2002ApJ...566..289B

The multivariate tool of principal component analysis (PCA) is applied to 23 fields in the FCRAO CO Survey of the Outer Galaxy. PCA enables the identification of line profile differences, which are assumed to be generated from fluctuations within a turbulent velocity field. The variation of these velocity differences with spatial scale within a molecular region is described by a singular power law, δv=cL^α, which can be used as a powerful diagnostic to turbulent motions. For the ensemble of 23 fields, we find a mean value α=0.62±0.11. From a recent calibration of this method using fractal Brownian motion simulations (Brunt & Heyer), the measured velocity difference-size relationship corresponds to an energy spectrum, E(k), which varies as k^–β, where β=2.17±0.31. We compare our results to both decaying and forced hydrodynamic simulations of turbulence. We conclude that energy must be continually injected into the regions to replenish that lost by dissipative processes such as shocks. The absence of large, widely distributed shocks within the targeted fields suggests that the energy is injected at spatial scales less than several parsecs.