Periodic changes in effluent chemistry at cold-water geyser: Crystal geyser in Utah

Highlights

• Changes in effluent chemistry coincide with the eruption periods at Crystal geyser.

• The Entrada Sandstone supplies more groundwater than previously hypothesized.

• The minor and major eruption periods are driven by different sources and mechanisms.

Abstract

Crystal geyser is a CO₂-driven cold-water geyser which was originally drilled in the late 1930’s in Green River, Utah. Utilizing a suite of temporal groundwater sample datasets, in situ monitoring of temperature, pressure, pH and electrical conductivity from multiple field trips to Crystal geyser from 2007 to 2014, periodic trends in groundwater chemistry from the geyser effluent were identified. Based on chemical characteristics, the primary sourcing aquifers are characterized to be both the Entrada and Navajo Sandstones with a minor contribution from Paradox Formation brine. The single eruption cycle at Crystal geyser lasted over four days and was composed of four parts: Minor Eruption (mEP), Major Eruption (MEP), Aftershock Eruption (Ae) and Recharge (R). During the single eruption cycle, dissolved ionic species vary 0–44% even though the degree of changes for individual ions are different. Generally, Na⁺, K⁺, Cl⁻ and SO₄²⁻ regularly decrease at the onset and throughout the MEP. These species then increase in concentration during the mEP. Conversely, Ca²⁺, Mg²⁺, Fe²⁺ and Sr²⁺ increase and decrease in concentration during the MEP and mEP, respectively. The geochemical inverse modeling with PHREEQC was conducted to characterize the contribution from three end-members (Entrada Sandstone, Navajo Sandstone and Paradox Formation brine) to the resulting Crystal geyser effluent. Results of the inverse modeling showed that, during the mEP, the Navajo, Entrada and brine supplied 62–65%, 36–33% and 1–2%, respectively. During the MEP, the contribution shifted to 53–56%, 45–42% and 1–2% for the Navajo, Entrada and Paradox Formation brine, respectively. The changes in effluent characteristics further support the hypothesis by Watson et al. (2014) that the mEP and MEP are driven by different sources and mechanisms.

Introduction

In the field of geothermal and CO₂-driven cold-water geysers, determining the physical and chemical mechanisms that cause periodic eruptions has been a common objective (Gouveia and Friedmann, 2006, Han et al., 2013, Hurwitz et al., 2012, Hurwitz et al., 2014, Ingebritsen and Rojstaczer, 1993, Karlstrom et al., 2013, Kieffer, 1989, Lu et al., 2006, Vandemeulebrouck et al., 2013, Watson et al., 2014). Both geothermal and cold-water geysers behave similarly in certain degrees; for example, they eject effluent with regular periodic eruption patterns, and in the end, will potentially transit to either fumaroles or springs once the subsurface energy is consumed. Regardless of these similarities, geothermal and cold-water geysers have their own characteristic eruption mechanisms. The primary difference is from the driving force; geothermal geysers are driven by the accumulation of thermal energy which triggers gas expansion and eruptions (Munoz-Saez et al., 2015). With respect to cold-water geysers, eruptions are not associated with a thermal effect, but result from a change in hydrostatic pressure. Reduction in hydrostatic pressure initiates a positive feedback system of CO₂ bubble exsolution, growth, and buoyant acceleration (Han et al., 2013, Lu et al., 2005, Watson et al., 2014). These multiple positive feedbacks drive geyser eruptions. Recently, in-geyser video observations and data collections have provided great insights into the operations of these geysers by identifying the configuration of geyser conduits/chambers and boiling/flashing depths (Belousov et al., 2013, Han et al., 2013, Hutchinson et al., 1997, Munoz-Saez et al., 2015, Watson et al., 2014).

Despite extensive work on determining how geysers operated physically, very few researchers focused on monitoring temporal variations in geyser effluent chemistry during eruption cycles (Hurwitz et al., 2012, Kampman et al., 2014, Ladd and Ryan, 2016, Noguchi and Nix, 1963, Shipton et al., 2004). The first attempt to monitor changes in effluent chemistry during an eruption cycle was conducted by Noguchi and Nix (1963) at Old Faithful geyser in Yellowstone National Park in June 23, 1962; the recorded eruption characteristics such as eruption height, duration, and interval were 35–52 m, 4 min, and 66.3 min, respectively. During the 4-min eruptions, variations in geyser effluent chemistry were distinct, which included immediate increase in Cl⁻, alkalinity, and SiO₂ concentrations followed by a gradual decay. Based on the field observation, Noguchi and Nix (1963) concluded that mixing of salty thermal groundwater and dilute groundwater occurred during the geyser eruption. More recently, Hurwitz et al. (2012) attempted to understand how chemical and physical relationships are tied together during geyser eruptions at Old Faithful, Daisy, Grand, Oblong, and Aurum in Yellowstone National Park. For this purpose, temporal variations in geyser effluent chemistry and eruption intervals were correlated, but no significant trends or strong correlations between effluent chemistry and geyser eruption patterns were noted. Additionally, based on observation of long-term data since 1884 (>120 years), Hurwitz et al. (2012) concluded that effluent chemistry from Old Faithful geyser has remained consistent despite the large variation in geyser eruption intervals.

Similar to the endeavor at Old Faithful geyser, Crystal geyser in Utah, a CO₂-driven cold-water geyser, was also investigated to determine variations in geyser effluent chemistry during eruptions. The eruption cycle at Crystal geyser has shown considerable variation since it was drilled in 1936; documentation of the eruption cycle can be found in Baer and Rigby (1978) and Gouveia and Friedmann (2006). The most extensive and recent characterization was found in Han et al., 2013, Watson et al., 2014 who inserted in situ pressure and temperature sensors within the geyser. While the eruption cycle at Crystal geyser has varied historically, the main constituents such as the minor eruption period (mEP), major eruption period (MEP) and recharge (R) maintained consistent; typical eruptions of the Crystal geyser in recent days were recorded (Supplementary video). Specific to the chemical observations of geyser effluent, Shipton et al. (2004) was the first to collect geyser effluent during an eruption, which revealed the reduction of Na⁺ and K⁺ concentrations during the eruptions. Later, Kampman et al. (2014) conducted further sampling of Crystal geyser’s effluent and used ternary mixing diagrams to show that deep Paradox Formation brine and groundwater from shallow Jurassic aquifers (Navajo and Entrada Sandstones) dynamically fed Crystal geyser. In addition to the temporal fluid sampling at Crystal geyser, Kampman et al. (2014) drilled a well (CO₂W55) 90 m north of the southern Little Grand Wash (LGW) fault trace and 285 m west of Crystal geyser. Fluid samples were collected over ∼325 m depth from the Entrada (98 m), Carmel (188 m), and Navajo Sandstones (206, 224, 276 and 322 m).

Building upon the initial effort at Crystal geyser, we here present dynamic and periodic variations of groundwater chemistry during multiple eruption cycles utilizing both in situ continuous measurements (pressure, temperature, electrical conductivity and pH) and geyser effluent samples from 2007 to 2014. Utilizing a suite of temporal groundwater chemistry datasets, systematic trends in geyser effluent chemistry have been revealed; Crystal geyser exhibited repeating dramatic changes in groundwater chemistry that coincided with its own unique eruption cycle. The evolving characteristics of effluent chemistry suggest that the source aquifers (Jurassic Entrada and Navajo Sandstone aquifers and deep Paradox Formation brine) are supplying groundwater to Crystal geyser throughout the eruption cycle. In addition, to determine the fractional contribution from the multiple aquifers during each eruption period, geochemical inverse modeling utilizing defined fluid end-members was conducted, which revealed new insights into the subsurface dynamics governing the eruptions at Crystal geyser.