Context :

Aims : We present Herschel/HIFI observations of 14 water lines in a small sample of galactic massive protostellar objects : NGC6334I ( N ) , DR21 ( OH ) , IRAS16272-4837 , and IRAS05358+3543 .
Using water as a tracer of the structure and kinematics , we aim to individually study each of these objects , to estimate the amount of water around them , but to also shed light on the high-mass star formation process .

Methods : We analyze the gas dynamics from the line profiles using Herschel-HIFI observations acquired as part of the WISH key-project of 14 far-IR water lines ( H _ { 2 } ^ { 16 } O , H _ { 2 } ^ { 17 } O , H _ { 2 } ^ { 18 } O ) , and several other species .
Then through modeling of the observations using the RATRAN radiative transfer code , we estimate outflow , infall , turbulent velocities , molecular abundances , and investigate any correlation with the evolutionary status of each source .

Results : The four sources ( plus previously studied W43-MM1 ) have been ordered in terms of evolution based on their SED : NGC64334I ( N ) \rightarrow W43-MM1 \rightarrow DR21 ( OH ) \rightarrow IRAS16272-4837 \rightarrow IRAS05358+3543 .
The molecular line profiles exhibit a broad component coming from the shocks along the cavity walls associated with the protostars , and an infalling ( or expansion for IRAS05358+3543 ) and passively heated envelope component , with highly supersonic turbulence likely increasing with the distance from the center .
Accretion rates between 6.3 \times 10 ^ { -5 } and 5.6 \times 10 ^ { -4 } M _ { \odot } yr ^ { -1 } are derived from the infall observed in three of our sources .
The outer water abundance is estimated to be at the typical value of a few 10 ^ { -8 } while the inner abundance varies from 1.7 \times 10 ^ { -6 } to 1.4 \times 10 ^ { -4 } with respect to H _ { 2 } depending on the source .

Conclusions : We confirm that regions of massive star formation are highly turbulent and that the turbulence likely increases in the envelope with the distance to the star .
The inner abundances are lower than the expected 10 ^ { -4 } perhaps because our observed lines do not probe deep enough into the inner envelope , or because photodissociation through protostellar UV photons is more efficient than expected .
We show that the higher the infall/expansion velocity in the protostellar envelope , the higher is the inner abundance , maybe indicating that larger infall/expansion velocities generate shocks that will sputter water from the ice mantles of dust grains in the inner region .
High-velocity water must be formed in the gas-phase from shocked material .