TRACING SOURCES AND QUANTIFYING EVAPORATION IN CENTRAL TEXAS SURFACE WATERS USING STABLE ISOTOPES AND CONDUCTIVITY

This study models water source mixing and evaporation in the Brazos River and its large tributary, the Navasota River, in Central Texas. The Brazos River from Waco to College Station, Texas is thought to generally gain baseflow from shallow groundwater while carrying dissolved salt largely from Lake Whitney. The Navasota flows from Lake Limestone through a shallow aquifer and into the Brazos River downstream of the study region. River water δ¹⁸O, δD, and conductivity were measured bi-weekly in Brazos County, Texas from January 2012 through August 2013. The Brazos shows a positive conductivity-δ¹⁸O relationship, varying seasonally. The Navasota shows a significant negative conductivity-δ¹⁸O trend during spring and summer when water conductivity increases from Lake Limestone (230 µS) to Brazos County (average 650 µS). The Brazos River δ¹⁸O-δD data fall along a local evaporation line (LEL) δD = 5.66 * δ¹⁸O – 2.47 (r² = 0.95), while the Navasota River LEL is δD = 6.27 * δ¹⁸O + 1.97 (r² = 0.89). These trends intersect the global meteoric water line (MWL) to give average rain δ¹⁸O of -5.33‰ for the Brazos River and -4.64‰ for the Navasota River, reflecting the more northern extent of the Brazos basin, while the greater slope of the Navasota LEL reflects higher average humidity. A simple mass-balance model predicts that water percent evaporation and δ¹⁸O enrichment during low flow are 2.8% and 0.40‰, meaning the alluvial aquifer is not necessarily dominating the Brazos River water supply during low flow. Percent evaporation of original rain sampled as river water was estimated with both Rayleigh distillation and the method from Gonfiantini (1986) based on Craig and Gordon (1969), while assuming 1) a closed system with an atmospheric exchange component and 2) the above δ¹⁸O and δD values of average rain for each river. The Gonfiantini (1986) model showed low sensitivity (< 5%) to humidity. Rayleigh distillation output is consistently ~5% greater than output from Gonfiantini’s method. Modeled percent evaporation of original rain in the Brazos River in the summer (JAS; 16-33%) is substantially greater than in the Navasota River (JAS; 3-23%). This may be due to greater humidity over the Navasota, or a greater relative contribution of groundwater to the Navasota than in the Brazos River.