CLAIRO Zoom-in 1: A modular sensor network to quantify air pollutant capture by greenery

The city of Ostrava, the capital of the Moravian-Silesian Region, is a significant industrial centre of the Czech Republic, where air pollution was a serious issue in the second half of the 20th century. Due to a major restructuring of the industry and a package of air quality related measures initiated by the City of Ostrava, the situation has significantly improved over time resulting in a nearly 90% pollution decrease in the city. The wide range of measures in place, targeting improved air quality, include among others:

• emission control measures implemented in the industrial sector,
• a household boiler replacement programme,
• the development of the public transport system (construction and reconstruction of tram and trolleybus tracks),
• intensive road cleaning that reduces secondary air pollution caused by vehicle traffic,
• construction of cycling infrastructure,
• the operation of a bike share system,
• expansion of pedestrian zones in the city,
• the conversion of municipal fleet to low-emission or electric vehicles,
• rehabilitation of parks and the planting of tolerant greenery on roads.

Despite the restructuring and the above measures, air quality still remains one of the city’s biggest environmental issues. The main sources of air pollution in Ostrava include metallurgical production, power generation, domestic heating, and transport. As the city is part of the Upper Silesian metropolitan area, an international industrial agglomeration, pollution arriving from the nearby industrial conurbation of Katowice in Poland increases the concentration levels of air pollutants in Ostrava. The situation is further aggravated in the city occasionally by local meteorological conditions in winter, which are characterised by frequent inversions that result in poor dispersion of pollutants.

One of the major pollutants in the city is particulate matter. The legal limit value for the annual mean for particulates was often exceeded in case of PM2.5 and PM10 over the period between 2004 and 2014. Exposure to particulates can lead among other to heart disease, stroke, lung cancer, aggravated asthma, chronic bronchitis, and kidney disease. Particles with a diameter less than 10 micrometres (PM10) are small enough to pass through the throat and nose and enter the lungs. Particles less than 2.5 micrometres in diameter (PM2.5) can penetrate deeply into the lung and can even get into the bloodstream reaching almost all organs.

Monitoring data in Ostrava have shown that exposure to benzo(a)pyrene is high, as for many years the legal limit has been exceeded at all locations where concentrations of this pollutant has been measured. Benzo(a)pyrene is a polycyclic aromatic hydrocarbon (PAH) that is formed during the incomplete combustion of organic matter. Sources of benzo(a)pyrene include residential wood burning, vehicle exhaust and industrial processes. It is a carcinogenic substance that can adversely affect the immune system, the reproductive system and the brain function.

In Ostrava altogether 18 sensor units are operated in the target areas of the project (in Radvanice and Bartovice, the most polluted neighbourhoods of the city), and one reference system at the premises of the Technical University of Ostrava. In addition, six sensor units are performing measurements in selected neighbouring cities. The sensor technology monitoring is planned for eight years, before, during and after greenery planting for the evaluation of its effects on the concentrations of various air pollutants. Before planting, data is gathered throughout an entire year to cover all seasons and weather conditions. Data on climate conditions, such as wind force, the direction of the wind, temperature, pressure, humidity, and rainfall is also collected to complement detailed information on air quality. The measurements have started in September 2019.

The sensor units are installed near the steelworks of Liberty Ostrava in the target area of CLAIRO that include the sites for the greenery planting. Ten units are in the direct vicinity of the ash pond of the steelworks and cover areas in front of and behind one of the green planting sites. Another eight sensor units are situated in the neighbouring residential area close to the other greenery site.

Continuous measurements are performed in the target areas of CLAIRO with an aim to provide a good overview of how the air quality situation develops throughout an entire year without greenery and to assess pollutant levels during different seasons. An additional benefit of continuous measurements is that they can indicate short-time peaks in concentration levels that can remain hidden in long-term averages.

On the basis of continuous measurements of air pollution and meteorological parameters undertaken by VSB, separate models for capture of air pollutants will be developed by the Silesian University in Opava for the existing vegetation, the designed new greenery, and the planted greenery. VSB provides the Silesian University in Opava with validated air quality data and information on the exact position of the sensor units for the developed models. Building on modelling it is possible to determine the most effective composition and structure for the new greenery in terms of pollutant capture.

The capture of air pollutants by existing greenery will be calculated from the dispersion, the renewed suspension, and deposition of airborne substances. The evaluation of air pollution situation, the meteorological parameters and natural conditions and the capture of pollutants by the existing greenery, will allow the development of a hypothetical model of the designed new greenery. The hypothetical model will be tested by continuous measurements of pollutant concentration levels. Building on previous models, a final innovative complex model will be developed that will include predictions for future capture of air pollution by the planted greenery.

In CLAIRO a modular sensor network is used that allows real-time simultaneous measurement of gas pollutants and particulate matter in the air. Within the network enviSENS sensor boxes are installed, each of which includes several sensors from various manufacturers that can measure concentrations of various air pollutants. The sensor boxes can run on a solar panel, or mains electricity. Apart from the measurement components (sensors) the boxes include a microcomputer, a battery, a connection to the solar panel, and a 220 V mains connection. The boxes are weather resistant and are easy to maintain. They are mounted in a height of 3-8 meters to ensure that the photovoltaic panel has access direct sunlight. Solar panels ensure that the units are energy self-sufficient.

The sensors allow fast, short-term measurements of both organic and inorganic substances. Concentrations of particulate matter in various fractions (PM10, PM2.5, and PM1), nitrogen dioxide (NO2) and ozone (O3), volatile organic compounds (VOC), and polycyclic aromatic hydrocarbons (benzo(a)pyrene) are monitored by the sensor units. In the boxes a new generation of sensor units are used, that enable monitoring of concentrations of benzo(a)pyrene, without the need for complicated laboratory analysis, as opposed to standard measurement processes. Currently the feasibility of benzo(a)pyrene sensory measurements is tested by VSB. The sensors capture data at very short intervals. The values are recorded every 10 seconds, and the average concentrations are sent in data packets every 5 minutes. Data from the sensors are collected in an integrated data logger and transmitted via LORA radio network or GPRS (General Packet Radio Service) data transfers to the central database at VSB.

Although the sensors are very accurate (particularly in case of PM10) and enable rapid measurement, their data cannot be used to assess compliance with limit values of air pollutants. The data accessed from the sensor stations is therefore used solely for the purposes of the CLAIRO project, which is monitoring changes in pollutant concentrations in relation to the ability of greenery to absorb airborne pollutants.

The system has high modularity and flexibility. It is a low-cost option compared to conventional measurement units. Its installation requires much lower initial investment compared to that of reference analysers, and its power consumption and operation cost is particularly low.

The application of the modular sensor network allows the creation of concentration maps for individual pollutants, calculation for modelling on the basis of measured meteorological parameters (e.g. wind speed and direction), as well as the indication of places and time periods with unusual concentration values.

Online data on concentrations of air pollutants that are transferred from the sensor boxes every 5 minutes, are stored in a structured database of the Intelligent Identification System (IIS) that was developed by VSB.

The system stores the transferred data, logs them into a database, and allows further data processing. It was originally developed to simplify the identification of sources of air pollutants on a local scale. IIS does not replace standard measurements, but aims to make refinements at a selected location in order to identify critical points in the area. Due to short-term concentrations enabled by the system, it is possible to search for time periods and locations of exceptional or unwanted concentrations of air pollutants.

The IIS consists of 30 sensor units, which are connected to two standard monitoring stations that are used for regular checking of the state of the sensors and for the validation of the sensory measurement results. The Czech national monitoring network of the Czech Hydrometeorological Institute (CHMI), that includes approximately 150 stations, covering the entire country, is suitable for assessments on a macro scale, but it often fails to respond to local problems. In contrast, the Intelligent Identification System builds on smaller local networks that supports local planning through enabling the detection of causes of air pollution, the selection of appropriate control measures for reducing emissions (such as traffic calming, the introduction of boiler subsidies, implementation of dedusting technologies, or air recuperation) and the identification of locations for their implementation. This is made possible by online measurement of pollutants (i.e. PAH and VOC) that are normally monitored only retrospectively with sample analysis in laboratories.

As the IIS network is mobile, it can be deployed to any specific area or pollution source. Due to the density of sensors, pollutant concentrations from local sources can be distinguished from background concentrations. In the concentration database, unusual data points or data sets, such as concentrations exceeding limit values, sharp increases, non-systemic changes of values and deviations are automatically marked by the system. The aim of IIS is the identification of critical points at local level that have the greatest contributions to emissions, and not compliance monitoring. The IIS checks the validity of measurements, it calculates the averages and finds the maximum values.

The sensor network of CLAIRO is part of the Intelligent Identification System, the 24 sensor boxes installed under CLAIRO complement the 30 original sensor units of IIS.

A huge amount of data is gathered during and after the lifetime of the project, the management of which requires the use of very powerful computing tools. Concentration values are recorded every 10 seconds, and the accumulated data is uploaded to the database at 5-minute intervals. In one day, the measurement of 8 substances by 20 sensor units results in more than 46 000 separate data values. Due to these enormous volumes of data, the database of IIS is stored on a supercomputer of the IT4Innovations National Supercomputing Centre.

IT4Innovations is part of the Technical University of Ostrava (VSB). IT4Innovations together with CESNET and CERIT-SC, two other infrastructures, constitutes e-INFRA CZ, the national research e-infrastructure of the Czech Republic. At present, IT4Innovations runs four supercomputers, which are located in a data room with an area of 500 m2. The operation of the computer system is sustained by an energy centre with a redundant power supply backed up by a diesel generator. The supercomputing centre is supported by a water-cooling system with a 500 kW of cooling capacity, that consist of a network of 5 cold-water and warm water circuits located under the elevated floor of the data room, and 15 cooling towers. The system ensures the recovery of the waste heat generated by the supercomputers. The data room also includes a fire protection system that actively reduces the oxygen concentration to 15% to effectively prevent fire.

Out of the four supercomputers hosted by VSB, Salomon is the largest one with an output of 2 PFlop/s. When it was launched in 2015, Salomon was the 40th most powerful supercomputer in the world. The database used by CLAIRO is stored on Salomon that consists of 1,008 computational nodes, each of which is equipped with 24 cores.

The data sent from the sensors is displayed in a map format on a platform created by the IT4Innovations, the Floreon geo-database. Floreon was developed to serve as an integrated operational platform for monitoring, modelling, prediction and support for crisis management in the Moravian-Silesian region. Originally, the platform was created to provide support in the field of hydrology, simulating and forecasting flood situations. Later on, its use was extended to environmental pollution modelling.

Floreon allows the selection of substances, sites and time periods. In platform, the sensor stations of CLAIRO are indicated by coloured circles on a base map. The colour code of the circles changes with concentration levels. The measured concentrations are displayed as numbers within the circles. The Floreon platform linked to the sensor units of CLAIRO shows concentrations of particulate matter (PM1, PM2.5 and PM10), NO2, O3, volatile organic compounds (VOCs) and polycyclic aromatic hydrocarbons (PAH). Users can access data from 25 sensor stations at locations in Ostrava, and in neighbouring cities, where pilot measurements are undertaken.

The platform displays also various detailed graphs illustrating basic statistical data from measurements.

The geo-database can be accessed at the airsens.eu website. The data will be accessible for any time in the period from 1 September 2019 up to 2027. Data from the Floreon platform is archived and can be retrieved at any time.

Measurements of air pollution will also be undertaken in other neighbouring cities in the Ostrava- Karvina Industrial Agglomeration in order to obtain background data, to draw comparisons with the Ostrava sites and to assess similarities and differences in air quality situation. The six towns identified as pilot areas are Trinec, Opava, Frýdek-Místek, Karviná, Havírov and Rychvald.

Six sensor units are available for measurements in the cities. The sensor boxes are moving from city to city and monitoring is undertaken in two cities at a time with three installed in each of them. In each of the collaborating cities, data will be collected with three sensor units over an eight months-long period that include both the winter and the summer seasons. The sensor boxes are placed in the most polluted areas of the towns, close to roads with heavy traffic and major junctions.

The secondary aim of measuring air pollution in neighbouring cities is to roll out the results of CLAIRO across the Ostrava- Karvina Industrial Agglomeration. Based on the evaluation of the data gathered a set of recommendations will be formulated for the selected cities on potential green space interventions and most appropriate plant species to be used in future greenery projects.

The monitoring activity outside Ostrava started in Trinec and Opava, where measurements were undertaken between January and August 2020. In Frýdek-Mistek and Karviná the steps towards the installation of the equipment and the launching of the measurements are underway.

A major challenge faced by the City of Ostrava is to reliably quantify the impacts of vegetation on air quality. For ensuring the precision of modelling and through this ensuring the accuracy of the predictions for quantified impacts, the application of the modular sensor network is used that allows real-time simultaneous measurements is essential. The innovative approach of the project involves the use of a new generation of sensor units that allows fast, short-term measurements. These sensor units enable online monitoring of concentrations of air pollutants, without the need for complicated laboratory analysis, as opposed to standard measurement processes.

The monitoring activity under CLAIRO evidenced the consistency of the measurements of sensors with those of reference analysers for particulate matter, the most significant pollutant in the target area. Although the system seems slightly less reliable in the case of nitrogen oxides and ozone, still the measurements are suitable for assessing the impact of traffic on concentration levels of these pollutants. The implementation of the project confirmed that it is easy to install and operate the sensor network and that the technology is suitable for complementary measurements. The system has been shown to be particularly useful for the detection of places and time periods with unusual concentration values and for the identification of local problems. Difficulties have been experienced mostly with data transmission, as the use of the current LORA radio network leads to data losses. In such cases, the data can be retrieved only through laborious offline methods. Switching to the more reliable GPRS standard would increase operation costs.

The meteorological conditions in the winter of 2019 and 2020 and the rest of 2020 were atypical with higher temperatures, with twice as much precipitation than the long-term average and strong winds, which helped the dispersion of air pollutants. The COVID-19 pandemic has also contributed to better air quality compared to other periods through reduced traffic and industrial activities. To correct potential deviations, the Technical University of Ostrava intends to compare the data measured during the end of 2019 and throughout 2020 with standard air quality data from the previous couple of years accessed from the stations of the Czech Hydrometeorological Institute.

It has been long debated whether the main source of air pollution in Radvanice and Bartovice is the steelworks, or it is mostly linked to domestic heating, local traffic and cross-border transfer. It is expected that the continuous measurements performed under CLAIRO will also enable a more accurate assessment of the impacts of the various polluting activities to the overall air quality in the area.

1 Health Effects Institute (2018) State of Global Air 2018. Special Report. Boston, MA:Health Effects Institute.

https://www.stateofglobalair.org/sites/default/files/soga-2018-report.pdf

2 https://www.health.nsw.gov.au/environment/air/Pages/particulate-matter.aspx

3 Chan, T., et al. (2018) Long-Term Exposure to Ambient Fine Particulate Matter and Chronic Kidney Disease: A Cohort Study. Environmental Health Perspectives. Vol. 126, No. 10.

4 Jeong, S. et al. (2019) PM2.5 Exposure in the Respiratory System Induces Distinct Inflammatory Signaling in the Lung and the Liver of Mice. Journal of Immunology Research. Vol. 2019. Article ID 3486841, pp. 11.

5 Williams, M. A., Salice, C., Reddy, G. (2015) Wildlife Toxicity Assessments for Chemicals of Military Concern. Chapter 24 - Wildlife Toxicity Assessment for Benzo[a]Pyrene. pp.421-437. ISBN 978-0-12-800020-5.

6 US EPA (2017) Toxicological Review of Benzo[a]pyrene. EPA/635/R-17/003Fa. U.S. Environmental Protection Agency. Washington, DC.

https://cfpub.epa.gov/ncea/iris/iris_documents/documents/toxreviews/0136tr.pdf

7 McCallister, M. et al. (2008) Prenatal Exposure to Benzo(a)pyrene Impairs Later-Life Cortical Neuronal Function. Neurotoxicology. 2008 Sep, 29(5): 846–854.

8 https://www.airsens.eu/

9 https://www.smartenvi.eu/?page_id=135

10 https://www.it4i.cz/

11 https://floreon.eu/website/Default.aspx