Bruintjes (1999) reviews cloud seeding experiments to enhance precipitation, including hygroscopic seeding. The following are excerpts from this article.

In the past ten years, a new approach to hygroscopic seeding has been explored in summertime convective clouds in South Africa as part of the National Precipitation Research Programme (Mather et al., 1997).

This approach involves seeding summertime convective clouds below cloud base with pyrotechnic flares that produce small salt particles on the order of 0.5 micron diameter in an attempt to broaden the initial cloud droplet spectrum and accelerate the coalescence process. The burning flares provide larger CCN (>0.3 micron diameter) to the growing cloud, influencing the initial condensation process and allowing fewer CCN to activate to cloud droplets.

The larger artificial CCN inhibit the smaller natural CCN from nucleating, resulting in a broader droplet spectrum at cloud base. The fewer cloud droplets grow to larger sizes and are often able to start growing by collision and coalescence with other cloud droplets within 15 minutes (Cooper et al., 1997), initiating the rain process earlier within a typical cumulus cloud lifetime of 30 minutes.

The development of this seeding approach was triggered by radar and microphysical observations of a convective storm growing in the vicinity of a large paper mill, which indicated an apparent enhancement of coalescence in these clouds as opposed to other clouds far away from the paper mill (Mather, 1991). Earlier observations by Hindman et al. (1977) also suggested a similar connection between paper mills and enhanced precipitation.

Mather et al. (1997) reported the results from a randomized cloud seeding experiment that was conducted from 1991 to 1996 in summertime convective clouds in the Highveld region of South Africa. The results of this experiment indicated that precipitation from seeded storms was significantly larger than from control (unseeded) storms (Figure 11 in Mather et al., 1997). The results were statistically significant at the 95% confidence level. Exploratory analyses indicated that seeded storms rained harder and longer than unseeded storms (Mather et al., 1997). Mather et al. (1997) also provided supporting microphysical evidence that supported the physical hypothesis. It is remarkable that statistical significance was reached on such a small sample set of 127 storms (62 seeded and 65 controls). The seeding signal was strong and readily detected, making the statistical tests very reasonable. Orville (1995) described the results from this experiment as perhaps the most significant scientific advancement in the past ten years in weather modification.

The calculations of Reisin et al. (1996) and Cooper et al. (1997) support the hypothesis that the formation of precipitation via coalescence might be accelerated by the salt particles produced by hygroscopic flares. These studies also found that for clouds with a maritime cloud droplet spectra hygroscopic seeding with the flares will have no effect, since coalescence is already very efficient in such clouds. However, the results from the calculations should be interpreted with caution since they oversimplify the real process of precipitation formation. Cooper et al. (1997) identified some of the shortcomings in the calculations related to mechanisms that broaden cloud droplet size distributions, sedimentation, and the possible effects on ice phase processes.

Bigg (1997) performed an independent analysis of the South African experiments and also found that the seeded storms lasted longer than the unseeded storms. Bigg (1997) also outlined some possible dynamic responses, which also were identified by Mather et al. (1997). Bigg suggested that the initiation of precipitation started at a lower height in the seeded clouds than in the unseeded clouds and that a more concentrated downdraft resulted closer to the updraft. The surface gust front generated by the seeded storm was thereby intensified and its interaction with the storm inflow enhanced convection.

The promising new results of the South African experiment, as well as the model calculations, led to the start of a new program in Mexico in 1996 using the South African hygroscopic flares in a similar fashion as in the South African program. The Mexican program, under leadership of NCAR, was conducted from 1996 to 1998 and included physical measurements and a randomized seeding experiment. Bruintjes et al. (2001) provides an overview of the experiment and preliminary results.

Previous excerpts from Bruintjes, R.T., 1999: A Review of cloud seeding experiments to enhance precipitation and some new prospects. Bull. Amer. Met. Soc., 80, 805-820.

The World Meteorological Organization subsequently sponsored a workshop of leading scientists in the world to evaluate this new technique and the experiments that were conducted in South Africa, Thailand and Mexico (Foote and Bruintjes, 2000). Two major points from this report can be summarized as follows:

• The recent experiments, if accepted, lead beyond the classical result in cloud physics linking cloud condensation nuclei and droplet spectra at cloud base to the efficiency of rain (for example, the probability that a cloud of a given depth will produce rain). Rather, these experiments suggest that CCN affect the total rainfall from a cloud, and apparently also the longevity of the cloud. This would have important practical implications not only for water resource issues, but also for such things as quantitative precipitation forecasting and global change (for example, a regional change in CCN might easily accompany a mean change in temperature).
• The final and perhaps the strongest conclusion of the workshop was that the experimental results were sufficiently exciting, and the topic sufficiently important, that a new international initiative should be launched to understand the physical processes taking place. It was recommended that a major cooperative field experiment employing modern instrumentation be planned and carried out in the near future.

Based on these exciting results, it was decided to evaluate this technique in the UAE. In addition, the WMO workshop also identified several areas where knowledge was lacking and where more research was needed. These areas included a better understanding of the natural and modified precipitation formation processes to support the statistical results, better characterization of the particles produced by the flares, and the necessity to extend the single-cloud, radar-evaluated results to area rainfall at the surface with associated hydrological impacts.

Read Executive Summary of the WMO Mazatlan Workshop Report

Request a copy of this report from NCAR/RAP.

Flare Characterization and Advantages

There are significant operational advantages to this form of hygroscopic seeding. The amount of salt required is much less, the salt particles are readily produced by flares, and the target area for seeding is an identifiable region at cloud base (updraft region) where the initial droplet spectrum is determined (Cooper et al., 1997).

Read about the NCAR/RAP Flare Facility

References