Both plasmaspheric hiss and chorus waves were observed simultaneously by the two Van Allen Probes in association with substorm-injected energetic electrons. Probe A, located inside the plasmasphere in the postdawn sector, observed intense plasmaspheric hiss, whereas Probe B observed chorus waves outside the plasmasphere just before dawn. Dispersed injections of energetic electrons were observed in the dayside outer plasmasphere associated with significant intensification of plasmaspheric hiss at frequencies down to ~20 Hz, much lower than typical hiss wave frequencies of 100–2000 Hz. In the outer plasmasphere, the upper energy of injected electrons agrees well with the minimum cyclotron resonant energy calculated for the lower cutoff frequency of the observed hiss, and computed convective linear growth rates indicate instability at the observed low frequencies. This suggests that the unusual low-frequency plasmaspheric hiss is likely to be amplified in the outer plasmasphere due to the injected energetic electrons.
 Plasmaspheric hiss is an incoherent, broadband whistler-mode emission, preferentially observed in the high-density region inside the plasmasphere or plumes with a frequency band typically ranging between ~100 Hz and ~2 kHz [Thorne et al., 1973; Meredith et al., 2004]. Although plasmaspheric hiss plays an important role in creating the slot region between the inner and outer radiation belt [Lyons and Thorne, 1973], its origin has been under intense debate over the past few decades. Two mechanisms were initially proposed for the origin of hiss; in situ growth of waves in space [Thorne et al., 1979; Church and Thorne, 1983; Solomon et al., 1988] and lightning generated whistlers [Sonwalkar and Inan, 1989; Green et al., 2005]. However, detailed ray tracing has suggested that plasmaspheric hiss originates from a subset of another type of whistler-mode wave, called chorus, which propagates from an equatorial source region outside the plasmasphere to higher latitudes and is subsequently refracted into the plasmasphere where it evolves into incoherent hiss [Bortnik et al., 2008, 2009]. More recent refinements in the ray tracing show that most of the characteristic features of plasmaspheric hiss can be explained using chorus as the embryonic source [Bortnik et al., 2011; Chen et al., 2012a, 2012b].
 Whistler-mode chorus waves are intense coherent electromagnetic emissions exhibiting discrete rising or falling tones and are predominantly observed outside the plasmasphere [Burtis and Helliwell, 1969; Tsurutani and Smith, 1974; Meredith et al., 2003; Li et al., 2011]. Chorus waves occur in the frequency band spanning 0.1–0.8 fce and typically occur in two distinct frequency bands (lower and upper band) with a gap in wave power at 0.5 fce [e.g., Burtis and Helliwell, 1969; Tsurutani and Smith, 1974; Koons and Roeder, 1990], where fce is equatorial electron cyclotron frequency. Chorus waves are believed to be excited due to cyclotron resonance with anisotropic energetic electrons (a few keV to ~100 keV) injected from the plasma sheet [Kennel and Petschek, 1966; Anderson and Maeda, 1977; Meredith et al., 2001; Omura et al., 2009].
 In this study, we report observations of enhanced plasmaspheric hiss and chorus waves measured simultaneously by the two Van Allen Probes during an electron injection event and provide evidence that the unusual low-frequency component of hiss is enhanced by a cyclotron resonant instability occurring in the outer plasmasphere.
2 Van Allen Probe Data Analysis
 Two identical Van Allen Probes with a perigee of ~1.1 RE, apogee of 5.8 RE, and inclination of 10° [Mauk et al., 2012] are well situated to measure whistler-mode waves in the inner magnetosphere. The Electric and Magnetic Field Instrument Suite and Integrated Science (EMFISIS) provides measurements of DC magnetic fields and a comprehensive set of wave electric and magnetic fields [Kletzing et al., 2013]. The waveform receiver (WFR) on the EMFISIS wave instruments measures wave power spectral density from 10 Hz up to 12 kHz and provides a full 3-D spectral matrix every 6 s along with a selection of burst modes which provide full waveforms with a sampling rate of ~35 kHz [Kletzing et al., 2013]. The high-frequency receiver (HFR) is designed to provide spectral information between 10 and 400 kHz, thus enabling measurements of the upper hybrid frequency, from which the plasma density can be calculated. Plasma density can also be inferred from the spacecraft potential measured by the Electric Field and Waves Instrument. Electron distributions with energy ranging from ~20 keV to ~300 keV, measured every 5.5 s (1/2 spin over 0–180° gyroangle) by Magnetic Electron Ion Spectrometer (MagEIS) [Blake et al., 2013] are used in this study to evaluate the source electrons believed to be responsible for whistler-mode wave generation.
3 Observational Results
 An intense injection event was observed associated with the substantial decrease in the AL index during ~15–18 UT on 30 September 2012, as shown in Figure 1. During this entire interval, Van Allen Probe A was located in the postdawn sector (Figure 4a) and remained inside the plasmasphere (the upper hybrid frequency shown in Figure 1b remained larger than ~100 kHz and thus the calculated plasma density was >100 cm−3). Broadband electromagnetic emissions with frequency from ~20 to ~2000 Hz were observed inside the plasmasphere (Figures 1c and 1d). Wave polarization properties (Figures 1e and 1f) indicate that these waves are plasmaspheric hiss rather than magnetosonic waves [e.g., Thorne, 2010], because they propagate in a quasi field-aligned direction and are predominantly circularly polarized with ellipticity close to 1. A strong intensification of waves was observed at ~14:55 UT and ~16:50 UT (indicated by the green dashed vertical lines) particularly at lower frequencies (20–100 Hz), associated with the increase in electron flux (> ~50 keV) measured by MagEIS (Figure 1g). Van Allen Probe B was located in the predawn sector (Figure 4a), and the plasmapause crossing occurred at ~4.2 RE (white arrow in Figure 1i) at ~16:18 UT. Note that while the plasmapause was located outside the apogee (> 5.8 RE) of the Probe A in the postdawn sector (~8 magnetic local time (MLT)) at ~16:18 UT, the plasmapause was compressed to a much lower value of ~4.2 RE in the predawn sector (~5 MLT), showing a highly asymmetric plasmasphere. Associated with the electron injection at ~16:50 UT (Figure 1n), Probe B observed intensification of both lower and upper band chorus outside of the plasmapause. A time-frequency spectrogram of wave magnetic spectral density obtained from waveform data (Figure 4b) clearly show the broadband structureless hiss observed on Probe A (top) and rising tone chorus elements with a gap at 0.5 fce observed on Probe B (bottom).
 We focus on the intensification of plasmaspheric hiss observed on Probe A during a shorter time interval (14:00–18:30 UT) in Figure 2. Plasma density inferred from the spacecraft potential is shown with the black line in Figure 2a and is consistent with the plasma density calculated from the upper hybrid line (Figure 1b). Interestingly, at ~14:55 UT and ~16:50 UT, a clear increase in electron fluxes up to ~200 keV was observed (Figure 2d) with positive anisotropy (Figure 2e). Here electron anisotropy at each energy bin is calculated using equation (2) in Chen et al. , and positive (negative) values indicate larger (smaller) electron fluxes near 90° pitch angle compared to those at small pitch angles. These two dispersed electron injections were associated with hiss intensification with wave amplitudes (from integrating power spectral density over 20–2000 Hz) up to ~330 pT (blue line in Figure 2a). Another interesting feature, although not the focus of this study, is the remarkable correlation between the hiss wave amplitude and plasma density (highlighted by gray blocks in Figures 2a–2c) particularly in the outer plasmasphere during 14:00–17:00 UT, consistent with the previous study by Chen et al. [2012d].
 Furthermore, the lower cutoff frequency (indicated by the magenta line in Figures 2b and 2c) decreased to ~20 Hz at ~16:50 UT in association with increased wave intensity particularly in the lower frequency range of 20–100 Hz. The electron minimum cyclotron resonant energy was calculated (black line in Figure 2d) for the corresponding lower cutoff wave frequency using in situ measurements of magnetic field intensity and plasma density and was found to agree remarkably well with the upper energy of injected electron fluxes measured by MagEIS in the time interval of 14:00–17:20 UT. Convective linear growth rates for parallel-propagating whistler-mode waves were calculated for various frequencies (Figure 2f) using the measured electron pitch angle distribution and plasma parameters, following the equations described in Appendix A of Summers et al. , which are similar to Kennel and Petschek  but include relativistic effects. The pitch angle distribution of electron phase space density (PSD) observed at 17:10:25 UT by MagEIS is shown with solid lines in Figure 3a, which shows peaks near 90° favorable for wave excitation. A least squares fit using was applied for each energy channel and is shown as the dashed lines, where f is PSD, α is the pitch angle, and An are the fitting parameters. These fitted electron distributions, which agree well with the observations, are used to calculate wave linear growth rates shown in Figure 2f. Electron PSD in momentum space is shown in Figure 3b, and the dashed black lines represent resonant ellipses for various wave frequencies. Note that since electron pitch angle distributions at energies below ~22 keV are not available (because Helium Oxygen Proton Electron [Funsten et al., 2013; Spence et al., 2013] was still in commissioning) during this event, wave growth rates calculated above ~137 Hz (0.02 fce) are not shown. Similarly, the convective growth rates in Figure 2f are only shown below this critical frequency, above which whistler-mode waves can resonate with electrons with energy <22 keV, not measurable by MagEIS. Nevertheless, linear wave growth rates show positive values at lower frequency (< ~100 Hz) after ~15 UT and the excited wave frequency decreased to ~20 Hz at ~17 UT, which is consistent with the observed wave spectrum (Figures 2b and 2c). However, this consistency breaks down after ~17:20 UT, in the region of <4.5 RE, which is also shown in the inconsistency between the calculated minimum resonant energy and upper energy of MagEIS electron flux enhancement (Figure 2d). This suggests that these intensified low-frequency hiss waves (20–100 Hz) during 14:55–17:20 UT (indicated by the red horizontal arrow) observed at relatively large L-shells are likely to be amplified due to the injection of energetic electrons which penetrated into the plasmasphere. However, hiss waves observed at <4.5 RE (after ~17:20 UT) are unlikely to be locally excited. Instead, they probably propagate to the observed region from a distant source, either from chorus waves outside the plasmasphere, as suggested by Bortnik et al. , or from hiss waves amplified due to a cyclotron resonant instability in the outer plasmasphere as discussed above. We discuss the origin of this low-frequency hiss in the outer plasmasphere in more detail in the next section.
4 Summary and Discussions
 We report simultaneous observations of whistler-mode waves by two Van Allen Probes. Van Allen Probe A, which was located in the postdawn sector (Figure 4a), observed very intense hiss waves (up to ~330 pT) with wave frequencies extending down to ~20 Hz (top panel in Figure 4b) in the outer plasmasphere, which are likely to be amplified due to the injection of energetic electrons into the plasmasphere. Van Allen Probe B, which was located in the predawn sector (Figure 4a), observed chorus waves (bottom panel in Figure 4b) outside the plasmasphere. We summarize the key process discussed in this study in the cartoon shown in Figures 4c and 4d. During quiet times, energetic plasma sheet electrons with energies of tens of keV to a couple of hundred keV drift from the nightside through dawn to the dayside but mostly remain outside of the plasmapause. At the onset of enhanced convection associated with electron injection, the plasmapause on the nightside moves to lower L-shells, while the plasmapause on the dayside remains in its original location. During this time interval, energetic electrons with energies above tens of keV injected from the nightside can gradient drift across the plasmapause at later MLT and penetrate into the outer portion of the dayside plasmasphere. This increase in electron flux in the outer plasmasphere is shown to be a plausible mechanism for hiss intensification with frequencies much lower than the typical plasmaspheric hiss frequency.
 The reported unusual low-frequency hiss (down to ~20 Hz) in the outer plasmapshere is unlikely to originate from chorus outside of the plasmapause, since a large ratio of the plasma to electron cyclotron frequency (fpe/fce > ~20 at L ~ 8) is required to excite chorus waves at such low frequencies (i.e., 0.01 fce at L ~ 8), which is rarely observed in the dayside magnetosphere. These low-frequency hiss waves are more likely to be amplified in the outer plasmasphere, where this large fpe/fce condition can be satisfied and where sufficient high-energy electrons can be injected [e.g., Solomon et al., 1988]. In order to amplify the low-frequency hiss in the outer plasmasphere, sufficient fluxes for anisotropic electron population at energy >~100 keV are needed inside the plasmasphere. More specifically, (1) substorm injection is sufficiently strong to produce an enhancement of >~100 keV electron fluxes at low L and (2) the shape of the plasmasphere needs to be highly asymmetric with a compressed plasmasphere on the nightside and an extended plasmasphere at larger MLTs in order for the injected high-energy electrons to encounter the plasmasphere as they drift around the Earth.
 Note that although the electron distributions measured by Probe A showed instability for generating these low-frequency hiss waves, the convective growth rates calculated based on the observed electron distributions seem to be insufficient (a few dB/RE) to directly lead to the observed wave intensity. We propose three possible explanations for linking the calculated small growth rates to the observed strong low-frequency hiss waves. First, the low-frequency hiss is not necessarily amplified exactly at the observed location, and wave intensification associated with the electron injection just inside the plasmasphere (but at a larger L-shell compared to the observed location) would be stronger due to larger electron fluxes. Second, the convective growth rates are calculated for a single pass, and the cumulative wave gains obtained by passing through this general region multiple times could lead to stronger wave power, as shown in Chen et al. [2012c]. Third, the measured electron pitch angle distributions were averaged over 11 s (1 spin period) with limited resolution in pitch angle and energy and may have already evolved to a quasi-steady state, thus, they may not accurately represent the initial electron distributions responsible for wave excitation. These possibilities, although beyond the scope of the present study, need further investigation and are left for a future study. Nevertheless, the observed low-frequency hiss is likely to be amplified in the outer plasmaphere associated with the injection of high-energy electrons into the plasmasphere and thus can provide an important source of hiss in addition to that which evolves from chorus waves.
 This work was supported by JHU/APL contracts 967399 and 921647 under NASA's prime contract NAS5-01072, NASA grants NNX11AD75G and NNX11AR64G, and NSF grant AGS-0840178. The analysis at UCLA was supported by the EMFISIS subaward 1001057397:01 and by the ECT subaward 13–041. We acknowledge Lunjin Chen and Yukitoshi Nishimura for helpful discussions. We thank the World Data Center for Geomagnetism, Kyoto for providing AU and AL indices used in this study.
 The Editor thanks two anonymous reviewers for their assistance in evaluating this paper.