NRDL WT-915 1956 Castle-Bravo fallout area exaggeration

Fig. 1 - the U.S. Naval Radiological Defense Laboratory (NRDL) January 1956 report WT-915 gave an exaggerated close-in fallout pattern, based on a pretty bad exaggeration of the effect of wind shear.

Fig. 2 - Dr Carl F. Miller, who was at Operation Castle, shot Bravo, showed in his 1963 Stanford Research Institute book Fallout and Radiological Countermeasures volume 1, that the effective Bravo wind shear was only 23 degrees.

Fig. 3 - Castle-Bravo wind hodographs for Bikini Atoll at shot time and Rongerik at 6 hours (when the cloud was actually in the vicinity of Rongerik) for the fallout hot line (55,000 feet cloud base down to the ground, as justified by Edward Schuert in USNRDL-TR-139) seem tat a glance to agree with the NRDL WT-915 fallout pattern centre-line or "hotline".  But they don't, because the wind hodograph is for a constant balloon rise rate (or conversely, constant fallout descent rate) and doesn't take account of the increased drag effect on fallout in lower altitude air, which slows them down.  This is why Miller's analysis in Fig. 2 is more important.  We need to weight the lower altitude wind components more substantially than those for higher altitudes, because the fallout takes longer to fall through the denser air at lower altitudes, and is therefore affected by the lower altitude winds for a longer period of time!

Fig. 4 - DTRIAC SR-12-001 photos of Castle-Bravo cloud at 16 and 30 minutes (backlit by the rising sun in the east), showing the upper part of the cloud stem headed north-east.  However, the lower part of the stem isn't being blown in that direction, and this is what's important for fallout which spends most of its time being blown by the winds in the denser lower altitude air.

Fig. 5 - April 1956 deep ocean sediment survey (pre-Redwing oceanographic survey, project 2.62 (this graph is included in an extract from Feenan D. Jennings' report Radioactivity background and oceanographic conditions in the Pacific Proving Grounds at the start of Operation Redwing, WT-1349, included here).  Note that the water surface bursts of Castle produced fallout that was a salt slurry which was trapped above the thermocline on landing in the ocean (i.e. in the top 100 metres layer of warm, buoyant water).  Only Castle-Bravo deposited solid fallout particles east of Bikini, which could have significantly contributed to the sediment gamma activity shown above (gamma decays per minute per gram of sediment).  The highlighted red circled core samples show evidence that the Bravo hotspot north of Naen island in Rongelap Atoll did not extend as far north as the massive hotspot shown in the NRDL WT-915. 
Bravo core sample C-9 about 60 nautical miles north of Ailinginae should be in the middle of the massive NRDL hotspot, yet it only had 6.5 gamma decays per minute per gram, compared to 210 at core sample C-10, which is on Miller's hotline for Bravo, and is roughly the spot near Naen island where the Lucky Dragon had stopped to fish. So Glasstone and Dolan kept to the cigar shaped Bravo pattern.

(WT-1317 shows that some Zuni surface burst fallout quickly descended below the thermocline, although this is very little.  Photos of fallout particles on lagoon rafts WT-615 for Mike shot shows why: solid fallout particles produced from coral were rapidly "hollowed out" by sea water, because they were mostly calcium hydroxide with only a thin shell of calcium carbonate.  Although this rinsed-out fallout activity is not chemically soluble in the sense of being dissolved ions, it is effectively colloidal, i.e. small micron-sized particles that don't sink rapidly like 100 micron-diameter solid fallout particles.  So most gets trapped above the thermocline.  But a small fraction of the activity is carried straight down to the ocean bed, as Zuni data shows.)

There are many issues with the NRDL WT-915 fallout pattern.  It exaggerates effective windshear, it has the wrong hotline (suspiciously similar the false Stokes law assumption in the 1957 edition of Glasstone's Effects of Nuclear Weapons that the fallout descent speed is independent of altitude), and it contains massive errors in the upwind fallout maps of Bikini Atoll for Castle tests which are copied without correction in the DASA-1251 1963 fallout patterns compendium (WT-915 figures 6.1 on page 76, 6.2 on page 78, 6.3 on page 79) which almost double the correct length of Bikini Atoll, and thus exaggerate upwind fallout pattern areas by nearly a factor of 4.  This has had serious consequences for planning civil defense rescue in the upwind and close-in blast areas after a nuclear explosion.  It has also "validated" incorrect fallout models that exaggerate the fallout near ground zero after large surface bursts.

Fig. 6 - 13.5 megaton water surface burst Castle-Yankee fallout pattern, following a similar trajectory to the NRDL Bravo hotline (WT-935).  This was obtained from a ship survey of the contaminated water.  This gives a better idea of the fallout dose contour areas than the NRDL pattern.  (Fig. 3.17 in WT-1344 compares the areas versus dose rate contour intensities for two land surface bursts and two water surface bursts of Redwing; they are very similar after the dose rates have been scaled to the same fission yield.)

The problems with lying or "exaggerating facts" to grab attention in order to bolster ethical appeasement or moral disarmament "education" programs, is that you ignore alternative problems, like the risk of a nuclear terrorist 9/11 type surprise attack, where the choice between evacuation or stay in place sheltering will depend on the width of the high dose rate contours (i.e. whether you can evacuate crosswind before getting a significant exposure to fallout).  There are always problems in introducing inaccuracies to promote arguments that can't be won by telling the plain unvarnished truth.

Fig. 7 - the RAND Corporation fallout area versus unshielded dose plot (verified by a good analysis of both the Nevada and Bravo surface burst fallout "hotspot" patterns).  For 100 kilotons fission yield (for a modern 200 kt weapon with 50% fission yield), 1 square nautical mile receives around 4,000 R outdoors over the first 48 hours, 10 square miles gets 1,000 R, 100 square miles gets about 200 R, and 1,000 square miles gets about 40 R.  These doses are reduced by a factor of 10 or more in most large modern city buildings, in central areas or the basement (which would give much better protection), away from windows and the roof.  The problem is that a protection factor of 10 still gives a 1 square mile area with 400 R to 48 hours, which would be lethal to many people.  This very intense area needs to be evacuated before such dangerous doses are accumulated.  The trade off is how much radiation is received while evacuating.  Most analyses suggest a few hours of shelter to allow some decay to occur, to reduce the transit dose while leaving the heavy fallout area.

Fig. 8 - the RAND Corporation fallout area for Castle-Bravo (terrain shielding not included), showing a smaller hotspot at the North of Rongelap Atoll with a centre line and wind shear that more closely matches Miller's analysis of the wind pattern.  The two smaller hotspots within Bikini Atoll were from crater throwout and stem fallout, respectively.

Fig. 9 - a far more realistic fallout pattern of the sort likely to result from nuclear terrorism, the 1952 Operation Hurricane 25 kt nuclear explosion inside a ship in shallow water, a simulated harbour.  Only 1.4% of the energy was released as thermal radiation, so no skin burns would occur.  Blast wind drag and flying debris impacts to standing personnel, as well as nuclear radiation, would be the major hazards.

Fig. 10 - fallout gamma radiation shielding in a typical modern city building (DCPA CPG 2-1AG 1973 Ch 6 Panel 18).