14.1-55.0 ; lighter yellow colors are closer to the Sun, whereas darker colors correspond to further heliocentric distances.
An estimate of the local spectral index for each frequency is obtained by performing a linear least-square fit of in a moving window consisting of 60 neighboring logarithmically spaced frequencies. The slope of a linear least-square fit in log space provides a measure of the local spectral index . Figure 1(b) shows the locally measured as a function of frequency for spectra in Figure 1(a). The observed spectra are consistent with three-power-law spectra often reported in the solar wind (Sahraoui et al., 2009; Alexandrova et al., 2008; Bowen et al., 2020c): each spectra has a characteristic low-frequency inertial range scaling , a transition range with steepened index , followed by flattening to an index at higher frequencies. The shape of the spectra is consistent but shifts to higher frequencies closer to the Sun. The index of the steep transition range spectra varies, and can have values of up to -10. The extremely steep transition range indices in this stream are predominantly associated with parallel spectra as shown in Duan et al. (2021). As we demonstrate, this steepening is also associated with significant ICW populations (Bowen et al., 2020a, 2023).
To determine the slope of the transition range we implement a piecewise, three power-law (3PL) fit, following methods developed in Bowen et al. (2020c). Each of the three ranges is modeled as a power-law spectra with indices , , and separated by break-point frequencies and , referring to the respective inertial-transition and transition-kinetic breaks. Fig. 1(c) shows the 3PL fits, corresponding to the spectra in Figure 1(a). Additionally, we apply the procedure to measure the local spectral index to the 3PL-fits, with the results shown in Figure 1(d). There is good agreement between the observations and the 3PL fits.
Break Scales & Cyclotron Resonance
We compare the measured break scales and to various physical ion-kinetic scales: the ion gyroradius, , where the ion thermal speed is and the ion gyrofrequency is ; and the ion inertial scale , where . The fundamental charge is given as , is the permeability of free space, is the proton temperature, is the average background magnetic field, is the background number density. The wavenumbers corresponding to the ion-inertial and ion gyroscale are and . We approximate the wavenumber corresponding to cyclotron resonant interactions as , where is the Alfvén speed (Leamon et al., 1998; Wicks et al., 2016; Woodham et al., 2018; Bowen et al., 2020a). This approximation for is derived via the cyclotron resonance condition between outward going ICW and protons flowing inwards in the plasma frame, (Leamon et al., 1998), and a low frequency limit to the ion-cyclotron dispersion for parallel propagating waves. The cold-plasma dispersion relation is
(1) |
For the ion-scales , , and , each wavenumber is converted to an effective spacecraft frequency first using the Taylor hypothesis, to associate ion-kinetic scales with spacecraft frequencies ,, and . For the cyclotron-resonant scale, , we also consider a correction to the Taylor hypothesis made by incorporating the Doppler shift equation
(2) |
to improve on our estimate of the frequency corresponding to cyclotron resonant interactions as
(3) |
which includes the contribution from evaluated at . For calculating , we assume that the angle between the mean magnetic field and flow direction, , is small such that and thus for parallel propagating ICWs.
Fig. 2(a) shows the computed kinetic-scale frequencies alongside the measured breaks from the 3PL fitting algorithm against . Fig. 2(b) shows distributions ratios of , and . These distributions provide a measure of agreement between break frequencies and the measured ion kinetic scales. Recent work has suggested that the cyclotron scale corresponds well with (Woodham et al., 2018; Vech et al., 2018; Duan et al., 2020; Lotz et al., 2023), here we find that is statistically lower than ; however, the corrected frequency well approximates the break. The transition range break additionally approximates , such that we cannot distinguish the inertial scale from the cyclotron resonant scale. We find that is within the transition range, which is consistent with previous results (Bowen et al., 2020c).
Importantly, and are estimates to a single cyclotron resonant scale corresponding to resonance between particles at the thermal speed , with low-frequency ICWs, . However, observations from the solar wind suggest that ICWs occur over a range of frequencies corresponding to (Bowen et al., 2020a). Empirical determination of circular polarization as a function of spacecraft frequency along with the cold plasma dispersion for parallel propagating ICWs Eq. (1) provides information regarding the wavenumbers at which cyclotron resonant waves occur (Bowen et al., 2020b).
A 64-scale Morlet wavelet transform is applied to each interval to compute a power spectral density with units nT2/Hz (Farge, 1992; Dudok de Wit et al., 2013; Bowen et al., 2020a). We extract circular polarization of the field via the reduced magnetic-helicity (Howes & Quataert, 2010),
(4) |
of the wavelet coefficients perpendicular to the mean field with left/right handed polarization represented by positive/negative . We calculate the left-handed polarized power spectra by filtering out power with . The normalized fractional left handed spectra are computed as . Very little right handed polarization is present in this stream.
Based on previous measurements of the phase speed of circularly polarized waves (Bowen et al., 2020b), which show a strong statistical preference for outward-propagation, we can assume that the waves are outward-propagating ICWs, travelling parallel to the mean field. The wavelet transform gives circular polarization as a function of the spacecraft frame frequency, . The use of the Doppler shift Eq. (2) in the parallel-propagating limit gives
(5) |
where we no longer assume either the Taylor Hypothesis or that . The combination of Eq. (5) with Eq. (1) determines the corresponding parallel wave-number for each spacecraft frequency (Bowen et al., 2020b). Figure 3(a) shows the normalized as function of computed from combining Eq. (1) with Eq. (5). At all radial distances the ICW population appears at , where ICWs are inherently dispersive, with a finite bandwidth in wave number ranging from approximately . The finite bandwidth indicates that, at each time, a range of particle velocities simultaneously satisfy the ICW resonance condition. It is important to note that the spacecraft reaction wheels, highlighted in teal lines in Figure 3(b-d), contribute circularly polarized power at higher (Bowen et al., 2020a).
Fig. 3(b) shows as a function of spacecraft frequency at each interval studied as a function of solar radial distance. We plot both the and breaks. The break agrees very well with the circularly polarized regime, while the break bounds the circular polarization at higher frequencies, indicating that the transition range is entirely circularly polarized. We additionally plot the cyclotron scale , which approximates the break.
Cyclotron Waves and Turbulent Dissipation
Using the cold-plasma dispersion and the Doppler shift Eqs. (1) & (2) and our observations of we measure both the internal energy density of the waves and the magnitude of the Poynting flux computed in an Heliocentric inertial frame (HCI), assuming the ICW propagation is entirely parallel the radial solar wind flow (Karpman, 1974; Shklyar & Matsumoto, 2009):
(6) | |||
(7) |
where and are the plasma frame phase and group velocities respectively. As the units of are nT2/Hz both and are defined as spectral densities. The internal energy and Poynting flux magnitude corresponding to each wavelet coefficient are obtained by multiplying and by the bandwidth of each wavelet , where the derivative and are obtained from Eq. 5. Figure 3 (c&d) show and at each frequency over the interval.
The total Poynting flux and internal energy can be integrated at each time to give total values and for each interval. Due to the sporadic noise that occurs at low frequencies, as well as high frequency contributions to polarization from the spacecraft reaction wheels, we zero out contributions to and where and .
We use and as proxies for the ICWs in each interval. Fig. 4 shows how these quantities relate to parameters associated with turbulence, and as well as the total RMS turbulent amplitude computed in each 128 s interval. We consider only intervals with in order to control for effects associated with anisotropy, which affect observational signatures of both the waves and the turbulence (Chen et al., 2010; Bowen et al., 2020a); of the total 6736 intervals, 3172 occur with . In each panel of Fig. 4 data is plotted with the color scale in Fig. 1 with lighter colors corresponding to intervals closer to the sun. We compute lines of best fit for data within , , , and , which are shown in each panel of Fig. 4. Fig. 4(a) shows that the transition range slope is strongly correlated with the level of circular polarization, (Bowen et al., 2023), with similar trends observed at all solar radii. Fig. 4(b) shows that the internal energy of the waves is anti-correlated with , such that steeper slopes contain greater amounts of ICW energy. This correlation is present at all , though at larger distances the maximum measured are similar to the lowest close to the sun, such that constraining the data by is necessary to see the correlations. Fig. 4(c) shows that the internal energy contained within the circularly polarized ion-scale transition range is globally a function of the amplitude of the turbulent fluctuations computed at the 128 s scale, consistent with previous results in Shankarappa et al. (2023).
ICW Heating Rates
Our analysis of the transition range reveals signatures of circularly polarized ion-cyclotron resonant waves; we apply well established principles of energy conservation and quasilinear heating to connect these waves to turbulent dissipation.
The Poynting theorem states that the change in internal energy of the ICW population is
(8) |
where is the dissipation rate of the waves and is a driving term. Assuming a steady state, we compute the three terms on the right side of Eq. (8). We approximate in the heliocentric inertial frame as the radial derivative
(9) |
We estimate using the quasilinear heating rate (Kennel & Engelmann, 1966) with the empirically observed spectrum of cyclotron waves
(10) |
For each SPANi measurement, we perform a drifting bi-Maxwellian fit in the proton core frame with the form
(11) |
which has anisotropic thermal speeds perpendicular and parallel the mean field ( and for the beam and core (subscripts, b and c): and a relative drift, , parallel to the mean magnetic field (Marsch, 2006; Klein et al., 2021). We compute the average fit parameters over the 128s interval such that an average gyrotropic is computed. Using , and the observed spectrum of cyclotron waves , we use the cold plasma dispersion, Eq. 1, with wave numbers defined from Eq. 5, to determine the volumetric heating rate in the local interval:
(12) |
with
(13) |
(Kennel & Engelmann, 1966). A gyrotropic differential volumetric heating rate can be defined from . We outline how to calculate and in Appendix A through using the -function to evaluate the integral over wave number and the subsequent numerical integration of over velocity coordinates.
Figure 5(a-b) shows , the volumetric heating rate as function of resonant parallel velocity, which is computed from integrating over perpendicular velocities, as a function of solar radius. Figure 5(a) shows positive , at resonant velocities less than the thermal speed (shown in yellow), corresponding to wave-absorption and consequent plasma heating. The term, which is the resolution of the numerical integration 1 km/s, is included to normalize the differential in terms of a volumetric heating rate W/m3. Figure 5(b) shows wave emission, which, consistent with Bowen et al. (2022), is significantly less energetically relevant than the regions with positive wave-absorption, which indicates net resonant-heating of the plasma.
Importantly, the use of a biMaxwellian model to approximate the distribution may not capture non-thermal features that resonate with the observed populations of waves (Dum et al., 1980; Viñas & Gurgiolo, 2009). In Appendix B, we implement two well understood non-parametric techniques to model the observed distributions, Hermite polynomials and radial basis functions, to verify the independence of our results from the biMaxwellian model. While there are slight differences in the level of heating predicted by each model, the qualitative interpretation of ICW resonance heating is found in each case and the average values closely follow that of the drifting biMaxwellian.
Connecting ICW Heating to the Turbulent Cascade
The importance of resonant heating can be understood by comparing the net heating rates to the turbulent cascade rate . Following (Bandyopadhyay et al., 2020) we compute the energy cascade rate of the turbulence assuming a von Kármán decay law.
(14) |
where are root mean square fluctuation amplitudes of the Elsasser variables The correlation length is determined from the solar-wind advected spatial-scale corresponding to the time-lag for which the correlation reduces below a factor . The coefficient is determined from numerical means (Hossain et al., 1995; Wan et al., 2012; Usmanov et al., 2014; Bandyopadhyay et al., 2018) and chosen to agree with Bandyopadhyay et al. (2020) and Wu et al. (2022). However, this value of is derived assuming low cross helicity, while the near-Sun solar wind is characterized by imbalance with (McManus et al., 2020). Note the units for in equation 14 are of J/s/kg; we define with units of W/m3, which are comparable to the volumetric heating rates in Equation 2, through multiplying by the background mass density
Measurement of via Eq. (14), which we assume approximates the total cascade rate , requires the Elsasser amplitude at the outer, energy containing scales, of the turbulence. Davis et al. (2023) have studied the evolution of the outer scales in this same stream and found that the outer-scale spectral break evolves from Hz close to 14 to Hz at . Accordingly, we compute the cascade rate in successive 30 minute intervals (corresponding to Hz) with overlap. In each 30 minute interval we then compute the average heating rate and .
Figure 5(c) shows , and as a function of radial distance. A surprisingly good agreement is found between and , while the divergence of the Poynting flux term is significantly smaller. The relatively small indicates that the wave dynamics are dominated by the source, , and dissipation, , and that the waves do not propagate significantly. This result is consistent with previous studies arguing for in situ driving of these waves (Bowen et al., 2020a; Vech et al., 2020; Liu et al., 2023).
We find a Spearman, non-parametric, ranked correlation between and of 0.79; we further compute the Pearson-correlation of the logarithmic quantities as 0.88. Both of these values indicate strong correlations. Furthermore, we fit each of and to a power-law in noted as and . We compute the local variations from the global power-law trends using and - as a measure of how closely the local quasilinear heating rate follows the turbulent energy cascade rate. Figure 5(d) shows the local fluctuations - against . We measure a Spearman ranked-correlation of 0.56.
Additionally, we interpolate , measured at a 30 minute cadence, onto the 128 second cadence of the measurements. Figure 5(e) shows a 2D-histogram of against the interpolated ; the distribution is column-normalized to the measured cascade rate. Good correlations are obtained, with a Spearman ranked-correlation of 0.58. These results suggest that characteristically, the quasilinear heating rate is proportional to the energy cascade rate with a constant of proportionality near unity. Figure 5(f) shows a 2D-histogram of , where the interpolated is again used, against solar radius ; the distribution is column-normalized to . Implicitly, this suggests that much of the turbulent cascade flux enters the ICW population such that with over the vast majority of the observed stream. There is some radial trend, with at and at larger distances. Values with suggest dissipation in excess of the turbulent cascade rate, we believe this is predominantly due to uncertainty in the measurements of and
3 Conclusions
The nature of the steepening of turbulent energy spectrum at ion-kinetic scales has long been debated and is an important signature in understanding collisionless turbulent dissipation (Denskat et al., 1983; Smith et al., 1990; Goldstein et al., 1994; Leamon et al., 1998). Studies of the location of the ion-kinetic scale break frequency suggest that the most likely candidate scale corresponds to resonance between outward going ICWs and the thermal ion population (Woodham et al., 2018; Vech et al., 2018; Duan et al., 2020; Lotz et al., 2023). Previous observations from 1 AU have suggested that ICW play a role in dissipating solar wind turbulence (Leamon et al., 1998; Lion et al., 2016; Woodham et al., 2019; Telloni et al., 2019; Zhao et al., 2021; Luo et al., 2022). In this Letter, we concretely demonstrate that ion-scale spectral steepening is associated with circular polarization. The break between the inerital and transition ranges follows the regime of ICWs, and the spectral break between the transition range and subion-scale turbulence bounds the circularly polarized waves exactly at higher frequencies. We suggest that the transition range corresponds to cyclotron waves that are associated with turbulent dissipation and ion-scale heating (Bowen et al., 2023). Our results show that internal energy and momentum of the waves is correlated with the amplitude of the turbulent fluctuations, Figure 3, suggesting that the energy contained in, and transported by, the waves originates from the turbulent fluctuations. These results are consistent with earlier observations from 1 AU suggesting that the level of steepening in the turbulent cascade may relate to the level of Alfvénicity and the turbulent cascade rate (Smith et al., 2006; Bruno et al., 2014).
In this Letter, we compute two measures of energy transfer: first, the quasilinear heating rate of kinetic scale ICWs, , and second, the von Kármán turbulent decay rate . The measurements of are obtained through integrating kinetic phase space densities measured by SPANi, while is determined from outer scale turbulent fluctuations. It is striking that these quantitative estimates of energy transfer, and , obtained via entirely different methods at vastly different spatial scales show significant correlations. Such strong correlations suggest that parallel-cyclotron resonance dissipates significant amounts of turbulent energy into the solar-wind ion populations. The correlation between energy flux and quasilinear heating rate is observed in both the global scaling of the quantities as well as in local variations: i.e., fluctuations in the turbulent cascade rate are typically accompanied by correlated variations in the quasilinear heating rate. These novel measurements, which support a radially extended cyclotron heating mechanism, are important in understanding how turbulent dissipation results in heating the expanding solar wind. These results further indicate significant progress on PSP’s objective “to trace the flow of energy that heats and accelerates the solar corona and solar wind” (Fox et al., 2016). Furthermore, the extended ICW heating, which is observed from 15 to 55 , may have significant implications for the role of ICW heating in the corona, (Hollweg & Johnson, 1988; Cranmer, 2000).
Recent simulations by Squire et al. (2022) suggest that the dissipation of strongly imbalanced Alfvénic turbulence may result in polarization signatures similar to those measured in the solar wind (Podesta & Gary, 2011; He et al., 2011; Huang et al., 2020). The Squire et al. (2022) simulations, and the underlying idea of a helicity barrier whereby imbalanced turbulence is prevented from cascading to small scales due to a conserved generalized helicity (Meyrand et al., 2021), provides a novel framework to understand the connection between turbulence and ion-cyclotron waves. These theoretical ideas have found support in observations showing that ion-scale ICWs preferentially occur when large-scale fluctuations are highly Alfvénic (Bowen et al., 2023) and that sub-ion scale intermittency depends largely on the level of ICWs present, which is in turn correlated to the cross-helicity, suggesting that ICWs play a role in the dissipation of imbalanced turbulence.
The idea that parallel propagating ICWs dissipate a predominantly perpendicular cascade challenges our current understandings of solar wind turbulence and dissipation. It has long been suggested that the turbulence should drive fluctuations into smaller perpendicular scales (Shebalin et al., 1983), and observations regularly indicate significant anisotropy perpendicular the mean field (Horbury et al., 2008; Chen et al., 2010; Duan et al., 2021). The hybrid simulations of the helicity barrier by Squire et al. (2022) suggest that oblique ICWs are the primary heating mechanism and that parallel ICWs are emitted as a secondary process (Chandran et al., 2010) through the Alfvén/ion cyclotron instability (Gary, 1993).
While the secondary emission of parallel ICWs from oblique-ICW heating provides a mechanism to generate parallel ICWs from turbulent heating, in such a mechanism, the plasma is heated by oblique ICWs but cooled by the parallel ICWs (since they are emitted as an instability). In contrast, our measurements indicate that the parallel ICWs are robustly heating the plasma, suggesting they should be directly driven by the turbulence, as opposed to generated via oblique ICW heating. If our present measurements, consistent with previous observations (Bowen et al., 2022), are correct, this presents a conundrum, given the difficulty of sourcing quasi-parallel waves from perpendicular turbulent structures. Our use of non-parametric representations of (Appendix B) further complicates this issue, as our measurement of ICW heating is not simply due to the incorrect parameterization of the SPANi observations with a drifting-bi-Maxwellian fit.
A possible source of systematic error is in the assumption of a cold-plasma. We assume a cold plasma dispersion (Stix, 1992), that is likely not entirely accurate for the solar wind. Implementation of warm-plasma dispersion solvers with arbitrary distribution functions (Verscharen et al., 2017; Walters et al., 2023) may enable greater understanding of parallel wave-generation. Furthermore, inclusion of beam populations (Verniero et al., 2020; Ofman et al., 2022) and -particles may affect instabilities (McManus et al., 2023) and may help further understanding of these dynamics.
Outside of uncertainties in the distribution function and wave-dispersion relations, a possible explanation for the in situ production of ICWs that result in solar wind heating could include the adiabatic evolution associated with expansion. Adiabatic evolution associated with expansion (Chew et al., 1956) should drive distribution functions that are unstable and emitting ICWs towards stability. The extent that expansion may enable ICW heating rather than cooling within the helicity-barrier framework, remains largely unstudied.
Furthermore, estimates of the cascade rate are notoriously difficult to estimate via single spacecraft measurements (Bandyopadhyay et al., 2018, 2020). Care must be taken in understanding these cascade rates, especially in strongly imbalanced states, in which a stationary energy flux may not exist, or a negative cascade rate may occur (Smith et al., 2009; Meyrand et al., 2021). This is especially important in the regime of imbalanced turbulence frequently observed by PSP (Meyrand et al., 2021; Squire et al., 2022). However, recently Wu et al. (2022) found good agreement between the Von Kármán decay rates with measured perpendicular proton heating rates, suggesting that the turbulent dissipation can be studied via these methods. While care has to be taken in further studies of cascade rates, the general correspondence between the measured heating rates and the von Kármán decay rates obtained in this present work and by Wu et al. (2022) is promising.
In any case, these robust observations of extended cyclotron resonant process in the inner-heliosphere suggest that kinetic scale waves are strongly coupled to the turbulent cascade and provide an important pathway to the dissipation of turbulence in the solar wind and corona.
4 Acknowledgements
T.A.B. acknowledges NASA Grant No. 80NSSC24K0272. The work of I.V. was supported by NASA grant No. 80NSSC22K1634. B.D.G.C. acknowledges the support of NASA grant 80NSSC24K0171.
References
- Alexandrova et al. (2008) Alexandrova, O., Carbone, V., Veltri, P., & Sorriso-Valvo, L. 2008, The Astrophysical Journal, 674, 1153, doi: 10.1086/524056
- Badman et al. (2023) Badman, S. T., Riley, P., Jones, S. I., et al. 2023, arXiv e-prints, arXiv:2303.04852, doi: 10.48550/arXiv.2303.04852
- Bale et al. (2019) Bale, S., Badman, S., Bonnell, J., et al. 2019, Nature, 1
- Bale et al. (2016) Bale, S. D., Goetz, K., Harvey, P. R., et al. 2016, Space Science Rev., 204, 49, doi: 10.1007/s11214-016-0244-5
- Bandyopadhyay et al. (2018) Bandyopadhyay, R., Oughton, S., Wan, M., et al. 2018, Physical Review X, 8, 041052, doi: 10.1103/PhysRevX.8.041052
- Bandyopadhyay et al. (2020) Bandyopadhyay, R., Goldstein, M. L., Maruca, B. A., et al. 2020, ApJs, 246, 48, doi: 10.3847/1538-4365/ab5dae
- Bavassano et al. (1998) Bavassano, B., Pietropaolo, E., & Bruno, R. 1998, JGR, 103, 6521, doi: 10.1029/97JA03029
- Boardsen et al. (2015) Boardsen, S. A., Jian, L. K., Raines, J. L., et al. 2015, Journal of Geophysical Research (Space Physics), 120, 10,207, doi: 10.1002/2015JA021506
- Bowen et al. (2023) Bowen, T. A., Chandran, B. D. G., Klein, K. G., et al. 2023, in 2023 XXXVth General Assembly and Scientific Symposium of the International Union of Radio Science (URSI GASS, 335, doi: 10.23919/URSIGASS57860.2023.10265538
- Bowen et al. (2020a) Bowen, T. A., Mallet, A., Huang, J., et al. 2020a, ApJS, 246, 66, doi: 10.3847/1538-4365/ab6c65
- Bowen et al. (2020b) Bowen, T. A., Bale, S. D., Bonnell, J. W., et al. 2020b, ApJ, 899, 74, doi: 10.3847/1538-4357/ab9f37
- Bowen et al. (2020c) Bowen, T. A., Mallet, A., Bale, S. D., et al. 2020c, PRL, 125, 025102, doi: 10.1103/PhysRevLett.125.025102
- Bowen et al. (2020d) Bowen, T. A., Bale, S. D., Bonnell, J. W., et al. 2020d, Journal of Geophysical Research (Space Physics), 125, e27813, doi: 10.1029/2020JA027813
- Bowen et al. (2022) Bowen, T. A., Chandran, B. D. G., Squire, J., et al. 2022, PRL, 129, 165101, doi: 10.1103/PhysRevLett.129.165101
- Broomhead & Lowe (1988) Broomhead, D., & Lowe, D. 1988, Complex Systems, 2, 321
- Bruno et al. (2014) Bruno, R., Trenchi, L., & Telloni, D. 2014, ApJL, 793, L15, doi: 10.1088/2041-8205/793/1/L15
- Chandran et al. (2010) Chandran, B. D. G., Pongkitiwanichakul, P., Isenberg, P. A., et al. 2010, ApJ, 722, 710, doi: 10.1088/0004-637X/722/1/710
- Chen et al. (2010) Chen, C. H. K., Horbury, T. S., Schekochihin, A. A., et al. 2010, Physical Review Letters, 104, 255002, doi: 10.1103/PhysRevLett.104.255002
- Chew et al. (1956) Chew, G. F., Goldberger, M. L., & Low, F. E. 1956, Proceedings of the Royal Society of London Series A, 236, 112, doi: 10.1098/rspa.1956.0116
- Cranmer (2000) Cranmer, S. R. 2000, The Astrophysical Journal, 532, 1197, doi: 10.1086/308620
- Cranmer (2014) —. 2014, ApJS, 213, 16, doi: 10.1088/0067-0049/213/1/16
- Davis et al. (2023) Davis, N., Chandran, B. D. G., Bowen, T. A., et al. 2023, arXiv e-prints, arXiv:2303.01663, doi: 10.48550/arXiv.2303.01663
- Denskat et al. (1983) Denskat, K. U., Beinroth, H. J., & Neubauer, F. M. 1983, Journal of Geophysics Zeitschrift Geophysik, 54, 60
- Duan et al. (2021) Duan, D., He, J., Bowen, T. A., et al. 2021, ApJL, 915, L8, doi: 10.3847/2041-8213/ac07ac
- Duan et al. (2020) Duan, D., Bowen, T. A., Chen, C. H. K., et al. 2020, ApJS, 246, 55, doi: 10.3847/1538-4365/ab672d
- Dudok de Wit et al. (2013) Dudok de Wit, T., Alexandrova, O., Furno, I., Sorriso-Valvo, L., & Zimbardo, G. 2013, Space Sci. Rev., 178, 665, doi: 10.1007/s11214-013-9974-9
- Dudok de Wit et al. (2022) Dudok de Wit, T., Krasnoselskikh, V. V., Agapitov, O., et al. 2022, Journal of Geophysical Research (Space Physics), 127, e30018, doi: 10.1029/2021JA030018
- Dum et al. (1980) Dum, C. T., Marsch, E., & Pilipp, W. 1980, Journal of Plasma Physics, 23, 91, doi: 10.1017/S0022377800022170
- Farge (1992) Farge, M. 1992, Annual Review of Fluid Mechanics, 24, 395, doi: 10.1146/annurev.fl.24.010192.002143
- Fox et al. (2016) Fox, N. J., Velli, M. C., Bale, S. D., et al. 2016, Space Science Reviews, 204, 7, doi: 10.1007/s11214-015-0211-6
- Gary (1993) Gary, S. P. 1993, Theory of Space Plasma Microinstabilities
- Goldstein et al. (1994) Goldstein, M. L., Roberts, D. A., & Fitch, C. A. 1994, Journal of Geophysical Research, 99, 11519, doi: 10.1029/94JA00789
- He et al. (2011) He, J., Marsch, E., Tu, C., Yao, S., & Tian, H. 2011, The Astrophysical Journal, 731, 85, doi: 10.1088/0004-637X/731/2/85
- He et al. (2015) He, J., Wang, L., Tu, C., Marsch, E., & Zong, Q. 2015, ApJ, 800, L31, doi: 10.1088/2041-8205/800/2/L31
- Hellinger et al. (2013) Hellinger, P., TráVníček, P. M., Štverák, Š., Matteini, L., & Velli, M. 2013, Journal of Geophysical Research (Space Physics), 118, 1351, doi: 10.1002/jgra.50107
- Heuer & Marsch (2007) Heuer, M., & Marsch, E. 2007, Journal of Geophysical Research (Space Physics), 112, A03102, doi: 10.1029/2006JA011979
- Hollweg & Isenberg (2002) Hollweg, J. V., & Isenberg, P. A. 2002, Journal of Geophysical Research (Space Physics), 107, 1147, doi: 10.1029/2001JA000270
- Hollweg & Johnson (1988) Hollweg, J. V., & Johnson, W. 1988, JGR, 93, 9547, doi: 10.1029/JA093iA09p09547
- Hollweg & Markovskii (2002) Hollweg, J. V., & Markovskii, S. A. 2002, Journal of Geophysical Research (Space Physics), 107, 1080, doi: 10.1029/2001JA000205
- Horbury et al. (2008) Horbury, T. S., Forman, M., & Oughton, S. 2008, PRL, 101, 175005, doi: 10.1103/PhysRevLett.101.175005
- Hossain et al. (1995) Hossain, M., Gray, P. C., Pontius, Duane H., J., Matthaeus, W. H., & Oughton, S. 1995, Physics of Fluids, 7, 2886, doi: 10.1063/1.868665
- Howes & Quataert (2010) Howes, G. G., & Quataert, E. 2010, The Astrophysical Journal Letters, 709, L49, doi: 10.1088/2041-8205/709/1/L49
- Huang et al. (2020) Huang, S. Y., Zhang, J., Sahraoui, F., et al. 2020, ApJL, 897, L3, doi: 10.3847/2041-8213/ab9abb
- Isenberg & Lee (1996) Isenberg, P. A., & Lee, M. A. 1996, JGR, 101, 11055, doi: 10.1029/96JA00293
- Jannet et al. (2021) Jannet, G., Dudok de Wit, T., Krasnoselskikh, V., et al. 2021, Journal of Geophysical Research (Space Physics), 126, e28543, doi: 10.1029/2020JA028543
- Jian et al. (2014) Jian, L. K., Wei, H. Y., Russell, C. T., et al. 2014, ApJ, 786, 123, doi: 10.1088/0004-637X/786/2/123
- Karpman (1974) Karpman, V. I. 1974, SSR, 16, 361, doi: 10.1007/BF00171564
- Kennel & Engelmann (1966) Kennel, C. F., & Engelmann, F. 1966, Physics of Fluids, 9, 2377, doi: 10.1063/1.1761629
- Kiyani et al. (2009) Kiyani, K. H., Chapman, S. C., Khotyaintsev, Y. V., Dunlop, M. W., & Sahraoui, F. 2009, Physical Review Letters, 103, 075006, doi: 10.1103/PhysRevLett.103.075006
- Klein et al. (2018) Klein, K. G., Alterman, B. L., Stevens, M. L., Vech, D., & Kasper, J. C. 2018, PRL, 120, 205102, doi: 10.1103/PhysRevLett.120.205102
- Klein et al. (2021) Klein, K. G., Verniero, J. L., Alterman, B., et al. 2021, ApJ, 909, 7, doi: 10.3847/1538-4357/abd7a0
- Leamon et al. (1998) Leamon, R. J., Smith, C. W., Ness, N. F., Matthaeus, W. H., & Wong, H. K. 1998, Journal of Geophysical Research, 103, 4775, doi: 10.1029/97JA03394
- Lion et al. (2016) Lion, S., Alexandrova, O., & Zaslavsky, A. 2016, The Astrophysical Journal, 824, 47, doi: 10.3847/0004-637X/824/1/47
- Liu et al. (2023) Liu, W., Zhao, J., Wang, T., et al. 2023, ApJ, 951, 69, doi: 10.3847/1538-4357/acd53b
- Livi et al. (2022) Livi, R., Larson, D. E., Kasper, J. C., et al. 2022, ApJ, 938, 138, doi: 10.3847/1538-4357/ac93f5
- Lotz et al. (2023) Lotz, S., Nel, A. E., Wicks, R. T., et al. 2023, ApJ, 942, 93, doi: 10.3847/1538-4357/aca903
- Luo et al. (2022) Luo, Q., Zhu, X., He, J., et al. 2022, ApJ, 928, 36, doi: 10.3847/1538-4357/ac52a9
- Marsch (2006) Marsch, E. 2006, Living Reviews in Solar Physics, 3, 1, doi: 10.12942/lrsp-2006-1
- Marsch & Goldstein (1983) Marsch, E., & Goldstein, H. 1983, J. Geophys. Res., 88, 9933, doi: 10.1029/JA088iA12p09933
- Marsch et al. (1982) Marsch, E., Schwenn, R., Rosenbauer, H., et al. 1982, J. Geophys. Res., 87, 52, doi: 10.1029/JA087iA01p00052
- Marsch & Tu (2001a) Marsch, E., & Tu, C. Y. 2001a, JGR, 106, 8357, doi: 10.1029/2000JA000414
- Marsch & Tu (2001b) —. 2001b, JGR, 106, 227, doi: 10.1029/2000JA000042
- McManus et al. (2020) McManus, M. D., Bowen, T. A., Mallet, A., et al. 2020, ApJS, 246, 67, doi: 10.3847/1538-4365/ab6dce
- McManus et al. (2023) McManus, M. D., Klein, K. G., Larson, D., et al. 2023, arXiv e-prints, arXiv:2310.14136, doi: 10.48550/arXiv.2310.14136
- Meyrand et al. (2021) Meyrand, R., Squire, J., Schekochihin, A. A., & Dorland, W. 2021, Journal of Plasma Physics, 87, 535870301, doi: 10.1017/S0022377821000489
- Ofman et al. (2022) Ofman, L., Boardsen, S. A., Jian, L. K., Verniero, J. L., & Larson, D. 2022, ApJ, 926, 185, doi: 10.3847/1538-4357/ac402c
- Parker (1958) Parker, E. N. 1958, ApJ, 128, 664, doi: 10.1086/146579
- Podesta & Gary (2011) Podesta, J. J., & Gary, S. P. 2011, The Astrophysical Journal, 734, 15, doi: 10.1088/0004-637X/734/1/15
- Richardson et al. (1995) Richardson, J. D., Paularena, K. I., Lazarus, A. J., & Belcher, J. W. 1995, Geophys. Research Letters, 22, 325, doi: 10.1029/94GL03273
- Roberts et al. (1987) Roberts, D. A., Klein, L. W., Goldstein, M. L., & Matthaeus, W. H. 1987, JGR, 92, 11021, doi: 10.1029/JA092iA10p11021
- Sahraoui et al. (2009) Sahraoui, F., Goldstein, M. L., Robert, P., & Khotyaintsev, Y. V. 2009, Physical Review Letters, 102, 231102, doi: 10.1103/PhysRevLett.102.231102
- Shankarappa et al. (2023) Shankarappa, N., Klein, K. G., & Martinović, M. M. 2023, ApJ, 946, 85, doi: 10.3847/1538-4357/acb542
- Shebalin et al. (1983) Shebalin, J. V., Matthaeus, W. H., & Montgomery, D. 1983, Journal of Plasma Physics, 29, 525, doi: 10.1017/S0022377800000933
- Shklyar & Matsumoto (2009) Shklyar, D., & Matsumoto, H. 2009, Surveys in Geophysics, 30, 55, doi: 10.1007/s10712-009-9061-7
- Smith et al. (2006) Smith, C. W., Hamilton, K., Vasquez, B. J., & Leamon, R. J. 2006, The Astrophysical Journal Letters, 645, L85, doi: 10.1086/506151
- Smith et al. (2009) Smith, C. W., Stawarz, J. E., Vasquez, B. J., Forman, M. A., & MacBride, B. T. 2009, PRL, 103, 201101, doi: 10.1103/PhysRevLett.103.201101
- Smith et al. (2012) Smith, C. W., Vasquez, B. J., & Hollweg, J. V. 2012, ApJ, 745, 8, doi: 10.1088/0004-637X/745/1/8
- Smith et al. (1990) Smith, W. C., Matthaeus, H. W., & Ness, F. N. 1990, in International Cosmic Ray Conference, Vol. 5, International Cosmic Ray Conference, 280
- Squire et al. (2022) Squire, J., Meyrand, R., Kunz, M. W., et al. 2022, Nature Astronomy, 6, 715, doi: 10.1038/s41550-022-01624-z
- Stix (1992) Stix, T. H. 1992, Waves in plasmas
- Telloni et al. (2019) Telloni, D., Carbone, F., Bruno, R., et al. 2019, ApJL, 885, L5, doi: 10.3847/2041-8213/ab4c44
- Tu & Marsch (1995) Tu, C. Y., & Marsch, E. 1995, Space Sci Rev, 73, 1, doi: 10.1007/BF00748891
- Tu & Marsch (1997) —. 1997, Sol Phys, 171, 363, doi: 10.1023/A:1004968327196
- Tu & Marsch (2001) —. 2001, JGR, 106, 8233, doi: 10.1029/2000JA000024
- Usmanov et al. (2014) Usmanov, A. V., Goldstein, M. L., & Matthaeus, W. H. 2014, ApJ, 788, 43, doi: 10.1088/0004-637X/788/1/43
- Vech et al. (2018) Vech, D., Mallet, A., Klein, K. G., & Kasper, J. C. 2018, The Astrophysical Journal Letters, 855, L27, doi: 10.3847/2041-8213/aab351
- Vech et al. (2020) Vech, D., Kasper, J. C., Klein, K. G., et al. 2020, ApJS, 246, 52, doi: 10.3847/1538-4365/ab60a2
- Verniero et al. (2020) Verniero, J. L., Larson, D. E., Livi, R., et al. 2020, ApJS, 248, 5, doi: 10.3847/1538-4365/ab86af
- Verniero et al. (2022) Verniero, J. L., Chandran, B. D. G., Larson, D. E., et al. 2022, ApJ, 924, 112, doi: 10.3847/1538-4357/ac36d5
- Verscharen et al. (2017) Verscharen, D., Chen, C. H. K., & Wicks, R. T. 2017, ApJ, 840, 106, doi: 10.3847/1538-4357/aa6a56
- Viñas & Gurgiolo (2009) Viñas, A. F., & Gurgiolo, C. 2009, Journal of Geophysical Research (Space Physics), 114, A01105, doi: 10.1029/2008JA013633
- Walters et al. (2023) Walters, J., Klein, K. G., Lichko, E., et al. 2023, ApJ, 955, 97, doi: 10.3847/1538-4357/acf1fa
- Wan et al. (2012) Wan, M., Matthaeus, W. H., Karimabadi, H., et al. 2012, PRL, 109, 195001, doi: 10.1103/PhysRevLett.109.195001
- Wicks et al. (2016) Wicks, R. T., Alexander, R. L., Stevens, M., et al. 2016, ApJ, 819, 6, doi: 10.3847/0004-637X/819/1/6
- Woodham et al. (2018) Woodham, L. D., Wicks, R. T., Verscharen, D., & Owen, C. J. 2018, The Astrophysical Journal, 856, 49, doi: 10.3847/1538-4357/aab03d
- Woodham et al. (2019) Woodham, L. D., Wicks, R. T., Verscharen, D., et al. 2019, ApJL, 884, L53, doi: 10.3847/2041-8213/ab4adc
- Wu et al. (2022) Wu, H., Tu, C., He, J., Wang, X., & Yang, L. 2022, ApJ, 926, 116, doi: 10.3847/1538-4357/ac4413
- Wüest et al. (2007) Wüest, M., Evans, D. S., & von Steiger, R. 2007, Calibration of Particle Instruments in Space Physics
- Zhao et al. (2021) Zhao, G. Q., Lin, Y., Wang, X. Y., et al. 2021, ApJ, 906, 123, doi: 10.3847/1538-4357/abca3b
Appendix A Numerical Determination of Quasilinear Heating Rate
Determination of and is obtained through integrating Eq 2 by parts to obtain
(A1) |
The function in Eq 2& A1 corresponds to the ICW resonance condition, and defines a single parallel wave number that resonates with particles at each parallel velocity. Change of variables yields a -function for resonant wave number as
(A2) |
where is the group speed evaluated at each . For positive , taken as outward going ICWs, there is no resonance unless , limiting our concern only to negative thermal speeds. We use the function to substitute for , thereby evaluating the integral over wavenumber, Eq A1 becomes
(A3) |
where the spectrum is interpolated from to , constructing . Eq A3 is entirely a function of and and can be calculated via a numerical integration over velocity space. We further identify the differential heating rate
(A4) |
To perform the numerical integral in Eq A3, we evaluate on a 1 km/s 1 km/s grid, which was found to be sufficient for convergence of Eq A3. The range of integration is set to -1000 km/s 0 km/s and 0 km/s 1000 km/s. We have removed contributions to the spectra with and , which for corresponds to an approximate range of thermal speeds between 0.02 and 7.5 . Figure 5(a&b) shows that the parallel thermal speeds hovers near 50 km/s over this entire interval. Thus, the 1000km/s limits are sufficiently large with respect to the thermal and Alfvén speeds for this stream that there are no significant contributions in portion of phase space out of the bounds of integration. The heating rate as a function of parallel resonant velocity is computed by integrating over perpendicular velocities.
These methods have previously been used to obtain heating rates of order of the turbulent cascade rate (Bowen et al., 2022) though this previously studied interval contained significantly less amounts of left-hand polarized waves than this stream. Our results shown in Fig 5 suggest that a great amount of the turbulent energy likely enters the plasma via cyclotron resonant heating.
Appendix B Application of Non-Parametric Representations
While in situ observations of the proton velocity distribution have long been recognized to have a mostly Gaussian shape (Marsch & Goldstein, 1983) and drifting biMaxwellian approximation has regularly been used to approximate ion velocity distributions (Marsch et al., 1982; Marsch & Tu, 2001b; Tu & Marsch, 2001; Heuer & Marsch, 2007; Klein et al., 2021; Verniero et al., 2022; McManus et al., 2023); however, it is clear that deviations from nonthermal structure in the velocity distribution can affect wave-particle resonant processes (Dum et al., 1980; Bowen et al., 2022; Walters et al., 2023). Furthermore, it is well known that the equilibrium kinetic contours for dispersive ICW resonance do not coincide with Maxwellian curvature (Isenberg & Lee, 1996), such that relaxation via ICW resonant processing cannot be entirely captured via bi-Maxwellian approximations (Heuer & Marsch, 2007). In this sense it is important to test the heating rates computed from the drifting biMaxwellian model in Equation 11 against heating rates computed from non-parametric models of the velocity distribution.
B.1 Hermite Polynomial Interpolation
Bowen et al. (2022) previously used linear least-square fits of orthogonal Hermite-polynomials to estimate non-Maxwellian features in the velocity distribution; following this previously developed method, we perform a linear least square fit to the average proton velocity distribution in each 128s interval using the Hermite polynomials , and Hermite functions
(B1) | |||
(B2) | |||
(B3) |
We fit an orthogonal set of Hermite functions to each velocity distribution through constructing the matrix , which consists of the transform coefficients corresponding to each two-dimensional Hermite polynomial combination. Through linear least square fitting of , with elements , to the observed velocity distribution, written as matrix such that and least square fit inversion gives:
(B4) |
where a singular value decomposition determines the pseudo-inverse of . The matrix then gives the best-fit coefficients for the Hermite polynomial composition.
We furthermore adopt errors on associated with Poisson counting, such that the uncertainty at each velocity coordinate is (Wüest et al., 2007). Poisson noise between energy bins is uncorrelated such that we construct the diagonal-weight matrix as with entries can be included in weighted linear least square fitting as such that
(B5) |
The coefficients are then used to construct using Equation B1.
Using , which is an analytic, differentiable funcion, it is possible to separately estimate without the assumption of a drifting bi-Maxwellian structure to the velocity distribution.
B.2 Radial Basis Function Interpolation
We further include another approximation to the velocity distributions through interpolation via radial basis functions (RBFs) (Broomhead & Lowe, 1988; Bowen et al., 2023) to model the distribution function via a summed set of interpolating functions given by
(B6) |
with , where is the center of each interpolating RBF, denoted by . The RBF method requires choice of a basis function, , which we choose to be an isotropic 2D bi-Maxwellian
(B7) |
The number of interpolating functions, , must be specified along with the thermal speed and central location, , of each of the implemented in the interpolation. In principle, does not need to be isotropic and can vary for each ; however, we determine that setting a Maxwellian-RBF at each measured SPAN energy bin with nonzero phase space density gives suitable results. Furthermore, we uniformly set km/s, which we find is suitable for interpolating the proton distribution.
The interpolated velocity distribution via radial basis functions is then given by
(B8) |
with . Determination of the weights is again performed through SVD estimation for the pseudoinverse giving a least square fit of to the observed velocity distribution. We again compute the RBF on each 128s average velocity distribution and include weighted errors in the least square fit corresponding to Poisson noise analogous to Eq B5.
B.3 Comparing Model Distributions
To verify the heating of the plasma via ICW resonant interactions, we recompute Equation 2 using the Hermite polynomial and radial basis function approximations to the proton distribution, and , we refer to the heating rates computed from Equation 2 via the drifting biMaxwellian, Hermite polynomials, and RBF interpolation as , , and respectively, we also compute the average heating rate of the three terms . Note that the ICW spectrum in unaffected by the choice of model as is a function only of the magnetic field strength and total plasma density, which are independent of the kinetic phase-space distribution.
The top row of Figure 6, panels (a-c), respectively show , , and against . While all computed heating rates have the same qualitative behavior, i.e. positive valued indicating the absorption of ICWs and plasma heating via cyclotron resonance, tends to systematically underestimate the heating rates predicted by , whereas tends to overestimate . The average of all heating rates closely follows , thus was chosen as the value reported in the body of the manuscript.
The bottom panels shows statistics on the quality of the various models. To compare the quality of the relative models we compute two parameters. Figure 6(d) shows the distribution of measured Pearson correlation coefficients between the logarithm of the observed distribution and the three models , , and . High levels of correlation indicate that the model correctly matches the observed distribution. The use of the logarithm of the distribution in computing the Pearson correlation coefficient allows for weighting of the entirety of phase space and includes the regions where ICW resonant interactions are most important; without the logarithmic weighting, the correlation coefficient is mostly dominated by the models performance at the peak of the distribution. We find is largest for the RBF interpolation and smallest for the parametric drifting biMaxwellian fit. Interestingly, the level of heating computed between these models remains relatively constant even though the values vary, this suggests that variations between the three models are mostly occurring in regions that are not resonant with the outward going ICWs. We speculate these differences are likely due to the models ability to approximate the ion-beam population.
We also compute the of each of the models. Given the nature of the nonlinear fitting there is lack of clear number of degrees of freedom. Furthermore errors in the model are not likely to obey Gaussian statistics. For these reasons hypothesis testing via the statistic is not possible. To determine the quality of the fits we compare the ratio of to the value compared for the biMaxwellian, . Figure 6(e) shows the distribution of the logarithmic ratio of and . The RBF model tends to perform much netter than the drifting biMaxwellian with a uniform . In terms of , the Hermite polynomial approximation tends to have a similar performance to the drifting biMaxwellian fit in this stream, though the Pearson correlation is significantly better than the biMaxwellian fit. Again, the ability of each of models to reproduce similar heating rates, but with significantly different suggests that the main difference in the models may be in reproducing the ion-beam. Our future work will investigate further the role of the ion-beam in these processes.
In any case, the use of these various parametric and non-parametric models for the observed proton distribution function all recover heating rates that suggest that ICW are absorbed into the core of the proton distribution, resulting in plasma heating via ICW resonance. These results suggest that the model used may not be inherently important in understanding ICW heating of the proton core.