The Stourhead LiDAR

Whatever next?

When Archaeology takes advantage of new techniques, whole new landscapes of information emerge.

One of the most exciting developments in recent years has been Light Detection and Ranging or LiDAR for short. Using a drone or an aircraft, pulsed light signals are sent using a laser. When linked to a scanner and a global position system (GPS), It can create an ultra-fine 3D record of the ground surface over wide areas.

The boundary of the National Trust’s Stourhead Estate in South Wiltshire. Whitesheet Hill is on the edge top right. Stourhead Park and mansion, garden, lake and Stourton village are lower centre. Park Hill Camp hillfort lies in woodland centre left.

In large surveys, millions of light points are plotted and tied to existing mapping with the GPS. Each point has its unique XYZ position… latitude, longitude and height above the datum level.

The total Stourhead survey area tilted slightly to show the contours covering roughly the same area as the map above. Whitesheet Hill on the right The hillfort faintly visible on the lower right hand edge. To the left, two valleys separate a ridge which has Park Hill Camp hillfort near the middle.

The Environment Agency has been using this technology for years and have made their data freely available. A quick visual link can be seen here https://houseprices.io/lab/lidar/map. This survey data was collected mainly to predict levels of flooding and consequently it tends to be concentrated along valleys and coasts. It has given good results but the detail tends to be at 1.0m resolution or in the better areas 0.5m. The best quality is 0.25m density of coverage.

There are still large gaps in the land area currently covered by LiDAR and therefore the National Trust is commissioning its own surveys at 0.25m.

In the South West, there are new surveys for the Bristol and North Somerset properties, the Bath and Dyrham properties and most recently the data has arrived for the Stourhead Estate in South Wiltshire. Bluesky collects the data and it is analysed by ArcHeritage who provide the baseline digital imagery in various forms as well as the core GPS files.Their report picks up many new sites which have now been uploaded onto the NT Historic Buildings Sites and Monuments Record. This is not the end: new archaeological sites can still be discovered by further manipulation of the data combined with other information sources.

The LiDAR data can be uploaded into the digital mapping system and then it can be overlaid as a layer on digitised historic maps, onto geophysical surveys and onto aerial photographs. It is so easy these days to zoom in an out of maps and also to fade one layer of information and then see another in direct relation to it.

Detail of Whitesheet Hill causewayed enclosure. Four round barrows can be seen along its bottom edge all with little dimples in the top where the owner of Stourhead Estate, Richard Colt Hoare, excavated them in he early 19th century. One at the lower left hand edge of the plot was cut by a chalk quarry in the 19th century. Close examination of the plot shows phases of trackways and faint embankments.

A great ability of LiDAR is to fell forests and woods (virtually) to see the ground surface beneath. Something impossible with air photography.

Imagine the light pulses from the aircraft like rain falling on the ground. Some will bounce off the tree tops (the first returns) but many will hit the ground below the tree canopy (the second returns). There are systems to filter out the first returns so that only the ground can be seen. It is why I always ask for surveys to be done in the winter when the leaves have fallen from the trees and the ground surface can be most clearly surveyed.

Park Hill Camp from the air surrounded by conifer plantations. In the last 10 years the National Trust has gradually removed the trees from the scheduled monument. The LiDAR defines the earthworks in a better way than can be seen by aerial photography when trees interrupt the view (see below)

Stourhead’s Park Hill Camp Iron Age hillfort has been covered in trees for many years making it difficult to see. Over a number of years, gradually, the National Trust has been clearing the woodland and bringing it back to grass. The LiDAR survey has enabled the ramparts and ditches to be clearly seen as well as showing its strategic position on the ridge top unimpeded by the conifer plantations that surround it.

The ramparts and ditches of Park Hill Camp Iron Age hillfort revealed by the LiDAR survey.

Another great thing: the LiDAR light point cloud is three dimensional and this enables a digital terrain model to be created. This can be viewed on its own or it is possible to drape aerial photographs and/or historic maps across it…as though the map or photograph has become a gigantic cloth thrown over the contours of the landscape. There is now the ability to screen- fly through the Stourhead landscape switching on or off other layers of information while weaving up the valleys or skimming over the hillfort ramparts.

Stourhead Lake (bottom) and the position of Park Hill Camp Iron Age hillfort clearly revealed on the ridge top between two valleys. This area is planted with trees and this vantage point of the fortification would not normally be appreciated.

During a bright winter day, low sunlight will traverse the landscape bringing different shadows in sharp relief and revealing new details. LiDAR analysis can introduce its own light source and the survey plot can be re-generated.. with the light source at any angle and direction. This shows up very faint archaeological earthworks when the light source is beamed from a particular direction.

The LiDAR survey shows the quality of surviving archaeology and reveals where conservation should be concentrated across the Stourhead Estate.

This image looks down to the arable land from the hillfort and causewayed enclosure on Whitesheet Hill. On the lower land, the earthworks have been almost levelled by modern ploughing but old quarries can be seen clustered, where stone outcrops on a low hill, and faint traces of prehistoric ‘celtic fields’ can be seen.

The Stourhead farmland, ploughed for many 100s of years, has lost much of its archaeology but the survey still shows traces of medieval and prehistoric agriculture and traces of buried enclosures suggesting settlement remains below the ploughsoil….(though much worn down buried pits and ditches will survive).

However, there is fine earthwork survival in Stouhead Park and on Whitesheet HIll.

The prehistoric earthworks on Whitesheet Hill show up very clearly: the Iron Age hillfort to the south, the Neolithic Causewayed enclosure in the middle and the other enclosure (also probably Neolithic) to the north and in between Late Bronze Age cross ridge dykes, Early Bronze Age round barrows and medieval pillow mounds all crossed by banks, trackways and quarries of various periods.

Whitesheet Hill: LiDAR shows the ramparts and ditches of the Iron Age hillfort (c.300BC) at the bottom. The three rectangles are modern reservoirs just east of the National Trust boundary. A cross-ridge dyke (c.1000 BC) divides the narrow downland ridge separating the hillfort from the Neolithic causewayed enclosure (c.3600BC) which has a Bronze Age round barrow (c.2000BC) built over its southern edge and across its north side runs the old cattle drove road from the the west towards Salisbury and then on to London. Further along the down to the north (the upper edge of the plot), is another faint enclosure (c.3000 BC) of similar size to the causewayed enclosure (this site has been ploughed in the 20th century but can be seen clearly on the LiDAR).

The parkland is a very precious survival. The ridge and furrow of medieval open field furlongs was fossilised when the park for the mansion house was created. This must have happened before 1722 which is the date of our earliest map of the park.

The 1722 Stourhead Estate map (Wiltshire and Swindon Record Office). The newly built Stourhead House is centre right. The north-east corner of the park is Spencers Mead. The strip fields and curving enclosure ditch shown on the LiDAR lie within this field and Slade Mead below. Buildings in red between these fields were demolished in the late 18th century, the building sites lie where there are earthworks shown on the LiDAR image below.

Near the Stourhead House and near the landscaped garden obelisk are two turbulent areas of earthworks, outside the areas of agriculture and therefore places already occupied ….before the open fields were created it seems.

One of these, east of the House, is likely to be the site of the medieval Stourton Castle..demolished when the present mansion was created in the early 18th century. The other area near the obelisk is a mystery… the LiDAR raises many new archaeological questions…. wonderful.

Next year the Cotswolds and Hidcote NT properties will have LiDAR We await the results with anticipation…what new Roman sites lies beneath Chedworth woods……

The site of the present Stourhead House is lower centre. The buildings show as triangles. Top right is the NE corner of the present park where a large oval enclosure (prehistoric?) underlies the regular furlong blocks of strip fields divided by trackways (this is just grass not visible on the ground). Centre right and NE of Stourhead House is an area of earthworks likely to be the site of the medieval Stourton Castle. At the left edge of the picture is another grouping of mixed earthworks, perhaps an early settlement to the right of the mound with the 18th century obelisk monument on it.

Stourhead? Where’s Stourton Castle?

In south Wiltshire, on the border with Dorset and Somerset was built a great and ancient house. It lay at the centre of a large estate and was known as Stourton Castle..

but there is a problem…

We don’t know where it is

We have a picture.

An enhanced drawing based on John Aubrey’s original sketch.

Drawn from a 1670 sketch by the antiquarian John Aubrey.

The place was massive and must have looked a bit like Lacock Abbey

Lacock Abbey in north Wiltshire. Stourton Castle was arranged around two courtyards like this and would have been of similar scale and outward appearance.

but it’s gone.. apparently without trace.

The story of the removal of Stourton Castle and the creation of Stourhead House has a touch of Poldark about it.

The Stourtons (old money Poldark) and the Hoares (new money Warleggan)

The Stourton family had taken their name from the village of Stourton (the farm by the River Stour), a place recorded in Domesday and at least Saxon in origin. The Stourtons claimed that their line went back to a mighty Saxon lord… Botulph.. and William Camden, writing in 1607, saw a ‘monstrous bone’ displayed in Stourton Castle… a leg of their legendary ancestor.

The surviving records trace the family back to the 12th century but the Stourtons only emerge as lords of the manor in the 13th century documents. During the 14th and 15th centuries, the family did very well and built up cash reserves via good marriages and military service in France.

Their manor house blossomed and flourished. Aubrey’s picture shows that it was built around two large courtyards and had a tall tower and shows parapets with military style battlements.

Things fell apart for the Stourtons in the 16th and 17th centuries. In the early 16th century, William Lord Stourton was working for Henry VIII in France and left the care of his estate to his trusted steward William Hartgill…he also looked after Lord Stourton’s  wife Elizabeth.

When Lord Stourton died in 1548, his hot-tempered son Charles inherited the Estate. He rode with a band of henchmen to Kilmington near Stourton and found his mother living at Hartgill’s house. A feud grew up between the men and eventually, in 1557, Lord Stourton kidnapped William and his son John, murdered them and buried their bodies in a cellar within the Castle.

As a catholic, with Queen Mary on the throne, Charles believed that he would get away with the murders. However, William Hartgill had friends and they made sure that the Castle was searched, the bodies found and Lord Stourton arrested. He was convicted and executed in Salisbury. His wife was forced to pay for her husband’s property which had been forfeited to the Crown following his trial. She was also separated from her eldest son John, who was only a child at the time.

The family backed the wrong side in the Civil War. In September 1644, Parliamentarian forces set fire to one of the gates, captured the house, ransacked it and made it untenable.The eldest son John was killed at the battle for Bristol and when the Royalist cause was finally lost, the estates were confiscated and heavy fines were imposed.

The family fortunes continued to decline and by 1686 the impoverished Stourton family had mortgaged their ancestral home and by 1704 it was for sale.

Enter the Hoare family who had made a fortune through banking. In 1720, they purchased the whole estate, demolished the castle and built a new flashy Palladian villa… quite the latest thing. They changed the name to Stourhead. The house at the source of the River Stour.

The Hoare family made sure the transformation was carried out quickly. They paid for a survey of their new property in 1722 and the Estate Map shows the new house completed… Stourton Castle was gone.

The 1722 map shows the new Stourhead House with its garden intruding slightly into the top left hand corner of the stable courtyard. The courtyard is below and right of the house. With a large gateway facing south. Was this the outer courtyard of the old house?

Finding the site has been difficult, the normal techniques have proved to be inconclusive and the quest for the Castle has become a great archaeological challenge.

It seemed simple at first. The 1880s 25 inch OS map marks a cross about 100m east of Stourhead House with the legend ‘site of Stourton Castle’ So we geophysed it and the results were very disappointing. Since then, year by year, we have surveyed around the house but nothing has been revealed.

When Meg did her student placement for the National Trust, I asked her to find the Castle and her MA dissertation tracked down the documentary references and descriptions of its chambers and halls and its chapel, which included a decorated tiled floor inlaid with the initials WS for William Stourton. Documents in the record offices of Cornwall, Wiltshire, Somerset and nearby Longleat House were examined. These built up the background: the estate, the farmland, the deer park and hunting lodge, the warren and the warrener’s lodge… bits about the repair of the great house and its approximate location… but nothing to pin it down.

We looked at the fabric of the stone-lined cellars of Stourhead House. Had they built the new house on the old?

The cellars of Stourhead House. Some reused stone but nothing to indicate that these were once the cellars of Stourton Castle.

We examined the stableyard to the south. This includes in its walls great chunks of reused? stone and a 16th century? doorway. Was this the remodeled outer courtyard of the Castle?. We dug a trench on its north side, hoping to find medieval walls leading to the inner courtyard.. just 18th century pottery above deep soil.

Our fieldwork in Stourhead Park. Earth resistance meter against the railings and our trench beside the stables courtyard to the left. The Hoare family’s 1720s mansion Stourhead House can be seen top right. The site of old Stourton Castle lies somewhere to the right of this picture.

So where is it hiding? A LiDAR laser survey of the parkland ground surface might help or more sensitive geophysics… perhaps ground probing radar. Everything seems to point near the cross marked by Ordnance Survey. A line of pre-Stourhead House chestnut trees are aligned north towards this point where there is a mound in the park. From the east, an old drove-way passes through Drove Lodge and runs as an earthwork into the park.

Our site is most likely to exist where these two alignments meet. Surely the backfilled cellars and extensive robbed out walls lie there or thereabouts.

We’ll keep looking.

“Shall we stack the naked acres with our CPIs”

March and April are the CPI months. The National Trust is a unique British organisation set up in 1895 to conserve places of historic interest and natural beauty. How do you do that? How can NT measure how successful it is? Where should it spend its limited resources? Is a Joshua Reynolds painting in a mansion more important than an orchid on the ramparts of an Iron Age hillfort? The answer is… the Conservation Performance Indicator (CPI).

Picture a room with a table, a powerpoint projector and a circle of chairs: perhaps a cafe on the Somerset coast or a barn in Dorset or even an old engine shed in Somerset. The players take their seats. The ecologist, the curator, the building surveyor, conservator, gardener, ranger, manager and…archaeologist. The main conservation features of the property are agreed and then three lists are made and the match begins. Which is most significant, which would have the greatest conservation impact if lost and which requires resources most urgently. For some places the top points are easier to allocate than others. Stonehenge landscape has archaeology at the top but what of the Stourhead Estate in Wiltshire.

Stourhead Lake and Pantheon aquatic archaeologists from NAS prepare to dive

Is the famous 18th century landscape garden with its grade I listed temples, lakes and bridges more important than the incredibly rare Pope’s cabinet in the house or does the scheduled Early Neolithic causewayed enclosure on Whitesheet Hill rank higher than the SSSI chalk downland sward that covers it ? (well, they tend to be mutually beneficial so taking care of the ecology generally benefits the archaeology).

Whitesheet Hill, Stourhead. A Bronze Age burial mound (c.2000 BC) above a causewayed enclosure (c.3600 BC)

Then ideal conservation objectives are set for each feature and actions are agreed which after 12 months will be scored depending on how much has been achieved. This is a good time to meet conservation colleagues and work together on integrated conservation management. This week the CPI scores for Kingston Lacy (Dorset), Stourhead (Wiltshire), Tyntesfield (Somerset) and Purbeck (Dorset) have been decided.