Time-variation of hydrothermal discharge at selected sites in the western United States: implications for monitoring

Abstract

We compiled time series of hydrothermal discharge consisting of 3593 chloride- or heat-flux measurements from 24 sites in the Yellowstone region, the northern Oregon Cascades, Lassen Volcanic National Park and vicinity, and Long Valley, California. At all of these sites the hydrothermal phenomena are believed to be as yet unaffected by human activity, though much of the data collection was driven by mandates to collect environmental-baseline data in anticipation of geothermal development. The time series average 19 years in length and some of the Yellowstone sites have been monitored intermittently for over 30  years. Many sites show strong seasonality but few show clear long-term trends, and at most sites statistically significant decadal-scale trends are absent. Thus, the data provide robust estimates of advective heat flow ranging from ∼130 MW in the north-central Oregon Cascades to ∼6100 MW in the Yellowstone region, and also document Yellowstone hydrothermal chloride and arsenic fluxes of 1740 and 15–20 g/s, respectively. The discharge time series show little sensitivity to regional tectonic events such as earthquakes or inflation/deflation cycles. Most long-term monitoring to date has focused on high-chloride springs and low-temperature fumaroles. The relative stability of these features suggests that discharge measurements done as part of volcano-monitoring programs should focus instead on high-temperature fumaroles, which may be more immediately linked to the magmatic heat source.

Introduction

In this paper we review what is known about the natural time-variation in mass and heat discharge from selected hydrothermal systems in the western United States. Our focus on the United States in part reflects the geographic emphasis of the U.S. Geological Survey (USGS), but also the fact that much of the reliable data on time-variation of hydrothermal discharge derives from studies done in the western United States during the latter part of the 20th century. These data were collected for various purposes, including basic understanding of water–rock interaction (e.g. Fournier et al., 1975, Ingebritsen et al., 1994), environmental-baseline monitoring (e.g. Norton et al., 1989, Sorey and Colvard, 1994, Sorey and Colvard, 1997, Sorey et al., 1994, Friedman and Norton, 2000), volcano monitoring (e.g. Farrar et al., 1985, Sorey et al., 1998), and water-quality monitoring as part of the USGS National Assessment of Water Quality. Much of the data collection was driven by mandates to collect environmental-baseline data in anticipation of geothermal development. The data provide a quantitative basis for assessing the seasonal to multi-decadal variability of some types of hot-spring discharge. The findings are relevant to: (1) the design of environmental-baseline monitoring programs intended to protect geothermal features; (2) the value of hydrothermal-discharge monitoring as a component of comprehensive volcano-monitoring programs; and (3) heat-budget studies of areas with significant advective heat loss (e.g. Ingebritsen et al., 1989, 1994; Manga, 1998, James et al., 1999), which often assume that heat discharge from spring systems is constant.

Two electronic spreadsheets are an integral part of this report. These are accessible through a link labeled ‘hydrothermal discharge in the western United States’ under http://water.usgs.gov/nrp/proj.bib/ingebritsen.html. They include full details of all measurements from high-chloride spring (highclspringdat.PDF) and fumarolic areas (acidsulfatedat.PDF), metadata with complete descriptions of the sites and methods, and basic time-series plots for each site (Skoustad et al., 1979, Farrar et al., 1989, Fishman and Friedman, 1989, Friedman et al., 1993, Howle and Farrar, 1996). We cite these spreadsheets in support of some particular points in the report; interested readers can use the spreadsheets to do their own complementary analyses. An index at the beginning of each spreadsheet facilitates cross-referencing with text, figures, and tables.

The scope of our discussion is mainly limited to the western United States and further limited to natural (non-anthropogenic) variations in mass and heat discharge. Geothermal development commonly leads to rapid and profound changes in thermal features — typically large and semi-permanent reductions in liquid discharge (Henley and Stewart, 1983, Turner, 1985, White, 1992, Glover and Hunt, 1996, Glover et al., 1996) and more temporary increases in steam discharge (Allis, 1981, Henley and Stewart, 1983, Ingebritsen and Sorey, 1985). Human impact on geothermal features in New Zealand and the western United States has recently been reviewed by Glover and Hunt, 1998, Sorey, 2000, respectively.

Three types of thermal features are commonly distinguished: high-chloride springs; geothermal fumaroles and associated acid-sulfate springs; and volcanic fumaroles (Fig. 1).

High-chloride springs emerge from liquid-dominated hydrothermal systems, generally have a near-neutral pH, and are commonly high in silica as well as chloride. Well-known examples in the western United States include the springs of Upper and Lower Geyser Basins, Yellowstone. Such features often occur in valleys near streams that eventually capture most of the thermal fluid, so their total discharge can often be gauged by measuring the solute flux in these adjacent streams. Discharge time-series from high-chloride spring systems in the western United States are relatively detailed and abundant.

Geothermal fumaroles (steam vents) and associated acid-sulfate springs are derived from steam up-flow generated by boiling of meteoric waters. Evolution of CO₂ and H₂S with the steam, and subsequent partial condensation and dissolution in shallow groundwater, leads to acid-sulfate springs with low pH, typically high dissolved CO₂, and variable sulfur (White et al., 1971). Well-known examples in the western United States include the thermal features of Lassen Volcanic National Park and Yellowstone's Mud Volcanoes. The total fumarolic and ‘steam-heated’ (acid-sulfate-spring) discharge is best measured by using total heat discharge as a proxy. Measurement of the multiple modes of heat discharge is time-consuming and difficult, and time series of fumarole and steam-heated discharge are sparse and rare, both in the United States and worldwide.

Volcanic fumaroles and associated springs occur where magmatic fluids reach, or nearly reach, the land surface. Whereas geothermal fumaroles are limited to <∼160°C — the temperature obtained by adiabatic decompression of saturated steam of maximum enthalpy — volcanic fumaroles often reach temperatures of hundreds of degrees Celsius. The pH and composition of any associated springs depends upon the degree of interaction between the magmatic steam, meteoric water, and wallrock (Hedenquist, 1995, Reed, 1997). Along the western rim of the Pacific Ocean, high-temperature volcanic fumaroles are relatively common, and associated springs are often high in both chloride and sulfur. In the United States, volcanic fumaroles occur along the Aleutian arc, on the Alaska Peninsula, and in the Cascade Range, but they are relatively rare and weak, and associated springs are usually low in chloride. Many volcanic fumaroles are relatively inaccessible and their surroundings potentially dangerous. There are few reliable measurements of total mass or heat flux except in special circumstances where magmatic steam condenses into crater lakes, which also act as calorimeters (e.g. Brantley et al., 1993, Varekamp and Rowe, 2000).

The same monitoring methods apply to both geothermal and volcanic fumaroles. Therefore, for purposes of the discussion that follows, we will consider only two broad categories of hydrothermal features: high-chloride springs (highclspringdat.PDF) and fumaroles (acidsulfatedat.PDF). We will touch on the distinction between geothermal and volcanic fumaroles again in Section 4 that concludes this paper.

The history of some developed hot springs in Europe indicates no dramatic changes in discharge, temperature, or chemical composition over millennia of casual human observation. Waring (1965), who published a global inventory of thermal springs, wrote that “(m)any hot springs have been described as remarkably uniform in temperature, flow, and mineral content,” and noted that “(a)s most observations of the temperature and flow of thermal springs have been made at intervals of many years, no trends in their changes have been established.”

In his monumental effort, Waring (1965) inventoried 1185 thermal-spring localities in the United States, mainly in California, Idaho, Nevada, Oregon, and Wyoming. Though he listed flow rates for most of the springs, a general lack of methodological detail leads us to regard the flow-rate data as non-quantitative. We will use Breitenbush Hot Springs, Oregon, as an example: for Breitenbush, Waring (1965) listed a flow rate equivalent to 60 l/s, and cited Langville et al. (1903) as a reference. This value, along with many others from Waring (1965), was repeated in USGS Circulars that assessed the geothermal resources of the United States in 1975 and 1978 (White and Williams, 1975, Muffler, 1978). However, Langville et al. (1903) listed no flow-rate data for Breitenbush; Waring (1965) provided no independent support for the 60 l/s value; and solute-inventory measurements made in the Breitenbush River in 1984–99 documented a flow rate of 12.5±1.5 l/s (n=6). The constancy of the data in the 1980s and 1990s, combined with the lack of supporting detail in Waring (1965), prompts us to discount the much higher reported value.

Most flow rates reported in the early literature were likely based on non-quantitative visual observation, sometimes combined with direct measurement of individual orifices. We place much more confidence in flow rates determined more recently by solute-inventory or heat-flux measurements. A few of the measurements reported by Day and Allen, 1925, Allen and Day, 1935 in the early 20th century are sufficiently detailed to provide an earlier point of comparison.