Cosmic radiation on Earth responds to heliospheric, geomagnetic,
atmospheric, and lithospheric changes. In order to use its signal for
soil hydrological monitoring, the signal of thermal and epithermal
neutron detectors needs to be corrected for external influencing
factors. However, theories about the neutron response to soil water, air
pressure, air humidity, and incoming cosmic radiation are still under
debate. To challenge these theories, we isolated the neutron response
from almost any terrestrial changes by operating bare and moderated
neutron detectors in a buoy on a lake in Germany from July 15 to
December 02, 2014. We found that the count rate over water has been
better predicted by a recent theory compared to the traditional
approach. We further found strong linear correlation parameters to air
pressure and air humidity for epithermal neutrons, while thermal
neutrons responded differently. Correction for incoming radiation proved
to be necessary for both thermal and epithermal neutrons, for which we
tested different neutron monitors and correction methods. Here, the
conventional approach worked best with the Jungfraujoch monitor in
Switzerland, while the approach from a recent study was able to
adequately rescale data from more remote neutron monitors. However, no
approach was able to sufficiently remove the signal from a major Forbush
decrease event, to which thermal and epithermal neutrons showed a
comparatively strong response. The buoy detector experiment provided a
unique dataset for empirical testing of traditional and new theories on
CRNS. It could serve as a local alternative to reference data from
remote neutron monitors.