What We Do in the Methane Show Case

Methane transport from the North Sea; comparing gas budgets between top-down and bottom-up approaches

Methane is a potent green house gas and increasing globally in the atmosphere. In the North Sea region methane is being released from natural sources like seeps which are fueled from shallow biogenic gas pockets or deep thermogenic gas reservoirs (Schneider v. Deimling et al., 2011; Mau et al., 2015; Römer et al., 2017), shallow water and intertidal anoxic sediment which temporarily get in direct contact with the atmosphere (Wadden Sea; Grunwald et al., 2009), river runoff from e.g. the Elbe river (Matoušů et al., 2019), but also from about 13,000 abandoned wells (Rehder et al., 1998; Vielstädte et al., 2017).

The North Sea is thus a representation of complex and various sources and sinks with internal (supply of gas through faults, biogenic methane generation in sediments) and external (tidal pumping, storm activities, river runoff) forcing. We choose the Methane in the North Sea Region as Show Case as it is a good example for 1) a typical comparison of bottom-up and top-down flux estimates, 2) it demands the compilation of data from various sources, 3) it involves the extrapolation and prediction of data in space and time, 4) and evaluated model/field data and their fit through visual and analytical data exploration. All these are typical generic tasks of data driven science in the marine and atmosphere Earth Sciences which we couple in this Show Case. Following the philosophy of Digital Earth of ‘thinking in workflows’ we divided the overall Show Case in a number of work flows, starting with the compilation of data from various sources (1), perform re-gridding tasks to make spatial data existing as grids comparable in space (2), extrapolate marine free gas release (bubbles) and the related methane flux from abandoned wells into the atmosphere (3), detect patterns in atmospheric data from EDGAR allowing for a better spatial and temporal extrapolation (4), predict the bottom-up methane flux from the North Sea into the atmosphere considering yearly oceanographic and weather variations (5), apply machine learning techniques to predict top-down atmospheric methane concentrations (6), test and use 4D visualization capabilities of the Digital-Earth Viewer (7) to explore data sets (8) and animate currents, fluxes, and comparisons interactively and finally transfer all this into new scientific and technological knowledge. Involved in this Show Case were partners from AWI, GEOMAR, HZG and KIT.

The figure below illustrates the 9 different workflows of the Methane in the North Sea Region Show Case. Click on the different workflows (1 to 9), data inputs (light blue) or approaches and methods (green) to see additional explanations.

References:
Grunwald et al., 2009: doi:10.1016/j.ecss.2008.11.021
Rehder et al., 1998: doi.org/10.1023/A:1009644600833
Römer et al., 2017: doi:10.1002/2017GC006995
Mau et al., 2015: doi.org/10.5194/bg-12-5261-2015
Schneider v. Deimling et al., 2011, doi:10.1016/j.csr.2011.02.012
McGinnis et al., 2006: doi:10.1029/2005JC003183
Matoušů et al., 2019, doi.org/10.1007/s00027-018-0609-9

1) The Emission Flux Workflow

In this workflow the global model ICON-ART (ICOsahedral Nonhydrostatic model - Aerosols and Reactive Trace gases) is used. ART is an online-coupled model extension for ICON that includes chemical gases and aerosols. Emission data from well established inventories like EDGAR and point information of the OSPAR commission are used as input for the model to simulate the global distribution of trace gases and detect local hotspots. After applying mathematical methods like the Pattern Algorithm on the model data we are going one step further and try to answer the question if emission fluxes can be determined from satellite measurements.

Sentinel 5P

The Copernicus Sentinel-5P (S5P) mission is a collaboration between European Space Agency (ESA) and the Netherlands Space Office (NSO) with the goal to replace current atmospheric monitoring instruments like SCIAMACHY and the Envisat satellite as both come to the end of their lifetimes. The S5P satellite was launched on October 13, 2017 (09:27 GMT) and has an altitude of 824 km with an orbit time of 101 minutes and a repeat cycle of 17 days or 227 orbits. This air quality and atmospheric chemistry monitoring mission aims to track changes in the composition of our earths atmosphere. Sentinel-5P carries the tropospheric measurement instrument TROPOMI, an advanced multispectral imaging spectrometer that scans trace gases like carbon monoxide, formaldehyde, methane, nitrogen dioxide, ozone and aerosols in the atmosphere with a swath width of 2600 km and delivers much information and data on these substances that affect our climate and the air quality.

Find more information and data access here: https://sentinel.esa.int/web/sentinel/missions/sentinel-5p

Reference:
H. J. Kramer. Copernicus: Sentinel-5p (precursor - atmospheric monitoring mission). https://directory.eoportal.org/web/eoportal/satellite-missions/c-missions/copernicussentinel-5p, Accessed: 2020-07-07, 2020.

OSPAR

The OSPAR commission has the goal to protect and conserve the North-East Atlantic ocean - in particular the North Sea - and its resources. Starting in 1972 OSPAR monitors the developement of offshore installations and publishes datasets containing the name, ID number, location, operator, water depth, production start, current status, category and function of the platforms. The latest version was published in 2017. The countries with oil and gas industry offshore installations are: Denmark, Germany, Ireland, the Netherlands, Norway, Spain and the United Kingdom. The Figure shows a map of Europe with OSPAR offshore platforms displayed on it.

Find data access here: https://odims.ospar.org/

Reference:
OSPAR. Offshore Installations. https://www.ospar.org/work-areas/oic/installations, Accessed: 2020-07-07, 2020.
NASA Visible Earth. https://visibleearth.nasa.gov/collection/1484/blue-marble, Accessed: 2020-07-07, 2020.

Weather Data

The ERA5 dataset is a global weather dataset provided by the European Centre for Medium-Range Weather Forecasts (ECMWF). It covers the atmosphere in 137 levels from the surface level to 80km high. The dataset covers the entire globe using a 30km grid. For our analysis we focus on the year 2018, but data is available since 1979 (1950 in a future release) to near real time. The original dataset provides hourly data, but we use daily aggregated data.

Reference:
Copernicus Climate Change Service (C3S) (2017): ERA5: Fifth generation of ECMWF atmospheric reanalyses of the global climate . Copernicus Climate Change Service Climate Data Store (CDS). https://cds.climate.copernicus.eu/cdsapp#!/home

EDGAR

The Emission Database for Global Atmospheric Research (EDGAR) is an inventory from the EC-JRC and Netherland’s Environmental Assessment Agency

(Saunois et al., 2016). The EDGARv4.3.2 inventory used as emission input for the simulations in this work covers sector- and country-specific

time series of 1970-2012 with monthly resolution and a global spatial resolution of 0.1° x 0.1° providing CH4, CO2, CO, SO2, NOx, C2H6, C3H8 and many other species. The basis for inventories like EDGAR are data from national emission inventory reports that are distributed in space and time via proxy datasets. The latter are based on national spatial data containing information about population density, the road network, waterways, aviation and shipping trajectories. A global 0.1° x 0.1° grid is used on which the emissions are assigned to, either emitted from a single point source (e.g. oil or gas platforms), distributed over a line source (e.g. shiptracks) or over an areal source (e.g. agricultural fields) always depending on the source sectors and subsectors. For this work point sources are from a high importance. These zero-dimensional sources are allocated to a single grid cell of the 0.1° x 0.1° grid with the average of all points that fall into the same cell.

Find more information here: https://edgar.jrc.ec.europa.eu/

Reference:
G. Janssens-Maenhout, M. Crippa, D. Guizzardi, M. Muntean, E. Schaaf, F. Dentener, P. Bergamaschi, V. Pagliari, J. G. J. Olivier, J. A. H. W. Peters, J. A. Van Aardenne, S. Monni, U. Doering and A. M. R. Petrescu. Edgar v4.3.2 global atlas of the three major greenhouse gas emissions for the period 1970–2012. Earth System Science Data Discussions, 2017:1–55, 2017. doi: 10.5194/essd-2017-79, 2017.
G. Janssens-Maenhout, V. Pagliari, D. Guizzardi, and M. Muntean. Global emission inventories in the Emission Database for Global Atmospheric Research (EDGAR) Manual (I) Gridding: EDGAR emissions distribution on global gridmaps. European Commission - Joint Research Centre - Institute for Environment and Sustainability, 2012.

ICON-ART

The global model ICON (ICOsahedral Nonhydrostatic model) is a joint development of German Weather Service (DWD) and Max Planck Institute for Meteorology (MPI-M). Due to its dynamical core which is based on the nonhydrostatic formulation of the vertical momentum equation simulations with a high horizontal resolution up to grid spacings of a few hundreds of meters are possible. ART (Aerosols and Reactive Trace Gases) is an online-coupled model extension for ICON that includes chemical gases and aerosols. One aim of the model is the simulation of interactions between the trace substances and the state of the atmosphere by coupling the spatiotemporal evolution of tracers with atmospheric processes.

Find more information here:
https://www.dwd.de/DE/forschung/wettervorhersage/nummodellierung/01numvorhersagemodelle/iconbeschreibung.html?nn=19880
https://www.imk-tro.kit.edu/5925.php

Reference:
Schröter, J., Rieger, D., Stassen, C., Vogel, H., Weimer, M., Werchner, S., Förstner, J., Prill, F., Reinert, D., Zängl, G., Giorgetta, M., Ruhnke, R., Vogel, B., and Braesicke, P.: ICON-ART 2.1: a flexible tracer framework and its application for composition studies in numerical weather forecasting and climate simulations, Geosci. Model Dev., 11, 4043-4068, https://doi.org/10.5194/gmd-11-4043-2018, 2018.

S5P Level 3 Processor

The Sentinel-5P Level-3 processor is a program written in python that processes a unstructured Level-2 product of the TROPOMI spectrometer to regular grid (Level-3 data) and interpolates it as shown in the left Figure. The unstructured parallelograms are displayed in black with their centers as circles. The red squares represent the new structured latitude/longitude grid. Each center of the unstructured parallelograms of the Level-2 data is allocated to exactly one of the Level-3 structured grid centers by first finding the next latitude and then the next longitude that intersects with a new grid center (blue arrows). The right Figure shows that the Sentinel-5P Level-3 processor is mapping the Level-2 data of the TROPOMI instrument in an appropriate way.

Reference:
European Space Agency. Sentinel-5p. https://sentinel.esa.int/web/sentinel/missions/sentinel-5p, Accessed: 2020-07-07, 2020.

1plus8 - Platform Pattern

The Latitude / Longitude information of the offshore platforms in the OSPAR dataset are brought to 0.1° x 0.1° grid as displayed in the left Figure. These are the locations where we are about to detect emission fluxes but with respect to transport and other uncertainties we need a larger area around the platforms. Adding the eight surrounding pixels to the platform locations we get the 1plus8 – Platform Pattern as displayed in the right Figure. This is the pattern we are using for the Pattern Algorithm.

Find data access here: https://odims.ospar.org/

Reference:
OSPAR. Offshore Installations. https://www.ospar.org/work-areas/oic/installations, Accessed: 2020-07-07, 2020.

Point Source Module

The Point Source Module of ART takes given emission fluxes as point sources of substances and adds them to new or existing tracers. Information like Latitude, Longitude, substance, source strength, start and end time are contained in xml files. With the Point Source Module of ART the sensitivity of tracers as well as the temporal and spatial accuracy of emission fluxes are improved.

Reference:
F. Prill, D. Reiner, D. Rieger, G. Zängl, J. Schröter, J. Förstner, S. Werchner, M. Weimer, R. Ruhnke, and B. Vogel. ICON Model Tutorial. Working with the ICON Model, Practical Exercises for NWP Mode and ICON-ART. Deutscher Wetterdienst, Karlsruhe Institute of Technology, Max-Planck-Institut für Meteorologie, 2019.

2) Marine Bubble Flow

One of the overarching scientific questions is to what extend gas bubble release from the seafloor and specifically from abandoned wells in the North Sea have an impact on atmospheric methane concentrations and how does this change through the seasons. For this evaluation well locations, the change of environmental parameters through a year as well as a bubble dissolution model are needed. Assumptions must be made on e.g. the bubble size distribution (as this determines the rising height and thus transport of methane in the water column) and the absolute amount of gas being released from each well.

Merging the different data sets allows to specify the amount of gas dissolved below the pycnocline and the remaining amount of gas being transported above this atmospheric mixing barrier. The compiled data set allows to interactively manipulate bubble sizes and flow rate at the about 13,000 well locations to see the impact and location of increased gas release towards the atmosphere.

Bubble Dissolution

Methane bubbles, once released from the seafloor, dissolve into the undersaturated water but at the same time strip gases from the water column into the bubble. This means, that the gas composition of the bubbles is changing while the bubble rises (e.g. McGinnis et al., 2006).

Depending on temperature, pressure and salinity gas exchange and equilibrium concentrations change. Knowing the environmental conditions, the initial bubble size and release depth allows to model gas dissolution and the transport of CH4 above the pycnocline where we anticipate full equilibration with the atmosphere - all excess methane from the seafloor that is transported as free gas above the pycnocline will be released into the atmosphere.

We used the GUI-based SingleBubble dissolution software (Greinert & McGinnis, 2009) to calculate bubble dissolution rates for bubbles between 4mm to 12mm diameter, 10m to 220m release depth, salinitiy between 10psu and 36psu and temperature from 2° to 30°C.

Reference:
Greinert, J. and McGinnis, D.F. (2009): Single Bubble Dissolution Model - The Graphical User Interface SiBu-GUI. Environmental Modelling and Software, 24, 1012-1013, doi:10.1016/j.envsoft.2008.12.011
McGinnis, D.F., Greinert, J., Artemov, Y., Beaubien, S.E., and Wüest, A. (2006): The fate of rising methane bubbles in stratified waters: What fraction reaches the atmosphere? Journal of Geophysical Research, 111, C09007, doi:10.1029/2005JC003183.

Well Locations

We created a dataset of well locations in the North Sea. The sources of the datasets are the national institutions involved in the regulation of the oil and gas sector in Denmark, the Netherlands, the United Kingdom and Norway. Depending on the source of data, there is information on the depth of the well, the purpose (e.g. exploration for oil or gas, production, …), whether it is active or not, when it was drilled and shut down. Overall there are 22268 wells in our dataset. On the image on the right you can see the individual wells (purple for Norway, blue for the UK, orange for the Netherlands, and green for Denmark).

Reference:
Danish Energy Agency (https://ens.dk/en/our-responsibilities/oil-gas/about-oil-and-gas)
NLOG from the Geological Survey of the Netherlands (https://nlog.nl/en/data)
Oil and Gas Authority UK (https://www.ogauthority.co.uk/data-centre/data-downloads-and-publications/well-data/)
Norwegian Petroleum Directorate (https://www.npd.no/en/facts/wells/)

NEMO Model

For the properties of the ocean we rely on data from the NEMO model provided by the HZG. Our dataset is a condensed version of the original dataset with adaptions for our use case. The dataset has weekly time steps for our focus year 2018. It covers the North Sea at the original grid cell size. We focus on three depth levels: The surface, the water column above the pycnocline and the water column below the pycnocline. For these levels, the following parameters are provided:

  • Temperature

  • Salinity

  • Zonal current

  • Meridional current

In addition to that, the dataset include the water depth. On the right-hand side you can see a plot of the sea surface temperature for the first week of the year.

Reference:
- CORRECT CITATION?