Paper available

The OceanFlux Greenhouse Gases thermal and haline handling methods are now available in this Biogeosciences Discussions paper

Observing gas transfer between ocean and atmosphere from space

 Short wind waves in the order of centimetres can be observed by satellite altimeters; their relation with gas transfer velocity through the sea surface is used to develop gas transfer algorithms for the world’s oceans.

 

Air-sea gas exchange has been studied for decades because of its role in predictions of climate change, the global carbon cycle and ocean acidification. The oceans absorb more carbon dioxide (CO₂) than they release, but there are major differences in net carbon uptake calculations. Gas transfer velocity, K, is a key variable and the commonly used K parameterizations based on empirical relations with 10 m wind speed (U₁₀) are not good enough on a global scale. An alternative is linking K with short wind waves (Fig. 1) as this relation is more direct than with wind and is measurable from space.

Figure 1

Satellite altimeters can measure the mean square slope of short wind waves, <s²>, and K increases with <s²>.  Two approaches using implicit relationships between <s²> and K have been reported. (1) Linear relations determined from wind-wave tank experiments¹ have been applied^{2, 3}. The same linear relationships are not likely to hold in the open ocean, however. (2) K has been obtained in the field from extrapolations of heat transfer velocity measurements^4,5. Although this method produces reasonable results it needs to be confirmed with in situ K measurements of gas. Our study is the first to correlate K measured at sea and the satellite altimeter signal directly.

An altimeter emits a radar wave and measures backscatter. Space base altimeters were developed to measure sea surface height but nadir looking altimeters can also measure <s²>.  As the wave steepness increases, and hence <s²>, the backscatter in altimeter view decreases. Altimeters on board satellites (ERS-1, ERS-2, TOPEX/POSEIDON, GEOSAT, JASON-1, JASON-2 and ENVISAT) have been measuring the Ku-band backscatter coefficient, σ_Ku, from the ocean surface for 20 years. A homogeneous and calibrated data set is available from the IFREMER site ftp://ftp.ifremer.fr/ifremer/cersat/products/swath/altimeters/waves/data/. For normal incidence <s²> should be proportional to 1/σ_Ku. We therefore correlated K measured during eight cruises around the world with concurring 1/σ_Ku from IFREMER’s data base.

The gas in question was dimethyl sulfide (DMS). K was scaled to a gas with a Schmidt number, Sc, of 660 and signified direct gas transfer through the unbroken surface. K best related to altimeter σ_Ku following K = C + (A/σ_Ku)², while K (cm/hr) was  roughly double U₁₀ (m/s)⁶. We can apply our calibrations for DMS to direct gas transfer of any other gas if we know the diffusivity (expressed by Sc) and solubility of that gas ⁶. For a gas less soluble than DMS, such as CO_2, we also need to add a small term to account for bubble mediated gas transfer^3,7. This term can be derived using the fraction of whitecapping, W,⁷ estimated from other Earth observation, EO, data (model or satellite)⁸. 

 

Disappointingly the K algorithm for DMS based on σ_Ku did not perform better than the traditional algorithm based on U₁₀ (in situ and altimeter). A likely explanation for this is that in the open ocean longer swell waves are usually present (Fig. 2) affecting the altimeter back scatter and muddying the signal of the short wind waves. A way to attenuate the contribution of longer swell waves is to subtract the back scattering signal of a second, lower frequency wave^2,4,5,9.

Figure 2



The Ku band signal frequency (wavelength) is 13.6 GHz  (2.1 cm). Recently IFREMER added C-band data (5.3 GHz / 5.5 cm) of the JASON-1 and -2 (Fig. 3) altimeters to the database. Using these data we explored two band algorithms and found a best fit for a relation of the form K = C + A(1/σ_Ku - B/ σ_C). We found some evidence that for small separation errors between altimeter overpasses and K sample stations (dx < 0.5° and dt < 2 hr) the dual-frequency algorithm reduced the uncertainty in the K estimation by ~0.5 cm/hr compared to both the single-band and wind speed (in situ and altimeter) parameterizations. This is an interesting discovery that will be investigated further.

Figure 3



It is described above how satellite altimeters can monitor air-sea gas transfer velocity, K, over the oceans and how our calibration for DMS can be applied to CO₂ using diffusivity and solubility of CO₂ and whitecap coverage (derived from other EO products). It is then possible to produce maps of CO₂ flux distributions from the product of K for CO₂ and air-sea CO₂ concentration differences.  It is expected that better K parameterizations and EO data will result in improved calculations of the total oceanic CO₂ budget. 

REFERENCES

1.    Bock, E. J., T. Hara, N. M. Frew, and W. R. McGillis (1999), Relationship between air-sea gas transfer and short wind waves, J. Geophys. Res., 104, 25,821-25,831.

2.    Glover, D. M., N. M. Frew, S. J. McCue, and E. J. Bock (2002), A multi-year time series of global gas transfer velocity from the TOPEX dual frequency, normalized backscatter algorithm, in Gas Transfer at Water Surfaces, Geophysical Monograph, vol. 127 edn., edited by M. Donelan, W. M. Drennan, E. Saltzman, and R. Wanninkhof, pp. 325-331, American Geophysical Union, Washington, DC.

3.    Fangohr, S. and D. K. Woolf (2007), Application of new parameterization of gas transfer velocity and their impact on regional and global marine CO₂ budgets, J. Mar. Syst., 66, 195-203.

4.    Frew, N. M. et al. (2004), Air-sea gas transfer: Its dependence on wind stress, small-scale roughness, and surface films, J. Geophys. Res., 109.

5.    Frew, N. M., D. M. Glover, E. J. Bock, and S. J. McCue (2007), A new approach to global air-sea gas transfer velocity fields using dual-frequency altimeter backscatter, J. Geophys. R., VOL. 112.

6.    Goddijn-Murphy, L. M., D. K. Woolf, and C. A. Marandino (2012), Space-based retrievals of air-sea gas transfer velocities using altimeters: Calibration for dimethyl sulfide, J. Geophys. Res., 117. doi:10.1029/2011JC007535

7.    Woolf, D. K. (2005), Parametrization of gas transfer velocities and sea-state-dependent wave breaking, Tellus, 57-B, 87-94.

8.    Goddijn-Murphy, L., D. K. Woolf, and A. H. Callaghan (2011), Parameterizations and algorithms for oceanic whitecap coverage, J. Phys. Oceanogr., 41(4), 742-756, doi: 10.1175/2010JPO4533.1.

9.    Chapron, B., K. Katsaros, T. Elfouhaily, and D. Vandemark (1995), A note on relationships between sea surface roughness and altimeter backscatter, in Air-Water Gas Transfer, 3rd International Symposium on Air-Water Gas Transfer, edited by B. Jähne and E. C. Monahan, pp. 869-878, AEON, Germany.

OceanFlux GHG teams up with UK NERC to produce video

Members from the OceanFlux GHG project team have participated in the NERC Coffee Break Science project to produce a short video describing some of the issues and work happening within OceanFlux GHG.

Coffee Break Science is a UK NERC funded science communication project aimed at producing short videos to communicate aspects of science to a general audience; the videos aim to use novel ways to explain aspects of science and they also explain why and how the work can benefit society.

Filming of the OceanFlux Coffee Break Science film has now been completed and we're hoping to have the final edited copy of the video very soon.
For now, here is a still image from the day of filming!

How to use the OceanFlux data

The project offers the ability to remotely process the data on the Cersat Cloud Facility at IFREMER.

This means that the user does not have to download the data archive but instead work and process them remotely.

Please find here how to open an account, set up a virtual machine matching and start playing with the OceanFlux Greenhouse Gases data.

The data

The relevant datasets (In Situ, models, satellite) are available now.

Please have a look at the details to get them on the website

"Phenomenal sea states and swell from a North Atlantic Storm in February 2011: a comprehensive analysis"

An extra-tropical cyclone in the North Atlantic in February 2011 provided the opportunity to study extreme wind and wave conditions in the open ocean and the subsequent swell field generated by a storm with extensive hurricane-force winds. In a paper soon to be published in the Bulletin of the American Meteorological Society (Hanafin et al, 2012, in print), with many of the OceanFlux-GHG researchers as co-authors, remarkable consistency was found between many different satellite and land-based observations and numerical models, considering the extreme conditions produced by this storm.

Annual average frequency of the low pressure centers with hurricane-force winds based on the NOAA OPC 6-hourly surface pressure analyses and QuikSCAT winds. The average was calculated based on data from September though May 2002-2009. The track of Quirin at 6 hourly intervals from 00Z on 13th February till 18Z on 14th February is overplotted. The size of the circle symbol at each time step reflects the surface area of winds ≥ 24.5 m/s and the colour represents the maximum wind speed.

 The spatial extent of the areas experiencing storm force and hurricane force winds compared very well between satellite-borne scatterometers and NCEP (National Center for Environmental Prediction) numerical model analysis fields. Wave heights up to 20.1m were observed by altimeters during the storm, which was the highest single measurement in a nine-year record, as was the along-track average of 16.2m over 533km. A series of hindcasts were run using the numerical wave model, WAVEWATCH-III ®, forced with NCEP analysis wind fields and comparisons with altimeters show that the model is capable of reproducing the extreme wave heights observed in the open ocean.

Top: altimeter H_s measured by 4 altimeters (Jason-1, Jason-2, ERS-2 and Envisat) on February 13^th (left panel) and February 14^th 2011 (right panel). The black square in the left (right) panel indicates the location of the most extreme sea states measured during these two days by the Envisat (Jason-2) altimeter, respectively.

Middle: Focus on the altimeter (black) H_s values estimated along the Envisat (left) and Jason-2 (right) tracks shown in figure 3 and indicated by the squares above, and computed from the WW3 model forced by ECMWF (red), NCEP (green) and NCEP+10% (blue) winds. A running average has been applied to the altimeter data (~5km resolution) to better match the resolution of the WW3 model (0.5°).

Bottom: Wind speed from different sources interpolated on the same Envisat (left) and Jason-2 (right) altimeter tracks. For both panels, black (green) lines give the altimeter (NCEP) wind speed. For the left (right) panel, the dashed red line gives the ASCAT scatterometer (Jason-2 radiometer) wind speed. On the left panel, the blue line gives the Oceansat-2 wind speed. All estimates have been computed at the spatial resolution of the NCEP fields. The dashed blue lines show the storm force (V ≥ 24.5 m/s) and hurricane-force (V ≥ 32.7 m/s) wind thresholds. A running average was again applied to the altimeter data to better match the resolution of the other data sources (~25km).

The extreme conditions observed during the storm generated swell of periods up to 25 seconds along the coasts of western Europe. The model hindcasts show that the model can also reproduce the swell field generated by the storm very well, as the model results were compared with ocean buoy data and seismic station observations around the Atlantic basin in the days following the storm as the swell made landfall. The time of arrival, the peak periods and the wave heights of the modelled wave fields were in very good agreement with the buoy and seismic station observations.

Top: Peak periods of the swell field: as calculated by WW3; from SAR observations; from wave buoy data; and from seismic buoy data. The background shows the output from the model at 12Z on the 15^th, as the longest swells were encroaching on the west coast of Scotland. The square symbols represent the wave buoy data, the size of the symbol signifying the H_s at the time of the maximum peak period observed and the color signifying the value of the peak period at this time. Beside each symbol is printed the time of arrival of the maximum peak period at each buoy. The circle gives the location of the SAR observations and diamond symbols represent the seismic stations, also colored according to the peak periods observed.

Bottom: Time series of the 3 hour median of the vertical ground displacement variance averaged over 20 minutes, from several stations around the North Atlantic, from February 14^th to 17^th. A timeseries of H_s from a buoy (OLERON) located off the west coast of France is also shown.

The Atlantic extra-tropical storm Quirin produced, on 14^th February 2011, wave heights that are expected to occur only about once a year over the globe, according to our hindcast results. Over a 12-year hindcast period, this storm ranked 3rd largest in terms of significant wave height in the North Atlantic. Waves from the center of the storm radiated as swell with very long periods, from 20 to 25 s, and were recorded around the northern and eastern Atlantic basin. Although the maximum values for wind and wave estimates are difficult to validate, the evidence presented in this study gives credence to the observed scales over which hurricane-force winds and sea state developed. Once the forcing wind field was adjusted to better match the satellite observations, a numerical wave model performed very well in reproducing the local sea state and swell field around the basin, given the extreme input conditions. We are encouraged by these results to report that our ability to both model and observe extreme wave events has improved greatly in recent years, while a novel look at century-old seismic records will help refine the climatology of such rare events.

Final workshop announcement

Please find the first annoucement of the Science Workshop of the project, which will be held at IFREMER (Brest) at the end of the project.

       The NOC team have finished quality controlling the next set of open ocean whitecapping (wavebreaking) in-situ measurements. These data will be used within OceanFlux GHG to help validate the model and Earth Observation derived estimates of whitecapping.

       The figure shows the percentage Whitecap coverage (W) versus wind speed (U10n) standardized to a height of 10 meters above the sea surface and neutral atmospheric stability. The red circles show W averaged as a function of wind speed and the dashed line shows an older empirical relationship (Monahan and Muircheataigh, 1980). All of these in situ data have been collected using digital camera systems mounted on a number of different research ships. 

Measurements were collected during the following projects:

Deep Ocean Gas Exchange Experiment (DOGEE)

High Wind Air-Sea Exchanges (HiWASE)

Sea Spray, Gas Fluxes and Whitecaps (SEASAW)

Waves, Aerosols and Gas Exchange Study (WAGES)

whitecapping in situ data quality control processing underway

The NOC team have recently finished quality controlling a large amount of open ocean whitecapping (wavebreaking) in-situ data. These data will be used within OceanFlux GHG to help validate the model and Earth observation derived estimates of whitecapping.  The attached image shows some example Whitecap coverage W (%) versus 10 m wind speed, with the dotted line showing an older empirical relationship which was derived in 1980. All of these in situ data have been collected using a camera system mounted on a number of different research ships. 

From surf to satellite

Oceanic whitecaps play a role in the uptake of carbon dioxide from the atmosphere, and hence in the Earth’s carbon cycle. Lonneke Goddijn-Murphy writes about monitoring their coverage on a global scale.

As a keen surfer I like to see whitewater, the white foam on the sea surface. From my desk at the Environmental Research Institute (ERI) in Thurso I can keep an eye on Dunnet Head, the most northerly tip of mainland Scotland, and the presence of whitewater where the cliffs meet the ocean is a good indicator for the possibility of a surf session later. As one can imagine, observing whitecaps from space is a more challenging business. But why would we want to use space technology to view whitecaps (other than for chasing surfing waves around the globe)?

My post-doc at the ERI is part of NCEO’s global carbon cycle research. It is well known that burning fossil fuels releases atmospheric carbon dioxide (CO2), a greenhouse gas, and that planting trees helps remove CO2 from the atmosphere. It might be less well-known that the oceans play an important role in the carbon cycle as well. The sea surface can emit or absorb CO2 gas depending on the region and conditions, but on the whole the world’s oceans take up more CO2 than they produce. Here at the ERI we study the physical controls on air-sea gas exchange, an area of expertise for Senior Research Fellow, David Woolf. This includes whitecaps because they enhance the absorption of CO2. We are interested in whitecap observations from satellites because, if we want to compute total CO2 fluxes, we need long term data on a global scale.

Whitecaps play an important role in various other physical processes, for example whitecaps are highly reflective, providing a cooling influence on the Earth’s climate. Whitecaps can also affect the colour of the sea surface, so that whitecap removal algorithms need to be applied to the remote sensing of ocean colour. A better understanding of whitecapping is avidly sought after by wave modellers, because whitecaps relate to energy dissipation of waves, the least known process of wave evolution. Whitecaps are presently used as a ‘tuning knob’ of any wave model, but what exactly are whitecaps made of ?

Whitecaps essentially consist of bubbles and foam, a product of breaking waves that generate turbulence and capture air at the sea surface. A common quantification of whitecaps is the fractional area coverage by whitecaps, W. Although whitecaps are known to distort remote sensing observations, it has appeared to be difficult to monitor W from satellites. A way around is to parameterize W using more assessable parameters. Because whitecaps are mainly wind driven and wind speed data are common in Earth Observation (EO), most W parameterizations are a function of wind speed. Unfortunately, the uncertainties in wind speed parameterizations are too big. This may not be surprising, as one can imagine many factors other than wind speed that affect whitecapping, such as wave height, the stage of wave development, the length of time the wind has been blowing and the interaction between waves. A range of different W parameterizations that take sea state factors into account have been proposed over the years, we tested several. 

One of our problems was finding high quality field measurements. W can be derived from the fraction of white area in an image of the sea surface taken from a ship or a stationary platform. In the pre-digital era these photos were printed and the whitecaps were cut out by hand to weigh on a scale. The weight of these paper cut-outs, divided by the weight of the initial paper, gave W. The drawback of this method is that you really need to average hundreds of images to achieve one useful W value, a bit too much to ask of the person holding the scissors! In present days this process is automated, so that hundreds of frames of digital video recordings can be easily analyzed and averaged in an objective manner. Our colleague Adrian Callaghan from the National University of Ireland, Galway (now at Scripps in San Diego, CA), who developed the Automated Whitecap Extraction technique, kindly gave us his shipboard W measurements.

The W retrievals were state of the art, but the dataset set did not contain information about the sea state. Also, because the wind speed was measured on a moving ship, its accuracy was questionable. EO data were needed to fill in the gaps. We obtained observations from the SeaWinds microwave scatterometer aboard NASA’s QuikSCAT (Quick Scatterometer) satellite. This scatterometer uses radar to measure near-surface wind speed and direction over the ocean under almost all weather and cloud conditions. For detailed information about the sea surface we went to the European Centre for Medium-Range Weather Forecasts (ECMWF). The ECMWF offers global meteorological data, produced by an assimilation of a coupled atmosphere–wave model with reliable observational datasets. The dataset we acquired contained 30 wind and wave variables, we used seven to describe sea surface conditions.

Combining all the field, satellite and re-analysis data, we found that accounting for the state of the sea surface improved W surprisingly little. This was disappointing, but not totally unexpected. An explanation could be that the assessed W parameterizations were just too simple. One of our conclusions was that developed waves relate to increased whitecapping, supporting the assumptions that W increases with wave age and height, and hence with swell. On the other hand cross-swell conditions, i.e. when the directions of wind and waves intersect, appeared to reduce whitecapping. These two counter-acting effects may explain the ongoing debate between wave modellers about whether the presence of swell does, or does not, dampen whitecapping. 

Interestingly, I have experienced both effects in the water while surfing; a bigger swell definitely means bigger whitewater to deal with and I have seen cross-winds blow out lovely waves. Our study might not have resulted in practically improved whitecap measurements from space, but it might have opened doors to a better understanding of wave breaking and wind-wave interaction. Currently we are re-examining alternative ways to measure whitecaps, or air-sea gas exchange, from space directly. But if you will excuse me now, I think I can see whitewater at the headland !

             Lonneke Goddijn-Murphy

Pictures about QuikSCAT :

http://www.nasa.gov/centers/jpl/images/content/92564main_craft1-browse.jpg

http://www.nasa.gov/centers/jpl/images/content/92566main_launch1-browse.jpg

http://winds.jpl.nasa.gov/images/g7_2_high.jpg

Example of ocean surface winds by QuikSCAT

http://winds.jpl.nasa.gov/images/olga_atlantic_high.jpg 

SMOS satellite data and high winds

Soil Moisture and Ocean Salinity (SMOS) is the European Space Agency’s water mission, an Earth Explorer Opportunity Mission belonging to the Living Planet Program. It was launched in November 2009. It aims to provide global and regular observations of soil moisture and sea surface salinity, which are crucial variables to understand and predict the evolution of the water cycle on our planet. 

As high wind observations are very often contaminated by heavy rain and clouds, Reul et al. (2012) show that the SMOS satellite L-band radiometer measurements present a unique opportunity to study the mesoscale evolution of surface winds and whitecap statistical properties under hurricanes and severe storms, and to complement existing active and passive observation systems. 

The SMOS mission currently provides multi-angular L-band brightness temperature images of the Earth. Because upwelling radiation at 1.4 GHz is significantly less affected by rain and atmospheric effects than at higher microwave frequencies, these new SMOS measurements offer unique opportunities to complement existing ocean satellite high wind observations that are often contaminated by heavy rain and clouds. To illustrate this new capability, SMOS data are presented over hurricane Igor, a tropical storm that developed to a Saffir-Simpson category 4 hurricane from 11 to 19 September 2010. Thanks to its large spatial swath and frequent revisit time, SMOS observations intercepted the hurricane 9 times during this period. Without correcting for rain effects, L-band wind-induced ocean surface brightness temperatures were co-located and compared to H*Wind analysis. The L-band ocean emissivity dependence with wind speed appears less sensitive to roughness and foam changes than at the higher C-band microwave frequencies. The first Stokes parameter on a 50 km spatial scale nevertheless increases quasi-linearly with increasing surface wind speed at a rate of 0.3 K/(m/s) and 0.7 K/(m/s) below and above the hurricane-force wind speed threshold (32 m/s), respectively. Surface wind speeds estimated from SMOS brightness temperature images agree well with the observed and modeled surface wind speed features. In particular, the evolution of the maximum surface wind speed and the radii of 34, 50 and 64 knots surface wind speeds are consistent with hurricane model solutions and H*Wind analyses. The SMOS sensor is thus closer to a true all-weather satellite ocean wind sensor with the capability to provide quantitative and complementary surface wind information of interest for operational hurricane intensity forecasts.

(a) Superimposed contours of the wind-excess L-band first Stokes brightness temperature parameter estimated from SMOS data during the 2010 hurricane Igor evolution from September 11 to 19. Contours are ranging from 4 K to 20 K by steps of 1 K. (b) Superimposed contours of the surface wind speed temporally interpolated at SMOS acquisition time from GFS/GFDL hurricane model and (c) H*WIND analysis. Contours are ranging from 15 m/s to 50 m/s by steps of 2.5 m/s. (d) Maximum sustained surface winds Vmax along Igor track from the National Hurricane Center Best Track ATCF system. The white dots indicate the location of the hurricane eye center at the SMOS acquisition time. From Reul et al. (2012)

 

SMOS passes over Igor for which rain rate estimates are available from other spaceborne sensors at an acquisition time less than half an hour from SMOS ones. (left) SMOS wind-excess first Stokes brightness temperature parameter Delta(Th + Tv)/2 in Kelvins. (a) 20:54 Z 11 Sept, (c) 21:16 Z 13 Sept, (e) 09:18 Z 15 Sept, (g) 10:05 Z 19 Sept. (right) Rain rate estimates in mm/h from (b) TRMM/ TMI 20:51 Z 11 Sept, (d) WindSat 21:16Z 13 Sept, SSMIS/F17: (f) 09:40Z 15 Sept and (h) 10:30 Z 19 Sept. The magenta dotted curve indicates the ATCF best-track. From Reul et al. (2012)

 

More information in

Reul, N., J. Tenerelli, B. Chapron, D. Vandermark, Y. Quilfen, Y. Kerr (2012), SMOS satellite L-band radiometer : a new capability for ocean surface remote sensing in hurricanes, J. Geophys. Res., 117, C02006, doi : 10.1029/2011JC007474.

Project has now started

The OceanFlux GHG Project was successfully started in November 2011 with the kick off meeting being held at ESA Esrin in Frascati, Italy. Work is now under way for the Reference Baseline document which will be the blueprint for the project.