Flaring is the common upstream oil and gas industry practice of disposing of unwanted combustible gas in turbulent diffusion flames open to atmosphere. Emissions from flaring remain uncertain and most measurement techniques rely on the assumption that combustion-derived species are well-mixed and therefore well-correlated in the plume. This thesis presents a spectroscopic measurement technique used to measure path-averaged species correlation in flare plumes to assess this assumption. Tunable diode laser absorption spectroscopy and line-of-sight attenuation techniques are used to measure H2O and soot, respectively. The techniques were validated using synthetic data from large eddy simulations of turbulent flare plumes. An experimental apparatus was subsequently developed and used to measure species correlation in lab-scale turbulent flare plumes. Results suggest that instantaneous path-averaged ratios of H2O and soot in the plume follow a skewed distribution, such that limited transects or short-duration sampling flare emission measurements would be subject to bias and uncertainty.