3. CIVILIAN AND MILITARY CASUALTIES NORTH VIETNAM JANUARY - SEPTEMBER 1966

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP78S02149R000200040009-6
• Release Decision: RIPPUB
• Original Classification: T
• Document Page Count: 20
• Document Creation Date: December 15, 2016
• Document Release Date: October 15, 2003
• Sequence Number: 9
• Publication Date: November 22, 1966
• Content Type: REPORT

Approved For Release 2003/10/29 : CIA-RDP78SO2149R000200040009-6
Civilian and Military Casualties North Vietnam
January - September 1966
Project 00.4985
22 November 1966
DIA review(s) completed.
Approved For Release 2003/10/29 : CIA-RDP78SO2049R0002
 Approved For Release 2003/10/29: CIA-RDP78S0
Table of Contents
Civilian and Military Casualties
A. Summary . . . . . . . . . . . . . . . . . . . . .
B. Fixed Targets . . . . 3
(1) Casualty Prediction Methodology for Targeted 4
Areas . . . . . . . . . . . . . . . . . .
(a) Target area . . . . ? ? ? ? ? ? ? . ' ' 4
(b) Spillover area . . . . . . . . . . . . . 5
(c) Peripheral area . . . . . . . . . . . . 5
Civilian Casualties
C. Small pre-briefed Non-JCS Targets . . . ? ? ? ?
Civilian-Military Casualties
D. Armed Reconnaissance . . . . . . . . . . . . . .
(1) Methodology . ? ? ? ? ? ? ? ? . ' ' ' ' '
(a) Civilian Casualties . . ? ? ? ? ? ? ? '
(b) Military Casualties . . . ? ? ? ? ?
Limitations of the Data . . . . . . . . . . . . .
(1+) Radio, Press and Official Claims .
DIA review(s) completed.
16
25X1
Approved For Release X003/10/29 : CIA-RDP78S02149R000200g40009-6
 Approved For Release 2003/10/29 : CIA-RDP78SO2149R000200040009-6
Civilian and Military Casualties
A. Summary
The number of casualties in North Vietnam which have resulted
from bombing attacks by the Unites States cannot be estimated with any
precision. However, the analysis
has provided sufficient evidence to suggest an
estimate. This estimate, given a 25 percent margin of error, reflects
the best information currently available. 0
The United States has continued to place restrictions on the air
offensive against North Vietnam in order to minimize civilian casualties.
North Vietnam, however, maintains that the Rolling Thunder program is a
vicious and unrestrained assault on hospitals, schools and the general
civilian population. In only one instance (September 1965: 40,000 killed
and 31,000 wounded) has Hanoi provided totals on the alleged number of
casualties inflicted by the air war. During 1966 North Vietnam made no
statements on the total number of casualties sustained by the population
as a result of U.S. bombings.
Although the re]ative number of estimated casualties per SDrtie has
remained stable, the total estimate of casualties inflicted on North
Vietnam in 1966 has increased over the 1965 estimates. There has also
been a significant shift in the distribution of casualties from Military
personnel to civilians. The changes in distribution and size of the
25X1
Approved For Release 2003/10/29 : CIA-RDP78S02149R0002000400b9-6 25X1
 25X1
Approved For Releas
estimate are primarily related to the decreased emphasis on attacks against
S targets (especially military facilities) and the substantial increase
in the armed reconnaissance program. Based on sample data, through the
first nine months of 1966) North Vietnamese casualties -- both civilian
and military -- are estimated to have been about 15,700, with some 40
percent killed and 60 percent wounded.
Estimated Casualties Resulting from Rolling Thunder, 1965-66
1965 1966
Total
Civilians
Fixed targets
2,000 700
2,700
Armed reconnaissance
4,000 11,200
15,200
6,000 11,900
17,900
Military
Fixed targets
4,300 400
4,700
Armed reconnaissance
2,900 3,400
6,300
7,200 3,800
11,000
GRAND TOTALS
13,200 15700
28,900
About 70 percent of the casualties in 1966 were civilians in contrast
to 45 percent in 1965. Armed reconnaissance accounted for 93 percent of
the total casualties in.1966 as versus 52 percent in 1965. A good number
of the civilians killed or injured 1966 were for the most part engaged
in maintaining the logistic supply lines in North Vietnam.
25X1
25X1
25X1
Approved For Releas+ 2003/10/29 : CIA-RDP78S02149R00020004g009-6
 Approved For Release
The casualty figures can at best be rationalized in the following
manner. In a country where over 350,000 people die each year of
natural causes the impact of 12,000 civilian casualties annually does
not appear to be an attrition of considerable proportions. In fact,
it is believed that the number of annual accidental deaths in North
Vietnam still exceed the current estimate of civilian deaths caused by
the Rolling Thunder Program.
Civilian and Military Casualties
B. Fixed Targets
In the past year the Rolling Thunder JCS fixed target program
has changed considerably in both size and scope. Some 2,000 attack
sorties were directed against JCS targets in 1966, as versus 8,700
flown in 1965. Generally, J'CS strikes in 1965 were directed against
facilities in areas with relatively low population densities, whereas
the growing scarcity of unstruck JCS targets this year has resulted in a
restrained bombi~g program aginst more lucrative targets in areas having
much higher population densities (iP.Hanoi and Haiphong) and in Xfestri4 t-S
against those facilities hit in 1965.
It is estimated that at a minimum some 700 civilian and 400 military
casualties were inflicted on North Vietnam as a by product of attacks on
JCS targets in 1966. In an absolute sense this compares favorably with
25X1
25X1
Approved For Rele4
 25X1
25X1
Approved For Release'
the 2,000 civilian and 4,300 military casualties estimated to be inflicted
by the 1965 JCS target campaign. It should be noted, however, that North
Vietnam incurred casualties at the rate of .5 per attach sortie against
JCS targets in 1965, and at the rate of .6 per attack sortie h 1966. The
increase in the sortie-casualty ratio undoubtedly reflects the above
mentioned change in the population densities of the new target systems
put under attack in 1966.
(1) Casualty Prediction Methodology for Targeted Areas
Three additive and one weighting factor were used
to determine casualty estimates for the individual fixed targets. DIA
analysts define the first three factors as being the immediate target
area, the spill over area, and the peripheral area. A fourth factor
is discussed later.
Military analysists decided that in theory all bombs dropped on
target should fall within a circular error of probability equal to 600
feet in diameter. Within this area a distinction is made between the
target and spill over area. An additional distinction is also made for
bombs falling outside the circular error of probability.
(a) Target area
To determine the maximum number of casualties
within the targeted area thinumber of people residing in the relevant
25X1
Approved For Releas
 25X1
Approved For Rele
area is multiplied by the ratio of the total average lethal area of the
bombs impacting within the target area. An account is also taken of the
weapons which overlap in effects.
(b) Spillover area
Probable maximum casualties in the spillover
area are found by multiplying the average lethal area of the bombs
impacting in the spillover area (in millions of square feet) by the
population density of the area (in people per million square feet).
(c) Peripheral area
The procedures used to estimate casualties in
the peripheral target areas in North Vietnam are based on World War II
data contained in the U.S. Strategic Bombing Survey of March 191+7. World
War II bombing of Japan caused some 550,000 casualties of which 250,000
were fatalities. A total of 2,100,000 buildings were destroyed by these
bombings. The Japanese expected their urban areas to be struck and their
air warning system was developed to provide early warning.
The buildings in North Vietnam are considered to be nearly equivalent
to those found in Japan during World War II. The North Vietnamese are
equally aware of U.S. bombing, but, since somdof the targets are
comparatively isolated, the North Vietnamese warning system may not be as
effective as the Japanese system. In North Vietnam, however, few incendiary
type bombs have been dropped, and targets generally are not located in as
Approved For Release 20 200040009-6
25X1
 25X1
25X1
25X1
25X1
Approved For Releas
densely a populated area. Consequently, it is considered reasonable to
base estimates of North Vietnamese casualties on the Japanese World War II
ratio. Thus, one may expect about .25 casualties per structure destroyed.
The minimum casualty figure is derived by anultiplying the number of
destroyed buildings outside of the spillover area by .25.
DIA casualty estimates for any given JCS facility are derived by
adding the estimates for casualties in thelimmediate target, spillover
and the peripheral areas. The immediate target and spillover casualty
estimates are theoretically maximum losses and must be deflated by a
factor of ton to get a maximum and minimum estimate of casualties. The
10.1 ratio is used because in theory unwarned personnel are 10 times more
vulnerable to casualties than are warned personnel under similar attack
conditions. Since the peripheral casualty casualty estimate is a minimal
figure it must be multiplied by 10 to establish a maximum and minimum
range.
initial attacks on TCS targets have inflicted casualties that are
nearly four times greater than the minimum. DIA post-strike estimate ,
of casualties. Consequently, the minimum DIA casualty estimates for the
respective targets were used as a basis for adjusted estimates. Attacks
Approved For Releas
25X1
25X1
 25X1
Approved For Releas
against previously unstruck JCS targets hit in 1966 were weighted by a
factor of four. Facilities hit in both years were credited with the
minimal estimate for purposes of the 1966 casualty figure. This was
done to take account of the probability that inhabitants residing in
areas previously attacked would either move out or construct adequate
shelter. The results of the exercise are shown below:
Estimated Civilian and Military
Casualties DRV: Inflicted by Strikes on JCS Targets 1966
JCS Fixed Target System
Military Civ
ilia
n Total
Barracks/supply depots/
ammunition depots
300
50
350
POL storage
50 1
00
150
Power plants
--
80
80
Airfields
20
--
20
Bridges
-- 1
80
180
Railyards
-- 2
00
200
Ports
30
90,
120
400 7
00
1,100
25X1
Approved For Release 2003/10x29 : CIA-RDP78SO2149R00020OP40009-6
 Approved For Release 2003110/29 : CIA-RDP78S02149R0001200040009-6 25X1
3. Civilian Casualties
C. Small pre-briefed Non-JCS Targets
In the past year it is estimated that some 1,200 casualties
were inflicted on the North Vietnamese civilian population as a con-
sequence of the armed reconnaissance attacks against small pre-briefed
non-JCS targets. These facilities were for the most part targets of
secondary importance but were attacked as fixed targets. Casualty
estimates were made on the basis as the four factor analysis explained
in the methodology in 3B1. Since no estimate of this type exists for
1965 it is impossible to put the current cas.ty estimate in perspective.
A tabulation of the targets and the respective casualties is presented
below.
Civilian Casualty Estimate
for Small Pre-briefed Non-JCS Targets, 1966
Ports 30
POL 180
Storage areas 800
Rail yards 100
Bridges 90
1,200
Approved For Release 2003/10/29 : CIA-RDP78SO2149R000200040009-6
 25X1
Approved For
3. Civilian-Military Casualties
D. Armed Reconnaissance
During 1966 the armed reconnaissance program over North Vietnam
changed considerably when contrasted with the operation conducted in
1965. Some 57,500 armed reconnaissance sorties were flown this year
as versus 25,300 in 1965. Since its inception the armed reconnaissance
program has been gradually expanded to cover nearly the entire expanse
of North Vietnam. The program remains, however, one with the limited
goals of minimizing civilian casu&ties and maximizing damage to the
infiltration-transportation network.
It is estimated that some 10,000 civilian and 3,400 military
casualties were inflicted on North Vietnam as a consequence of the
armed reconnaissance-program in 1966. This represents a considerable
increase over the estimated 4,000 civilian and 2,900 military casualties
in 1965. Relatively, however, the casualty/sortie rate has remained
stable. In 1966 casualties were inflicted at the rate of .23 per sortie
as versus .27 in 1965. The stability of the sortie/casualty rate
undoubtedly reflects improved North Vietnamese air warning and civilian
defense measures and the continued efforts by the USAF and USN to mini-
mize casualties as the tempo of the airwar increases.
(1) Methodology
(a) Civilian Casualties
Three factors were employed to estimate civilian
casualties inflicted by the armed reconnaissance program in 1966.
Approved For Releas - 149R000200040009-6 25X1
 25X1
Approved For Relea4
A daily casualty factor was derived
for a representa-
tive route package. This factor was then expanded to cover all of the
six route packages in North Vietnam. Modifying indexes were calculated
to reflect the varying population densities and sorties flown in the
respective route packages in 1966. The adjusted daily route package
casualty figures were then multiplied by the number of days of air
operation in the specific route packages. This exercise is shown in
the tables below.
25X1
Approved For Release 2003/10/29 : CIA-RDP78SO2149R000200040009-6
 25X1
Approved For Release 2003/10/29 : CIA-RDP78S02149R000200040009-6
Estimated Civilians Killed By Air DRV: Inflicted by Armed Reconnaissance, Jan-Sep 1966
Route
Total
Package
(KBA Ratio-day) X (Population Density Index) X
(Armed.Recon Index) =
KBA Value Day X (# Days Air Operations) =
Civilians KRA*
I
(3.5)
( .40)
(2.03),
=2.84
243
=690
II
(3.5)
(1.00)
(l.oo)
=3.50
243
=840
III
(3.5)
( .63)
( .76)
=1.67
243
=4oo
Iv
(3.5)
(2.10)
( .46)
=3.38
243
=8oo
v
(3.5)
( .23)
( .20)
_ .16
243
= 30
VI A+B
(3.5)
(1.31)
( .29)
=1.32
183
=240
Total 3,000
Approved For Release 2003/10/29 : CIA-RDP78S02149R000200040009-6
 Approved For Release 2003110129: CIA-RDP78S02149R0002(2 Xi1l009-6
Estimated Civilians Wounded by Air DRV: Inflicted by Armed Reconnaissance, Jan-Sep 1966
Route
) X
(
BA R
ti
ulation Density Index X
Po
AR Index = WWBA Value D X
Days Air Operations) = Total Civilians WBA*
Package
o
a
W
p
day
i
(8)
( .4o)
(2.03) =
6.50
243
=
1,580
II
(8)
(1.00)
(1.00) =
8.00
243
=
1,950
III
(8)
( .63)
( .76) =
3.83
243
=
goo
Iv
(8)
(2.10)
( .46) =
7.73
243
=
1,880
V
(8)
( .23)
( .20) =
.37
243
=
9o
VI AB
(8)
(1.31)
( .29) =
3.04
183
=
560
Total 72000
Approved For Release 2003/10/29 : CIA-RDP78S02149R000200040009-6
 Approved For Release 2003/10/29 : CIA-RDP78SO2149R000200040009-6
Average daily casualties for the thirty day period were
2.2 killed by air (KBA) and 6.7 wounded by air (WPA).
an atrocity factor was also
included in the daily casualty figures. It was assumed that one atrocity
bombing occurred in each route package each month. An average atrocity
figure (2.6/day) was calculated
assuming half the casualties were killed and half wounded. The
final daily casualty rates for civilians as a consequence of the armed
reconnaissance were estimated to be 3.5 KBA and 8.0 WBA in route package
two. 13
Approved For Release 2003/10/29 : CIA-RDP78SO2149R000200040009-6
25X1
25X1
25X1
25X1
 Approved For Release 2003/10/29 : CIA-RDP78SO214PR000200040009-6
.(b) Military Casualties
The problems encountered in deriving a credible esti-
mate of military casualties inflicted by armed reconnaissance were
25X1 insoluble.
A methodology employing tabulated results of the armed reconnaissance
program weighted the destroyed and damaged figures with DIA estimates of
casualties per truck and for other facilities. In two specific cate-
gories, trucks and watercraft, the casualty weights were decreased
25X1
It should be pointed out, however, that at best this estimate represents 25X1
a very crude estimate.
Small Targets of a Military and Military Associated
Nature Attacked on Armed Reconnaissance
Destroyed
and Damaged Military Casualties
Barracks 80 x .1
= 10
Supply warehouses 138 x .1
W 15
Misc. buildings 4,640 x .1
= 470
Radar Com. 78 x .1
= 10
Truck parks 92 x .1
= 10
AAA sites 568 x .1
= 60
Trucks 3,096 x .5
=1560
Watercraft 2,071 x .5
=l040
RR cars 2,234 x .1
225
3,400*
- 14 -
Approved For Release 2003/10/29 : CIA-RDP78SO2149R000200040009-6
 25X1
25X1
Approved
newspaper articles, radio broadcasts, and protests to the ICC.
49R000200040009-6
E. Limitations of the Data
In a country such as North Vietnam where the casualty figures
of the French-Indochinese are still held as a state secret there appears
to be little to gain from deducing casualty figures from NorthVietnamese
- 15 -
25X1
Approved For Release 2003/10/29 : CIA-RDP78SO2149R000200040009-6
 Approved For Release 2003/10/29 : CIA-RDP78SO2149R000200040009-6
25X1
(4) Radio, Press and official claims
The North Vietnamese radio and press in 1966 continued to
give vague and emotional accounts of US air strikes anddamage. There
have been no cumulative statements of casualties from Hanoi since Sep-
tember 1965 when Egyptian journalists were reportedly told that 40,000
North Vietnamese had already been killed and 35,000 wounded. The Hanoi
news media was, as expected, of no value in estimating casualties.
The few recent official claims of casualties indicate that
the toll in human lives isn't too high. A letter in May 1966 from the
- 16 -
Approved For Release 2003/10/29 : CIA-RDP78SO2149R000200040009-6
 Approved For Relea
Hanoi Red Cross to the International Committee of the Red Cross in
Geneva specifically cited 239 civilian casualties inflicted since
31 January 1966, although implying many more. In October 1966 Hanoi
also claimed that 300 school children and 30 teachers had been killed
by the US aircraft.
- 17 -
25X1
Approved For Release 2003/10/29 : CIA-RDP78SO2149R000200040009-6
 Approved For Release 2003/10/29 : CIA-RDP78SO2149R000200040009-6
Approved For Release 2003/10/29 : CIA-RDP78SO2149R000200040009-6