ISSN:
1573-0581
Keywords:
thermal conductivity
;
heat flow
Source:
Springer Online Journal Archives 1860-2000
Topics:
Geosciences
,
Physics
Notes:
Abstract Heat flow data provide constraints on the thermal structure and evolution of the oceanic lithosphere. Because precise determination of the heat flux requires that both the thermal gradient and the thermal conductivity be well determined we have examined the thermal conductivities used in a new Pacific Basin heat flow data set. ∼43% of the ∼1600 heat flow determinations rely on values estimated by various methods, rather than directly measured. Although the measured and estimated conductivities have comparable means, the measured conductivities have a standard deviation ∼50% larger than the estimated, suggesting that the estimated values underestimate the actual variation. We investigate the limitations of using such estimates by examining factors controlling the variations of measured conductivity values. We find that the variation between the closest adjacent sites increases with increasing separation, such that sites within 200 km are on average noticeably closer in conductivity than sites further apart. Contributing to this effect may be the variation of conductivity with lithology (with mean conductivity highest for carbonate oozes, intermediate for deep-sea clays, and least for siliceous oozes) and a possible trend of decreasing conductivity with increasing seafloor depth. Tests with the measured data suggest that the best method for estimating conductivity is using the mean value measured within 200 km. The mean of a larger geographical region is a somewhat poorer predictor, and using the oceanwide mean and the value at the nearest site are poorer still. Approximately 29% of the estimated values were not based on measurements from a reference site. For most others, the reference site was the nearest measurement from the same cruise, typically a large distance away. For those sites where conductivity was not measured, 78% had measured conductivity within 200 km and were reestimated using the local mean, whereas the remaining 22% were reestimated using the regional mean. The resulting change in the estimated conductivity averaged ∼9% using the local mean and ∼6% using the regional mean. We suggest that such a procedure be used to improve the utility of the heat flow data set, as an alternative to discarding the large fraction of the available data that does not incorporate measured conductivities.
Type of Medium:
Electronic Resource
URL:
http://dx.doi.org/10.1007/BF00366281
Permalink
