Surface bursts

Figure 4.17. Side-on blast peak overpressure due to (a) a TNT surface burst. (Kingery and Panill 1964) and (b) a free-air burst of TNT (Glasstone and Dolan 1977).

TNT-blast data for hemispherical surface bursts are used to determine the blast effects due to the equivalent charge. These blast data are based on the Army, Navy, and Air Force Manual (1990). 

Surface Bursts of Gas-Filled, Massless, Spherical Pressure Vessels... 

This energy measure is equal to Brode s definition of the energy, multiplied by a factor 2. The reason for the multiplication is that the Brode definition applies to free-air burst, while Eq. (6.3.15) is for a surface burst. In a fiee-air burst, explosion energy is spread over twice the volume of air. 

Blast parameters for surface bursts of gas-filled pressure vessels have not been investigated thoroughly. Parameters presently used are derived from investigations of free-air bursts. 

W,. = effective charge weight in pounds of TNT for estimating surface burst effects with free air. 

W = required vapor capacity in pounds per hour, or any flow rate in pounds per hour, vapor relief rate to flare stack, Ibs/hr W(. = charge weight of explosive, lb Wj. = effective charge weight, pounds of TNT for estimating surface burst effects in free air W, = required steam capacity flow or rate in pounds per hour, or other flow rate, Ib/hr Whe = hydrocarbon to be flared, Ibs/hr Wtnt equivalent charge weight of TNT, lb Wl = liquid flow rate, gal per min (gpm)... 

For Class D hazards, the company has defined the evaluation case event to be 8 x 106 Btu (8.4 x 106 kj) energy release as a hemispherical ground level explosion, unless a comprehensive analysis defines a lesser event as the evaluation case. For a VCE evaluation, this is further defined as the release and vaporization of 10,000 lb. (4,500 kg) of ordinary hydrocarbons, or the release and vaporization of 6,600 lb. (3,000 kg) of fast-burning materials [fundamental burning velocity >24 in/sec (>60 cm/sec)] within 5 minutes, when the largest connection to a tank or vessel is broken. [On a TNT basis, this is equivalent to a surface burst of approximately 2 tons (1,800 kg) TNT, calculated on the basis of 4% efficiency for ordinary hydrocarbons or 6.6% efficiency for fast-burning materials.]... 

The sources of error indicated above were avoided in a series of experiments carried out by Donnan and Barker, which in principle resemble those made by Lewis, so that only a brief reference to them is necessary. The dissolved substance was nonylic acid, and a drop method. The results could be reproduced with very great accuracy, i.e., to a fraction of one drop in 300—500 drops. Adsorption was produced at a surface air-liquid, air being passed through the solution in bubbles of known size and number, so that the total active surface could be calculated. The bubbles, on reaching the surface, burst, hence the excess of solute carried by them remained in the surface very effective precautions were used to prevent diffusion backwards from this portion into... 

Some sources such as the tri-service manual (Ref. 30) include sets of blast parameter curves for spherical free-air explosions and separate sets of curves for hemispherical surface burst explosions. This is superfluous except at very small scaled distances, because the free-air curves can be used for both situations by simply using a higher effective charge weight for surface bursts. 

Account for a surface or near-surface burst by first calculating a new effective free-air charge weight, We, as... 

Kingery, C. N., and Bulmash, G., "Airblast Parameters from TNT Spherical Air Burst and Hemispherical Surface Burst," Ballistic Research Laboratory Technical Report ARBRL-TR-02555, Aberdeen Proving Ground, MD., 1984. 

The loads from external near-surface burst explosions are based on hemispherical surface burst relationships. Peak pressure (P psi) and scaled. impulse Ci/W psi/lb ) are plotted vs. scaled distance (R/W ft/lb ). Roof and sidewall elements, side-on to the shock wave, see side-on loads (P and i ). The front wall, perpendicular to the shock wave, sees the much higher reflected shock wave loads (P and i ). An approximate triangular pressure-time relationship is shown in Figure 5a. The duration, T, is determined from the peak pressure and impulse by assuming a triangular load. Complete load calculations include dynamic loads on side-on elements, the effect of clearing times on reflected pressure durations, and load variations on structural elements due to their size and varying distance from the explosive source. 

FIG. 23-45 Incident overpressure vs. scaled distance, surface burst. (The + points are from Kingery and PanniU, Menu Report 1518 BRL. Adapted from Department of Army, Navy, and Air Force TM5-1300, NAVFAC P-397, AFM 88-22.)... 

As a description of the calculational method, a 100-KT near-surface burst over CaSi03-CaAl2Si208-Si02 eutectic was simulated. The particle size distribution is given in Table II. 

Detonations which produce particle populations of the first category are land surface bursts, land subsurface bursts, vented underground bursts, and tower bursts. 

Land Surface Burst. The particle population clearly consists of two distinct components—crystalline particles and glass particles. The crystalline particles are local soil material which entered the fireball at a late time and hence were not melted. 

Land Subsurface Burst. Everything which was said above about land surface burst applies exactly to the aerial cloud particle population produced by a land subsurface burst in which an aboveground fireball appears. However, a third component of the particle population is found. This component appears to result from soil material which interacted with the fireball at high temperature but which was separated from the fireball early, before the temperature had fallen below the melting point of the soil materials. The particles in this component have diameters ranging from tens of microns to several centimeters and have densities which are apt to be quite low compared with those of the original soil components. The relative abundance of radionuclides in this component is quite constant from sample to sample and is independent of particle size. If we indicate by subscript 1 this third component and by 2,3 the aerial cloud components, radionuclide partitioning can be described by a series of equations of the forms... 

Tower Burst. If the energy of the detonation is sufficient to vaporize the entire tower mass, the particle population is like that described for the land surface burst. If, however, the entire tower is not vaporized, the particle population will consist of three identifiable components— the crystalline and glass components of the surface detonation plus a metal sphere population which arises from melted (not vaporized) tower materials resolidifying as spheres. Such spheres are metallic rather than metal oxide and exhibit the density and magnetic properties of the tower material. The size range of the spherical component is from a few microns to perhaps a few hundred microns diameter. If we indicate by... 

Detonations of the second category—those which produce particles by condensation from the vapor state—include airbursts and water surface bursts. 

Water Surface Burst. When the entire platform of a water surface detonation is vaporized, the primary particle population is exactly like that described under airburst. However, the particles of the primary population act as condensation nucleii for the late-time condensation of sea salts. The salt particles are hygroscopic and eventually dissolve and leave the primary population behind. However, particle transport is affected by the sea salt particle growth which temporarily, at least, produces larger particles than does an airburst. 

The following discussion gives an example of the data treatment required to characterize the population for three cases—a land surface burst, a land subsurface detonation, and an airburst detonation. These three examples cover the complete range of types of solutions to the characterization problem. 

Table III. Parameters of Distribution Functions (Land Surface Burst)...

As in the case of the land surface burst, complete characterization of the particle population requires only that particle mass, a volatile species, and a refractory species distribution with particle size be determined. All other isotopic distributions may be deduced from the istotope partition calculations described above. In the subsurface detonation, the earliest aerial cloud sample was obtained in the cloud 15 minutes after detonation. The early sample was, therefore, completely representative of the aerial cloud particle population. In Figure 5 the results of the size analysis on a weight basis are shown. Included for comparison is a size distribution for the early, local fallout material. The local fallout population and the aerial cloud population are separated completely from the time of their formation. 

In discussing the fission-product composition of fallout samples it is advantageous to choose some fission product as a reference nuclide, j, and express the composition of the other fission products i by a set of fij ratios. For local fallout from land-surface bursts the choice of 95Zr as reference nuclide has proved convenient. A ratio of particular interest is 7 89,95 since 89Sr and 95Zr generally fractionate from each other about as severely as any pair of nuclides. Thus, r89,95 indicates approximately the maximum extent of fractionation that will be observed in the sample. 

The results of correlation studies reported in Refs. 2, 4, and 7 indicated that the fractionation behavior of most fission products was remarkably similar for coral surface bursts, bursts on the surface of deep and shallow seawater, and bursts at altitudes sufficiently great to avoid entrainment of soil or water in the fireball and cloud. Furthermore, the correlations showed no clear-cut dependence on the explosive yield of the device. This report extends the treatment to a near-surface event on silicate soil. 

Transient Test Air Bursts Coral Surface Burst Silicate Surface Bursts... 

The Specific Activity of Nuclear Debris from Ground Surface Bursts as a Function of Particle Size... 

Samples from the following events were analyzed Castle Bravo, Castle Koon, Redwing LaCrosse, Redwing Zuni, Redwing Tewa, and from Johnie Boy. Koon samples from three different altitudes were analyzed but only one of the Johnie Boy samples (842L). Additional Johnie Boy samples have been analyzed by Russell (6). Details of all events and samples are reported elsewhere in this volume (5), except for LaCrosse and for Tewa. LaCrosse was a coral surface burst of about 40 kilotons whose cloud topped out at 12,000 meters. The sample analyzed (054) was collected at 6500 meters and at 2.6 hours after the event. 

The specific activities of radionuclides in debris from ground surface bursts are generally complicated functions of the particle size, in accordance with the composite nature of the debris. 

The specific activities of radionuclides in samples from surface burst debris are not necessarily unique functions of the particle size. The uniqueness is determined by the degree of mixing in the cloud, and is, therefore, expected to be a function of the time elapsed between the shot time and the sampling time. 

The size distributions of the particles in cloud samples from three coral surface bursts and one silicate surface burst were determined by optical and electron microscopy. These distributions were approximately lognormal below about 3/x, but followed an inverse power law between 3 and ca. 60 or 70p. The exponent was not determined unequivocally, but it has a value between 3 and 4.5. Above 70fi the size frequency curve drops off rather sharply as a result of particles having been lost from the cloud by sedimentation. The effect of sedimentation was investigated theoretically. Correction factors to the size distribution were calculated as a function of particle size, and theoretical cutoff sizes were determined. The correction to the size frequency curve is less than 5% below about 70but it rises rather rapidly above this size. The corrections allow the correlation of the experimentally determined size distributions of the samples with those of the clouds, assuming cloud homogeneity. 

T,he size and mass frequency distributions of the particles in clouds from ground surface bursts have an essential bearing on predictions of the fallout field resulting from such nuclear explosions. Fallout models that are still in use employ size distributions which have been derived... 

In this paper we report on the results of size distribution measurements of cloud samples from three coral surface bursts and one silicate surface burst and present the results of the calculations of the sedimentation correction. 

Table II. Sampling Data for Some Cloud Samples from Ground Surface Bursts...

---