FORMALDEHYDE METHODS COMPARISON IN THE REMOTE LOWER TROPOSPHERE DURING THE MAUNA-LOA PHOTOCHEMISTRY EXPERIMENT 2

Five methods for the measurement of CH2O vapor were compared under rem ote tropospheric conditions. The techniques included an aqueous-scrubb er enzyme fluorescence method (URIcoil), TDLAS (UNI), 2,4-dinitropheny lhydrazine (DNPH) impregnated-cartridges (BNL/WSUcart), an aqueous-scr ubber DNPH mettled (BNL/WSUcoil), and an unpublished aqueous-scrubber immobilized-enzyme fluorescence method (NCARbound). This was part of t he Mauna Loa Observatory Photochemistry Experiment 2 (MLOPEX 2) which was performed in four similar to 30-day intensives. In MLOPEX 2a (fall 1991), the URIcoil technique indicated higher concentrations relative to the BNL/WSUcoil which were higher than UNI. The limited number of NCARbound measurements for MLOPEX 2a were greater than the other measu rements and least reliable. During MLOPEX 2b (winter 1992), URIcoil, U NI, BNL/WSUcoil, and NCAR-bound gave comparable measurements of CH2O w ith the latter again being consistently higher than the first three me thods and its performance and reliability was improved over MLOPEX 2a. URIcoil and BNL/WSUcoil were comparable in MLOPEX 2c (spring 1992) an d 2d (summer 1992), whereas, BNL/WSUcart was consistently higher than both for these intensives. The major source of difference between tech niques was ascribed to variations in the analytical-procedural field b lanks. Calibration differences were the most likely cause of URIcoil b eing higher than BNL/WSUcoil and BNL/WSUcoil being higher than UNI dur ing MLOPEX 2a. Lower free troposphere median concentrations were betwe en 100 and 150 ppt for all seasons and without a seasonal trend. These values were a factor of 2 to 3 lower than model estimates for this st udy site and were in keeping with earlier work.