treat different channels not as independent sensors

not all info is reported on both channels (e.g. 'humidity')
we should calculate averages between channels and see if there is a disagement between them

Algo described here: https://cfpub.epa.gov/si/si_public_file_download.cfm?p_download_id=540979&Lab=CEMM