Atmospheric Deposition (Proceedings of the Baltimore Symposium, May 1989). IAHS Publ. No. 179.
TRITIUM DEPOSITION OVER THE CONTINENTAL
S T A T E S , 1953-1983
R o b e r t L. Michel
U.S. Geological Survey, 431 National Center, Reston, VA
UNITED
22092, USA
ABSTRACT
Tritium was produced in great quantities by
atmospheric nuclear weapons testing, the majority of which
occurred in the mid-1950's and early 1960's. After production,
the tritium atom is incorporated in the water molecule and
can be used as a tracer of hydrologie and océanographie
phenomena.
A major limitation in using tritium as a tracer for
hydrologie processes is the limited d a t a on its input. Since the
early 1960's the U.S. Geological Survey has monitored tritium
deposition with monthly analyses of tritium in precipitation by
a network of 14 stations throughout- the United States.
Tritium deposition has now been calculated for this series of
network stations for 1953 to 1983. For years when d a t a were
not available for a given station, estimates of tritium
deposition were made from correlations with data collected at
other stations. Depositions are given in TU-meters (1 TU = 1
tritium atom per 10 18 hydrogen atoms) on a cumulative basis
for all stations in the network. Depositions were largest in the
midwest and exceeded 7500 TU-m/cm at the Chicago station.
Depositions were smallest in the Southwest due to either the
low precipitation or the influence of moisture from the Pacific
Ocean. At Menlo Park, California and Albuquerque, New
Mexico depositions were about l / 5 t h of those found in
Chicago. Approximately 60% of all deposition occurred
between 1961 to 1965 and over 90% of the deposition occurred
by 1970. By 1983, more than 70% of all tritium deposited at
network stations from 1953 to 1983 had been removed from the
environment by radioactive decay. Major peaks in deposition
occurred in 1954, 1958, and 1963; a small increase in deposition
from Chinese testing occurred in the late 1970's.
From the network d a t a and the d a t a of other laboratories, a
map of cumulative tritium deposition over the continental
United States has been constructed for 1953 to 1983. The total
tritium deposition on the continental United States was
calculated to be 12 ± 2 kg during 1953 to 1983.
INTRODUCTION
Tritium, a radioisotope of hydrogen with a half-life of 12.43 years, has been
used as a hydrologie tracer for the past three decades (IAEA, 1981).
Tritium is produced naturally by interactions of cosmic rays with
atmospheric molecules, where it oxidizes rapidly and enters the hydrological
cycle. Tritiated water has a residence time of less than a year in the
atmosphere, eventually falling out as precipitation or through molecular
exchange. Tritium follows the pathway of water through its flow and
mixing processes exactly and its radioactive decay and distribution can be
used to estimate flow and mixing rates.
The other factor t h a t makes tritium a useful tracer is its transient
nature. In the 1950's and early 1960's the n a t u r a l levels of tritium were
109
110
Robert L.
Mchel
overwhelmed by the introduction of tritium through nuclear weapons
testing. Tritium concentrations in precipitation increased 2 to 3 orders of
magnitude, especially in the northern hemisphere, due to the locations of
the tests and atmospheric mixing. The large input was reflected in
continental waters and the surface ocean where some concentrations
increased to the 1000 TU and 50 TU level, respectively. Since the test ban
treaty, small amounts of anthropogenic tritium are still released to the
environment from nuclear reactors and atomic weapons tests by
nonsignatory powers (Carter and Moghessi, 1977). These releases have a
limited effect on tritium concentrations in precipitation.
Tritium
concentrations in present-day precipitation are tending toward the prebomb or background levels. Many studies consider the source of tritium as
a spike centered in the early 1960's t h a t is now being distributed through
the hydrological system. Currently, tritium d a t a are best used to study
processes t h a t occur on a timescale of 10 to 100 years.
For most studies, a knowledge of the input of tritium during the bomb
period is important. There are large variations in tritium concentrations in
precipitation and, frequently, no d a t a are available for a given location. In
this paper, the deposition of tritium is analyzed at a series of stations where
long-term d a t a are available. Using these d a t a and the Ottawa correlation
to estimate concentrations for periods where d a t a are missing, deposition
records are constructed for these stations through the period of the bomb
transient. The results are then expanded to obtain estimates of bombtritium deposition for all areas of the continental United States.
DATA
In the early 1960's, the U.S. Geological Survey began routine analyses of
tritium in precipitation at several sites within the continental United States.
Monthly composites (three month composites at some stations) of rainfall
were collected and forwarded to the U.S. Geological Survey Tritium
Laboratory in Washington, D.C. (Reston, Virginia since 1972), for analysis.
Depending on the tritium concentration expected, some samples were
counted by direct liquid scintillation without enrichment. Accuracy for
most samples was within 3 percent. This monthly network sampling
program has continued to the present time at most stations, giving
continuous tritium deposition records for the past 20 to 25 years. Results
of the precipitation samples have been published by the International
Atomic Energy Agency (1981, 1983, 1986). Analyses of the pre-1970 d a t a
have also been published (Stewart & Farnsworth, 1968; Stewart &
Wyerman, 1970; Wyerman, Farnsworth & Stewart, 1970). These papers
show tritium concentrations in precipitation across the continental United
States during the immediate post-bomb period and discuss its effects on
runoff.
During the pre-bomb and early nuclear period, limited tritium d a t a are
available for the North American continent. The only consistent d a t a set
for this period is from Ottawa, Canada, where measurements are available
from 1953 to the present. If a long term tritium record exists at another
North American station, a correlation can be made with the Ottawa record
and this correlation can be used to fill in gaps in the station of interest.
The Ottawa correlation was developed and presented by the International
Atomic Energy Agency (1981). A least-squares regression yields an equation
of the form,
C; = a C 0 t t + b
W
where a and b are obtained from periods where d a t a are available and
Tritium
Deposit
111
C; and Cott a r e tritium concentrations at the station of interest and
Ottawa, respectively. The a and b terms are assumed to be constant with
time for any given station. This correlation can be used to estimate tritium
concentrations at U.S. Geological Survey stations when d a t a are lacking,
which is usually the 1953 to 1960 time span. All tritium estimates used in
this paper are derived from the constants supplied by the International
Atomic Energy Agency (1981). For years when tritium concentrations are
very low, there are certain stations where the Ottawa correlation does not
yield a reasonable estimate due to the nature of b (i.e. a possible negative
concentration). For these cases, a minimum tritium concentration equal to
the level found in precipitation in the 1980's is used. These concentrations
are always very low, so any error introduced will be small.
T a b l e 1 Deposition at network stations, 1953-83
STATION
DEP
REM
%1963-1964
YA
ALB
BIS
BOS
CAHAT
CHI
LINC
MAD
MENPK
OCALA
PORT
STL
SLC
WACO
WASHDC
1524
7191
7131
4382
7681
6720
7160
1350
3208
3591
5835
5481
2352
5159
442
2014
1993
1306
2232
1890
2041
426
915
1034
1688
1551
677
1564
44
44
34
42
40
38
39
25
41
42
38
44
36
39
62-83
62-83
63-83
60-83
60-79
62-83
63-81
61-83
61-83
63-83
63-83
63-83
61-83
62-83
DEP=DEPOSITION(TU-M) FOR 1953-1983; REM= TRITIUM NOT DECAYED(TU-M) BY
1987; %63-64=PERCENT OF TRITIUM DEPOSITED IN 1963-64; YA=YEARS STATION
ACTIVE BETWEEN 1953-83; ALB=ALBUQUERQUE, NM; BIS=BISMARK, ND;
BOS=BOSTON, MA; CAHAT=CAPE HATTERAS, NC; CHI=CHICAGO, ILL;
LINC=LINCOLN, NEB; MAD=MADISON, WI; MENPK=MENLO PARK, CA;
OCALA=OCALA, FL; PORT=PORTLAND, OR; STL=SAINT LOUIS, MO; SLC=SALT
LAKE CITY, UT; WACO=WACO, TX; WASHDC=WASHINGTON, DC
It is possible to calculate tritium deposition at the U.S. Geological Survey
network sites for the period of bomb testing of 1953-83 using measured
tritium data and the O t t a w a correlation. Table 1 gives a record of
cumulative tritium deposition in TU-meters for these stations during the
1953-83 period. The TU-meter is found by multiplying the precipitation, in
meters, measured at a station during the year, by the weighted average
tritium concentration for t h a t year:
1 T U - m e t e r = 3.3xl0" 1 7 gm/cm 2 = 0.32 pcuries/cm 2
^
Also listed in Table 1 is the decay corrected tritium deposition for each
station and the percentage of tritium t h a t was deposited during 1963-1964.
Depositions throughout the network range from a high of more than 7500
TU-m at Chicago to about 1500 TU-m at Albuquerque, New Mexico, and
Menlo Park, California. Certain trends are evident for both concentration
and total deposition. Because tritium is mixed from the stratosphere into
112
Robert L.
Mchel
the troposphere mainly in northern latitudes (30°-60° N), a north-south
gradient exists with largest depositions found in the north.
Large
depositions also occur in the midcontinental stations. Large depositions
also are found at the East Coast sites relative to sites at the same latitudes
on the West Coast. P a r t of this difference can be attributed to the
comparatively low rainfall in parts of the West Coast. However, the source
of air masses involved in storms also has a bearing on the deposition
differences. On the West Coast, most storms are oceanic in origin and have
spent little time over continental land masses. The tritium concentrations
in water from these storms are primarily determined by exchange with the
surface ocean which has a relatively low tritium concentration. On the East
Coast, storms are strongly influenced by air masses traveling across the
continent. Exchange with and input of water vapor from the stratosphere
will substantially increase tritium concentrations in the troposphere. As a
result, stations on the East Coast will record higher depositions than
stations at the same latitude on the West Coast. Deposition also is low in
some mid-continental areas, such as Albuquerque, New Mexico, and Waco,
Texas, where a scarcity of precipitation is the main cause of low deposition.
Approximately 60% of all tritium deposition occurred during the five year
period 1961-65 as is demonstrated graphically in Figure 1 which shows the
tritium deposition occurring in each year from 1953-1983 as a percentage of
the total. The percentage is calculated as follows:
% = 100*D t /D,
(3)
where D t is the deposition in TU-m for a given year and D is the total
deposition in TU-m for 1953-1983. Three stations are shown representing
the West Coast (Portland, Oregon), the mid-continental area (Lincoln,
Nebraska) and the East Coast (Boston, Massachusetts). Three peaks are
present (1954, 1958, 1963) representing the fallout from the three major
periods of nuclear weapons testing. The largest peak in the 1963-1964
period is more than double the 1958 peak. About 40% of all tritium
depositions occurred in 1963-1964 and more than 7 5 % of the total
deposition occurred before 1965. After the 1963 bomb peak, tritium
deposition declined quickly, until, by 1970, only about 1% of the total
tritium deposition occurred annually. A small peak occurs in the late 1970's
due to small nuclear tests by China and France. This is reflected in an
increase in total deposition at most stations, but the peak deposition from
these tests is less than 5% of the previous deposition peak. For many
purposes, such as the study of reservoirs with exchange timescales on the
order of the tritium transient, deposition can initially be considered as a
spike occurring in the early 1960's.
It should be noted t h a t Table 1 lists tritium deposition by precipitation,
but many other factors influence the effective tritium input into-a system.
For any given body of surface water, molecular exchange can be an
important source of tritium input. Water molecules from tritium-rich
atmospheric water vapor enter the liquid phase at the surface of a water
reservoir while water molecules from the surface water go into the vapor
phase in an equilibrium exchange. Because most surface waters have
tritium concentrations lower than t h a t of overlying atmosphere, a net flux
of tritium to the surface water occurs. Over the continental surface, this
process will be important only for basins with a large water surface area.
Evapotranspiration will have a major effect on the deposition, as major
amounts of tritium can be returned to the atmosphere soon after
deposition. Because of its short half-life, decay is a significant factor when
studying the tritium transient. On a thirty year timescale, decay would
Tritium Deposit
113
remove about 70% of all tritium deposited at the network stations during
1953-1983. If decay were the only factor influencing the tritium burden,
about 50% of the remaining tritium would have been deposited during
1961-1965.
32
"i—rn—i—i—i—i—i—r
30
28
X PORTLAND, OR.
26
&
LINCOLN, NEB.
•
BOSTON, MA.
24
22
?
<)
\-
w
o
n
o
20
18
_i
<
16
tu.
(1
1-
14
n
«•:
ee
12
X «
1953 1955
f
1960
* sà
** t
1970
1975
i * g
1980
1983
YEAR
Figure 1 Percent deposition per year for 1953-83.
DEPOSITION OVER THE CONTINENTAL UNITED STATES
On the basis of network data and data from other laboratories, it is
possible to construct the tritium deposition patterns over the continental
United States since the advent of nuclear testing. To construct the
deposition patterns, the United States was divided into a series of boxes 2°
in latitude and 5° in longitude. To determine the tritium deposition for
any box, all that is needed is the amount of precipitation and the weighted
tritium concentration for each year. The average rainfall for a box can be
114
Robert L.
Mchel
obtained from the yearly precipitation records of the National Weather
Service. Since most boxes contain one or more long-term weather stations,
reasonable estimates of the precipitation in each box can be found.
As noted, the coverage for tritium concentrations is incomplete. In
addition to the U.S. Geological Survey stations, other stations have been
operated by other laboratories. D a t a from these stations are available from
the International Atomic Energy Agency. Although records at these
stations are frequently less complete than those from the U.S. Geological
Survey network, most were active during the period of maximum deposition
and the few years following. It was during this period t h a t both
concentrations and concentration gradients were highest. Thus the other
stations' d a t a are available during the most critical period of the bomb
transient. Approximately 2 5 % of the boxes are directly covered by tritiummonitoring stations during p a r t of the bomb-testing period. Data from a
network of Canadian stations also are available to help establish patterns
in tritium concentrations
in precipitation.
Estimates of
tritium
concentrations in precipitation can be made for all areas of the United
States during the 1953-1983 period. Maps of tritium concentrations in
precipitations for some years have previously been published (Stewart &
Farnsworth, 1968). Tritium deposition patterns in the United States are
shown in Figure 2.
Figure 2 Deposition of Tritium
represent sampling stations.
in
TU-meters
for
1953-83.
Circles
Precipitation
patterns
produce differences
between
patterns
of
concentration and deposition. This is most evident in the northwest where
precipitation of 2000mm or more per year results in high deposition. The
arid conditions of the Southwestern United States result in low depositions
Tritium
Deposit
115
despite high tritium concentrations in precipitation. This is evident as a
tongue of low tritium deposition extending into the Great Basin from the
southwest. A tongue of high deposition is present in the Midwest extending
to the Gulf Coast. With some exceptions, tritium deposition in the
continental United States shows a southwest-to-northeast gradient across
the country. In the midwestern and eastern United States, cumulative
deposition exceeded 5000 TU-m (approximately 1.6xl0~ 5 ci/meter 2 ).
F r o m these data, the total deposition of bomb tritium on the continental
United States can be calculated. It is found t h a t 12 i 2 kg of tritium was
deposited as precipitation on the 48 contiguous States since the beginning of
nuclear-weapons testing. As earlier noted, a small additional amount also
was deposited in some water bodies by molecular exchange. Because of this
decay, évapotranspiration, and runoff, only a fraction of this tritium is still
present in continental waters.
A C K N O W L E D G E M E N T S The d a t a in this paper were compiled at the
U.S. Geological Survey Tritium Laboratory over the past three decades.
Ted Wyerman and F r a n k Brookman were the sources of most of the data.
REFERENCES
Carter, M.W. & Moghessi, A.A. (1977) Three decades of nuclear testing.
Health Physics 33, 35-57.
International Atomic Energy Agency (1981) Statistical treatment of
environmental isotope d a t a in precipitation. Technical Report Series
#206.
International Atomic Energy Agency (1983) Environmental isotope d a t a No.
7. Technical Report Series #226.
International Atomic Energy Agency (1986) Environmental isotope d a t a No.
8. Technical Report Series #264.
Stewart, G.L. & F a r n s w o r t h , R.K. (1968) United States tritium rainout and
its hydrologie implications. Wat. Resour. Res. 4, 273-289.
Stewart, G.L. & Wyerman, T.A. (1970) Tritium rainout in the United States
during 1966, 1967, 1968. Wat. Resour. Res. 6, 77-87.
Wyerman, T.A., F a r n s w o r t h , R.K. & Stewart G.L. (1970) Tritium in the
United States, 1961-1968. Radiol. Health Data and Rep. 11, 421-439.