The impact of aerosol fluorescence on the measurement of water vapor by UV (355nm emission) Raman lidar in the upper troposphere and lower stratosphere (UTLS) is investigated using the long-term records of three high-performance Raman lidars contributing to the Network for the Detection of Atmospheric Composition Change (NDACC). Comparisons with co-located radiosondes and aerosol backscatter profiles indicate that laser-induced aerosol fluorescence in smoke layers injected into the stratosphere by pyrocumulus events can introduce very large and chronic wet biases above 15km, thus impacting on the ability of these systems to accurately estimate long-term water vapor trends in the UTLS. In order to mitigate the fluorescence contamination, a correction method based on the addition of an aerosol fluorescence channel was developed and tested on the water vapor Raman lidar TMWAL located at the JPL Table Mountain Facility in California. The results of this experiment, conducted between 27 August and 4 November 2021 and involving 22 co-located lidar and radiosonde profiles, suggest that the proposed correction method is able to effectively reduce the fluorescence-induced wet bias. After correction, the average difference between the lidar and co-located radiosonde water vapor measurements was reduced to 5, consistent with the difference observed during periods of negligible aerosol fluorescence interference. The present results provide confidence that after a correction is applied, long-term water vapor trends can be reasonably well estimated in the upper troposphere, but they also call for further refinements or use of alternate Raman lidar approaches (e.g., 308nm or 532nm emission) to confidently detect long-term trends in the lower stratosphere. These findings may have important implications for NDACC's water vapor measurement strategy in the years to come.