Spatiotemporal rock moisture distribution at the medieval cave town of Uplistsikhe, Georgia

Moisture and salts cause considerable damage to built and rock-hewn heritage. Rock moisture is a key factor for numerous decay processes, but there is limited knowledge of salt and moisture distribution because measurements of spatial and temporal moisture distribution still remain challenging. The medieval cave town of Uplistsikhe (Georgia) is hewn out of very soft Lower Miocene sandstone and is a typical example of a heritage site suffering from progressive decay. We present data on moisture and salt distribution derived from a multi-method approach, including microwave sensor monitoring (MW-mon; continuously over 2 yrs), microwave handheld sensors (MW), 2D-resistivity profiles (ERT), rock sampling by drilling, and salt extraction by paper pulp poultices (PPP).

Microwave monitoring was applied for the first time (to our knowledge) in a long-term monitoring of heritage sites. We used equipment from hf-sensor (Germany) with two types of microwave reflectivity sensors penetrating approx. 7 cm and 13 cm deep, respectively. The sensors were installed inside and outside of two prominent caves (Grand Hall and Long Hall). MW, ERT, PPP and drilling were carried out in four caves (the two mentioned plus Blackberry Hall and Teatron). Careful laboratory calibration using samples from the site was necessary to produce quantitative results for MW-mon, MW and ERT.  

MW-mon showed pronounced annual fluctuation with highest moisture saturation occurring in summer. The moisture maximum in the caves lags 2 months behind the spring precipitation maximum and might be partly caused by air humidity condensation amplified by salts. Heavy rainfall events cause additional moisture pulses by seeping through the rock or by capillary rise. Spatial moisture distribution derived from MW shows relatively dry rock outside the caves and different patterns of moisture ingress into the caves: Capillary rise from the base, ingress through fractured or otherwise water-permeable areas of the roofs or back walls. The spatial patterns are confirmed by ERT; however, calculated moisture saturation differs between MW and ERT due to electrical conductivity effects of salty pore water.

All drill samples from the caves are significantly saltier on the respective surfaces, which points to the rate of evaporation being smaller than the outward migration of salts. Outside the caves, flaking of thicker layers (several cm) point to deeper layers of salt concentration caused by higher evaporation from the surface; flaking at the "lips" above the caves is probably also amplified by stronger temperature and moisture fluctuations. The main ions everywhere are  Ca²⁺ and SO₄^2- (subordinate K⁺) while at the strongly flaking surfaces of Grand Hall, Na⁺, Cl^- and NO₃^2- are also present. Summing up, the results show very diverse and complex patterns of moisture and salt distribution at an apparently homogeneous site.