Imaging and identifying Irish Sea shipwrecks

WP3.2 Science and the Sea - StoryMap #1

Unpath'd Waters

September 15, 2023

About Unpath’d Waters

The UK has a rich maritime heritage, stretching back over 23,000 years. It’s impossible to tell the story of our islands without talking about our relationship with the sea. This maritime past is becoming increasingly important. People are more aware of our exploitation of the sea and topics such as colonialism, slavery and immigration.

Unpath’d Waters aims to increase interaction with the UK’s maritime heritage by making it easier to research and easier for the public to discover and share stories in new ways.

Work Package 3.2: Science and the Sea

A lack of high-quality remotely sensed data and limited research means our maritime heritage remains poorly understood. Thousands of shipwrecks around the UK coastline have been identified incorrectly or remain unknown. The 'Science and the Sea' work package is developing new methods to help solve this problem. These methods will accelerate maritime research activity and support the development of research tools designed to improve the management of UK wreck sites. Focussing on vessels lost in the Irish Sea, we’re developing methods to refine wreck site identification and methods for integrating marine scientific data with the historic record and maritime collections.

Remote sensing and maritime archaeology

The past two decades have witnessed remarkable advances in seafloor mapping, with high-resolution acoustic imaging now routinely used in archaeological studies. Technological and methodological advances in acoustic imaging and digital rendering now permit shipwreck sites and individual artifacts to be imaged at centimetric resolution in tens or hundreds of meters of water.

Multibeam echosounder data of SS *Chirripo,* which sank at the mouth of Belfast Lough in 1917 after she struck a mine. Data: Ulster University and INFOMAR

Multibeam-derived pointcloud of SS Chirripo, Belfast Lough. The cloud contains > 20,000,000 individual data points. Data: Ulster University and INFOMAR.

In response to the 1992 Valetta European Convention on the Protection of the Archaeological Heritage and the 2001 UNESCO Convention on the Protection of the Underwater Cultural Heritage , archaeologists are encouraged to adopt non-destructive techniques for wreck investigations and consider in situ preservation a priority. Therefore, remote sensing techniques play an increasingly important role in our work. These techniques offer rapid, non-invasive, high-resolution, and cost-effective solutions in all phases of wreck investigations—from initially locating the site, to identifying and characterizing the site, through site management.

MBES data of RMS *Leinster*. In the final weeks of the First World War, RMS Leinster was torpedoed by German submarine UB-123. 571 people were lost in the sinking, resulting in the greatest ever loss of life in the Irish Sea. UB-123 subsequently sank in a minefield in the North Sea with all 36 crew. Data: Ulster University and INFOMAR

What techniques do we use in seabed mapping?

Multibeam echosounders (MBES) are active sonar systems used to map the seafloor and detect objects in the water column and on the seafloor. These systems use a transducer array to send and receive sound pulses that map the seafloor. A multibeam array is usually mounted directly on the ship’s hull, but can be deployed from other surface and underwater vehicles.

NOAA ocean charting operations - multibeam echosounder animation

National initiatives in seabed mapping

Although humans have been exploring the seabed for hundreds of years, seabed mapping has been revolutionised in the last few decades with the advent of MBES systems, and increasing computing power in terms of hardware and software. In response to our need to collect baseline data to effectively manage our offshore resources, many countries around the world embarked on ambitious seabed mapping initiatives.

Ireland is a good example, having undertaken deep-water survey operations to designate its maritime territory since 1996. The country now has 700,000 km² of the seafloor within the Irish designated area (10 times the land area) mapped in high resolution. Its aim is to map Ireland’s entire seabed by 2026 , becoming the first country in the world to have systematically mapped its marine territory.

INFOMAR - Mapping the Irish Seabed.

In 2022, the UK established the Centre for Seabed Mapping (CSM), with the aim of enabling a more unified approach to seabed mapping in the UK. This will be achieved through increased coverage, quality and access to seabed mapping data collected from public money or in collaboration with industry.

UK Centre for Seabed Mapping (CSM) | UK Hydrographic Office

The Irish and UK governments are both committed to the SEABED 2030 initiative, in support of the UN’s Sustainable Development Goal 14: ‘to conserve and sustainably use the ocean, seas and marine resources for sustainable development.’ Seabed 2030 is committed to building the necessary technical, scientific, and management framework to compile all available seabed mapping information into a seamless digital map of the world’s ocean floor by 2030.

Unpath'd Waters MBES data

The MBES data we use in Unpath'd Waters were collected by the Centre for Applied Marine Sciences (Bangor University) using the research vessel Prince Madog. We collected data from over 150 wreck/obstruction sites that are listed on the UK Hydrographic Office database, including 60 sites in the Bristol Channel, 26 sites off the coast of South Wales and 63 located between the Isle of Man and Northern Ireland.

This new dataset includes 129 sites not previously thought to have been surveyed using MBES, as well as data from more detailed surveys undertaken over several wreck sites (previously surveyed by Bangor University in 2020) that are of potentially historical importance within our Unpath’d Waters project study area to the west of the Isle of Man.

How can we use remotely sensed data to identify unknown wrecks?

Innes McCartney's (2022) book ' Echoes from the Deep ' demonstrates how multibeam echosounder data can be used to identify unknown wrecks. McCartney refined techniques whereby high-resolution digital elevation models (DEMs) derived from multibeam surveys are used to represent the archaeological record, and comparison is made between these DEMs and documentary resources (e.g. ship's plans, historic texts) to identify previously unknown wrecks.

MBES data from a wreck which is dimensionally, positionally and archivally consistent with landing craft *(pictured below)*. The signature of the wreck in the MBES data displays features similar to those seen on the plan below (McCartney, 2022).

The Mk III *LCT 326* was of the type seen in this image (wrecksite.eu).

What impact does MBES data resolution have on wreck identification?

The spatial resolution of mutlibeam echosounder data really matters when it comes to identifying unknown wrecks. The images below show the wreck of SS Chirripo (Belfast Lough) at 5 m, 1 m and 0.1 m resolution: - 5 m resolution is typical of legacy MBES data sets available through online data portals such as INFOMAR in Ireland and the UK Marine Data Portal . - 1 m resolution is typical of wreck surveys in moderate water depths - 0.1 m resolution data is typical of high-definition wreck surveys

1 m resolution data contain 25 times more information for the end user than the 5 m data. 10 cm resolution data contain 100 times more information for the end user than the 1 m data.

What about the temporal resolution of multibeam echosounder data?

Repeat (time-lapse) surveys can tell us about the dynamics of a site, whether sediment is moving through site, and whether wrecks are being covered or uncovered. Linked hydro- and sediment-dynamics are important in terms of site formation and the preservation of UCH. In extreme cases, hydrodynamic forcing can cause a wreck to move or collapse.

The DEMs below show the wreck of SS Apapa (North Wales): (i) surveyed on the 19 May 2014, (ii) 12 May 2017, and (iii) a difference model showing the deposition (red) and erosion (blue) of the seabed over that time period. Multibeam data: Bangor University.

SS Apapa was travelling from West Africa to Liverpool when she was torpedoed by U-96 off Point Lynas, Anglesey on 28 November 1917 with the loss of 77 passengers and crew.

Can multibeam echosounder data tell us anything else about wreck sites?

The primary output from multibeam echosounder surveys are DEMs of the seafloor, like this one showing two WWI wrecks side-by-side in Belfast Lough, SS Chirripo and SS Tiberia. Both sites display complex erosional and depositional patterns around them.

We can derive a local relief model (LRM) from the DEM at any length scale we wish using focal statistics. This image shows a LRM derived at a length scale of 10 m, the scale at which the depositional and erosional features typically form in this environment. This method successfully separates the depositional (in red) and erosional (in blue) features around both sites, giving us an insight into the dominant hydro- and sediment-dynamic processes acting at the sites.

Using multibeam echosounders we can also map the scattering function (the backscatter) of the seabed. Hard and coarse beds (e.g. gravel) scatter a lot of sound, and produce high intensity returns. Soft and smooth beds (e.g. mud) scatter little sound, producing low intensity returns.

When we integrate backscatter data with sediment samples we can derive substrate (grain size) maps of the seafloor using image classification techniques. Sediment type is one of the key variables that controls the preservation of UCH.

Next up

Read about the preservation modelling and risk assessment work we have completed as part of the 'Science and the Sea' work package in StoryMap #2 .