Login
Miniature temperature data loggers, metrological traceability of measurements, thermal properties of rocks. An open-air laboratory to study one of the most important triggering factors of rockfalls in the European Alps: temperature; permafrost and a metrology lab at 3000m.
A growing interest of the climatological community is addressed towards mountain environment and mountain monitoring sites. A number of observatories have been providing relevant observational data for decades. Some mountain observatories even take part in the group of WMO recognised centennial stations. Mountain sites are usually involved in scientific research and normally provide high quality data. Climate change has a stronger impact on the high mountain environment: in the Alps glacier retreat is quicker than expected, permafrost thawing occurs at increased depths, rockfalls and instabilities show greater frequency. A multitude of measurements are performed in the Alps to monitor and record such amplified climate signals. Comparability and data quality can be achieved only through establishing traceability and evaluating uncertainties. Metrology now supports such climate studies with dedicated calibration procedures, transportable systems and specific uncertainty analysis.
Rock temperature in the Alps: a new measuring approach
In a context of climate change that affects the glacial/periglacial environments of the Alps, landscapes are evolving and new hazards are emerging (Fig 1).
Changes in rock extreme temperature values are the most important conditioning and triggering factors for rockfalls that occur in these environments (Fig 2).
In order to improve the knowledge about the thermal properties of rocks responsible for the development of rockfalls in the Alps, in 2016 we started a series of temperature measurements. The research was carried out in our open-air laboratory, located in the Bessanese glacial basin in Balme, Italy.
The main objective of the research is to develop the heat transfer model for each type of rock. The models will be applied to estimate rock temperature in non-instrumented sites, where only the lithotype and the air temperature are known. These models will be used to assess temperature trends in near real time and will be applied to the mountain slopes affected by rockfalls.
For this research we adopted a new measuring approach. This approach consists of the application of an innovative methodology integrated by the use of advanced technologies. The main innovative aspect of our approach is the goal to achieve fully documented measurement traceability. This is obtained in three specific steps:
1) Selection of good quality instruments and sensors, taking into account the condition of use, the possibility to store data for long periods (one year or more), with the possibility to be calibrated several times; 2) Definition of specific calibration methodologies, procedures and the sensors calibration with associated calibration curve and uncertainty; and 3) Evaluation of field measurement uncertainty. In this research field, the metrological traceability of measurements is fundamental for data comparability in space and in time, as reported in our previous 2014 article for AWE International: Tracing Glacier Retreat.
To measure rock temperatures we used 20 MadgeTech MicroTemp Data Logger (MT). The MT is a miniature, submersible, self-contained temperature data logger, 66mm long and 18mm in diameter. Such thermometers work in the temperature range -40°C ÷ 80°C, with a resolution of 0.1°C and can be easily inserted into the rock. MTs can measure and record data for more than one year with user-replaceable batteries. After a measurement period, the MTs can be recalibrated, both in the laboratory and on-site.
Taking into account the MTs calibration adopted, which minimises the difference from the calibration conditions, including the tests, characterisation and calibration results, a tentative measurement uncertainty budget was evaluated. The overall measurement uncertainty, excluding drift and shocks, accounts to 0.098°C which is rounded to 0.1°C, and also according to the instrument resolution, which becomes a correct and evaluated indicator of the measurement uncertainty.
After calibration, the MTs have been inserted in different types of rocks, with different slope/aspect conditions and at different depths (10, 30 and 50 cm). The MTs inserted in the rocks have been smeared with white silicon grease to favour the contact between the rock and the MT, excluding the thin layer of air, obtaining the physical state that the thermal inertia of the MT is equal to the thermal inertia of the rock and guaranteeing a correct thermal conductivity (Fig. 3). This is also the condition met in a liquid bath during calibration, thus making the calibration curve and associated uncertainty more representative of the measurement conditions. Data acquisition by the MTs started on July 20, 2016 and is still ongoing, without missing data. Temperatures are recorded every 30 minutes (instant readings): in this way, it is possible to synchronise rock and AWS air temperature data and consequently make accurate comparisons. To monitor the glacial basin and the environmental conditions (in particular snow cover and solar irradiation) of the monitoring sites, we have installed a 360° webcam on the left moraine of the glacier, at 2775 m a.s.l. (https://bessanese.panomax.com/).
For air temperature data, we use the Rifugio Gastaldi Automatic Weather Station (AWS), 2659 m a.s.l., localised close to temperature monitoring sites. The AWS is active since 1988 (owned by ARPA Piemonte). The climate of the Bessanese glacial basin (observation period 1990-2019) shows a mean annual air temperature of 0.5°C. The hottest month is August (8.1°C) and the coldest one is February (-5.9°C). The extreme minimum and maximum temperature observed during the observation period are respectively -25.9°C (27 February 2018) and 21.2°C (11 June 2017). Also in this glacial basin there is a clear warming trend. The annual mean temperature is increasing by 0.5°C/10 years (0.8°C/10 years in summer).
Furthermore, the main physical properties of the different types of rocks (bulk density, specific heat, thermal conductivity and colour) have been determined in our “Criosfera-Clima” laboratory. These parameters will be used for the development of the heat transfer model.
Figure 4 shows an example of rock and air temperature records, acquired during two consecutive sunny days in calcschist rock. The trends observed in calcschists are similar to those of the other lithotipes, but several differences in the quantitative values have been found. That’s why it is necessary to acquire data for different lithotypes. This figure highlights the differences between air and rock temperature, but also it highlights the differences among the rock temperature recorded at different depths. These differences are important both in terms of maximum values reached and in terms of time required to reach these values, and are crucial for the definition of the heat transfer model.
“reference network can originate in a harmonised way, for generating long term rock temperature datasets”
This brief report shows a first example of the application of procedures aimed at the metrological characterisation of the sensors used for measuring rock temperature in glacial/periglacial environments. This is a novelty in studies of rockfall that occur at high elevation sites. In this research field, many instrumented experimental sites in the Alps have been equipped during the last 20 years, following independent approaches and different solutions. This results in a low level of data comparability, very different technical methods adopted, total absence of measurement guidelines or discussed dedicated calibration procedure. Full understanding and evaluation of measurement uncertainty is not present in literature, and rarely reported results include at least instrumental uncertainty evaluation. When general principles will be formally adopted, a reference network can originate in a harmonised and comparable way, for generating long term rock temperature datasets.
Field calibration campaign for permafrost temperature sensors
The permafrost, the portions of Earth’s surface where water is in solid form for two consecutive years, influences energy exchanges, hydrological processes, water availability, and gas emissions. Evaluating the degradation of permafrost is one of the major challenges in understanding global warming and its impact on the cryosphere1.
The Global Cryosphere Watch, a World Meteorological Organization programme, is promoting actions towards data quality and traceability, to achieve comparability of observations from different permafrost stations. In response to this, within European Project MeteoMet2, a transportable system for on-site calibrations of permafrost temperature sensors was studied, developed and tested in field to monitor temperature profiles in permafrost boreholes. This objective has been motivated by needs and background provided by operators in cryospheric measurements. For instance, the transport itself, from remote and hard-to-reach areas to the laboratory can cause mechanical shocks to the sensors; the overall effect of the quantities of influence cannot be fully reproduced in laboratory during the calibration process. Strong cold winds, high radiation, low temperatures, quick changes in meteorological conditions are examples of factors not encountered in laboratory, where the room conditions must be controlled and stable. A calibration carried out in the normal operating conditions of the whole equipment allows the calibration curve to be more representative of the measuring process.
Measurement needs and adopted techniques
Monitoring of permafrost is normally performed by inserting thermometers chains at various depths in boreholes drilled in the frozen soil. Detecting the 0°C transition due to the thawing process of the water in the frozen permafrost soil is of key importance in understanding permafrost active layer thickness and its degradation. Moreover, it is not uncommon for permafrost chains to show thermal trends which are comparable in amplitude with sensors’ drift. For this reason, calibration of instruments and sensors, accurate measurements of the permafrost properties and improved data quality become fundamental aspects for achieving full traceability of the measurements and more reliable knowledge on the evolution of this component of the cryosphere.
The transportable calibration system for permafrost temperature sensors has been designed to meet the target calibration uncertainty when operating in field, also in difficult conditions.
The temperature sensors are usually calibrated by putting them in direct contact with the liquid medium inside the thermostatic bath, we found this method unsuitable for our purposes. The large number of sensors (25~30) put simultaneously into the bath introduces a series of uncertainties: the resistive sensors heat themselves due to the electric current; large batch of sensors affects its thermal homogeneity and the stability of the bath could be insufficient. For all these reasons, a copper comparator block was realied to be placed inside the thermostat volume. The block was equipped with slots of different sizes in order to accommodate as many sensors as possible, including reference thermometers.
Three reference sensors (Platinum Resistance Thermometers – PRT 100 Ω) were used in conjunction with a high-accuracy resistance bridge. The PRTs were calibrated at the Italian National Metrology Institute (INRiM) laboratories by comparison against a Standard Platinum Resistance Thermometers with an evaluated expanded uncertainty at 0.7mK. The PRTs are checked before and after the field calibration for possible drifts.
On-site calibration activity
The on-site calibration campaign was planned in August 2018 in cooperation between INRiM and ARPA (Agenzia Regionale Protezione Ambiente), for the temperature sensors used at the Sommeiller permafrost monitoring station, located at about 3,000 m a.s.l in N-W Italy (Figure 5).
The Sommeiller station consists of three boreholes of 5m, 10m and 100m depth, equipped with thermometric chains composed by 2, 12 and 22 thermistors, respectively, with a resistance of 100 kΩ at 25°C. The permafrost temperature curves show a degradation of the permafrost base at approximately 58m of depth since 2014, with a variation of the temperature of about 0.25°C from 2012 to 20163 (figure 7a – left). Thirty-two permafrost temperature sensors (coded PS) were extracted from the 10m and 100m boreholes and inserted in the copper comparator block, along with the three PRTs (figure 6). Five temperature points centered at the freezing point of water were selected: (-5, -3, 0, 3, 5) °C. Data was considered valid for the analysis when temperature stabilities were measured as 2-5 mK by PRT and maximum 14 mK by PS over one hour and then recorded for 30 min. This procedure was repeated at each point (nominal value).
Figure 6 Inside the shelter: preparing the 32 thermometers for the calibration in the liquid bath, by comparison with reference standards.
The calibration uncertainty budget, both statistical and instrumental (considering the associated probability distribution) includes: PRT calibration uncertainty and repeatability; reference standard self-heating; thermometry bridge resolution; copper comparing block homogeneity; temperature stability, resolution, residuals and interpolation of sensors in calibration.
The total calibration uncertainty in field accounted for less than 0.05°C for all PSs. Moreover, a simultaneous calibration allows a direct comparison of the sensors at reduced uncertainty. Thanks to the uniformity of the copper block system, and to the limited thermal noise even in the presence of large numbers of PSs, the uncertainty in the comparison between them is reduced to 0.03°C. This allows reducing the uncertainty in the evaluation of the temperature gradient along the borehole. Together with the traceability obtained using calibrated references and by calibrating each PS, this process allows very accurate measurements of temperature values and profiles.
A calibration curve was then evaluated for each sensor with complete uncertainty budget and is now used as post-processing algorithm. The permafrost temperature curves show now a degradation of the permafrost base at approximately 68 m of depth since 2014 (Figure 7b – right). The evolution of the permafrost thawing and active layer is now robustly evaluated, through documented data traceability and in terms of comparability among the measuring points. The system can be adapted to almost any kind of thermometers used to measure temperature in liquid and solid matrixes (ice, soil, permafrost, seawater) and the procedure can be easily implemented for other locations, including permafrost monitoring stations in the Arctic4. This initiative can be expanded to further agreed processes for calibration and uncertainty evaluation to also benefit the data quality achievable by the Cryonet stations network supervised by the GCW of the WMO. It will also form the basis for the development of best practices by the Global Cryosphere Watch, for inclusion of guidance material in the revision of the WMO guide n.8 on instruments and methods of observations.
Summary
Climate change is significantly modifying the glacial/periglacial environments and one of the most evident consequences is the increase of hazards and risks and degradation of Alpine ice including permafrost. For improve knowledge on climate trends and associated risks in these environments, rapid progress in research activities is necessary. Standardised methodologies, increasing data acquisition, evaluating and reducing uncertainties and improvement of data comparability for sharing results are the key missions of the growing collaboration between metrology and climate scientists. This is the main vision of the MeteoMet initiative.
References
1. Riseborough D, Shiklomanov N, Etzelmüller B, Gruber S, Marchenko S. Recent advances in permafrost modelling. Permafr Periglac Process. 2008 Apr;19(2):137–56. doi.wiley.com/10.1002/ppp.615
2. Merlone A, Sanna F, Beges G, et al. The MeteoMet2 project—highlights and results. Meas Sci Technol. 2018 Feb 1;29(2):025802. doi: 10.1088/1361-6501/aa99fc
3. Paro L, Guglielmin M, Merlone A, Coppa G, Musacchio C, Sanna F. Permafrost station at Sommeiller Pass (NW Italy): from the monitoring to reference site and methods. 5th Eur Conf Permafr. 2019;(March):831–2.
4. Musacchio C, Merlone A, Viola A, Vitale V, Maturilli M. Towards a calibration laboratory in Ny-Ålesund. Rend Lincei. 2016;27:243–9. doi: 10.1007/s12210-016-0531-9
