Lagrangian methods for atmosphere and ocean sciences

Lagrangian trajectories are currently used for a vast range of purposes in ocean and atmosphere sciences. Examples include studying the connectivity of ocean basins, forecasting the spreading of ash clouds, mapping global ocean diffusivities, observing the deep ocean, or tracing plastics and other forms of pollutants in the ocean, etc. There is thus a need for numerical models capable of simulating Lagrangian particles in the ocean and atmosphere as well as accurate methods for analysing the data from surface drifters, floats, and simulated particles.

This session aims at bringing together scientists working on all sorts of Lagrangian methods, e.g. observed or simulated particles in the atmosphere and ocean, and a variety of use cases e.g. studying oceanic mixing/diffusivity, tracing pollution in the atmosphere or ocean, iceberg tracking etc. We welcome presentations on e.g.:

* Connectivity and pathways of air- or water-masses in the atmosphere and the ocean
* Quantifying water mass transformations and fluxes between regions in the ocean
* Development of Lagrangian particle-tracking algorithms and approaches to model particles with active behaviours, e.g. icebergs, fish, ash clouds, plastics etc.
* New methods and tools to analyse observed or simulated Lagrangian particles, e.g. diffusivity, spreading rates, etc.
* New developments in in-situ observations such as balloons, surface drifters or floats.

Co-organized by AS5/CL5.2/OS4
Convener: Joakim KjellssonECSECS | Co-conveners: Sara BerglundECSECS, Kristofer Döös, Bror Jönsson
vPICO presentations
| Fri, 30 Apr, 09:00–12:30 (CEST)

vPICO presentations: Fri, 30 Apr

Chairpersons: Bror Jönsson, Joakim Kjellsson
Aitor Aldama Campino, Kristofer Döös, Sara Berglund, Dipanjan Dey, Joakim Kjellsson, and Bror Jönsson

The latest version of the TRACMASS trajectory code, version 7.0 will be presented. The latest version includes several new features, e.g. water tracing in the atmosphere, generalisation of the tracer handling, and improvements to the numerical scheme. The code has also become more user friendly and easier to get started with. Previous versions of TRACMASS only allowed temperature, salinity and potential density to be calculated along the trajectories, but the new version allows any tracer to be followed e.g. biogeochemical tracers or chemical compounds in the atmosphere. 

TRACMASS calculates Lagrangian trajectories offline for both the ocean and atmosphere by using already stored velocity fields, and optionally tracer fields. The code supports most vertical coordinate systems, e.g. z-star, z-tilde, sigma, and hybrid sigma-pressure coordinates. Hence, TRACMASS supports a range of atmosphere and ocean models such as ECMWF IFS, NEMO, ROMS, MOM, as well as reanalysis products (e.g. ERA-5) or observations (e.g. geostrophic currents from AVISO satellite altimetry). The fact that the numerical scheme in TRACMASS is mass conserving allows us to associate each trajectory with a mass transport and calculate the Lagrangian mass transport between different regions as well as construct Lagrangian stream functions. 

A short course on how to set up, configure and run the TRACMASS code will be given separately, SC5.17.

How to cite: Aldama Campino, A., Döös, K., Berglund, S., Dey, D., Kjellsson, J., and Jönsson, B.: TRACMASS 7.0 - A Lagrangian trajectory code for atmosphere and ocean sciences, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-1690,, 2021.

Christian Kehl, Daan Reijnders, Reint Fischer, Roel Brouwer, Raoul Schram, and Erik van Sebille

Lagrangian simulations contribute to the study and comprehension of particulate-matter transport, its dissolution and dispersion in the oceans. Parcels is an open-source, Python-based module for Lagrangian ocean simulations. It is a known tool in the oceanographic community that has been applied to a variety of case studies, such as the tracing of microplastics, the backtracking of ocean floor plankton, and the migration of fish. In this module, particles are advected over time according to a selected flow field, where those particles can represent particulate-matter, biota or other objects with physical, hydrodynamic or biogeochemical properties. In this contribution, we present the substantial extensions of Parcels with respect to usability, physics modelling aspects of particle advection, and computational aspects of versatile, scalable and efficient simulations.

Specifically, a suite of simple, concise notebook tutorials are tailored to novice user, covering step-by-step simulation setup instructions, whereas self-contained special-issue tutorials address advanced- and proficient user requirements. The considerable expansion of supported OGCM flow field input formats (e.g. MITgcm, POP and MOM5, among others) is a major interest in Parcels v2.2 for our steadily-growing user base.

The new version further integrates previously-published physics methods into practical lagrangian particle simulations. As such, we implement an analytical advection scheme in addition to existing Runge-Kutta advection schemes. Furthermore, two-dimensional advection-diffusion is upgraded with the Milstein stochastic integration scheme and improved documentation. Those capabilities enable a more consistent modelling of diffusion- and uncertainty-dominated fluid transport processes.

The case studies performed with previous versions indicate increased computational demands. Simulations are run over long decadal time scales as well as over day-periods with sub-second temporal increments, involving multiple basins and global scenarios, while also modelling increasingly complex particle processes. Overall, our developments respond to the big-data requirements of modern oceanographic studies, which include the aspects of (i) high record volume (i.e. large number of particles), (ii) high dimensionality in multi-variate records, (iii) high spatial resolution, (iv) high temporal resolution, (v) high scenario (i.e. case study) variability and (vi) the prevention of numerical error accumulation over long simulation time scales.

The novel features of Parcels v2.2 are illustrated on distinct case studies within our contribution, in order to connect the technical features to their impact on particulate-matter ocean transport studies.

How to cite: Kehl, C., Reijnders, D., Fischer, R., Brouwer, R., Schram, R., and van Sebille, E.: Parcels 2.2 - An increasingly versatile, open-source Lagrangian ocean simulation tool, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-1033,, 2021.

Ezra Eisbrenner and Léon Chafik

Knowledge about water-mass properties is critical to understanding how ocean climate variability impacts the shelf seas. Disentangling the origin of shelf sea water-masses and associated driving mechanisms is, therefore, a significant step towards improving the predictive skill related to water-mass evolution. Especially more conservative water-mass properties, even of surface waters, have the potential to reveal links between the shelf seas and large-scale ocean circulation regimes when traced back to their origin. The northern North Sea for example as the main gateway for water-masses to one of Europe's largest shelf sea areas is largely supplied by water-masses from the open North Atlantic, a connection which can be seen from, e.g., sea surface salinity.

The aim of this study is to identify the origin of northern North Sea water-masses and distinguish pathway variability relative to the subpolar gyre regimes. This is done using Lagrangian trajectories, calculated using satellite-derived velocity fields. The results of the Lagrangian statistics mainly indicate that on inter-annual time-scales the North Atlantic subpolar gyre strength largely influences the water-masses found in the North Sea. The relation is found to originate from varying pathways and therefore origin. We conclude that on inter-annual time scales the subpolar gyre strength is a good proxy and skillful predictor of water-mass variability in the North Sea.

How to cite: Eisbrenner, E. and Chafik, L.: A satellite-based Lagrangian view on the origin of water-masses in the northern European shelf seas, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-2220,, 2021.

Philippe Miron, Maria J. Olascoaga, Francisco J. Beron-Vera, Kimberly L. Drouin, and M. Susan Lozier

The North Atlantic Deep Water (NADW) flows equatorward along the Deep Western Boundary Current (DWBC) as well as interior pathways and is a critical part of the Atlantic Meridional Overturning Circulation. Its upper layer, the Labrador Sea Water (LSW), is formed by open-ocean deep convection in the Labrador and Irminger Seas while its lower layers, the Iceland–Scotland Overflow Water (ISOW) and the Denmark Strait Overflow Water (DSOW), are formed north of the Greenland–Iceland–Scotland Ridge.

In recent years, more than two hundred acoustically-tracked subsurface floats have been deployed in the deep waters of the North Atlantic.  Studies to date have highlighted water mass pathways from launch locations, but due to limited float trajectory lengths, these studies have been unable to identify pathways connecting  remote regions.

This work presents a framework to explore deep water pathways from their respective sources in the North Atlantic using Markov Chain (MC) modeling and Transition Path Theory (TPT). Using observational trajectories released as part of OSNAP and the Argo projects, we constructed two MCs that approximate the lower and upper layers of the NADW Lagrangian dynamics. The reactive NADW pathways—directly connecting NADW sources with a target at 53°N—are obtained from these MCs using TPT.

Preliminary results show that twenty percent more pathways of the upper layer(LSW) reach the ocean interior compared to  the lower layer (ISOW, DSOW), which mostly flows along the DWBC in the subpolar North Atlantic. Also identified are the Labrador Sea recirculation pathways to the Irminger Sea and the direct connections from the Reykjanes Ridge to the eastern flank of the Mid–Atlantic Ridge, both previously observed. Furthermore, we quantified the eastern spread of the LSW to the area surrounding the Charlie–Gibbs Fracture Zone and compared it with previous analysis. Finally, the residence time of the upper and lower layers are assessed and compared to previous observations.

How to cite: Miron, P., Olascoaga, M. J., Beron-Vera, F. J., Drouin, K. L., and Lozier, M. S.: Identification and quantification of the North Atlantic Deep Water pathways, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-10358,, 2021.

Sara Berglund and Kristofer Döös

Water mass transformation is an important part of the Ocean circulation. Lagrangian trajectories can be used to connect pathways with water mass properties such as temperature and salinity. Here, we will introduce the Lagrangian divergence of heat and salt that can be computed using Lagrangian trajectories. This is a new method that can be used to determine where water masses are changing temperature or salinity geographically.
Further, the following two examples on how to use the Lagrangian divergence will be given:

(1) In the Atlantic Ocean water flows northward and transform from warm and saline to cold and fresh. The Lagrangian divergence has been used to show that this cooling and freshening is confined to the North Atlantic Subtropical Gyre.

(2) Waters in the upper limb of the Southern Hemisphere Conveyor Belt circulation converts from cold and fresh to warm and saline as it travels from the Southern Ocean to the tropics. The Lagrangian divergence shows that this warming and salinification are confined to the Antarctic Circumpolar Current, the southern subtropical gyres, and the equator. In this study, the Lagrangian divergence are separated by the mixed layer depth, which distinguishes if a change in heat and salt is driven by internal mixing or air--sea interactions.

How to cite: Berglund, S. and Döös, K.: The Lagrangian divergence of heat and salt: A new method to determine water mass transformations, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-1246,, 2021.

Jonathan Wiskandt, Siren Ruehs, Franziska Schwarzkopf, and Arne Biastoch

The upper limb of the Atlantic Meridional Overturning Circulation (AMOC) is supplied in the South Atlantic from Drake Passage (DP) and Agulhas Leakage (AL). The relative contributions from DP and AL influence the stratification as well as the properties of the upper limb return flow and potentially impact the formation of deep water in the North Atlantic.
While early studies suggested a clear dominance of the AL contribution, recent studies indicate that the DP contribution is not negligible. Here, we use a set of Lagrangian experiments in the eddy-resolving (1/20 degree) ocean model INALT20 to analyze the inflow from DP into the South Atlantic in more detail. We find that the majority of water, that enters the subtropical South Atlantic across 30° S from DP, originates from the upper 2000 m of the northern branch of the ACC that follows the Sub Antarctic Front (SAF). Before  entering the South Atlantic, the majority of theses particles turn northward east of DP and follow the SAF through the Brazil Malvinas Confluence, where the SAF meets the Sub Tropical Front. In or parallel to the South Atlantic Current, particles cross the basin and become part of the subtropical gyre to follow the Benguela Current northward. We further compare pathways, volume transports, transit times and thermohaline properties of particles entering through DP and leaking into the South Atlantic to those from particles not leaking into the South Atlantic. These analyses help exploring potential recipes for building a timeseries of “Drake Passage leakage”, complementary to the already established Agulhas Leakage timeseries.

How to cite: Wiskandt, J., Ruehs, S., Schwarzkopf, F., and Biastoch, A.: Assessing "Drake Passage Leakage" in an eddy resolving ocean model, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-2821,, 2021.

Leo Berline, Andrea Doglioli, Anne Petrenko, Stephanie Barrillon, Boris Espinasse, Frederic Le Moigne, François Simon-Bot, Thyssen Melilotus, and François Carlotti

In the upper layers of the Ionian Sea, young Mediterranean Atlantic Waters (MAW) flowing eastward from the Sicily channel meet old MAW. In May 2017, during the PEACETIME cruise, fluorescence and particle content sampled at high resolution revealed unexpected heterogeneity in the central Ionian. Surface salinity measurements, together with altimetry-derived and hull-mounted ADCP currents, describe a zonal pathway of AW entering the Ionian Sea, consistent with the so-called cyclonic mode in the North Ionian Gyre. The ION-Tr transect, located ~19-20°E- ~36°N turned out to be at the crossroad of three water masses, mostly coming from the west, north and from an isolated anticyclonic eddy northeast of ION-Tr. Using Lagrangian numerical simulations, we suggest that the contrast in particle loads along ION-Tr originates from particles transported from these three different water masses. Waters from the west, identified as young AW carried by a strong southwestward jet, were intermediate in particle load, probably originating from the Sicily channel. Water mass originating from the north was carrying abundant particles, probably originating from northern Ionian, or further from the south Adriatic. Waters from the eddy, depleted in particles and Chl-a may originate from south of Peloponnese, where the Pelops eddy forms.

The central Ionian Sea hence appears as a mosaic area, where waters of contrasted biological history meet. This contrast is particularly clear in spring, when blooming and non-blooming areas co-occur.

High resolution measurements reveal a high heterogeneity in properties such as particles abundances. To interpret these distributions, combination of multiparametric in situ measurements with remote sensing and Lagrangian modeling appears necessary.

How to cite: Berline, L., Doglioli, A., Petrenko, A., Barrillon, S., Espinasse, B., Le Moigne, F., Simon-Bot, F., Melilotus, T., and Carlotti, F.: Particle transport in the central Ionian Sea, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-11870,, 2021.

Christina Schmidt, Franziska Schwarzkopf, Siren Rühs, and Arne Biastoch

The exchange of water between the Indian Ocean and South Atlantic with their different thermohaline properties via Agulhas Leakage is important for the meridional overturning circulation. Agulhas Leakage as well as the output of ocean general circulation models in general can be analysed using a Lagrangian approach with a variety of different tools available. Here, Agulhas Leakage is estimated with both the newly developed tool Parcels and the well established tool Ariane, and different designs of the Lagrangian experiment are analysed. In a hindcast simulation with the eddy-rich ocean sea-ice model INALT20 (1/20° horizontal resolution) under the new JRA55-do forcing, Agulhas Leakage increases from the early 1960s to mid 1980s, but there is no clear trend afterwards, which is in contrast to earlier studies using hindcast simulations under the CORE forcing. During the transit from the Agulhas Current at 32°S to the Cape Basin, a cooling and freshening of Agulhas Leakage waters occurs especially in the western part of the Retroflection, resulting in a density increase as the thermal effect dominates. The average transport, its variability, trend and the transit time from the Agulhas Current to the Cape Basin of Agulhas Leakage is simulated equally with the Lagrangian tools Ariane and Parcels, emphasising the robustness of our method.

How to cite: Schmidt, C., Schwarzkopf, F., Rühs, S., and Biastoch, A.: Characteristics and robustness of Agulhas Leakage estimates: an inter-comparison study of Lagrangian methods, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-12107,, 2021.

Yonggang Liu, Robert H. Weisberg, Lianyuan Zheng, and Katherine Hubbard

A short-term forecast tool is developed to help federal, state, and local end users monitor and manage harmful algal blooms on the west coast of Florida. The short-term forecasts are based on the West Florida Coastal Ocean Model (WFCOM) that downscales from the deep ocean, across the continental shelf and into the estuaries, and the Tampa Bay Coastal Ocean Model (TBCOM) that has resolution high enough to include all of the inlets connecting Tampa Bay, Sarasota Bay and the Intra-Coastal Waterway with the adjacent Gulf of Mexico. Observed Karenia brevis cell concentration data are uploaded daily into the WFCOM and TBCOM to generate 3.5 day forecasts of the bloom Lagrangian trajectories on the shelf and in the estuaries. This provides information where red tide may go in the next few days. Noting that the spatial red tide sampling is limited and blooms may be patchy, a more general and user-friendly map is produced to show where a red tide bloom may occur along the coast over the next several days. The tracking tool displays modeled bloom trajectories at the surface and the bottom with five categories of cell concentrations (present, very low, low, medium, and high, each differing approximately by an order of magnitude). The performance of the Lagrangian trajectory model is evaluated with satellite-tracked surface Lagrangian drifters using a skill score that is defined from the normalized cumulative Lagrangian separation (NCLS).

How to cite: Liu, Y., Weisberg, R. H., Zheng, L., and Hubbard, K.: Short-Term Forecast of Harmful Algal Blooms on the West Florida Shelf, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-7819,, 2021.

Alberto Baudena, Enrico Ser-Giacomi, Donatella d'Onofrio, Xavier Capet, Cedric Cotté, Yves Cherel, and Francesco d'Ovidio

Oceanic Lagrangian Coherent Structures have been shown to deeply influence the distribution of primary producers and, at the other extreme of the trophic web, top predators. However, the relationship between these structures and intermediate trophic levels is much more obscure. In this work we contribute to address this knowledge gap by comparing acoustic measurements of mesopelagic fish concentrations to satellite-derived fine-scale Lagrangian Coherent Structures in the open ocean. The results demonstrate that higher fish concentrations occur more frequently over stronger Lagrangian Coherent Structures. Quantile regression analyses reveal that Lagrangian Coherent Structures represent a limiting condition for high fish concentrations. Therefore, while the presence of a fine-scale feature does not imply a concomitant fish assembly, increased fish densities are more likely to be observed over these structures. Finally, we discuss a model representing fish movement along Lagrangian features, and specifically built for mid trophic levels. Even though it was not possible to validate it with the available data, its results, obtained with realistic parameters, are consistent with the observations. These findings may help to integrate intermediate trophic levels in trophic models, which can ultimately support management and conservation policies.

How to cite: Baudena, A., Ser-Giacomi, E., d'Onofrio, D., Capet, X., Cotté, C., Cherel, Y., and d'Ovidio, F.: Fine-scale structures as spots of increased fish concentration in the open ocean, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-5540,, 2021.

Julio Antonio Lara-Hernández, Cecilia Enríquez-Ortiz, Jorge Zavala-Hidalgo, Abigail Uribe-Martínez, and Eduardo Cuevas-Flores

The possible fate of pelagic sargassum in the Mexican Caribbean during aug-2018, sep-2018, and apr-2019 is analyzed using a particle-tracking model coupled to diverse datasets of wind [ERA5 reanalysis and the NCEP Climate Forecast System (CFSv2)] and ocean current velocities (HYCOM experiments of high and lower resolution). Advection of particles was computed considering 0, 1, 2, or 3 % of the wind magnitude and either surface currents (0 m) or the averaged currents from the surface to 5 m depth. For each day of the three months, virtual particles were initially located at the vertices of a uniform mesh within the Mexican Caribbean and subsequently tracked for 10 days. Results revealed that the percentage of the wind magnitude accounted for the transport had the greatest impact on the number of particles that ran aground in the Mexican Caribbean: with a higher percentage of the wind magnitude more particles reached the land. The depth of the layer of the ocean currents used in the transport was also important in the results: particle stranding was higher when only surface currents were used. On the other hand, the different data sources had less influence in the results: the simulations using CFSv2 winds resulted in more stranding of particles than those using ERA5 winds, although the differences were relatively small. The number of stranded particles was virtually insensitive to the selection of the ocean data resolution (i. e. HYCOM of high or lower resolution). In general, virtual particles located closer to the coast and further south in the Mexican Caribbean showed the highest probability of running aground on the shores of the Mexican Caribbean. The arrival time depended on the distance from the shore and the wind magnitude. With the wind and current conditions of the three months used for the study, particles located less than 50 km from the shore usually required less than 3 days to run aground. Particles between 50 and 200 km from the shore usually had an arrival time between 3 and 10 days. The dynamics of the particles were similar during each of the months. However, the greatest differences corresponded to apr-2019, when shifting winds and northerlies were observed. This provides an insight of the variations that most likely would result for different months and years. However, sargassum arrivals are expected to occur during the summer, hence these results are relevant for the local preparedness of managing strategies for massive sargassum stranding in the Mexican Caribbean.

How to cite: Lara-Hernández, J. A., Enríquez-Ortiz, C., Zavala-Hidalgo, J., Uribe-Martínez, A., and Cuevas-Flores, E.: Transport dynamics of pelagic sargassum in the Mexican Caribbean: sensitivity studies to the wind and depth of the transporting ocean layer, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-3688,, 2021.

Noémie Ehstand, Reik Donner, Cristóbal López, and Emilio Hernández-García

In the past decades, boreal summers have been characterized by a number extreme weather events such as heat waves, droughts and heavy rainfall periods with significant social, economic and environmental impacts. One of the most outstanding examples occurred in the summer of 2010 when an anomalously strong heatwave persisted over Eastern Europe for several weeks while extreme rainfalls struck Pakistan, leading to the country’s worst floods in record history. Both events were related to the presence of an anomalously persistent atmospheric blocking situation - that is a large-scale, nearly stationary, atmospheric pressure pattern - over Eastern Europe.

The high impact of blocking events has motivated numerous studies. However, there is not yet a comprehensive theory explaining their onset, maintenance and decay and their prediction remains a challenge.

In this work, we employ a Lagrangian dynamics based, complex network description of the atmospheric transport to study the connectivity patterns associated with atmospheric blocking events. The network is constructed by associating nodes to regions of the atmosphere and establishing links based on the flux of material between these nodes during a given time interval, as described in Ser-Giacomi et al. [1]. One can then use the tools and metrics developed in the context of graph theory to explore the atmospheric flow properties. In particular, we demonstrate the ability of measures such as the network degree, entropy and harmonic closeness centrality to trace the spatio-temporal characteristics of atmospheric blocking events.

[1] E. Ser-Giacomi, V. Rossi, C. López, E. Hernández-García, Chaos 25(3), 036404 (2015)


This research was conducted as part of the CAFE Innovative Training Network (Climate Advanced Forecasting of sub-seasonal Extremes, which has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No. 813844.

How to cite: Ehstand, N., Donner, R., López, C., and Hernández-García, E.: Detection and tracking of atmospheric blocks: a Lagrangian flow network approach, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-1527,, 2021.

Dipanjan Dey and Kristofer Döös

The water-mass sources and their variability responsible for the South Asian summer monsoon precipitation were investigated using Lagrangian atmospheric water-mass trajectories. The results indicated that water-masses from the Central and South Indian Ocean are the dominant contributors to the total South Asian summer monsoon rainfall, followed by the contribution from the local recycling, the Arabian Sea, remote sources and the Bay of Bengal. It was also found that although the direct contribution originating from the Bay of Bengal is small, it still provides a route for the water-masses that come from other regions. The outcomes further revealed that the water-masses originating from the Central and South Indian Ocean are responsible for the net precipitation over the coastal regions of the Ganges-Brahmaputra-Meghna Delta, Northeast India, Myanmar, the foothills of the Himalayas and Central-East India. Water-masses from the Arabian sea are mainly contributing to the rainfall over the Western coast and West-Central India. Summer monsoon precipitation due to the local recycling is primarily restricted to the Indo-Gangetic plain. No recycled precipitation was observed over the mountain chain along the West coast of India (Western Ghats). The inter-annual variability of the South Asian summer monsoon precipitation was found to be mainly controlled by the water-masses from the Central and South Indian Ocean.

How to cite: Dey, D. and Döös, K.: Lagrangian tracing and analysis of the South Asian summer monsoon precipitation, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-1450,, 2021.

Kristofer Döös, Sara Berglund, Dipanjan Dey, Aitor Aldama Campino, and Christophe Menkes

The hydrological cycle of the tropical Pacific Ocean is traced with Lagrangian water mass trajectories in the coupled ocean-atmosphere system.
The cycle consists of one half in the atmosphere and one half in the ocean, where the two halves are connected by the evaporation and precipitation regions at the sea surface.
The atmospheric part of the water cycle is traced backward from the precipitation at the sea surface of the Warm Pool to the evaporation regions in the eastern tropical Pacific.
Reversely, the ocean part of the cycle is also traced from the precipitation to the evaporation regions with water mass trajectories, with emphasis on the part that recirculates within the Tropical Pacific.
The air circulation of the Walker Cell is superimposed on the ocean-atmosphere water cell both in the zonal-vertical space as well as in the hydrothermohaline space. This reveals how the ocean and atmosphere are connected, which are, to some extent, governed by the Clausius-Clapeyron relationship in the evaporation regions. 

The Lagrangian trajectories are computed with the trajectory code TRACMASS, where the atmospheric water parcels are advected with the 3D water mass fluxes based on a new water mass conservation method, which includes precipitation.

How to cite: Döös, K., Berglund, S., Dey, D., Aldama Campino, A., and Menkes, C.: The hydrological cycle of the Walker Cell, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-2580,, 2021.

Jezabel Curbelo, Gang Chen, and Carlos R. Mechoso
The evolution of the Northern Hemisphere stratosphere during late winter and early spring of 2020 was punctuated by outstanding events both in dynamics and tracer evolution. It provides an ideal case for study of the Lagrangian properties of the evolving flow and its connections with the troposphere. The events ranged from an episode of polar warming at upper levels in March, a polar vortex split into two cyclonic vortices at middle and lower levels in April, and a remarkably deep and persistent mass of ozone poor air within the westerly circulation throughout the period. The latter feature was particularly remarkable during 2020, which showed the lowest values of stratospheric ozone on record.
We focus on the vortex split in April 2020 and we examine this split at middle as well as lower stratospheric levels, and the interactions that occurred between the resulting two vortices which determined the distribution of ozone among them. We also examine the connections among stratospheric and tropospheric events during the period.
Our approach for analysis will be based on the application of Lagrangian tools to the flow field, based on following air parcels trajectories, examining barriers to the flow, and the activity and propagation of planetary waves. Our findings confirm the key role for the split played by a flow configuration with a polar hyperbolic trajectory and associated manifolds. A trajectory analysis illustrates the transport of ozone between the vortices during the split. We argue that these stratospheric events were linked to strong synoptic scale disturbances in the troposphere forming a wave train from the north Pacific to North America and Eurasia.
Reference: J. Curbelo, G. Chen,  C. R. Mechoso. Multi-level analysis of the northern polar vortex split in April 2020 during development of the Arctic ozone hole. Earth and Space Science Open Archive. doi: 10.1002/essoar.10505516.1
Acknowledgements: NSF Grant AGS-1832842, RYC2018-025169 and EIN2019-103087.

How to cite: Curbelo, J., Chen, G., and Mechoso, C. R.: Lagrangian analysis of the northern polar vortex split in April 2020 during development of the Arctic ozone hole, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-13597,, 2021.

Johanna Knauf, Joakim Kjellsson, and Annika Reintges

We study the impact of ocean horizontal resolution on storm tracks over the North Atlantic Ocean using the FOCI-OpenIFS climate model and the TRACK storm-tracking algorithm. We find that increasing ocean resolution from 1/2° to 1/10° reduces a cold bias over the North Atlantic which leads to a northward shift of the storm tracks, in particular in winter and spring seasons. 

Most climate models with non-eddying oceans, i.e. horizontal resolutions of 100 km or higher, suffer from a cold SST bias in the North Atlantic. Refining the horizontal resolution from 1/2° to 1/10° allows for a distinct Gulf Stream extension and better representation of the major current systems which reduces this cold bias. The associated warming of the ocean surface with increasing resolution also warms the troposphere and leads to a northward shift in the tropospheric eddy-driven jet. Overall, the increased ocean resolution thus improves the ocean circulation as well as the atmospheric circulation. 

We use two metrics to evaluate the storm track activity in the simulations. We calculate 2-8 day bandpass-filtered mean sea-level pressure (MSLP) and eddy heat flux (v’T’) which is an Eulerian metric that shows variability of low- and high-pressure systems as well as their associated heat flux, but says nothing about the genesis, lysis or life time of individual storms. We also use the TRACK storm-tracking algorithm with 12-hourly MSLP data to produce trajectories of individual storms, which allows us to study individual storms. 

The Eulerian approach using MSLP variance and eddy heat fluxes clearly shows a northward shift of the storm tracks as the ocean resolution is increased. Overall, the northward shift leads to reduced biases compared to ERA-Interim reanalysis. Storm-track trajectories show higher storm track and storm genesis densities around 60°N with the higher ocean resolution. Interestingly, a higher ocean resolution also results in longer life time of storms. We speculate that this is due to enhanced air-sea interactions where cyclones are fed more energy from the eddy-resolving ocean than from the non-eddying ocean.

How to cite: Knauf, J., Kjellsson, J., and Reintges, A.: Impact of ocean resolution on storms in the North Atlantic region, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-6061,, 2021.

Chairpersons: Sara Berglund, Joakim Kjellsson, Kristofer Döös
Daan Reijnders, Erik Jan van Leeuwen, and Erik van Sebille
To identify barriers to transport in a fluid domain, community detection algorithms from network science have been used to divide the domain into clusters that are sparsely connected with each other. In a previous application to the closed domain of the Mediterranean Sea, communities detected by the Infomap algorithm have barriers that often coincide with well-known oceanographic features. We apply this clustering method to the surface of the Arctic and subarctic oceans and thereby show that it can also be applied to open domains. First, we construct a Lagrangian flow network by simulating the exchange of Lagrangian particles between different bins in an icosahedral-hexagonal grid. Then, Infomap is applied to identify groups of well-connected bins. The resolved transport barriers include naturally occurring structures, such as the major currents. As expected, clusters in the Arctic are affected by seasonal and annual variations in sea-ice concentration. An important caveat of community detection algorithms is that many different divisions into clusters may qualify as good solutions. Moreover, while certain cluster boundaries lie consistently at the same location between different good solutions, other boundary locations vary significantly, making it difficult to assess the physical meaning of a single solution. We therefore consider an ensemble of solutions to find persistent boundaries, trends and correlations with surface velocities and sea-ice cover.

How to cite: Reijnders, D., van Leeuwen, E. J., and van Sebille, E.: Ocean Surface Connectivity in the Arctic: Capabilities and Caveats of Community Detection in Lagrangian Flow Networks, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-201,, 2021.

Stefano Berti and Guillaume Lapeyre

Oceanic motions at scales larger than few tens of km are quasi-horizontal due to seawater stratification and Earth’s rotation and are characterized by quasi-two-dimensional turbulence. At scales around 300 km (in the mesoscale range), coherent vortices contain most of the kinetic energy in the ocean. At scales around 10 km (in the submesoscale range) the flow is populated by smaller eddies and filamentary structures associated with intense gradients (e.g. of temperature), which play an important role in both physical and biogeochemical budgets. Such small scales are found mainly in the weakly stratified mixed layer, lying on top of the more stratified thermocline. Submesoscale dynamics should strongly depend on the seasonal cycle and the associated mixed-layer instabilities. The latter are particularly relevant in winter and are responsible for the generation of energetic small scales that are not trapped at the surface, as those arising from mesoscale-driven processes, but extend down to the thermocline. The knowledge of the transport properties of oceanic flows at depth, which is essential to understand the coupling between surface and interior dynamics, however, is still limited.

By means of numerical simulations, we explore Lagrangian pair dispersion in turbulent flows from a quasi-geostrophic model consisting in two coupled fluid layers (representing the mixed layer and the thermocline) with different stratification. Such a model has been previously shown to give rise to both meso and submesoscale instabilities and subsequent turbulent dynamics that compare well with observations of wintertime submesoscale flows. We focus on the identification of different dispersion regimes and on the possibility to relate the characteristics of the spreading process at the surface and at depth, which is relevant to assess the possibility of inferring the dynamical features of deeper flows from the experimentally more accessible (e.g. by satellite altimetry) surface ones.

Using different statistical indicators, we find a clear transition of dispersion regime with depth, which is generic and can be related to the statistical features of the turbulent flows. The spreading process is local (namely, governed by eddies of the same size as the particle separation distance) at the surface. In the absence of a mixed layer it rapidly changes to nonlocal (meaning essentially driven by the largest eddies) at small depths, while in the opposite case this only occurs at larger depths, below the mixed layer. We then identify the origin of such behavior in the existence of fine-scale energetic structures due to mixed-layer instabilities. We further discuss the effect of vertical shear and address the properties of the relative motion of subsurface particles with respect to surface ones. In the absence of a mixed layer, the properties of the spreading process are found to rapidly decorrelate from those at the surface, but the relation between the surface and subsurface dispersion appears to be largely controlled by vertical shear. In the presence of mixed-layer instabilities, instead, the statistical properties of dispersion at the surface are found to be a good proxy for those in the whole mixed layer.

How to cite: Berti, S. and Lapeyre, G.: Lagrangian pair dispersion in upper-ocean turbulent flows with mixed-layer instabilities, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-254,, 2021.

Nikolas Aksamit, Alex Encinas Bartos, and George Haller

Lagrangian coherent structures (LCS) provide a means to understand persistent flow features in an objective manner. There has been great success identifying and harnessing hyperbolic, elliptic, and parabolic structures in both oceanic and atmospheric flows. These approaches (e.g. FTLE, PRA, LAVD) rely on well resolved velocity information for the computation of the gradient of the flow map or vorticity deviation. Thus, for sparse data, such as that available from ocean drifters or atmospheric balloons, the quality of these methods quickly deteriorates. On the other hand, all elementary features of individual particle paths, such as velocity, acceleration, looping number, curvature and trajectory length, are non-objective, i.e., depend on the observer. To bridge this gap between LCS and sparse data, we derive measures of local material stretching and rotation that are computable from individual trajectories without reliance on other trajectories or on an underlying velocity field. Both measures are quasi-objective: they approximate objective (i.e., observer-independent) coherence diagnostics in frames satisfying a certain condition. We illustrate with several examples how our quasi-objective coherence diagnostics highlight elliptic and hyperbolic LCS, even from very sparse unstructured trajectory data. This approach shows great potential for expanding the possibilities of LCS applications through its simplicity, performance with sparse data, and enhanced computational efficiency.

How to cite: Aksamit, N., Bartos, A. E., and Haller, G.: Quasi-Objective Coherent Structures from Single Lagrangian Trajectories, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-3243,, 2021.

Thomas Meunier, Paula Pérez Brunius, Javier Rodríguez Outerelo, Heather Furey, Amy Bower, Andrée Ramsey, Paula García Carrillo, and Argelia Ronquillo

The Deep Water Horizon oil spill has dramatically impacted the Gulf of Mexico from the seafloor to the surface. While dispersion of contaminants at the surface has been extensively studied, little is known about deep water dispersion properties. This study describes the results of the Deep Water Dispersion Experiment (DWDE), which consisted in the release of surface drifters and RAFOS floats drifting at 300 and 1500 dbar in the Gulf of Mexico. We show that surface diffusivity is elevated, and decreases with depth. The separation dependence of relative diffusivity follows a Richardson law at all depths. Time dependence of dispersion suggests a Richardson regime near the surface and a mixed Richardson/ballistic regime in depth at scales of [10-100 km]. Finite Scale Lyapunov Exponents and pair separation Kurtosis suggest the existence of a Lundgren regime at scales smaller than the Rossby radius near the surface, and at smaller scales in depth.

How to cite: Meunier, T., Pérez Brunius, P., Rodríguez Outerelo, J., Furey, H., Bower, A., Ramsey, A., García Carrillo, P., and Ronquillo, A.: A Deep Water Dispersion Experiment in the Gulf of Mexico, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-5777,, 2021.

Gary Froyland, Ryan Abernathey, Michael Denes, and Shane Keating

Transport and mixing properties of the ocean's circulation is crucial to dynamical analyses, and often have to be carried out with limited observed information. Finite-time coherent sets are regions of the ocean that minimally mix (in the presence of small diffusion) with the rest of the ocean domain over the finite period of time considered. In the purely advective setting (in the zero diffusion limit) this is equivalent to identifying regions whose boundary interfaces remain small throughout their finite-time evolution. Finite-time coherent sets thus provide a skeleton of distinct regions around which more turbulent flow occurs. Well known manifestations of finite-time coherent sets in geophysical systems include rotational objects like ocean eddies, ocean gyres, and atmospheric vortices. In real-world settings, often observational data is scattered and sparse, which makes the difficult problem of coherent set identification and tracking challenging. I will describe mesh-based numerical methods [3] to efficiently approximate the recently defined dynamic Laplace operator [1,2], and rapidly and reliably extract finite-time coherent sets from models or scattered, possibly sparse, and possibly incomplete observed data. From these results we can infer new chemical and physical ocean connectivities at global and intra-basin scales (at the surface and at depth), track series of eddies, and determine new oceanic barriers.

[1] G. Froyland. Dynamic isoperimetry and the geometry of Lagrangian coherent structures. Nonlinearity, 28:3587-3622, 2015

[2] G. Froyland and E. Kwok. A dynamic Laplacian for identifying Lagrangian coherent structures on weighted Riemannian manifolds. Journal of Nonlinear Science, 30:1889–1971, 2020.

[3] Gary Froyland and Oliver Junge. Robust FEM-based extraction of finite-time coherent sets using scattered, sparse, and incomplete trajectories. SIAM J. Applied Dynamical Systems, 17:1891–1924, 2018.

How to cite: Froyland, G., Abernathey, R., Denes, M., and Keating, S.: Surface and deep ocean connectivity inferred from robust extraction of coherent sets in ocean flow using models and sparse, scattered, and incomplete float data with transfer operator and dynamic Laplacian methods., EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-6710,, 2021.

Enrico Ser-Giacomi, Alberto Baudena, Vincent Rossi, Mick Follows, Ruggero Vasile, Cristobal Lopez, and Emilio Hernandez-Garcia


The study of connectivity patterns in networks has brought novel insights across diverse fields ranging from neurosciences to epidemic spreading or climate. In this context, betweenness centrality has demonstrated to be a very effective measure to identify nodes that act as focus of congestion, or bottlenecks, in the network. However, there is not a way to define betweenness outside the network framework. Here we introduce the “Lagrangian betweenness”, an analogous quantity which relies only on the information provided by trajectories sampled across a generic dynamical system in the form of Finite Time Lyapunov Exponents, a widely used metric in Dynamical Systems Theory and Lagrangian oceanography. Our theoretical framework reveals a link between regions of high betweenness and the hyperbolic behavior of trajectories in the system. For example, it identifies bottlenecks in fluid flows where particles are first brought together and then widely dispersed. This has many potential applications including marine ecology and pollutant dispersal. We first test our definition of betweenness in an idealized double-gyre flow system. We then apply it in the characterization of transport by real geophysical flows in the semi-enclosed Adriatic Sea and the Kerguelen region of the highly turbulent Antarctic Circumpolar Current. In both cases, patterns of Lagrangian betweenness identify hidden bottlenecks of tracer transport that are surprisingly persistent across different spatio-temporal scales. In the marine context, high Lagrangian betweenness regions represent the optimal compromise between the heterogeneity of water origins and destinations, suggesting that they may be associated with relevant diversity reservoirs and hot-spots in marine ecosystems. Our new metric could also provide a novel approach useful for the management of environmental resources, informing strategies for marine spatial planning, and for designing observational networks to control pollutants or early-warning signals of climatic risks. 

How to cite: Ser-Giacomi, E., Baudena, A., Rossi, V., Follows, M., Vasile, R., Lopez, C., and Hernandez-Garcia, E.: Lagrangian betweenness: detecting fluid transport bottlenecks in oceanic flows, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-7894,, 2021.

Emilio Hernández-García, Alberto Baudena, Enrico Ser-Giacomi, Cristobal Lopez, and Francesco d'Ovidio

In the context of tracer transport in the ocean, we introduce a quantity, the crossroadness [1], which allows identifying the optimal disposition of a set of locations in order to monitor a given ocean surface region. The optimization is performed so that these sites observe the largest amount of water coming from the region and, at the same time, monitor waters coming from separate parts of the ocean. These are key criteria when deploying a marine observing network. Considering surface circulation, crossroadness measures at any location the extent of the ocean surface which transits in its neighborhood in a given time window. When the analysis is performed backward in time, this method allows us to identify the major sources which feed a target region. The method is first applied to a minimalistic model of a mesoscale eddy field, and then to realistic satellite-derived ocean currents in the Kerguelen area. In this region, we identify the optimal location of fixed stations capable of intercepting the trajectories of 43 surface drifters. We then illustrate the temporal persistence of the stations determined in this way. Finally, we identify possible hotspots of micro-nutrient enrichment for the recurrent spring phytoplanktonic bloom occurring there. Promising applications to other fields, such as larval connectivity or contaminant detection are discussed.

[1] A. Baudena, E. Ser-Giacomi, C. López, E. Hernández-García, F. d’Ovidio, Crossroads of the mesoscale circulation, Journal of Marine Systems 192, 1-14 (2019).

How to cite: Hernández-García, E., Baudena, A., Ser-Giacomi, E., Lopez, C., and d'Ovidio, F.: Optimal monitoring of the ocean surface by observing the transport crossroads, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-8732,