readme_wolfe2019_manuscript_data.txt This document describes the data used in "Mapping Hydroxyl Variability Throughout the Global Remote Troposphere via Synthesis of Airborne and Satellite Formaldehyde Observations," Wolfe et al., PNAS (2019). Users should read said paper before manipulating this dataset. FILES DESCRIBED atom_integrated_columns.csv omi_hcho_and_xoh.csv REVISION NOTES R1 20190123 Modification following DAAC review, including: - Added underscores to some variable names - Moved order-of-magnitude scaling from readme to data files - Better control of decimal places - Fixed x2 error in scale_diel_oh_atom[N] and scaling error in hcho_omi_atom[N] R0 20181225 Initial Submission -------------------------------------------------------------------- CONTENTS 1. PI CONTACT INFO 2. FAIR USE 3. atom_itegrated_columns.csv 4. omi_gridded_hcho_and_xoh.csv --------------------------------------------------------------------- 1. PI CONTACT INFO Glenn Wolfe NASA Goddard Space Flight Center 301-614-6008 glenn.m.wolfe@nasa.gov -------------------------------------------------------------------- 2. FAIR USE These data are made freely available to the public and the scientific community in the belief that their wide dissemination will lead to greater understanding and new scientific insights. The availability of these data does not constitute publicationof the data. We rely on the ethics and integrity of the user to assure that we receive fair credit for this work. If the data are obtained for potential use in a publication or presentation, the PI should be informed at the outset of the nature of this work. If the data are essential to the work, or if an important result or conclusion depends on the data, co-authorship may be appropriate. This should be discussed at an early stage in the work. Manuscripts using the data should be sent to the PI for review before they are submitted for publication so we can insure that the quality and limitations of the data are accurately represented. -------------------------------------------------------------------- 3. atom_integrated_columns.csv COLUMN INFORMATION Note: missing values indicated by -999 for all columns. Variable Name Units Description profile_number NA Unique identifier for each profile, consisting of 5 digits. First digit is mission number. Second/third digits are flight number (per mission). Last two digits are profile number (per flight). date yyyymmdd UTC date at midpoint of profile. UTC_sec_of_day s UTC seconds of day at midpoint of profile. mission_number NA ATom mission number. flight_number NA ATom flight number. Starts at 1 for each mission. lat_start deg Latitude at beginning of profile. lat_mid deg Latitude at middle of profile. lat_stop deg Latitude at end of profile. lon_start deg Longitude at beginning of profile. lon_mid deg Longitude at middle of profile. lon_stop deg Longitude at end of profile. hcho cm^-2 Profile-integrated formaldehyde concentration. oh cm^-2 Profile-integrated hydroxyl concentration. p_oh cm^-2 Profile-integrated hydroxyl production rate. k_hcho s^-1 Profile-integrated formaldehyde loss frequency (photolysis + OH loss). j_hcho s^-1 Profile-integrated formaldehyde photolysis frequency. hcho_err cm^-2 1-sigma uncertainty in hcho. oh_err m^-2 1-sigma uncertainty in oh. p_oh_err cm^-2 1-sigma uncertainty in p_oh. k_hcho_err s^-1 1-sigma uncertainty in k_hcho. j_hcho_err s^-1 1-sigma uncertainty in j_hcho. omi_hcho cm^-2 OMI formaldehyde columns averaged over ATom profiles. omi_hcho_err cm^-2 1-sigma uncertainty in omi_hcho. DESCRIPTION This file contains ATom-integrated vertical profiles used in Figs. 1-3 of Wolfe et al. (2019). The following description is mostly from the Methods section of that paper. ATom 1-second merge files and ATHOS files were pulled from the ATom DAAC on 20181203 (https://doi.org/10.3334/ORNLDAAC/1581). 1-second observations were averaged to the native OH time base (30 seconds for ATom 1, 20 seconds for ATom 2) and converted from mixing ratios to number density. Rate coefficients are taken from 2015 JPL Handbook recommendations. All 1-sigma (68% confidence level) measurement and rate coefficient uncertainties are propagated through to column-integrated values. Measurement uncertainties are treated as random unless specifically stated otherwise in native data files. Rate coefficient uncertainties are treated as systematic. ATom profiles were executed “en route” along the flight track (as opposed to spirals) and typically spanned 200 – 450 km of horizontal distance and variable altitude ranges (typical minima and maxima of < 1 km and > 6 km). For each profile, concentrations and reaction rates are averaged to an altitude grid of 0 - 10 km with 200 m spacing using GPS altitude above mean sea level as the vertical coordinate. Missing points in each gridded profile are filled with linear interpolation (for points within the min and max sampling altitude) and constant-value extrapolation (using the mean of the last two adjacent valid points) to the surface and/or 10 km. Following trapezoidal integration, columns are filtered for solar zenith angle < 80°, profile sampling altitude range at least 0.6 - 8 km, fraction of missing HCHO or OH observations < 30%, fraction of interpolated gridded OH < 40%, and fraction of extrapolated HCHO < 25%. These limits are determined by visual identification of outliers in Fig. 2 a-b and inspection of vertical profiles. Three additional profiles are removed due to intermittent cloud cover, which tends to decouple OH and HCHO. Two profiles over the central U.S. are also removed. Total column OH production is calculated by summing column-integrated rates for O3 photolysis, HO2 + NO, HO2 + O3, and H2O2 photolysis. All rates are calculated directly from observations, and O(1D) is assumed to be in steady state for deriving OH production from O3 photolysis. A small linear correction(slope = 1.06, intercept = 8.6E-10 cm^-2 s^-1) is also applied to account for OH sources beyond the listed reactions. The slope for this correction is based on simulated rates from a 0-D box model (see manuscript), while the offset is chosen to force the intercept of a plot of p_oh vs oh to pass through zero. We add an additional systematic error of 6% to p_oh_err for this correction. k_hcho and j_hcho are calculated by integrated HCHO loss rates and dividing by profile-integrated HCHO. OMI Observations: global HCHO columns are taken from NASA’s operational OMI product provided by the Smithsonian Astrophysical Observatory (OMI-SAO v003), available at https://mirador.gsfc.nasa.gov/. Level 2 daily vertical column densities are screened for cloud fraction < 0.3, solar zenith angle < 70 deg, and a main data quality flag of 0 before averaging to a spatial resolution of 0.5 deg x 0.5 deg. Fitting uncertainties for each grid are calculated using the individual uncertainties for each pixel using a weighted average (see manuscript for formula). Daily gridded columns are averaged over each campaign (29 July – 23 August 2016 for ATom 1, 26 January – 21 February 2017 for ATom 2) using a fitting error-weighted mean. Uncertainty is estimated as follows: 1) for each grid cell, calculate the fitting error-weighted standard deviation of the mean when averaging over each campaign; 2) find the median of all such values over the remote Pacific reference sector region (140 to 160 deg W, all latitudes). This estimate (1E15 cm^-2) is a reasonable upper limit as it may include some influence from true atmospheric variability. For each ATom profile, OMI retrievals for the relevant mission are sampled and averaged over all grid cells intersected by the flight path. OMI columns are scaled using ATom-derived air mass factors and corrected for the fraction of total-column HCHO residing below 10 km as determined by retrieval a priori profiles (see manuscript for further information on air mass factors). ------------------------------------------------------------------- 4. omi_gridded_hcho_and_xoh.csv COLUMN INFORMATION Note: missing values indicated by -999 for all columns. Variable Name Units Description lat deg grid center latitude. lon deg grid center longitude. land boolean grid land mask. 1 = land, 0 = water. hcho_omi_atom1 cm^-2 OMI HCHO vertical column density, gridded/averaged over ATom 1. j_hcho_gmi_atom1 s^-1 GMI column-average formaldehyde photolysis frequency for ATom 1. k_hcho_calc_atom1 s^-1 Column-average formaldehyde total loss frequency for ATom 1. xoh_atom1 cm^-3 24-hour tropospheric column mean OH concentration for ATom 1. xoh_err_rand_atom1 cm^-3 1-sigma random uncertainty in xoh_atom1. This error is solely from OMI HCHO uncertainty. xoh_err_syst_atom1 cm^-3 1-sigma systematic uncertainty in xoh_atom1. Includes contributions from fitting coefficients and j_hcho. tp_height_atom1 km GMI output average tropopause height for ATom 1. scale_diel_oh_atom1 NA Scaling factor to convert 1300 local time to 24-hour average OH. Calculated using GMI OH. air_mass_atom1 kg GMI output column air mass for ATom 1. Identical descriptions apply to variables with suffix _atom2. But they are for ATom 2. Obviously. DESCRIPTION This file contained gridded data and model output used in Figs. 1, 4, and 5 of Wolfe et al. (2019). Spatial resolution is 0.5 deg latitude x 0.625 deg longitude grid. Values are temporal averages over ATom 1 (29 July – 23 August 2016) or ATom 2 (January – 21 February 2017). OMI HCHO is derived from L2 retrievals as described in Sect. 3 of this document. Averaged column densities are linearly interpolated from 0.5 x 0.5 to 0.5 x 0.625 resolution. 1-sigma uncertainty is estimated as 1E15 cm^-2 for all grid cells and is presumed random. Model products, including jhcho_gmi, tpheight, scale_diel_oh, and air_mass, are taken from a Global Modeling Initiative (GMI) simulation. The simulation is described in supplementary information of the manuscript. The variables k_hcho_calc are derived from j_hcho_gmi using ATom scaling factors as described in the manuscript. Both are provided here for convenience. The variables xoh are calculated from ATom scaling factors, OMI HCHO, GMI jhcho, and GMI scaling factors using a linear transformation as described in the manuscript. xoh and related quantities are only valid for non-land grid cells. The variables tp_height and scale_diel_oh are used to convert OH vertical column density at the OMI overpass time (1330LT, but model output is hourly so we round to 1300LT) to 24-hour tropospheric mean concentration (xoh) and are provided here for convenience. xoh_err includes propagated uncertainties from HCHO columns, jhcho_gmi, and ATom scaling factors. Averaging of xoh should include weighting by tropospheric mass, e.g. xoh_avg = sum(xoh * air_mass) / sum(air_mass). Random and systematic uncertainties in xoh must be treated separately when averaging. (N = total number of valid points in average) random error = sqrt(sum(xoh_err_rand^2)) / N systematic error = sqrt(sum(xoh_err_syst^2)) / sqrt(N) total error = sqrt(random^2 + systematic^2) In some instances it may also be appropriate to include air mass weighting in the above formulae. ------------------------------------------------------------------- END OF DOCUMENT