Use of correlation and stepwise regression to evaluate physical controls on the stable isotope values of Panamanian rain and surface waters

To investigate the climatic controls on the stable isotope composition of surface waters from Panama, we measured δ18O and δD on river and lake waters collected from throughout the isthmus. We statistically analyzed the isotope data using correlation (n=162 samples) and multiple stepwise regression (n=148 samples). Variables are δ18O, δD, deuterium-excess (dx), latitude, longitude, sample elevation, stream head elevation, median stream elevation above sample site, distance from the Caribbean, stream length above sampling site, estimated mean annual precipitation, and pH. Rainfall isotope data from Panama City were analyzed for temporal variability. Temporally, rainfall δ18O values are negatively correlated with precipitation amount. Spatially, surface water δ18O values decrease with distance from the Caribbean, an indication of progressive rainout as air masses traverse the isthmus. The Panama surface water line is defined as δD=(7.6±0.09)×δ18O+(10.1±0.6)δD=(7.6±0.09)×δ18O+(10.1±0.6), statistically identical to that from Costa Rica. Respective Panama meteoric water lines based on monthly and annual data are δD=(7.4±0.07)×δ18O+(5.5±0.3)δD=(7.4±0.07)×δ18O+(5.5±0.3), and δD=(8.2±0.60)×δ18O+(10.6±3.3)δD=(8.2±0.60)×δ18O+(10.6±3.3). We conclude that the dominant control on rain and surface water δ18O values associated with the Central American monsoon are temporal and spatial amount effects. Surface water δ18O values also show significant correlations with drainage basin parameters such as median and stream head elevation, and latitude and longitude. An equation derived from stepwise multiple regression relates δ18O values to various physical parameters and explains 74‰ of the observed isotopic variation. Weaker correlations were found between deuterium excess and physiographic variables. Our results support the use of spatial sampling of surface waters for use as a rainfall δ18O proxy, and will benefit ongoing paleoclimatic research in humid tropical regions.