That really shouldn't be the case - the specs with which cellular
basestations have to comply are fairly strict, and they do test them
for compliance, and they usually overdesign them anyway. For GSM and
UMTS you can actually download the specs for the basestations and the
handsets from the 3gpp website (free of charge).
Here is a list of the specifications:
http://www.3gpp.org/ftp/Specs/html-info/SpecReleaseMatrix.htm
And here is the RF specification for 3G basestations:
http://www.3gpp.org/ftp/Specs/latest/Rel-6/25_series/25104-6g0.zip
It shows they are allowed to emit -36dBm in any given 100kHz channel
over the frequency range 30MHz to 1GHz, but I would be surprised if
they came anywhere close to breaking that spec.
Have you verified with a spectrum analyser or a careful experiment
with your receiver and some fixed attenuators to make sure that it is
not intermodulation or cross modulation distortion in your LNA? If
the interference drops by more dBs than the wanted signal drops when
you insert say a 6dB attenuator inline between your antenna and your
Rx LNA input, then that points to problems with the receiver linearity
that could be solved with a band pass filter for 2m. If everything
drops by the same number of dB - interference and wanted signal alike
- then it points to genuine spurious emissions.
Chris