Water is a dominant gas component in the envelope of oxygen-rich
evolved stars. It is an important coolant of the  circumstellar gas
and can be photodissociated by interstellar UV photons in the
periphery of the circumstellar envelope to produce OH. Therefore, H2O
plays a major role in the thermal balance and in the photo-chemistry
inside oxygen-rich circumstellar envelopes. However, the abundance of
H2O is not well known. This is because the detection of H2O lines
from the ground is hindered by atmospheric water. So far only a few
non-thermal radio lines have been detected using ground-based radio
telescopes. The determination of the abundance of H2O based on its
maser transitions is rather delicate, because the maser intensity is
highly variable. Our model calculations have shown that H2O infrared
lines are thermal. Our goal is to observe thermal emission lines of
H2O to determine the abundance of this molecule, and to investigate
the physico-chemistry in oxygen-rich envelopes, using a model treating
radiative transfer coupled with thermal balance and photodissociation.
The observed intensities depend mainly on the distance and the mass
loss rate of the star and on the abundance of water. We select a
sample of nearby (within 350 pc) visible oxygen-rich AGB stars to
search for 3 infrared H2O transitions with the LWS Fabry-Perot. A
bright M supergiant, S Per (2300 pc), and a nearby S type star, Chi
Cyg, are also included in this sample to investigate the effects of
luminosity and chemical composition on the H2O abundance. The
sources exhibit at least one of the maser emission lines of H2O, OH
and SiO and their mass loss rates are available. The three selected
water lines at 179, 108 and 75 microns are expected to be strong and
represent the minimum number to provide adequate constraints on the
intermediate shell layers (100-300 K). Similarly, we estimate 8
targets to be the minimum number needed to adequately sample the
range of stellar properties.