Does anyone have any ideas how we could measure the full spectrum of total temperature in an air flow up to, say, 100kHz? We know how to get the deterministic or periodic variation due to blade passing and up to 50kHz, but the random turbulent heat transfer u'T' remains ellusive.