Use of tritium time series to estimate physical parameters of hydrologic systems

Tritium has been used as a tracer for the physical movement of water for the last five decades. After the major increases in the tritium world inventory produced by atmospheric nuclear weapons testing, its value as a tracer was quickly recognized. The tritium transient has been applied to a large variety of environmental studies in hydrology, oceanography and atmospheric circulation. Tritium data frequently yield information on physical processes which can be difficult or impossible to obtain by other methods. As the tritium transient enters its sixth decade and the tritium concentrations are declining, the interpretation of tritium measurements has become more ambiguous. The tritium/3He methodology has been developed and is of use in groundwater studies, but is of limited values for surface waters studies. One method, which is frequently overlooked, is the use of measurements made over long periods of time, either on a routine basis, or sporadically. With the use of a proper input function, the physical meaning of these data can often be unraveled. A monthly tritium input function has been developed for a 2 deg. latitude by 5 deg. longitude grid for the continental United States for the period 1953-2001. The tritium concentrations for the grid are derived from a series of precipitation collection stations that have operated during the tritium transient. For coordinates where no station was available, the data is interpolated between the nearest stations. No station operated prior to 1959 and many stations were discontinued during the 1970s and 1980s. To overcome these gaps in the data, correlations were carried out with the Ottawa, Canada station to f ill in gaps in the data for 1953 to 1987. Due to possible influences from local nuclear sources in Ottawa after 1987, correlations were carried out with the Vienna, Austria station for the period 1987-2001. Correlation coefficients relative to the Vienna station are given for each section on the grid for each month so that when more data becomes available on the IAEA website (http://isohis.IAEA.org) researchers can use it to extend the input function in time. This input function can then be used in modeling tritium time series. An historical surface water data set has been compiled in cooperation with the IAEA consisting of surface water tritium concentrations measured since the 1940s. This data set consists of over 6000 surface tritium measurements made world-wide from prior to the beginning of atmospheric nuclear testing to 2004. There are many long-term data sets available in this group which furnish valuable data bases for understanding physical processes within the river basins. These data show the response of the surface water systems to the tritium transient over the past several decades. For many hydrologic systems, tritium concentrations in river water will be more important than concentrations in precipitation. This is frequently the case in studies involving irrigation and urban studies. The response of the typical surface water system (Missouri River at Nebraska City, Iowa) to the tritium transient compared to that of the precipitation within the basin is shown. The rise of tritium concentrations in surface waters is less than that found in precipitation, but the drop in concentrations is much slower. This is due to the residence times of the various reservoirs that contribute to the base flow of the river system. Mean residence times calculated for a series of different systems have shown variations from 1-2 years to approximately 20 years. These differences will have important implications for the response of different basins to pollution, and changes in land use and climate. The seasonal signal from release of spring melt and rains is also evident in the tritium concentrations for rivers for three to four years after the peak of the transient in 1963. These increases in concentration indicate the fraction of young water that makes up the flow during that period. As tritium concentrations in precipitation decreased this seasonal signal disappeared and c annot be used to determine the fraction of young water at the present time. These results indicate the importance of obtaining previous data for proper assessment of current data