Water seepage in Hal Saflieni Hypogeum, Malta : tracing the source by chemical analysis

Abstract

The Hal Saflieni Hypogeum is a unique prehistoric underground temple, entirely hewn out of the limestone bedrock in the three successive storeys, which dates to the fourth millenium RC The preservation of this site, anti !Q.Y' particularly that of red ochre decorations which are present here, has since its discoery in 1902 been threatened by the infiltration of water from the surrounding area. This study was aimed at identifying the source(s) of wafer which causes the seepages. Twenty two sampling sites were chosen from various areas in the Hypogeum. Samples, taken over a period of 5 months, included water droplets from the ceiling and accumulated water from pools within the temple. Regular observations of the presence, or otherwise, of water were also made. The sampled water was analyzed for sodium, potassium, magnesium, calcium, sulphate, chloride and nitrate and for the presence of faecal coliforms. The same chemical analysis was also performed on solutions obtained by equilibrating limestone from the Hypogeum with demineralized water. Most of the seepage water sampled was similar to local tap water while some samples were substantially more dilute in their mineral content. This indicated that the main source of the seepage in the Hypogeum was tap water with a minor contribution from rain water. High nitrate content and the presence of faecal coliforms indicated that some of the sites were sourced, at least in part, by sewage water. In the sites where the seepage was perennial, the rate of infiltration was highest during the dry summer months. Sites of intermittent seepage were also mostly active during the summer time. As the temple is located in a heavily built-up residential area, we conclude that water seepage in the Hypogeum probably occurs in response to episodic irrigation of nearby domestic gardens using, mostly, tap water, but also reservoired rain water. The increased rate of seepage during the dry summer months is consistent with the concurrent increase in irrigation input.