Medicine Lake Volcano:
A Single Long Traverse to Determine Regional Deformation

A 20-km traverse across the summit area of Medicine Lake volcano was measured during August 1988. This first-order class II survey repeated a second-order survey made in 1954 by the National Geodetic Survey. After it was recognized that substantial subsidence had occurred between the two surveys and that the deformation field extended well beyond the 1988 traverse, a more amibitious leveling effort was conducted in the summer of 1989. The 1989 survey crossed Medicine Lake volcano and the Stephens Pass area, where intense swarms of shallow earthquakes occurred in 1978 and 1981. Closure for the 1989 survey was 19 mm around the 193-km circuit, well within NGS requirements for first-order class II surveys (5 mm * L^1/2 = 69 mm; Vanfcek and others, 1980). Approximately 80% of the 1954 benchmarks along the route were recovered, and missing marks were replaced to maintain an average spacing of 1.5 km.

The most obvious feature of the vertical displacement data is an area of pronounced subsidence (maximum 389 +/- T502, corresponding to an average rate of 11 mm/yr) in the summit area of Medicine Lake volcano, near the center of the caldera rim and across the entire volcanic edifice. In addition, there is evidence in several places along the leveling route for localized faulting, presumably associated with historical crustal extension in the western Basin and Range tectonic province (for example, near M500, Y500, V501, and C500).

The apparent inverse correlation between topography and vertical displacements raises the possibility that the calculated displacements are contaminated by systematic error in one or both of the leveling surveys. For example, if the leveling rods used for the 1954 survey were poorly calibrated or the rod corrections were incorrectly applied, error would have accumulated as a function of elevation. As a result, the difference between the 1954 and 1989 surveys would indicate elevation-dependent displacements. This possibility is currently under investigation, but at present the calculated displacements are throught to be correct because (1) the magnitude of the rod-correction error that would be required to explain the displacements is implausibly large (389 +/- 43 mm in 600 m elevation change is approximately 6 x 10^-4); (2) the correlation between topography and displacements breaks down in detail, particularly in the Stephens Pass area; and (3) a general correlation between topography and vertical displacements makes sense if the displacements are caused by isostatic loading, magmatic deflation, or some other volcanogenic process.

The area of localized subsidence near C500 (about 20 cm with respect to nearby marks) is within the epicentral area of the 1978 Stephens Pass earthquake swarm, which was accompanied by ground breakage (Cramer, 1978; Bennett and others, 1979). Although only one bench mark is involved, both its undisturbed appearance and proximity to the 1978 epicenters suggest that displacements was caused by faulting rather than bench-mark instability or cultural disturbance. Unfortunately, several nearby marks were destroyed by road work during the period between the leveling surveys. Thre grabenlike feature near V501 probably is associated with a swarm of shallow earthquakes that occurred just north of Tennant during 1981. No historical seismicity has been recognized that might correlate with the area of localized faulting near M500 and Y500, but that area is cut by many young faults that bound the Whitehorse Mountains and horst-and-graben terrain to the west.

The cause of the 1954-1989 subsidence at Medicine Lake volcano is still under investigation, but almost surely there are both tectonic and volcanic aspects to the problem. Crustal thinning owing to Basin and Range extension, isostatic loading by the weight of the volcanic pile, and processes within a possible magma chamber beneath the summit area (Evans and Zucca, 1988) may each play a role.