Cloud Forecasting

Discussion of UV Index hinted that accurate cloud forecasts are useful, and last week clouds were discussed. This week, cloud forecasts are discussed.

Very similar categories are used for standard meteorological observations and plotted surface weather charts, except for NOC, which I include. If a person is observant, (s)he can often notice that cloudy skies are not always overcast, very small breaks often existing. A person should be aware that opaque sky coverage is referred to, though categories can be defined to include semi-transparent clouds. Because thin clouds such as cirrostratus can often transmit as much as 80 % of incident solar energy, such a distinction should be made. When observing cloudiness, I use 20^ths of sky cover, though such determination is difficult and sky conditions often transient. METAR code is used for reporting weather observations at airports, though such reports are often translated to plain language for the public. Consider if you wish, METAR reports for DET (Detroit City Airport (MI)) for 10 MAR 1997. They indicate an area of alto types of clouds moving over the region, developing and advected by westerly winds aloft, as GOES 8 satellite images from NOAA/NESDIS & UCAR illustrate (1) (2) (3).

When forecasting cloudiness, a person should consider :

Detail of cloud forecasts is often determined by how far in the future such are valid for. Those for several days from present must (obviously) often be general, while those near present can be quite specific. To predict cloudiness, predicting formation of weather systems (e.g., surface Highs and Lows, upper air trofs) is very helpful. Specific cloud types and characteristics tend to be related to these, because of forces on air in their vicinity.

Satellite imagery is helpful for identification of clouds and their characteristics, and most useful for forecasting near-future cloudiness. You may notice some resemblance between clouds associated with a real cyclone and those indicated in the previous drawing. Images are obtained from 2 basic satellite types - geostationary and polar-orbiting. 3 types of imagery are mainly used to monitor cloudiness - visible (VIS), infrared (IR), and water vapor (WV). Each is helpful for specific purposes, but VIS images tend to most clearly depict cloudiness. You may notice differences regarding how cloudiness is depicted on each image, WV images tend to indicate moisture (or lack of it) in the middle to upper troposphere (thus, subsiding dry air intrusions to cyclone areas, inhibiting cloudiness). High resolution is very helpful for images, some meteorological satellites' being 1 km, but most available ones 4 or 8 km. Sounding data is very useful, but balloon soundings quickly become old, as our atmosphere is constantly changing. Satellite soundings can sometimes be used though, providing current atmospheric profiles.

When forecasting cloudiness, a person should be aware of how the diurnal (day/night) cycle influences it. Such is much more evident during warm weather, cumulus clouds often growing quite large during afternoon & evening, dissipating during night. Many people are familiar with morning fog, especially after rain occurred the previous night. Other more subtle things occur, such as evening cooling of clouds causing them to subside and dissipate if the environmental lapse rate is favorable, perhaps causing some light snow showers to fall out of them during winter as snow crystals precipitate out; and mountain-valley flows causing local cloudiness (or lack of it). Cloudiness tends to be most abundant during late afternoon and early evening, and least so during late morning, solar energy often heating clouds, causing them to dissipate. Some local effects should be considered though; such as lake breezes, which can cause clear weather to prevail near shores during afternoon, and lake-effect cloudiness, which can be responsible for some quite hefty snows downwind of the Great Lakes during very cold weather.

For times longer than several hours in the future, computer model forecasts are more helpful to forecast cloudiness than imagery (ETA model panels & details). With an initial set of observations (mainly balloon soundings), forces which cause vertical air motions, which greatly influence cloudiness, can be calculated. Much of cloud forecasting is presently done with computer models, forecasters using such info as guidance, while using other info (e.g., satellite images, knowledge of local effects, and the diurnal cycle) to often improve such forecasts. Model output parameters & statistics are often used as further guide to forecasting cloudiness. Model forecasts shown are for Detroit, MI (DTW, not DET) for 10 MAR 1997. (I chose DET for the observations, being more reliable than those from DTW.)

Images for the eastern Great Lakes almost a year ago (20 MAR 1996) illustrates a storm system as depicted on a surface chart, satellite image, and computer model forecasts (ETA) (1) (2). A surface Low developed south of the Ohio Valley states, developing northward and even slightly westward. You may notice that the surface chart and computer model forecast do not correspond exactly - model forecasts are not always correct. The surface Low formed over Ohio much earlier than 7 UTC, but the ETA model forecast did not indicate it there until 12 UTC. Thus, radar imagery shows heavy precipitation significantly farther west over southern MI than the computer forecast indicated. 7-8 inches of snow occurred during the early morning of 20 MAR near Detroit, MI, after several hours of steady rain, with some 50 mph NNE winds gusts.

Such major events are often discussed, but some of the most challenging forecasts are for areas of clouds which are unnoticed by many people, but can be important for some activities & applications, such as the UV Index, solar car racing (which requires solar energy for power), viewing comet Hale-Bopp, etc. For such events, trying to be very specific regarding cloudiness is quite helpful. Consider if you wish, 10 MAR NGM model relative humidity GIFs for DET, comparing them with surface observations and satellite images previously shown. The model forecast indicates moisture moving over the region associated with a weak surface Low, but cloudiness was not exactly as portrayed. Thus, some knowledge of cloud behavior and interpretation is required to use computer model forecasts effectively.

I plan to discuss seasons next week, as Vernal Equinox approaches.

Text is copyright of Joseph Bartlo, though may be used with proper crediting.