Potential Constraints on “Background” Oxidation in Chemistry–Climate Models Using AToM-1 and -2 Observations of Formaldehyde Variability

The hydroxyl radical (OH) is a primary sink for many pollutants and reactive greenhouse gases, but chemistry-climate model simulations vary widely in their projection of abundances and trends of OH. We hypothesize that variations in formaldehyde (HCHO) reflect variations in OH, and in particular the average of OH as expressed in the calculated methane (CH₄) loss rate, in remote atmospheric “background” regions where its near-exclusive source is as an intermediate of CH₄ oxidation. To explore this idea, we first define these background regions using two different model-based approaches. The first focuses on regions of the atmosphere where HCHO is linearly correlated with the CH₄ loss rate. In this method, we use a 400-year simulation with emissions from the year 2000 repeated perpetually in the GFDL CM3 chemistry-climate model. The second defines background by calculating the ratio of NOx and CO averaged over 2004-2012 in two months characteristic of the ATom-1 and 2 deployments between a GEOS-Chem simulation with anthropogenic emissions turned off and an identical simulation with emissions from all sectors. Background by this definition is thus where natural emissions contribute most to the total concentration, or where this ratio is highest (specifically, > 0.5 for both NOx and CO). In addition, we use ATom measured acetonitrile concentrations to identify plumes with fresh fire influence. We investigate observed relationships along ATom-1 and 2 flights between HCHO and methane loss rates in background regions versus those strongly influenced by fire plumes. Within background regions, a moderate correlation (r² ≈ 0.6) is found between HCHO and the methane loss rate in the low- to mid-troposphere, where the majority of CH₄ oxidation occurs. This relationship is robust across both of our definitions of background and largely reflects OH variability. However, the HCHO-methane loss relationship is modulated by NOx:NOy partitioning, which reduces the correlation strength at high ratios. Additionally, we explore the HCHO-methane loss relationship in the context of the primary drivers of OH variability (e.g. J_O3→O1D, H₂O and CO:NOx) and contrast our findings between background regions and air masses influenced by biomass burning. Last, we investigate whether a chemistry climate model frequently applied to project future changes in atmospheric composition and climate (GFDL AM3 nudged to reanalysis winds) captures these relationships. Through this study we aim to better understand variability in OH, a major sink of many pollutants that affect air quality and climate change.