The hydroxyl radical (OH) is the dominant methane (CH₄) sink and one of the largest uncertainties in the CH₄ budget[1]. Recent work suggests that satellite proxy data can provide new constraints on OH trends, variability, and sensitivity to changes in its drivers[2-6]. Such proxies are needed because direct OH observations are limited in time and space[7].

Traditional methods of determining OH trends and variability, such as chemistry transport models (CTMs)[8, 9] or methyl chloroform (MCF) inversions[10-13], also have strong limitations.

Differences in CTMs, including in their chemical mechanisms[14], lead to large disagreements in OH sensitivities, trends, and variability[8, 15, 16]. MCF inversions are underconstrained[13], leading to a wide disparity in results[10], and only yield information at hemispheric to global scales. Satellite-constrained OH products offer results at finer spatiotemporal resolution (Fig. 1), allowing for understanding of regional impacts on the atmospheric oxidative capacity.

We propose to use a combination of machine learning, chemistry model output, and satellite proxy data to constrain OH in the lower troposphere and in the tropospheric column over the global oceans, providing new insight into the CH₄ lifetime. This builds on previous work (Fig. 2), constraining tropical OH columns. Key goals include: 1) expanding the methodology outlined in Fig. 2 to the global oceans and to lower tropospheric OH, where CH₄ oxidation maximizes; 2) determining trends and variability of OH at multiple spatiotemporal scales; 3) evaluating the inversions, this methodology highlights the large spatial variability in the sensitivity of OH to changes in its chemical drivers in various photochemical environments and determining the impact on CH₄ lifetime.

We will rigorously evaluate the methodology and derived OH sensitivities using in situ observations and box modeling, where possible. Likewise, we will thoroughly characterize uncertainties resulting from the satellite retrievals (e.g. random and absolute errors, cloud biases) and their impacts on OH.

Results from this work could provide insight for policy makers as to where and what emissions reductions (e.g., NO_x) could have the largest impact on reducing CH₄. This will be achieved by providing constraints on:

• The relative impacts of internal climate modes and anthropogenic influences on OH variability.
• OH sensitivity to changes in its chemical drivers in different photochemical environments.
• The horizontal and vertical variability in OH and its subsequent impacts on CH₄ loss.

In addition, this work will highlight necessary improvements in the atmospheric observation system to better constrain CH₄ lifetime by identifying:

• Potential regions for intensive field campaigns for results validation.
• Needed improvements to satellites and retrievals to better constrain OH in the future.