Graphs Contradict the Story
Engineer Chris Kapsambelis was scratching his head recently about the Plymouth Sound Compliance Monitoring Report. How, he wondered, could it be possible that the ambient lowest noise level for the cranberry bogs (L90–usually a quiet night time) was right up at the 50 dBA point? The quiet level shown here by a blue line is actually louder than some of the periods when the turbines were operating.
In the Tech Environmental Report, the Appendix Section, I came across this graph, and others like it, which I am having difficulty understanding. The note on the upper left corner states that the increase over ambient is only 0.6 dBA applying the DEP method. The blue line at the center, close to the 50 dBA level is described as the Ambient L90 background sound level with the turbines OFF.
Here is my problem. By casual inspection, the L50 metric, applied to the plot for turbine ON noise, would yield a level close to the L90 Ambient background level. Does that mean that the noise is softer when the turbines are running? Is the noise that falls below the L90 line on the graph inaudible? What kind of instrument is there that can record sound that one cannot hear?
We need to ask the DEP if their method for wind turbine compliance testing yields sensible results.
For more on L90 and other noise measurements, Kapsambelis explains in his essay Acoustics and Wind Turbine Noise.
Acoustician Stephen Ambrose issued an independent peer review citing numerous problems with the compliance report.
Wind Wise ~ Massachusetts 
The graph indicates that the L90 of the sound with the turbines running is lower than the L90 with the turbines OFF. Since this is physically impossible, the test procedure approved by MassDEP and followed by Tech Environmental, is producing obvious nonsensical results