In February-March 1995, an aircraft campaign to explore the global distribution of sprites was based in Lima, Peru (Heavner et al.(1995),Sentman et al.(1995a)). Although Amazonia is one of the most active lightning regions of the world only ~20 sprites were observed during a two week period. Similar aircraft campaigns over the central United States observed hundreds of events. The observed low activity level is likely due in part to logistical and observational challenges of the particular observations: a border war between Peru and Ecuador was being fought during the campaign, and a system such as the U.S. National Lightning Detection network was not available to identify regions of strong positive lightning activity. The NLDN data greatly improves the identification of likely sprite producing regions of storms. Effects of the nearly horizontal magnetic field of the Earth near the equator is another possible factor for the apparent low frequency of sprite observations in low latitude regions. Theorists postulating a runaway electron beam associated with sprites and blue jets generally agree that a horizontally aligned Earth's magnetic field will deacrease the formation of the runaway electron beams (Gurevich et al.(1996),Lehtinen et al.(1997),Taranenko and Roussel-Dupré(1997)). The proposed QE heating models do not generally address any dependence on the orientation of the local Earth's magnetic field.
Research has also been underway in Australia during southern hemisphere summers since 1997 (Hardman et al.(1998)) and recently observations of sprites occurring above Japanese winter thunderstorms have been made (Fukunishi et al.(1999)). Intensified cameras aboard the space shuttle have observed at least one sprite above thunderstorms in Africa, as well as several above South America (Boeck et al.(1998)).
In order to approach the questions of global rates and distributions of sprites, elves, and blue jets, more observations are required. The identification of a global synoptic detection method of these events would enhance such measurements. Attempting to estimate the global occurrence of sprites is an elusive problem. The latest estimate of the global cloud to ground lightning flash rate of between 10 s-1 and 14 s-1 (Mackerras et al.(1998)). Positive flashes are less than 10% of the total lightning flashes, and not every positive flash produces an optically detectable sprite. Therefore a preliminary upper limit for the global occurrence of sprites is estimated at 1 per second. However, this is based on the assumption that the global sprite distribution is similar to global lightning distribution. While current observations are biased geographically, lightning and sprites do not appear to necessarily have the same spatial distribution.
Based on the estimates of one sprite per second globally and one GJ energy per sprite, the global energy depositionin the mesosphere is approximately 2 uW/m2. comparing this with Figure 1(c) of Mlynczak (2000), sprites may be .1% of the mesospheric energy budget. The more significant effects are likely to be on smaller spatial scales. Often an active thunderstorm produces several hundred sprites in a few hours, which leads to more significant local heating.
Blue jets are rarely recorded by ground based observers and even aircraft campaigns record many more sprites than blue jets/starters. However, the dearth of jet observations may be partially explained by Rayleigh scattering and observational difficulties described earlier. The long duration of blue jets and the observation of strong N2+(1NG) emissions indicate that blue jets may be an important energetic process in the stratosphere. Elves are associated with both positive and negative cloud-to-ground lightning, so the global elves occurrence rate is probably higher than the global sprite rate.