Concept & Methodology
CiROCCO aims to collect data from low-cost, stand-alone, electronic sensing nodes, in addition to higher-quality, off-the-self, sensing stations and medium-cost portable sun photometers through mobile monitoring campaigns. The collected data are fused with remote sensing data and further analysed. A mixture of wireless communication means including satellite, and local Wireless Sensor Networks is used for the collection of data. A series of services are developed, based on the needs of the local communities, as presented in Figure 1, namely:
- Renewable Energy Systems planning
- Air quality early-warning system for human health
- Land use and ecosystem management
- Modelling of Green House Gases and particles emissions, to validate the quality of the collected data and drive the sustainability and exploitation path of the overall system.
Figure 1: CiROCCO Concept
CiROCCO will install the same sensing nodes and data collection equipment in all the areas covered, aiming to produce consistent time-series data in a standardised way. The CiROCCO system will demonstrate how an extended network infrastructure (supplemented by LoRaWAN and satellite communicationss) can be used to transmit data from sensors in hard-to-reach areas to the central monitoring system. Multiple nearby stations can be connected to WiFi, Bluetooth, Zigbee, LoraWAN networks, where an edge device aggregates data, pre-processes it and transmits (either raw or pre-processed) using 4G or radio waves depending on the network coverage, data to other existing platforms and monitoring services. To retrieve data from the edge device, the satellite Internet of Things network from consortium partners will be employed. All the data will be treated in the Stream Handler Platform offered by the consortium partners. This platform provides the hooks for interconnecting, storing, transforming and processing data, as well as training, validating and executing Machine Learning and Deep Learning algorithms, resulting to a full Big Data solution with AI Capabilities. It is a high-performance (low latency and high throughput) distributed streaming platform for handling real-time data based on Apache Kafka, and it can efficiently ingest and handle massive amounts of data into processing pipelines, for both real-time and batch processing. The Stream Handler Platform and its underlying technologies can support any type of data-intensive ICT services (Artificial Intelligence, Business Intelligence, etc.) from cloud to edge.
The key capabilities and features offered by the platform are:
Real-time monitoring and event-processing
Interoperability with all modern data storage technologies and popular data sources Distributed messaging system
High fault-tolerance – Resiliency to node failures and support of automatic recovery; Elasticity – High scalability, Security (encryption, authentication, authorization)
The selection of a suitable low-cost sensor that can actually capture the variability of the targeted environmental parameter is a crucial first link in the chain of validated sensor data generation. In the framework of CiROCCO, this decision is highly relevant for the air quality parameters. Particulate Matter (PM) is the main air quality output that corresponds to the design and objectives of the project. Given that the focus is placed on desert areas, it is consequential that coarse dust particles have to be measured, thus requiring the ability of the low-cost sensors to sample and quantify PM10 particles along with smaller size fractions (e.g., PM2.5). Coarse dust particles transported from desert areas are considered a major issue for the attainment of the EU PM10 standard, especially in Southern European countries, and there are specific provisions for the identification and exclusion of such events when assessing compliance. However, regardless of regulatory issues, ambient dust exerts some impact on human health and major impacts on welfare, ecosystems and the climate, through various pathways. These impacts are expected to be magnified in near-desert areas, which frequently experience extreme ambient concentrations. CiROCCO aims at developing sophisticated calibration models that will extend between simple linear equations and advanced models incorporating physicochemical aerosol properties and AI models. In the framework of CiROCCO, the optimal approach for sensor calibration will be assessed in a pre-deployment calibration phase using advanced instrumentation for particle size distributions and atmospheric composition (including beta and optical reference-grade monitors, as well as aerosol mass spectrometry).
A critical upgrade, in terms of expanding the observational capabilities from space, constitutes the utilisation of the recently developed Michigan Institute for Data Science (MIDAS) dataset capable of addressing the highly demanding needs of the scheduled applications within the framework of the CiROCCO project. In addition, Lidar climatology of Vertical Aerosol Structure (LIVAS) dataset will be utilised providing vertical profiles of dust aerosols, among other aerosol types, which can be accurately identified thanks to the coincident depolarisation measurements. Therefore, the combination of MIDAS and LIVAS datasets ensures an accurate representation of dust burden in horizontal and vertical terms. Moreover, the establishment of the regional networks with the low-cost sensors offer a unique opportunity to evaluate MIDAS and LIVAS products over areas not well covered by global or European aerosol networks. For assimilating the MIDAS dust optical depth, a method will be applied which seeks a solution that minimises differences from the background field and the observations within an assimilation window. The process will incorporate a successive correction method to improve the fit between the observations and background fields. During each successive correction procedure, a modified Barnes scheme will be employed to weight and blend observation increments with the updated background field until the appropriate fine-scale structure is developed. The assimilated fields (dust analyses) will then be evaluated using data from the “low-cost” sensors. To test and validate the assimilation methodology, several sensitivity tests will be performed covering all the pilot areas of the project (Egypt, Cyprus, Serbia and Spain).