Volcanic ash cloud forecasts for hypothetical eruptions
NOTE: this web server is not maintained in an "operational" environment and should not be relied upon for 24/7 access.
Volcanoes may be added to or removed from this list without notice, which may cause direct links to model output for a certain volcano to fail. Timely delivery of data and products from this server through the Internet is not guaranteed (see the disclaimer).
This web page is under contstruction and changes may be made at any time.
If one of these volcanoes has erupted, do not use the charts above; see the products from the appropriate VAAC or use the VAAC links below for more information.
These charts are for planning purposes only.
The table above contains volcanoes in and near the Washington and Anchorage VAACs' area of responsibility. The charts are updated four times a day whether the volcanoes have erupted or not. Because the three runs are created separately, if comparisons are made between them, please verify they have the same eruption date/time.
The time of the hypothetical eruption is twelve hours after the initialization time of the NCEP forecast. The forecast duration is 18-hours. The output is one-hourly averaged concentration every 6 hours. For Runs 1 and 3, the ash column top (12192 m = 40,000 ft) and eruption duration (1 hour) are arbitrary; for Run 2, the column top and duration are not arbitrary.
Run 1 is the traditional unit mass run, with model output concentration given relative to the unit mass source. Run 2 produces the dispersion output using the range of eruption source parameters (ESP) defined by the USGS (Mastin et al., 2009). Run 3 uses a unit mass as does Run 1, however it is an ensemble of 27 members. Results from the Run 3, individual 27 members, are shown, as well as probabilistic ensemble output.
There is always uncertainty in forecasting volcanic ash dispersion in real-time. Basic model inputs such as ash column top height and eruption duration (start and stop time) may or may not be known well. The mass of ash in the eruption is unknown and difficult to obtain. The uncertainties of three-dimensional meteorological data (winds, etc.) depend on the meteorological observations used to initialize the model, and the numerical analysis and forecast. An individual model run does not portray any of the model input uncertainty; however, the output for Runs 2 and 3 in this table span some of the input uncertainty and may be useful in interpreting the output from Run 1.
More information on Runs 1, 2, and 3 are given below.
Run 1 is the traditional unit mass run, with model output concentration given relative to the unit mass in the source (e.g. 1 kg). The resulting concentrations are very small (e.g. 1.0E-17 kg / cubic meter) because of the use of the unit mass. With such an output, typically the forecaster bases the choice of which concentration contour delineates the "edge" of the ash on the current satellite analysis or past history of the particular volcano. The contours correspond to the large-medium-small-default ash reduction levels in the NOAA volcanic ash dispersion model output products (e.g. latest products). For example, for the default case all 4 areas outlined by the contours represent ash, but for a large ash reduction only the greatest (inner) concentration contour delineates the ash region. If the mass of the eruption of the ash particle sizes modeled is known, the forecast concentration could be computed by multiplying the given output concentration by the eruption mass.
Output from the deterministic run is also shown as particle plots, color-coded by height, including a vertical cross section of the plume as viewed perpendicular to the dashed line. These are instantaneous plots, compared to the 1-hour average of the concentration plots.
Run 2 is actually several separate model runs. The USGS Volcano Hazards Program has assigned one of 11 eruption types to the world's volcanoes based on recent history. The assigned types are the most likely anticipated future eruption. The USGS also identified eruption source parameters (ESP) for each of these eruption types. The parameters are eruption height, duration, mass flow rate, volume of erupted material and mass fraction of fine ash particles. When these ESP are used, the model output concentration is the actual forecast concentration since an ash mass is input, unlike the unit mass source runs. Excluding the one type for submarine volcanoes, the two main classes of eruption types are defined by magma type (M, mafic; or S, silicic). Then these are classified by general scale ("standard": M0/S0, which is the same as M2/S2; "small": M1/S1, "medium": M2/S2, and "large": M3/S3) or other possible geological effects (pyroclastic flow or "co-ignimbrite cloud" - S8, and "lava dome [collapse]" called "brief": S9). Due to the uncertainties in the ESP by the scale within each mafic of silicic categorization, the output from several eruption scales should be considered for planning purposes. For more information see the report: "Mastin, L.G., Guffanti, M., Ewert, J.E., and Spiegel, J., 2009, Preliminary spreadsheet of eruption source parameters for volcanoes of the world: U.S. Geological Survey Open-File Report 2009-1133, v. 1.2, 25 p."
One way to use this is to look at the ESP forecast hour plot at a given time. If there is an eruption, starting at the given time, and it is has the characteristics of ESP-small, the plume will be as shown for S1/M1. If it has the characteristics of the ESP-large eruption, the plume will be as shown for S3/M3. Then see the time series output using the appropriate ESP-scale link. The ESP classification for the particular run is given in the title at the top of the charts and again at the bottom. Also at the top is given the assigned value. Note that the "brief" category does not pertain to some volcanoes, and the values of the concentration contour lines and the map domains may vary among the various eruption scales.
Run 3 uses a unit mass as does Run 1. Use of This ensemble is one way to account for some uncertainty in the meteorological model output that is input to the dispersion model. The meteorological ensemble is described here (Draxler, R.R. 2003, Evaluation of an ensemble dispersion calculation, Journal of Applied Meteorology, Vol. 42, February, 308-317). In brief each member of the ensemble is computed by assuming a plus or minus 1-gridpoint shift in the horizontal, and a plus or minus 1500 m shift in the vertical, of the particle position, with respect to the meteorological grid. For the volcanic ash application the vertical shift is much larger than that in the Draxler (2003) because of the generally smaller variability in winds at the higher altitudes than the lower altitudes. The ensemble contains 27 members, including the baseline "control" member with no shift in the particle position. Post-processing sorts the 27 concentrations in ascending order at each concentration grid point. The five types of ensemble output given in the table above are:
This ensemble output can be used in several ways. For example, the animation of the number of members with concentration greater than zero shows the degree of variability among the 27 members. The yellow area is the region where at least 15 members have the ash cloud. The larger this area, the less the uncertainty in the meteorological forecast, and vice versa.
Inquiries should be addressed to: