Conference Agenda

For centuries, waterways were strategic infrastructure for communication, transportation, agriculture, mobility and trade of ancient Empires, both in Western and Eastern countries. These waterways have endured up to present and are now embedded in modern cities and landscapes wherein, while they remain water infrastructures, they also represent a cultural heritage (CH) resource and landscape asset to safeguard from modern challenges, e.g. urbanization, development, mass tourism. These processes induce a multitude of physical transformations and changes at surface that, in a relatively short time, can sum up and cause a complete modification of land use and skyline in the buffer area surrounding the waterways and the heritage assets (e.g. historical buildings, green spaces, archaeological areas) distributed along them. Given the high temporal dynamics of these processes and the spatial scale of the waterways, Earth Observation (EO) data are crucial to detect changes at the rate of the transformations, and long and dense time series need to be processed and analysed. While the existing Copernicus and Chinese missions offer the opportunity to exploit a plethora of data to achieve this scientific purpose, Artificial Intelligence (AI) and Deep Learning (DL) methods are required to make the change detection task more cost-effective and replicable, and generate value-added maps.

The new Dragon-6 STAI4CH project will follow on from successful achievements by Dragon-5 projects such as SARchaeology in demonstrating the capabilities of optical and SAR data to provide crucial information for archaeological and CH mapping and monitoring applications. Building upon those achievements, STAI4CH will make a step change towards the development and implementation of novel AI and DL methods to assess anthropogenic impacts on CH sites. In particular, STAI4CH will aim to demonstrate that AI and DL can be effectively implemented on medium to very high resolution optical and radar EO imagery acquired by operating ESA, TPM and Chinese missions (e.g. Sentinel-1/2, SDGSAT-1, SkySat, ICEYE), in order to detect recent and current human-induced impacts on CH and, as such, be used to support decision-making for CH preservation. The achievable benefits will be showcased by focusing on waterways and waterscapes of natural or manmade origin, which have been recognized as UNESCO World Heritage Sites (WHS): the Tiber river in Italy, part of the Historic Centre of Rome UNESCO WHS; the Grand Canal WHS in China, the longest artificial river in the world; and the Ahwar (marshes) of Southern Iraq WHS, refuge of biodiversity and the relict landscape of the Mesopotamian cities.

The developed algorithms will improve the change detection performance compared to conventional single-source methods, and provide evidence of the direct correlation between detected Land Cover Changes (LCC) and ongoing anthropogenic activities. Through DL models, combinations of optical/radar EO data will be processed to extract LCCs as objective proxies to depict human activities of potential threat for local CH. The spatio-temporal mining method will enable the generation of LCC maps and geodatabases highlighting “preserved” and “at-risk” areas, acting as AI-based prototypes that heritage bodies may use for planning and mitigation purposes. In this respect, the comparison between the three use-cases will allow an evaluation of the commonalities and differences in conservation challenges, and adaptability of AI-based methods to suit CH applications at different spatial and temporal scales, and in different types of environment. The cases will be developed on the Open Geospatial Engine (OGE), which is a cloud computing platform for EO data analysis.

On the 21^st of May 2023, the Macau Science Satellite (MSS-1) was successfully launched at the Jiuquan satellite launch centre in the Gansu province of China and is currently operating in orbit in good condition. The MSS-1 satellite is equipped with over ten pieces of state-of-the-art scientific equipment and is designed to monitor Earth’s magnetic field in the medium and low latitudes in both scalar and vector forms with a special focus on the South Atlantic magnetic anomalous region with unprecedented resolutions; simultaneously it measures the solar activities and the interplanetary magnetic field for detecting the variation of the space weather. All scientific payloads have undergone precise calibrations and begun to produce high-quality magnetic observational data that will continue for the next five to ten years.

The geomagnetic field, which is generated by Earth’s dynamo action within Earth’s outer core region and interacts and balances with solar wind at approximately eleven Earth’s radii, provides the ultimate protection of Earth’s ecosystem and the sustainability of the civilizations of human beings against the harmful radiations from the deep space. The variation of the geomagnetic field reveals the interactions of a series of physical and chemical processes occurring in Earth’s dynamical systems, i.e., the dynamics of the geodynamo system, the space electromagnetic environments, the mantel/ocean magnetic induction and the crustal magnetic field induction and magnetisation processes. Therefore, the geomagnetic field is a crucial physical quantity to study for understanding the inner workings of Earth’s interior and the surrounding near-Earth environment.

Given the highly accurate geomagnetic observations on a global scale obtained by MSS-1 and other sources, e.g., the Swarm mission by ESA, a series of challenging but rewarding research projects may be carried out in the studies of geomagnetism. We plan to divide the big scientific project into a few smaller ones and carry out them in a parallel manner on an international basis within the Dragon6 program. We will focus our studies on 1) the short-term and secular variation of the geodynamo system, 2) the mantel conductivity and induction process 3) the ocean tide & current induced magnetic field 4) the crustal magnetisation and the induced magnetic field and 5) the magnetic field generated by the space currents. A physically constrained geomagnetic model, namely the Macau Magnetic Model (M3), will be created for accurately describing the variation of the geomagnetic field in a broad range of spatiotemporal scales. In comparison with the conventional geomagnetic models, the new approach combines a set of significant physical processes as the constraint for modelling the geomagnetic field. It is expected to be accurate up to O(0.1)～O(1) nT with a spatial resolution of up to a hundred kilometres.