This is probably because of the use of a wastewater treatment database to model your water discharge impacts.
Wastewater treatment databases have an output of water in "Emissions to water" that have a negative characterization factor in the water consumption impact category. This is because this water is discharged (made available to the environment) thus increasing the amount of freshwater inside your system boundaries.
Play around with the database by changing the place to where the water is discharged (change from river to ocean, for example) to see how these values change.
and what could be the reason if -ve water depletion is associated with the A1 stage (Raw material stage), is it because of the elimination of water while Raw material processing?
When emissions exhibit a negative value in an impact category, it signifies that your model is beneficial rather than detrimental to the environment considering that particular impact category.
Nevertheless, the accuracy of this conclusion depends on several factors, such as the boundaries defined in the study and the correctness of the modeling process itself.