Infrared thermography reveals weathering hotspots at the Požáry field laboratory

Abstract

Evaluating physical properties and mechanical parameters of rock slopes and their spatial variability is challenging, particularly at locations inaccessible for fieldwork. This obstacle can be bypassed by acquiring spatially-distributed field data indirectly. InfraRed Thermography (IRT) has emerged as a promising technology to statistically infer rock properties and inform slope stability models. Here, we explore the use of Cooling Rate Indices (CRIs) to quantify the thermal response of a granodiorite rock wall within the recently established Požáry Test Site in Czechia. We observe distinct cooling patterns across different segments of the wall, compatible with the different degrees of weathering evaluated in the laboratory and suggested by IRT observations of cored samples. Our findings support previous examinations of the efficacy of this method and unveil correlations between cooling phases in the field and in the laboratory. We discuss the scale-dependency of the Informative Time Window (ITW) of the CRIs, noting that it may serve as a reference for conducting systematic IRT field surveys. We contend that our approach not only represents a viable and scientifically robust strategy for characterising rock slopes but also holds the potential for identifying unstable areas.

Keywords: Cooling rate index; Informative time window; Infrared thermography; Porosity; Rock mass; Slope stability.

Figure 1

(a) Location of the study area in Czechia, Europe; (b) the Požáry Test Site is identified by a red rectangle (map sources: Wikimedia Commons and Czech Cadastral Office); (c) overview of the west-facing rock wall before the installation of the monitoring system. The maps were produced in ArcMap v. 10.8 (ArcGIS Desktop, ESRI, https://www.esri.com/en-us/home); post-processing was done in CorelDRAW 2019 (https://www.coreldraw.com/); the photograph of the site was taken by O. Racek.

Figure 2

Thermograms were recorded at the slope scale where substantial thermal anomalies appeared, suggesting different behaviours within the rock mass. ROIs were selected based on the visual anomalies of the acquired images.

Figure 3

IRT time-lapse monitoring of cooling phases of cored samples collected at the ROIs and analysed in a climate chamber. Air A, B and C refer to the chamber’s air temperature throughout the three distinct experiments.

Figure 4

(a) Cooling phase of the three ROIs (A, B, C) after the peak temperature (15:50) till the lowest temperature; (b) detail of the first four hours (300 min) of the cooling phase and the time of the inflection point for all the ROIs, reached at 18:00.

Figure 5

Changes in CRIs, computed on a 10-min basis from the beginning of the cooling (15:50) at the different ROIs. The red-coloured dashed line shows the ITW or inflection point after which the use of CRI progressively yielded useful and comparable information in the field. Especially, the red line signs the peak in CRI values after 130 min (18:00) from the peak in temperature, while the yellow line signs the end of 60 min as a straight comparison with the laboratory tests from the peak in temperature. The definition of ITW for a longer period after the inflection has a rigorous mathematical basis, established by the progressively linear decrease of CRI values (Fig. S6).

Figure 6

(a) Linear regressions of CRIs and porosity in the field. It corresponds to the values along time series CRI(t) from t = 130 min to t = 300 m after peak of temperature; (b) comparison of field and laboratory results in the 130–300 min interval for the field. The best correlation was found between CRI_lab60min and CRI_field300min. CRI for laboratory samples is computed uniformly for t = 60 min, which revealed the best correlation.

Figure 7

Cooling rate indices vs. porosity for (a) the field and (b) the laboratory.

Figure 8

Regressions between CRI_60min and rock sample properties.

Figure 9

Performance of measure porosity in the field and predicted porosity from the method proposed (n_measured / n_predicted) using a combination of Eqs. (3) and (4). The performance is plotted as a ratio, and values close to the unit represent predictions fitting the input values.