UCLA

©2015. American Geophysical Union. All Rights Reserved. A novel technique capable of inferring wave amplitudes from low-altitude electron measurements from the Polar Operational Environmental Satellites (POES) spacecraft has been previously proposed to construct a global dynamic model of chorus and plasmaspheric hiss waves. In this paper we focus on plasmaspheric hiss, which is an incoherent broadband emission that plays a dominant role in the loss of energetic electrons from the inner magnetosphere. We analyze the sensitivity of the POES technique to different inputs used to infer the hiss wave amplitudes during three conjunction events with the Van Allen Probes. These amplitudes are calculated with different input models of the plasma density, wave frequency spectrum, and electron energy spectrum, and the results are compared to the wave observations from the twin Van Allen Probes. Only one parameter is varied at a time in order to isolate its effect on the output, while the two other inputs are set to the values observed by the Van Allen Probes. The results show that the predicted hiss amplitudes are most sensitive to the adopted frequency spectrum, followed by the plasma density, but they are not very sensitive to the electron energy spectrum. Moreover, the standard Gaussian representation of the wave frequency spectrum (centered at 550-Hz) peaks at frequencies that are much higher than those observed in individual cases as well as in statistical wave distributions, which produces large overestimates of the hiss wave amplitude. For this reason, a realistic statistical model of the wave frequency spectrum should be used in the POES technique to infer the plasmaspheric hiss wave intensity rather than a standard Gaussian distribution, since the former better reproduces the observed plasmaspheric hiss wave amplitudes. Key Points Inferred amplitudes are very sensitive to frequency spectrum and plasma density All estimates remain within a factor of 0.48 to 2.09 of the observed hiss waves Statistical frequency spectra should be used instead of standard Gaussian models