CAMS 81 – Global and Regional emissions · where T is the sea surface temperature (in C), taken from the ECMWF IFS ° meteorological data. The formula is valid between -2°C and

ECMWF COPERNICUS REPORT

Copernicus Atmosphere Monitoring Service

CAMS_81 – Global and Regional emissions

D81.3.10.1-M24: DMS emissions from oceans, including methodology

Issued by: MET Norway / Michael Gauss

Date: 08/09/2019

Ref: CAMS81_2017SC2_D81.3.10.1-M24-201908_v1.docx

CAMS81_2017SC2_D81.3.10.1-M24_201908_DMSemis_v1

This document has been produced in the context of the Copernicus Atmosphere Monitoring Service (CAMS). The activities leading to these results have been contracted by the European Centre for Medium-Range Weather Forecasts, operator of CAMS on behalf of the European Union (Delegation Agreement signed on 11/11/2014). All information in this document is provided "as is" and no guarantee or warranty is given that the information is fit for any particular purpose. The user thereof uses the information at its sole risk and liability. For the avoidance of all doubts, the European Commission and the European Centre for Medium-Range Weather Forecasts has no liability in respect of this document, which is merely representing the authors view.


CAMS81_2017SC2 - DMSemis Page 3 of 12

Contributors

MET NORWAY M. Gauss



Table of Contents

1. Summary 5

2. Introduction 6

3. Calculation 6

3.1 Input data 6 3.1.1 DMS concentrations 6 3.1.2 Meteorological data 6 3.2 Flux formulas and data processing 7 3.3 Comments on the resolution 8

4. Data files and results 8

4.1 List of files and their content 8 4.2 Results and example output 9 4.3 Differences between the previous data set and the new version 10

5. Conclusions and future work 11

6. References 11



1. Summary A new version of the data set containing global gridded emissions of dimethyl sulfide (DMS) from the world ocean has been generated. The data set replaces the version that was submitted as deliverable D81.3.10.1-M12, and this report replaces the D81.3.10.1-M12 deliverable report. The emission data are based on observed DMS concentration data published by Lana et al. (2011), flux formulas described in Nightingale et al. (2000), and meteorological data computed by the Norwegian Meteorological Institute using the ECMWF-IFS model version Cy40r1. The emission data are provided as daily means for the 19-year period from 2000 to 2018 (one file per year) on 0.5° x 0.5° resolution in a regular longitude-latitude grid. During the lifetime of CAMS_81, the emission data will be updated annually, as new input data become available. The DMS emission data is part of the CAMS-GLOB-OCE inventory and is available at the ECCAD data base along with documentation and references. Users of the data should, in addition to CAMS, make reference to the publications of Lana et al. (2011) and Nightingale et al. (2000). Important updates since the previous version of the data set (D81.3.10.1-M12) include the temperature dependence of the kinetic transfer velocity, the use of a fully consistent meteorological data set, the use of ECMWF IFS sea surface temperature, and the extension of the data set from 2015 to 2018 (i.e. from 16 to 19 years of data). The present document for D81.3.10.1-M24 can be seen as a stand-alone document replacing the report written for D81.3.10.1-M12.



2. Introduction As specified in the CAMS_81 work description, emissions of DMS have been calculated from observed DMS concentrations in ocean water and using meteorological parameters based on ECMWF data. The formula relating sea-air fluxes of DMS to observed concentrations and meteorological data is based on a publication by Nightingale et al. (2000) and has been used in the Norwegian Earth System model for several years. In the frame of CAMS_81, the formula has been applied by the Norwegian Meteorological Institute to calculate DMS fluxes over the 2000-2018 period in a consistent manner, i.e. by using the same flux formula, the same ECMWF-IFS model version, and the same DMS data source for the entire period. The calculation will be described in the next section, while example results are shown in Section 4. Plans for future work are outlined in Section 5.

3. Calculation The formula adopted by CAMS_81 to estimate DMS fluxes from the ocean to the atmosphere requires DMS concentrations in the ocean as well as meteorological input. This section describes where we obtained these input data from and presents the formulas we have used.

Input data

3.1.1 DMS concentrations For the calculation of DMS fluxes we need DMS concentrations in the ocean water. These concentrations were provided on 1°x1° spatial resolution by Lana et al., 2011 and described there in detail. The data are derived from a large number of measurements performed during the 1980-2009 period and accessible from a web page of the Surface Ocean Lower Atmosphere Study (SOLAS), see https://www.bodc.ac.uk/solas_integration/implementation_products/group1/dms/. One of the zip files to be downloaded there is ‘dmsclimatology.zip’ and contains DMS concentration files on csv format, one file per month, in units of nmol(DMS)/L. The data are given as monthly means averaged over the period of measurements, but there is no inter-annual variation in the file. Hence, the concentrations are to be considered as average values representative of the 1989-2009 period. For more details about the DMS concentration data see Lana et al. (2011) and the SOLAS web page mentioned above.

3.1.2 Meteorological data The method to calculate DMS emissions also requires the u and v components of 10-meter wind speed, as well as the Sea Surface temperature as input. Being the national weather service of Norway, the Norwegian Meteorological Institute has experience in running the ECMWF-IFS model to generate consistent data sets over multi-year time periods. Currently, global meteorological data are available to this project as netCDF files on 0.5°x0.5° spatial and 3-hourly temporal resolutions with global coverage for every year within the 1990-2018 period. Data for all these years have been generated

http://www.solas-int.org/

https://www.bodc.ac.uk/solas_integration/implementation_products/group1/dms/

https://www.ecmwf.int/en/forecasts/documentation-and-support/changes-ecmwf-model/ifs-documentation



with IFS version Cy40r1. The multi-year meteorological data set can thus be considered as being consistent in the sense that it has been created by using essentially the same method/software for every year within the 2000-2018 period, for which DMS emissions have been calculated.

Flux formulas and data processing The formulas to calculate ocean-atmosphere fluxes from the input data described in Section 3.1 are based on equations by Nightingale et al. (2000). Within the frame of CAMS_81 these formulas have been implemented by MET Norway in a FORTRAN routine, which reads all required input data, applies the formulas and writes the DMS emission data to netCDF files, one per year. Fluxes are generally calculated as F = Kw * delta_C, where Kw is the gas transfer velocity in water and delta_C is the sea-air concentration gradient. For DMS the concentration gradient can be approximated by its concentration in water (CwDMS), because the air concentration (CaDMS) is negligible compared to the water concentration, thus: F= kDMS * delta_C = kDMS * (CADMS – CWDMS) = kDMS (CWDMS/H - CWDMS) ~ - kDMS CWDMS In atmospheric science we define fluxes out of the water (and into the atmosphere) as positive, so the minus sign in the above formula is omitted. CWDMS is taken from the climatology, while for kDMS we proceed as follows : The 10-meter wind speed U is calculated from its u and v components as U = sqrt (u2+v2) unit [m/s] and the gas exchange coefficient k600 for CO2 in freshwater at 20°C, also called piston velocity, is calculated as k600 = 0.222 * U^2 + 0.333 * U unit [cm/hr] The temperature-dependent Schmidt number SC for DMS is calculated with a 4-degree formula from Wanninkhof (2014) : SCDMS = 2855.7-177.63*T+6.0438*T^2-0.11645*T^3+0.00094743*T^4 where T is the sea surface temperature (in °C), taken from the ECMWF IFS meteorological data. The formula is valid between -2°C and +40°C. In rare cases, the sea surface temperature is lower. In these cases the value for -2°C is taken. The gas exchange coefficient for DMS is calculated as kDMS = k600 * sqrt(k600/SCDMS) unit [cm/hr]



and finally the flux of DMS is: FDMS = kDMS*2.778e-15* MDMS*CWDMS unit [kg(DMS)/m2/s] Where MDMS is the molecular weight of DMS (62.13 g/mol), CWDMS is the concentration of DMS in sea water (in nmol/L). The unit conversion factor 2.778e-15 is needed to get FDMS in the desired unit of kg(DMS)/m2/s. Explanation of unit conversion factor: FDMS kg(DMS)/m2/s = kDMS cm/hr * 2.778e-15 * 62.13 g/mol *CDMS nmol/1000cm3 FDMS kg(DMS)/m2/s = [k600/100/3600] m/s * [62.13/1000] kg/mol *[CDMS*1000/1e9] mol/m3 Unit conversion factor: 1/100/3600 * 1/1000 * 1000/1e9 = 2.778e-15 Before writing out to the data files, the fluxes are averaged over one day. The sums are checked by multiplying fluxes with the areas of each grid cell, integrating over the globe, and comparing the results with numbers in Lana et al., 2011. Examples are given in Section 4.

Comments on the resolution Given the fact that the inputs used for creating the emission data have different resolutions (see previous subsections) it is not obvious which temporal and spatial resolutions should be used in the resulting emission data. Furthermore, the level of uncertainty is very different between, say, 10-meter wind speed (relatively certain, with high temporal resolution) and DMS concentrations in the ocean water (relatively uncertain, with low temporal resolution). We provide daily means on a resolution of 0.5° x 0.5° resolution. Obviously, this relatively high resolution does not reflect the (limited) accuracy of some of the input data, but it does correspond to the spatial resolution of the meteorological data and thus retains some of the high variability in wind speed, to which the gas exchange coefficient of DMS respond quite strongly. Nevertheless, depending on the user needs, and considering the uncertainties in the input data, a coarser resolution can be provided in the next version of the data, or in the frame of deliverable D81.3.10.1-M36 next year. This could be desirable for example in order to reduce file size.

4. Data files and results

List of files and their content The files delivered to CNRS are named ‘DMS_emissions_YYYY.nc’, where YYYY is the year number (‘2000’, ‘2001’, etc.). At CNRS they are included in the CAMS-GLOB-OCE archive.



Each of the files contains longitude values, latitude values, and the emissions of DMS from ocean water in units of kg(DMS)/m2/s. In grid cells with no available data or over land, the emissions are set to zero. In grid cells containing both water and land surface, any non-zero flux value should only be applied to the fraction of the grid cell that is covered by ocean.

Results and example output Averaged over the 2000-2018 period, the annual global emissions of DMS calculated with the method and input data described in Section 3 amounts to 26.5 Tg(S)/yr, which is near the lower end but still within the uncertainty range of 24.1 to 40.4 Tg(S)/yr estimated by Lana et al. (2011). Figure 1 shows the seasonal variation in emissions for DMS in the two hemispheres. The emissions show a clear seasonal variation, with largest emissions during the summer season. Emissions are larger in the Southern Hemisphere than in the Northern Hemisphere. Figure 2 shows the inter-annual variation in total emissions in the Northern and Southern Hemispheres. It has to be kept in mind, however, that the ocean concentration data have no inter-annual variation, and they cover only the period up to 2009. The variability seen after 2009 in the plot is due to meteorological variability. Nevertheless it is interesting to note that DMS emissions tend to increase during the plotted period.

Figure 1: Monthly emissions of DMS, averaged over the 2000-2018 period, integrated over the Northern (blue) and Southern (red) Hemispheres. Units: Tg(S)/month.



Figure 2: Annual emissions of DMS in the Southern (SH) and Northern (NH) Hemispheres and the sum (Global) for the 2000-2018 period. Unit: Tg(S)/year.

Differences between the previous data set and the new version The differences between last year’s (D81.3.10.1-M12) and this year’s (D81.3.10.1-M12) data sets are, apart from the extension to 2018):

• Update in the use of temperature dependence of the kinetic transfer velocity: this was the main reason for the decrease in emissions with respect to last year’s data set.

• use of a fully consistent meteorological data set, based solely on ECMWF-IFS Cy40r1: this concerned only the years 2012 and 2013 as all the other years were already based on ECMWF-IFS Cy40r1 in last year’s data set. The differences between Cy38r2 and Cy40r1 (mainly related to cloud microphysics in the free troposphere) are much smaller than the uncertainty in other input parameters or in the flux calculation itself. The effect of this change on total DMS emissions was well below 1%;

• the use of ECMWF IFS sea surface temperature instead of WOA data: this was done to retain the information on spatial and temporal variability in the ECMWF IFS data set. The effect on total DMS emissions was well below 1%;



• use of a 4th degree formula (Wanninkhof, 2014) instead of a 3rd degree formula (Saltzmann et al., 1993) to calculate the temperature dependence of the gas exchange coefficient. This had also a relatively small effect on total DMS emissions (below 1%).

5. Conclusions and future work Data files containing global gridded emissions of DMS from the world ocean have been generated for the 2000-2018 period and are provided as part of CAMS_81 deliverable D81.3.10.1-M24. The files should be available on the ECCAD server from September 2019. In the event that errors are found in the data, corrected files will be provided as versions v2.2, v2.3, etc. during the period September 2019 to August 2020. Users of the data should, in addition to CAMS, make reference to the publications of Lana et al. (2011) and Nightingale et al. (2000). For deliverable D81.3.10.1-M36 (due in August 2020), possible updates in the available input data (concentration climatology) or in the flux formula will be taken into account. In any case, the data set will be extended by one year (2019) for D81.3.10.1-M36. The versions and updates provided for deliverable D81.3.10.1-M24 will be labeled v3.1, v3.2, etc. As a final remark, users of these DMS emission data may, as an alternative, consider implementing the method described in Section 3 directly in their own model, in order to calculate DMS emissions dynamically with their own meteorology/climate data and their model’s resolution and atmospheric DMS concentrations.

6. References Lana, A., T. G. Bell, R. Simo, S. M. Vallina, J. Ballabrera-Poy, A. J. Kettle, J. Dachs, L. Bopp, E. S. Saltzman, J. Stefels, J. E. Johnson, and P. S. Liss, 2011: An updated climatology of surface dimethlysulfide concentrations and emission fluxes in the global ocean. Global biogeochemical cycles, 25, https://doi.org/10.1029/2010GB003850 Nightingale, P., G. Malin, C. Law, A. Watson, P. Liss, M. Liddicoat, J. Boutin, and R. Upstill-Goddard, 2000: In situ evaluation of air-sea gas exchange parameterizations using novel conservative and volatile tracers. Global biogeochemical cycles, 14, https://doi.org/10.1029/1999GB900091 Saltzman, E.S., et al., 1993: Experimental determination of the diffusion coefficient of dimethylsulfide in water. J Geophys Res 98(C9), https://doi.org/10.1029/93JC01858 Wanninkhof, R., 2014: Relationship between wind speed and gas exchange over the ocean revisited. Limnol. Oceanogr. Methods, 12: 351–362, https://doi.org/10.4319/lom.2014.12.351

https://doi.org/10.1029/1999GB900091


atmosphere.copernicus.eu copernicus.eu ecmwf.int

ECMWF - Shinfield Park, Reading RG2 9AX, UK Contact: [email protected]

Documents

CAMS 81 – Global and Regional emissions · where T is the sea surface temperature (in C), taken from the ECMWF IFS ° meteorological data. The formula is valid between -2°C and