Comparability Of Meteoric Water Lines: Daily, Monthly, Or Annual Data?

The stable isotopes of hydrogen (δ2H) and oxygen (δ18O) have been widely applied in hydrology. The equation relating δ2H to δ18O in precipitation, also defined as a meteoric water line, can provide a reference point for interpretation of stable isotopic compositions of a range of water samples in an area. Influenced by many spatial and meteorological factors, the slopes and intercepts of meteoric water lines vary depending on location.

It is not difficult to determine a meteoric water line using an ordinary least squares regression in Microsoft Excel or any other office software. However, is there any comparability of meteoric water lines based on different sampling frequencies?


For example, in a sampling site, there are 1000 rain hours during 200 events in 150 days of 11 months in one year, and we have isotope data for each rain hour. To determine a comparable meteoric water line, should we select 1000 original data, or precipitation weighted values using 200 event-based data, or 150 daily or 11 monthly data?

During recent years, stable isotopes in event-based (and even minute-based) precipitation have been widely investigated, often based on sampling durations of approximately one or two years. In many studies, event-based data over a relatively short period is used directly to determine a meteoric water line and for comparison with meteoric water line using monthly data. The potential influence of irregular sample frequency or short duration is not always considered because of the limited data availability.

Shengjie Wang, associate professor of hydrology from Northwest Normal University, and collaborators focused on this question based on their network of stable isotopes in precipitation across the Tianshan Mountains in arid central Asia. Approximately 1000 event-based samples at 23 stations were taken into consideration using traditional and other regression methods, and the precipitation weighted approach was considered.


Generally, the use of precipitation weighting is important in arid areas. If we have only a short-term, event-based precipitation isotope record spanning, for example, a year, especially in an arid climate, the weighted regression can be considered as a strong alternative.

Meteoric water lines derived using a weighted regression with event-based data are generally consistent with applying traditional regression of ordinary least squares to monthly data. But this is less true for the more arid sites, where low event numbers mean that aggregation into monthly composites does not provide effective precipitation weighting.

In circumstances where only short-term datasets are available, there is a benefit to be gained by using event data rather than monthly data. A precipitation weighted regression using event-based data can provide a wider range of values to constrain the regression than using a small number of a monthly data point. However, the importance of capturing complete seasonal cycles still remains; therefore, meteoric water line regressions should only be applied to multiples of complete 12-month periods.

These findings are described in the article entitled Meteoric water lines in arid Central Asia using event-based and monthly data, recently published in the Journal of Hydrology. This work was conducted by Shengjie Wang and Mingjun Zhang from the Northwest Normal University, Catherine E. Hughes and Jagoda Crawford from the Australian Nuclear Science and Technology Organisation, and other collaborators.


  1. Wang S, Zhang M, Hughes CE, Crawford J, Wang G, Chen F, Du M, Qiu X, Zhou S. (2018). Meteoric water lines in arid Central Asia using event-based and monthly data. Journal of Hydrology, 562, 435-445,

About The Author

Shengjie Wang is an associate professor of physical geography at the College of Geography and Environmental Science, Northwest Normal University, China. The underlying focus of his research agenda is on the application of stable water isotopes in hydrology and ecology across China as well as arid central Asia.