The New Volcanic Ash Satellite Retrieval VACOS Using MSG/SEVIRI and Artificial Neural Networks: 2. Validation

Volcanic ash clouds can damage aircrafts during flight and, thus, have the potential to disrupt air traffic on a large scale, making their detection and monitoring necessary. The new retrieval algorithm VACOS (Volcanic Ash Cloud properties Obtained from SEVIRI) using the geostationary instrument MSG/SEVIRI and artificial neural networks is introduced in a companion paper. It performs pixelwise classifications and retrieves (indirectly) the mass column concentration, the cloud top height and the effective particle radius. VACOS is comprehensively validated using simulated test data, CALIOP retrievals, lidar and in situ data from aircraft campaigns of the DLR and the FAAM, as well as volcanic ash transport and dispersion multi model multi source term ensemble predictions. Specifically, emissions of the eruptions of Eyjafjallajökull (2010) and Puyehue-Cordón Caulle (2011) are considered. For ash loads larger than 0.2g/□m and a mass column concentration-based detection procedure, the different evaluations give probabilities of detection between 70% and more than 90% at false alarm rates of the order of 0.3–3%. For the simulated test data, the retrieval of the mass load has a mean absolute percentage error of ~40% or less for ash layers with an optical thickness at 10.8m of 0.1 (i.e., a mass load of about 0.3– 0.7g/□m, depending on the ash type) or more, the ash cloud top height has an error of up to 10% for ash layers above 5km, and the effective radius has an error of up to 35% for radii of 0.6– 6m. The retrieval error increases with decreasing ash cloud thickness and top height. VACOS is applicable even for overlaying meteorological clouds, for example, the mean absolute percentage error of the optical depth at 10.8m increases by only up to ~30%. Viewing zenith angles > 60∘ increase the mean percentage error by up to ~20%. Desert surfaces are another source of error. Varying geometrical ash layer thicknesses and the occurrence of multiple layers can introduce an additional error of about 30% for the mass load and 5% for the cloud top height. For the CALIOP data, comparisons with its predecessor VADUGS (operationally used by the DWD) show that VACOS is more robust, with retrieval errors of mass load and ash cloud top height reduced by >10% and >50%, respectively. Using the model data indicates an increase in detection rate in the order of 30% and more. The reliability under a wide spectrum of atmospheric conditions and volcanic ash types make VACOS a suitable tool for scientific studies and air traffic applications related to volcanic ash clouds.

Online treatment of eruption dynamics improves the volcanic ash and SO₂ dispersion forecast: case of the Raikoke 2019 eruption

In June 2019, the Raikoke volcano, Kuril Islands, emitted 0.4–1.8 × 109 kg of very fine ash and 1–2 × 109 kg of SO2 up to 14 km into the atmosphere. The eruption was characterized by several phases or puffs of different duration and eruption heights. Resolving such complex eruption dynamics is required for precise volcanic plume dispersion forecasts. To address this issue, we coupled the atmospheric model system ICON-ART (ICOsahedral Nonhydrostatic – Aerosols and Reactive Trace gases) with the 1-D plume model FPlume to calculate the eruption source parameters (ESPs) online. The main inputs are the plume heights for the different eruption phases that are geometrically derived from satellite data. An empirical relationship is used to derive the amount of very fine ash (particles < 32 µm), which is relevant for long range transport in the atmosphere. On the first day after the onset of the eruption, the modeled ash loading agrees very well with the ash loading estimated from AHI (Advanced Himawari Imager) observations due to the resolution of the eruption phases and the online treatment of the ESPs. In later hours, aerosol dynamical processes (nucleation, condensation, coagulation) explain the loss of ash in the atmosphere in agreement with the observations. However, a direct comparison is partly hampered by water and ice clouds overlapping the ash cloud in the observations. We compared 6-hourly means of model and AHI data with respect to the structure, amplitude, and location (SAL-method) to further validate the simulated dispersion of SO2 and ash. In the beginning, the structure and amplitude values differed largely because the dense ash cloud leads to an underestimation of the SO2 amount in the satellite data. On the second and third day, the SAL values are close to zero for all parameters indicating a very good agreement of model and observations. Furthermore, we found a separation of the ash and SO2 plume after one day due to particle sedimentation, chemistry, and aerosol-radiation interaction. The results confirm that coupling the atmospheric model system and plume model enables detailed treatment of the plume dynamics (phases and ESPs) and leads to significant improvement of the ash and SO2 dispersion forecast. This approach can benefit the operational forecast of ash and SO2 especially in case of complex and non-continuous volcanic eruptions like the Raikoke 2019.

Modelling the size distribution of aggregated volcanic ash and implications for operational atmospheric dispersion modelling

We have developed an aggregation scheme for use with the Lagrangian atmospheric transport and dispersion model NAME, which is used by the London Volcanic Ash Advisory Centre (VAAC) to provide advice and guidance on the location of volcanic ash clouds to the aviation industry. The aggregation scheme uses the fixed pivot technique to solve the Smoluchowski coagulation equations to simulate aggregation processes in an eruption column. This represents the first attempt at modelling explicitly the change in the grain size distribution (GSD) of the ash due to aggregation in a model which is used for operational response. To understand the sensitivity of the output aggregated grain size distribution (AGSD) to the model parameters we conducted a simple parametric study and scaling analysis. We find that the modelled AGSD is sensitive to the density distribution and grain size distribution assigned to the non-aggregated ash at the source. Our ability to accurately forecast the long-range transport of volcanic ash clouds is, therefore, still limited by real-time information on the physical characteristics of the ash. We assess the impact of using the AGSD on model simulations of the Eyjafjallajökull 2010 ash cloud, and consider the implications for operational forecasting. Using the time-evolving AGSD at the top of the eruption column to initialise dispersion model simulations had little impact on the modelled extent and mass loadings in the distal ash cloud. Our aggregation scheme does not account for the density of the aggregates; however, if we assume that the aggregates have the same density of single grains of equivalent size the modelled extent of the Eyjafjallajökull ash cloud with high concentrations of ash, significant for aviation, is reduced by ~3 %. If we assume that the aggregates have a lower density (500 kg m−3) than the single grains of which they are composed and make-up 75 % of the mass in the ash cloud the extent is 1.2 times larger.

Simulation of ash clouds after a Laacher See-type eruption

Dated to approximately 13 000 years ago, the Laacher See (East Eifel volcanic zone) eruption was one of the largest midlatitude Northern Hemisphere volcanic events of the Late Pleistocene. This eruptive event not only impacted local environments and human communities but probably also affected Northern Hemispheric climate. To better understand the impact of a Laacher See-type eruption on NH circulation and climate, we have simulated the evolution of its fine ash and sulfur cloud with an interactive stratospheric aerosol model. Our experiments are based around a central estimate for the Laacher See aerosol cloud of 15 Tg of sulfur dioxide (SO2) and 150 Tg of fine ash, across the main eruptive phases in May and a smaller one in June with 5 Tg SO2 and 50 Tg of fine ash. Additional sensitivity experiments reflect the estimated range of uncertainty of the injection rate and altitude and assess how the solar-absorptive heating from the fine ash emitted in the first eruptive phase changed the volcanic clouds' dispersion. The chosen eruption dates were determined by the stratospheric wind fields to reflect the empirically observed ash lobes as derived from geological, paleoecological and archeological evidence linked directly to the prehistoric Laacher See eruption. Whilst our simulations are based on present-day conditions, and we do not seek to replicate the climate conditions that prevailed 13 000 years ago, we consider our experimental design to be a reasonable approximation of the transport pathways in the midlatitude stratosphere at this time of year. Our simulations suggest that the heating of the ash plays an important role for the transport of ash and sulfate. Depending on the altitude of the injection, the simulated volcanic cloud begins to rotate 1 to 3 d after the eruption. This mesocyclone, as well as the additional radiative heating of the fine ash, then changes the dispersion of the cloud itself to be more southward compared to dispersal estimated without fine ash heating. This ash-cloud-generated southerly migration process may at least partially explain why, as yet, no Laacher See tephra has been found in Greenland ice cores. Sulfate transport is similarly impacted by the heating of the ash, resulting in stronger transport to low latitudes, later arrival of the volcanic cloud in the Arctic regions and a longer lifetime compared to cases without injection of fine ash. Our study offers new insights into the dispersion of volcanic clouds in midlatitudes and addresses a likely behavior of the ash cloud of the Laacher See eruption that darkened European skies at the end of the Pleistocene. In turn, this study can also serve as significant input for scenarios that consider the risks associated with re-awakened volcanism in the Eifel.

Using multi-scale modeling and observations to link the eruption source parameters to the dispersion of volcanic clouds in case of the Raikoke eruption 2019

&lt;p&gt;&lt;span&gt;The Raikoke volcano emitted about &lt;/span&gt;&lt;span&gt;0.4-1.8 x 10&amp;#8313; kg &lt;/span&gt;&lt;span&gt;of ash and &lt;/span&gt;&lt;span&gt;1-2 x 10&amp;#8313; kg&lt;/span&gt;&lt;span&gt; of SO&lt;/span&gt;&lt;sub&gt;&lt;span&gt;2 &lt;/span&gt;&lt;/sub&gt;&lt;span&gt;up to 15 km into the atmosphere. However, the eruption was characterized by several puffs of different &lt;/span&gt;&lt;span&gt;time periods&lt;/span&gt;&lt;span&gt; and eruption heights. &lt;/span&gt;&lt;span&gt;Here, we&lt;/span&gt;&lt;span&gt; use the ICON-ART model in a model setup in which we resolve the phases of the Raikoke eruption. &lt;/span&gt;&lt;span&gt;W&lt;/span&gt;&lt;span&gt;e&lt;/span&gt;&lt;span&gt; calculated the eruption source parameters &lt;/span&gt;&lt;span&gt;(ESPs) online &lt;/span&gt;&lt;span&gt;by coupling &lt;/span&gt;&lt;span&gt;ICON-ART&lt;/span&gt;&lt;span&gt; to the 1-D plume model FPlume. &lt;/span&gt;&lt;span&gt;The input heights for the different eruption phases needed for FPlume are geometrically derived from GEOS-17 satellite data. An empirical relationship &lt;/span&gt;&lt;span&gt;is used to derive&lt;/span&gt;&lt;span&gt; the amount of very fine ash (particles &lt;32&amp;#181;m) which is relevant for long range transport in the atmosphere. &lt;/span&gt;In the first hours during and after the eruption, the modeled ash loading agrees very well with the observed ash loading from Himawari-8 due to the resolution of the eruption phase and the online calculation of the ESPs. In later hours, aerosol dynamical processes (nucleation, condensation, coagulation) explain the loss of ash in the atmosphere in agreement with the observations. However, a direct comparison is partly hampered by water and ice clouds overlapping the ash cloud in the observations. In case of SO&lt;sub&gt;2&lt;/sub&gt;, we compared 6-hourly means of model and Himawari data with respect to the structure, amplitude, and location (SAL-method). In the beginning, the structure and amplitude values differed largely because the dense ash cloud directly after the eruption leads to an underestimation of the SO&lt;sub&gt;2&lt;/sub&gt; amount in the satellite data. On the second and third day, the SAL values are close to zero for all parameters indicating a good agreement of model and observations. We argue that representing the plume phases and ESPs in ICON-ART by FPlume enhances ash and SO&lt;sub&gt;2&lt;/sub&gt; predictability in the first days after the eruption, especially in case of non-continuous volcanic eruptions like the Raikoke eruption 2019.&lt;/p&gt;

Exploring forecast variability during volcanic ash cloud events

&lt;p&gt;Atmospheric transport and dispersion models are used by Volcanic Ash Advisory Centers (VAACs) to provide timely information on volcanic ash clouds to mitigate the risk of aircraft encounters. Inaccuracies in dispersion model forecasts can occur due to the uncertainties associated with source terms, meteorological data and model parametrizations. Real-time validation of model forecasts against observations is therefore essential to ensure their reliability. Forecasts can also benefit from comparison to model output from other groups; through understanding how different modelling approaches, variations in model setups, model physics, and driving meteorological data, impact the predicted extent and concentration of ash. The Met Office, the National Centre for Atmospheric Science (NCAS) and the British Geological Survey (BGS) are working together to consider how we might compare data (both qualitatively and quantitatively) from the atmospheric dispersion models NAME, FALL3D and HYSPLIT, using meteorological data from the Met Office Unified Model and the NOAA Global Forecast System (providing an effective multi-model ensemble). Results from the model inter-comparison will be used to provide advice to the London VAAC to aid forecasting decisions in near real time during a volcanic ash cloud event. In order to facilitate this comparison, we developed a Python package (ash-model-plotting) to read outputs from the different models into a consistent structure. Here we present our framework for generating comparable plots across the different partners, with a focus on total column mass loading products. These are directly comparable to satellite data retrievals and therefore important for model validation. We also present outcomes from a recent modelling exercise and discuss next steps for further improving our forecast validation.&lt;/p&gt;

The Development of Volcanic Ash Cloud Layers over Hours to Days Due to Atmospheric Turbulence Layering

Volcanic ash clouds often become multilayered and thin with distance from the vent. We explore one mechanism for the development of this layered structure. We review data on the characteristics of turbulence layering in the free atmosphere, as well as examples of observations of layered clouds both near-vent and distally. We then explore dispersion models that explicitly use the observed layered structure of atmospheric turbulence. The results suggest that the alternation of turbulent and quiescent atmospheric layers provides one mechanism for the development of multilayered ash clouds by modulating vertical particle motion. The largest particles, generally μ>100 μm, are little affected by turbulence. For particles in which both settling and turbulent diffusion are important to vertical motion, mostly in the range of 10–100 μμm, the greater turbulence intensity and more rapid turbulent diffusion in some layers causes these particles to spend greater time in the more turbulent layers, leading to a layering of concentration. The results may have important implications for ash cloud forecasting and aviation safety.

FALL3D-8.0: a computational model for atmospheric transport and deposition of particles, aerosols and radionuclides – Part 2: Model validation

This paper presents model validation results for the latest version release of the FALL3D atmospheric transport model. The code has been redesigned from scratch to incorporate different categories of species and to overcome legacy issues that precluded its preparation towards extreme-scale computing. The model validation is based on the new FALL3D-8.0 test suite, which comprises a set of four real case studies that encapsulate the major features of the model; namely, the simulation of long-range fine volcanic ash dispersal, volcanic SO2 dispersal, tephra fallout deposits and the dispersal and deposition of radionuclides. The first two test suite cases (i.e. the June 2011 Puyehue-Cordón Caulle ash cloud and the June 2019 Raikoke SO2 cloud) are validated against geostationary satellite retrievals and demonstrate the new FALL3D data insertion scheme. The metrics used to validate the volcanic ash and SO2 simulations are the structure, amplitude and location (SAL) metric and the figure of merit in space (FMS). The other two test suite cases (i.e. the February 2013 Mt. Etna ash cloud and associated tephra fallout deposit, and the dispersal of radionuclides resulting from the 1986 Chernobyl nuclear accident) are validated with scattered ground-based observations of deposit load and local particle grain size distributions and with measurements from the Radioactivity Environmental Monitoring database. For validation of tephra deposit loads and radionuclides, we use two variants of the normalised root-mean-square error metric. We find that FALL3D-8.0 simulations initialised with data insertion consistently improve agreement with satellite retrievals at all lead times up to 48 h for both volcanic ash and SO2 simulations. In general, SAL scores lower than 1.5 and FMS scores greater than 0.40 indicate acceptable agreement with satellite retrievals of volcanic ash and SO2. In addition, we show very good agreement, across several orders of magnitude, between the model and observations for the 2013 Mt. Etna and 1986 Chernobyl case studies. Our results, along with the validation datasets provided in the publicly available test suite, form the basis for future improvements to FALL3D (version 8 or later) and also allow for model intercomparison studies.