How well do state-of-the-art techniques measuring the vertical profile of tropospheric aerosol extinction compare?

[1] The recent Department of Energy Atmospheric Radiation Measurement (ARM) Aerosol Intensive Operations Period (AIOP, May 2003) yielded one of the best measurement sets obtained to date to assess our ability to measure the vertical profile of ambient aerosol extinction sigma(ep)(lambda) in the lower troposphere. During one month, a heavily instrumented aircraft with well-characterized aerosol sampling ability carrying well-proven and new aerosol instrumentation devoted most of the 60 available flight hours to flying vertical profiles over the heavily instrumented ARM Southern Great Plains (SGP) Climate Research Facility (CRF). This allowed us to compare vertical extinction profiles obtained from six different instruments: airborne Sun photometer (AATS-14), airborne nephelometer/absorption photometer, airborne cavity ring-down system, ground-based Raman lidar, and two ground-based elastic backscatter lidars. We find the in situ measured sigma(ep)(lambda) to be lower than the AATS-14 derived values. Bias differences are 0.002-0.004 Km(-1) equivalent to 13-17% in the visible, or 45% in the near-infrared. On the other hand, we find that with respect to AATS-14, the lidar sigma(ep)(lambda) are higher: Bias differences are 0.004 Km(-1) (13%) and 0.007 Km(-1) (24%) for the two elastic backscatter lidars (MPLNET and MPLARM, lambda = 523 nm) and 0.029 Km(-1) (54%) for the Raman lidar (lambda = 355 nm). An unnoticed loss of sensitivity of the Raman lidar had occurred leading up to AIOP, and we expect better agreement from the recently restored system. Looking at the collective results from six field campaigns conducted since 1996, airborne in situ measurements of sigma(ep)(lambda) tend to be biased slightly low (17% at visible wavelengths) when compared to airborne Sun photometer sigma(ep)(lambda). On the other hand, sigma(ep)(lambda) values derived from lidars tend to have no or positive biases. From the bias differences we conclude that the typical systematic error associated with measuring the tropospheric vertical profile of the ambient aerosol extinction with current state-of-the-art instrumentation is 15-20% at visible wavelengths and potentially larger in the UV and near-infrared.