This page allows on-demand calculation of the estimated modern mean annual and monthly deuterium and oxygen isotope composition of precipitation at a specified location. 95% confidence intervals for the estimated annual values can also be obtained. It is intended as a tool to facilitate the use of stable isotope data in hydrological, ecological, forensic, and geological studies, and to help standardize the interpolation of precipitation stable isotope data. Estimates are calculated from a global data set according to an algorithm developed by Bowen & Wilkinson (2002) and refined by Bowen & Revenaugh (2003) and Bowen et al. (2005). These papers are the primary references for the OIPC, provide description of the methods, and should be cited in any publication or report that uses isotopic estimates generated using this site. More background on this project can be found here. The data used by the OIPC are derived primarily from the International Atomic Energy Association/World Meteorological Organization Global Network for Isotopes in Precipitation, and their work should also be recognized.
The OIPC consists of a simple HTML form that allows user entry of site location and elevation parameters and returns estimated δ2H and δ18O values for the specified location. Enter the latitude and longitude in decimal degrees, with East and North positive and West and South negative. Altitude should be entered in meters above sea level. Please note that the OIPC provides limited checks on the accuracy of your site locations...any inaccuracies in the latitude/longitude/altitude values you enter on the form will affect the isotopic estimates returned. If you would like to obtain quantitative estimates of the 95% confidence intervals for interpolated mean annual values, check the appropriate box...however, generating the confidence intervals increases the required computational time by an order of magnitude, so be prepared to wait!
A number of problems and challenges related to using the existing isotopes in precipitation data are described by Bowen & Revenaugh (2003). Of primary concern, the un-evenness of sampling through time has the potential to create local artifacts in the interpolated isotopic fields. A robust interpolation algorithm should help reduce such artifacts, but they may still exist. The confidence intervals calculated by our program give some sense of where these effects may be significant. The values provided here are estimates of the modern (post-1950) mean condition, and may not agree with specific instantaneous or time-averaged measurements of the isotopic composition of precipitation at a given site.
The OIPC continues to grow and evolve, providing new capabilities and incorporating a growing database. At intervals, changes are made that will affect isotope ratio estimates for any given site, hopefully with the result that estimates are continually improved. I have established a system of versioning in order to document these changes and allow users to know which version they have used. Version numbers are shown in the OIPC form window, and are keyed to the notes provided here on each version.
Version 3.2: minor data additions and updates. These have been incorporated in some of the raster data products but are not yet implemented in the OIPC itself.
Version 3.1 incorporates fixes to 2 bugs that produced incorrect H isotope ratio estimates for January and February in the monthly module of version 3.0 and performance improvements reducing the time required for monthly calculations.
Version 3.0 represents a major update to the OIPC database. Station number updates are shown below. Many thanks to everyone who has collected and contributed data to the collection, and to Annie Putman for assistance with data compilation.
In addition, the seasonal cycle correction algorithm implemented in OIPC 2.0 has been improved by weighting offset terms for the relative uncertainty of the annual and monthly estimates at any given location. This has the effect of reducing the offset applied at locations where available data more tightly constrain the actual value for a given month.
| Dataset | Old station count | New station count |
|---|---|---|
| H annual | 476 | 1492 |
| O annual | 500 | 1560 |
| H monthly (average) | 401 | 609 |
| O monthly (average) | 411 | 609 |
Processed data used by OIPC3 can be downloaded here.
Version 2.2 incorporates a minor correction allowing monthly OIPC output to display correctly in non-Microsoft browsers.
Verson 2.1 includes a minor code correction to the monthly OIPC re-enabling prediction for sites with non-integer latitude/longitude coordinates. This should not affect any of the other calcuations.
The OIPC has been migrated to a new server at Purdue University and updates have been applied to resolve problems with the monthly interpolation algorithm. The first change should be largely transparent to users, and should result in improved performance for most calculations. Although the new system seems to be working well at this point, there is always a possibliity that we'll run into some problems during the transition period. I ask that you bear with me, and please contact me to report any problems.
The second change has been made in an attempt to improve monthly average predictions by drawing on the much denser annual average dataset, and will affect the monthly OIPC estimates at some sites. As users have pointed out, OIPC monthly and annual predictions are incompatible for some regions, particularly those where the number and location of monthly and annual data stations is very different. In an attempt to resolve these inconsistencies and improve the overall accuracy of the monthly predictions, I have included an offset term that is applied to 'correct' all monthly estimates. In short, the offset term for a location <x,y> is calculated as the difference of the OIPC-predicted mean annual isotope ratio at <x,y> and the precipitation-weighted annual average isotope value calculated from the OIPC monthly predictions at <x,y>. This offset is added to each month's prediction in order to 'shift' the seasonal cycle to center on the annual average prediction. Conceptually, what I have done is assumed that monthly predictions give an accurate estimate of the relative values for each month, but that the annual prediction gives a more accurate indication of the absolute values at a given site.
We continue to increase the number of non-GNIP data in the OIPC database. At this time, these are mostly mean-annual observations. In addition to data for the United States culled from Kendall & Coplen (2001) and Friedman et al. (1992), we have added stations from local networks in Italy (Longinelli & Selmo, 2003) and Siberia (Kurita et al., 2004). Citations for these data sources can be found here, and the current distribution of stations can be seen here.
The OIPC has been updated to allow calculation of monthly average δ2H and δ18O values. The monthly interpolation model is of the same fundamental form as that used for mean annual estimates, but operates using long-term monthly average water isotope data from the GNIP database. Details and a complete description of the model are given in a paper in Oecologia (Bowen et al., 2005). Confidence intervals are not currently available for monthly values.
Other updates to the OIPC include new checks for errant lat/lon/alt values.
Original release, offering mean annual predictions and confidence intervals only.
