4 Isallobaric Effects

4.1 Cross-Isobaric Flow


What is the importance of isallobaric winds in forecasting marine boundary layer winds and sea state?


Key Points

Wind Forecasting Implications

view section (click to show/hide)
Why Isallobaric Effects are More Important Over Water

Because there is generally less friction over water than land, the isallobaric component of the wind is often of greater significance in maritime regions. In addition, it is important to forecast isallobaric effects well in advance because the water responds rapidly to any change in wind direction or speed leading to a newly generated wind wave group and a confused sea state. Keep in mind that mariners are at the mercy of the sea state and frequently cannot quickly seek the "safe shelter" easily available on land.

Large-Scale Isallobaric Effects

You must account for isallobaric effects in order to correctly forecast the wind field in the marine boundary layer. Isallobaric wind accelerations are going to be most prominent on synoptic scales in association with of rapidly moving systems, and/or systems experiencing rapid changes in intensity. In these cases, the wind direction must be adjusted to reflect the accelerations from prominent pressure rise regions toward significant pressure fall regions, and the wind speed may need to be adjusted upward or downward.

Small-Scale Isallobaric Effects

Mesoscale (e.g., squall lines and associated wake lows and outflow boundaries) will also have an isallobaric component to the wind, but these are difficult to account for in observations or by NWP predictions if the phenomenon (a squall line, for instance) is not well predicted by the model, thus critical structure may be missing even in a model forecast with convection in the right place. In addition, these types of mesoscale events can have extreme pressure gradients over very short distances and time periods, and the parcel accelerations tend to be directly toward low pressure rather than along isobars. Formulas for isallobaric and geostrophic wind overestimate wind speeds due to short parcel time in the acceleration region—parcels never come into a large-scale balance, they quickly pass through. Convective features generally are not handled well by the NWP operational models and therefore are usually best diagnosed using satellite and radar data over maritime regions. However, NWP models can be useful as evidenced by the fact that in the past few years the GFS patterns of low-level convergence and precipitation over tropical oceans during the first 6 to 12 hours of the forecast have matched up well against satellite images.

Rapid Cyclogenesis

Be alert for situations where rapid intensification may be favored, such as when cyclonic vorticity advection and/or warm advection contribute to the development of a cyclone which then moves over a warm current (e.g. Gulf Stream ). Keep in mind that rapidly moving systems can magnify the isallobaric effects. Monitor pressure change fields, wind velocity with data from surface buoys, ship observations, or satellite-derived winds, to determine areas where the isallobaric component may be large. In data sparse maritime regions, NWP model isallobaric fields may need to be used, provided the model has a good handle on the situation.

Eta model 6-hour pressure change and MSLP 24 and 30 hour forecasts valid 1200 UTC and 1800 UTC March 8 2004
Eta model 6-hour pressure change (rust solid/dashed lines, hPa/6-hr) and MSLP (light blue solid lines, hPa) 24- and 30-hour forecasts valid 8 March at 200 UTC and 1800 UTC, respectively. Rapid cyclogenesis is occurring as well as movement of the cyclone to the northeast. The isallobaric wind component will be perpendicular to the lines of constant pressure change directed toward falling pressure and proportional to the allobaric (pressure tendency) gradient. A strong pressure-change couplet is a clear sign that the isallobaric wind component will affect the resulting total wind.
How to Make Adjustments to the Total Wind

Compare the observed and model-derived pressure-change fields (e.g., 3-,6-,12-hr pressure change). If they are not in agreement, mentally (or if possible, interactively) adjust the model wind field and associated isallobaric component to more closely depict the true wind field.

It may help to view a plot of the mean sea level pressure, surface winds, and 3-, 6-, or 12-hour pressure change. The isallobaric wind component is directed perpendicular to the isallobars, toward falling pressure (i.e., outward from areas of increasing pressure and into areas of falling pressure). One can make adjustments by overlaying the MSLP and isallobaric fields and assessing the influence of the isallobaric component on the total wind field as follows:

  • If the isallobaric and gradient components are nearly aligned, increase the wind speed above the gradient wind speed.
  • If the isallobaric and gradient components are nearly opposite, decrease the wind speed below the gradient wind speed.
  • If the two components are perpendicular, then adjust for a cross-isobar component due to isallobaric effects.
Eta model 6-hour pressure change, MSLP, wind speed, and wind barb 24 and 30 hour forecasts valid 1200 UTC and 1800 UTC March 8 2004
Eta model 6-hour pressure change (rust solid/dashed lines, hPa/6-hr), MSLP (light blue solid lines, hPa), wind speed (image, kt), and wind barb 24- and 30-hour forecasts valid 8 March at 1200 UTC and 1800 UTC, respectively. Notice that the isallobaric contribution (vector perpendicular to isallobars) to the total wind increases the wind speed where the vectors are more parallel to the gradient wind (parallel to isobars) and causes larger cross-isobar flow where they are more perpendicular.

Next, determine if the rate of change in the pressure field is well captured by the model. If the model is under forecasting the rate of cyclogenesis (or other pressure changes), the magnitude of the isallobaric component will be greater than predicted. Again, mentally (or interactively) adjust the speed of the isallobaric component and compare it with the gradient wind in order to see the effect of the two components toward the total wind.

Wave Forecasting Implications

view section (click to show/hide)
Confused Sea States and Implications to Mariners

Forecasts of isallobaric winds are important to marine users because of the confused seas that sometimes develop. In strong cyclongenesis cases, the primary synoptic wind field will have already developed a large energetic wind wave field. When a significant isallobaric component is superimposed, its relatively short duration and fetch create short, steep waves moving in a different direction from the primary wave field that had just been created. This can lead to a confused sea state where two relatively new wave fields exist in the same area, but have differing directions of movement. This situation is hazardous to small vessels because relatively steep waves are coming from multiple directions, making it difficult for the vessel to keep from being over turned by a large wave hitting the vessel from the side instead of the bow. Even large vessels may be impacted (e.g., the 1975 sinking of the Edmund Fitzgerald on Lake Superior) if the pressure change is rapid enough to generate a strong isallobaric wind component and large wave heights result that differ in direction from the wind waves just previously generated by the storm system.

Ship in seas with single wave group.
Examine the animation of a single wave group where synoptic scale winds have been generating the primary waves being felt by a ship. Note how the ship is being affected as it moves over the water and compare this to the "Confused Seas" animation below.

Ship in a confused seas.
Examine the animation of confused seas where an existing wave group from synoptic winds begins to be affected by a new wind direction influenced by isallobaric effects. The motion of the ship as it encounters these 2-wave groups at the same time shows a much rougher ride as the ship travels through the water.

Isallobaric Effect on Wave Height

Large isallobaric effects on a synoptic scale are more likely to develop significant wave heights than those seen during mesoscale events such as squall lines or outflow boundaries, which have relatively short duration and fetch lengths. Constructive interference of multiple wave groups leads to a significant wave height that is greater than the individual wave height of either group. Hence it may be necessary to upgrade advisories and/or warnings as required for any increase in wave height that results from confused seas.

Remember that in addition to wind speed, the wave height is dependent on the fetch length and duration of a particular wind direction where the short duration and fetch limit wave generation. Given a high storm velocity, a strong isallobaric component of the wind will result in short fetch lengths and durations.

Duration limitation to wave generation
Examine the animation of wave generation with a moving low pressure area. The gray ovals represent justtwo of the many fetch areas for this system. The animation shows how the waves generated can not reach their height potential as they are not able to stay within the fetch area due to the system movement. Imagine the even shorter duration for wave generation if this system was experiencing a large isallobaric wind component due to rapid cyclogenesis.


The exception to this is when the storm system is moving in a straight path and the parallel winds in the right quadrant of the storm are resonant with the wave generation taking place in that area. This is referred to as a dynamic or "trapped" fetch.

Conceptual animation showing wave growth within regions of dynamic fetch.
View an animation of how large waves can develop from dynamic fetch.


U.S. Coast Guard, 1977: Marine Board casualty report, SS Edmund Fitzgerald; Sinking in Lake Superior on 10 November 1975 with loss of life. U.S. Coast Guard Marine Board of Investigation Report and Commandant's Action, Rep. USCG 16732/64216, 121 pp.

Hultquist, T., 2006: Reexamination of the 9-10 November 1975 "Edmund Fitzgerald" Storm Using Today's Technology. Bull. Amer. Meteor. Soc., 87, 607-622.

NWP Considerations

view section (click to show/hide)

In general, the primitive, equation-based NWP models handle isallobaric wind effects fairly well. However, model forecasts may be adversely impacted by poor initialization that may lead to errors in forecast location and intensity. Inadequate horizontal and vertical resolution and parameterization schemes may limit NWP models in their ability to predict rapid cyclogenesis or anticyclogenesis or, more often, may result in predicting spurious or misplaced cyclogenesis events in situations that appear synoptically favorable. To assess model performance, verify model output with ship and buoy data or satellite-derived winds, when available. The best tool for determining specific isallobaric adjustments to marine wind forecasts is upstream surface reports and analysis of 2-3 hour pressure tendencies. Three-hour pressure trends are available on AWIPS through MSAS (MAPS). Two-hour pressure trends are available from the Storm Prediction Center (SPC) mesoanalysis website.

Rapid Cyclogenesis

Synoptically, isallobaric winds are modeled well except in the case of polar lows, which can be too small for models such as the GFS. However, there can be an output availability issue since the GFS uses 6-hour increments, which prevents the more desirable comparison of observations with 3-hr isallobars. Further, the isallobaric component locally dominant in lake/sea and land breeze situations is generally not well forecast by model surface wind forecast products. Land/sea breeze is more complex in nature because winds first flow toward lower pressure, then turn over time. So the ageostrophic flow rotates around an inertial circle (Durran 1993, Egger 1999, Lundquist 2003, Ostdiek et. al. 1997) until the pressure gradient changes due to changing thermal forcing (e.g. day to night, night to day).  

NWP model forecast errors on fast moving and rapidly intensifying systems have been documented and are the subject of other training. See "Cyclogenesis: Analysis utilizing Geostationary Satellite Imagery," a training session developed by the Virtual Institute for Satellite Integration Training (VISIT) ( http://rammb.cira.colostate.edu/visit/cyclo.html ). This training includes several examples of cyclogenesis over the ocean.

For the Great Lakes a good resource to start with is the reexamination of the 9-10 November 1975 event that sank the Edmund Fitzgerald (Hultquist 2006). Also see COMET's NWP module "Understanding Data Assimilation" regarding how the model assimilation during rapidly moving or intensifying cyclones can cause poor initialization, and hence a poor forecast.

Climatologically, more storms deepen rapidly off the U.S. East Coast than over land, thus climatologically, the errors associated with rapid deepening (e.g., pressure gradient, pressure change, isallobaric wind, pressure heights, etc.) are therefore bigger off the coast. There are several reasons for this. The coastal/offshore cyclogenesis is a result of both the climatological jet position and ocean/land winter baroclinic contrast. In addition, the added effect of the synoptic baroclinic zone combined with the west wall of the Gulf Stream amplifies the baroclinicity. The Gulf Stream by itself also contributes to rapid cyclogenesis by creating a deep layer of weak static stability, often manifest by convection, which allows dynamic coupling between potential vorticity at low levels associated with fronts and low-level diabatic heating and potential vorticity at high levels such as associated sharp gradients on the tropopause. As for NWP forecast errors, the cyclogenesis problem is amplified by having poor data coverage offshore. This is more of a problem when the cyclone's path has been over the ocean for a long time, such as with cyclones developing in the Gulf of Alaska. In this area data for model initialization does exist but there is little information about the combined vertical structure of the mass (i.e., pressure and virtual temperature structure) and wind field. In other words the vertical structure of remote sensing observations is limited. For example, cloud drift winds are only at a single cloud-top level and scatterometry derived winds are only avaliable at the surface. In addition vertical temperature and moisture structure can be derived from remotely sensed radiances, but only in clear skies. Hence this "mass" information might be available but the vertical wind field from cloud drift winds would be absent. For model initialization, both the mass and wind fields are needed for accurate predictions; having one or the other will simply not suffice. Off of the U.S. East Coast rapid cyclogenesis begins when a mid- to upper-level disturbance over the land passes over an offshore frontal zone. The air mass on the eastern side of the frontal zone has traveled through areas where rich vertical sounding information does not exist. This in turn leads to a poor model initialization of environmental conditions on this side of the front.

It's important to quantify the relative error in the model winds before deciding how to adjust the wave field. A speed error will require a different adjustment than a directional error or a combination of the two, which often occurs if the position of the storm is not well forecast. If there is not a position error, look for where the observed pressure gradient is larger than the model forecast and where the pressure change over a short period is largest. Larger pressure gradients will lead to increased wind speed and areas of rapidly changing pressure with time will often have a larger cross-isobar component. Strong cyclones tend to be closer to cyclostrophic than to geostrophic. For example, hurricanes are very nearly cyclostrophic. The Coriolis part of the gradient wind can be removed and the result is almost the same at those speeds and pressure gradients, as noted in much work by the Hurricane Research Division. Cyclostrophic balances the wind speed squared divided by the radius of curvature against the pressure gradient. So doubling the pressure gradient would increase wind by square root of 2, which is about 1.4 (i.e., a 40% increase).


Durran, D. R., 1993: Is the Coriolis force really responsible for the inertial oscillation? Bulletin of the American Meteorological Society, Boston , MA , 74 , 2179-2184.

Egger, J., 1999: Inertial oscillations revisted. Journal of the Atmospheric Sciences, Boston , MA, 56 , 2951-2954.

Hultquist, T.R., M. Dutter, D. Schwab: Reexamination of the 9-10 November 1975 " Edmund Fitzgerald " Storm Using Today's Technology Bulletin of the American Meteorological Society, Boston, MA, 87 , 607-622

Lundquist, J. K., 2003: Intermittent and Elliptical Inertial Oscillations in the Atmospheric Boundary Layer. J.Atmos.Sci., 60 , 2661-2673, doi:10.1175/1520-0469(2003)060(2661:IAEIOI)2.0.CO;2.

Ostdiek, V., W. Blumen , 1997: A dynamic trio: inertial oscillation, deformation frontogenesis, and the Ekman-Taylor boundary layer. Journal of the Atmospheric Sciences, Boston , MA , 54 , 1490-1502.