Quantification and Modelling Approaches of Geoarchaeological Processes – the Course, Construction and Collapse of the Carolingian Canal Fossa Carolina

Quantification and modelling approaches of geoarchaeological processes – The course, construction and collapse of the Carolingian canal Fossa Carolina

Von der Fakultät für Physik and Geowissenschaften der Universität Leipzig genehmigte

DISSERTATION

zur Erlangung des akademischen Grades Doctor rerum naturalium Dr. rer. nat.

vorgelegt von M. Sc. Geogr. Johannes Schmidt geboren am 08.12.1988 in Jena

Gutachter: Prof. Dr. Christoph Zielhofer

Prof. Dr. Hans-Rudolf Bork

Tag der Verleihung: 19.04.2021

Dedicated to Luni

„Denken and Wissen sollten immer gleichen Schritt halten. Das Wissen bleibt sonst tot and unfruchtbar“

oder

„Die Natur muss gefühlt werden“

Alexander von Humboldt

Abstract

The European watershed runs across Europe and divides the river basins into a northern component and a southern component. Therefore, the river basins drain either northwards into the North Sea, Baltic Sea or North Atlantic, or southwards into the Mediterranean or Black Sea. There is no natural, fluvial connection between the North and South components. Between the cities of Weißenburg and Treuchtlingen in Bavaria, the Central European Watershed separates the catchment areas of the Rhine-Main System and the Danube System. The Swabian Rezat has its spring at the steep slope of the Frankenalb and runs northwards towards Weißenburg and drains into the Rhine-Main system. South of the European watershed flows the Altmühl, which drains into the Black Sea via the Danube. In the early Middle Ages, shipping routes were of the highest geostrategic interest and were used for military purposes in addition to the movement of people and goods. It was probably for these reasons that Charlemagne built a canal in 792-793 AD to provide a navigable connection (also known as Fossa Carolina). The Swabian Rezat and Altmühl rivers have a linear distance of ~2 km and the natural features (water supply, sedimentary architecture of the valley landscape, geological conditions) are almost ideal for the technical possibilities for canal construction in the early Middle Ages. The canal is of outstanding engineering quality. In the meantime, the canal has been filled in and is only visible in some places.

In addition to archaeological, historical and geoscientific work, especially from the 20th century, an interdisciplinary research project has been investigating various aspects of the canal in since 2010. The focus of the research so far has been on proving the actual course of the canal, the archaeological documentation of specific finds and features of the canal and its surroundings, the history of the sedimentation of the canal, and the investigation of structures accompanying the construction. Furthermore, research was carried out on the archaeological and historical classification of the structure in the history of canal constructions. The decoding and comparison of geoarchaeological finds and processes is characterised by qualitative approaches. The description and documentation of a phenomenon and the way in which it manifests itself are usually the main focus. In recent years, however, methods have been developed from the disciplines of archaeology and archaeometry, geosciences and geoinformatics which serve to quantify geoarchaeological processes in the narrower sense in relation to monuments, but also in the broader sense in relation to landscapes. The aim of this work is to quantify individual processes of planning, construction and the collapse of the Fossa Carolina.

High-resolution digital elevation models usually have the disadvantage that anthropogenic structures produce artefacts in the data which, although they represent the current surface, do not allow for natural- landscape-related, hydrogeographic calculations. A modelling of the pre-modern relief has been achieved by precise elimination of anthropogenic structures such as roads, buildings, railway lines, etc. The data basis is provided by official geodata of the State Office, OpenStreetMap and the historical cadastre

boundaries recorded in the beginning of the 19th century. For the semi-quantitative verification and validation of the modelling results, paleo-surfaces were recovered from drillings. The high-resolution pre- modern digital elevation model for the area around the Fossa Carolina allowed the calculation of the shortest hypothetical canal course. This is similar to the built original, conspicuously S-shaped and proves the topographical dependence of the canal construction. Hydrogeographic indices, such as the topographic wetness index, were used to justify the deviations of the actual canal course from the ideal line. It has demonstrated that the Carolingian builders had a comprehensive knowledge of the natural features of the area. They deliberately avoided wet areas with high groundwater levels and organic sediments, which would have been too unstable for the earthworks of the canal.

Approaches to combine different geoarchaeological and geophysical methods and are not unknown, but mostly qualitative in nature. The previous investigations and analyses at the Fossa Carolina were combined by a new integrative and quantitative approach. Geometric and stratigraphic results from 3 archaeological excavation sections, 39 core drillings, as well as 2 direct push sensing transects with 105 colour logs were included. The resulting numerical 3D-model of the state of maximum excavation of the approx. 3 km long canal allows for the first time the calculation of the total excavation volume. Almost 300,000 m³ of material were moved for the construction. In contrast, the excavation dams that are still visible today, have a total volume of only approx. 120,000 m³. Therefore, almost two thirds of the total excavated material are no longer preserved in the dams, but have been eroded or removed since the end of the construction. Moreover, the excavated volume is not evenly distributed over the entire course of the canal, but the main part is located in the area of the Central European watershed. The 3D-model also allows the calculation of the trench bottom level of the longitudinal section of the canal. Thereby the summit zone of the stepped canal was identified.

The basis of the infilled sediment of the canal trench is formed by sediments that have initially slipped back from the excavation walls (so-called backfills). Stratigraphically, they are accumulated directly on the canal trench bottom. A systematic evaluation of the geochemical and stratigraphical findings could show that the backfills can be declared as initial collapse sediments. Radiocarbon dating of macro remains date the backfills around the construction period 792/793 AD. Dendrochronological analyses also show that the timber used in the Fossa Carolina were felled and used in the years 792 AD and 793 AD. The spatial analysis of the felling ages also shows that construction progressed from north to south. The absence of still water sediments on the canal trench floor and the timber serve as a chronological framework and provides relative dating of the backfill sediments. Subsequently, the application of the 3D-modelling approach was transferred to the backfill sediments. The resulting spatial and quantitative distribution of the backfills along the entire canal course shows the concentration of these sediments in the Central Sections around the Central European Watershed. Sediment budgeting, as applied in geomorphology, could now be applied adaptively to the backfills. Nearly 35,000 m³ of backfill sediments were calculated in the Central and West- East Sections. This corresponds to approx. 15 % of the total excavation volume in this area. Therefore, these sections can be designated as a collapse zone. According to contemporary written sources, the canal II

was abandoned due to unstable dams and back-sliding material. For the first time, the quantitative results provide a scientific proof for this hypothesis.

In summary, this work is an example of the application of numerical and quantitative methods in the field of geoarchaeology and shows the value of these methods in understanding geoarchaeological processes and their reconstruction.

III

Kurzzusammenfassung

Die Europäische Hauptwasserscheide verläuft quer durch Europa and trennt die Flusseinzugsgebiete in eine Nordkomponente and eine Südkomponente. Die Einzugsgebiete entwässern demnach entweder nach Norden in die Nordsee, Ostsee, oder den Nordatlantik oder aber nach Süden in das Mittelmeer oder das Schwarze Meer. Es gibt keine natürliche, fluviale Verbindung zwischen der Nord- and der Südkomponente. Zwischen den Städten Weißenburg and Treuchtlingen in Bayern trennt die Europäische Hauptwasserscheide die Einzugsgebiete des Rhein-Main Systems and des Donau Systems. Die Schwäbische Rezat entspringt an der Frontstufe der Frankenalb and verläuft nordwärts in Richtung Weißenburg and entwässert in das Rhein-Main System. Südlich der Europäischen Wasserscheide fließt die Altmühl, welche über die Donau in das Schwarze Meer entwässert. Im Frühmittelalter waren Schifffahrtswege von höchstem geostrategischem Interesse and dienten neben dem Personen- and Warenverkehr auch der militärischen Nutzung. Vermutlich aus diesen Gründen ließ Karl der Große 792/793 AD einen Kanal anlegen, welcher eine schiffbare Verbindung herstellen sollte. Luftlinie kommen sich Schwäbische Rezat and Altmühl auf knapp 2 km nahe and die naturräumliche Ausstattung (Wasserdargebot, sedimentäre Architektur der Tallandschaft, geologischer Untergrund) ist für die technischen Möglichkeiten des Frühmittelalters nahezu ideal. Der Kanal ist von herausragendem ingenieurstechnischem Niveau. Inzwischen ist der Kanal verfüllt and nur an manchen Stellen weiterhin sichtbar.

Neben archäologischen, historischen and geowissenschaftlichen Arbeiten vor allem aus dem 20. Jahrhundert, konnte ein interdisziplinäres Forschungsprojekt seit 2010 den Kanal in diversen Fragen untersuchen. Die Fokusse der bisherigen Forschungen lagen auf dem Nachweis des tatsächlichen Kanalverlaufs, der archäologischen Dokumentation konkreter Funde and Befunde des Kanals and seines Umfelds, der Verfüllungsgeschichte des Kanals, sowie der Untersuchung baubegleitender Strukturen. Weiterhin gab es Forschungen zur archäologischen and historischen Einordnung des Bauwerks in die Kanalbaugeschichte. Die Entschlüsselung and der Vergleich geoarchäologischer Funde and Prozesse ist von qualitativen Ansätzen geprägt. Die Beschreibung and Dokumentation eines Phänomens and die Art and Weise der Ausprägung stehen dabei meist im Vordergrund. In den letzten Jahren sind jedoch aus den Disziplinen der Archäologie and Archäometrie, der Geowissenschaften sowie der Geoinformatik Methoden entwickelt worden die dazu dienen geoarchäologische Prozesse, im engeren Sinne denkmalsbezogen, aber auch im weiteren Sinne landschaftsbezogen zu quantifizieren. Ziel dieser Arbeit ist die Quantifizierung einzelner Prozesse der Planung, Konstruktion and Kollapses des Karlsgrabens.

Hochauflösende Höhenmodelle haben stets den Nachteil, dass anthropogene Strukturen Artefakte in den Daten produzieren, die zwar die heutige Oberfläche abbilden, aber keine naturlandschaftsbezogenen, hydrogeographischen Berechnungen zulassen. Eine Modellierung des Vor-modernen Reliefs wurde durch

die gezielte Beseitigung von anthropogenen Strukturen wie Straßen, Gebäude, Bahnlinien, etc. erreicht. Die Datengrundlage bilden offizielle Geodaten des Landesamts, OpenStreetMap sowie die historische Uraufnahme aus dem beginnende 19. Jahrhundert. Für die semiquantitative Überprüfung and Validierung des Modellierungsergebnisses wurde anhand von erbohrten Paläooberflächen vorgenommen. Das hochauflösende vor-moderne Digitale Höhenmodell für das Gebiet um die Fossa Carolina ermöglichte die Berechnung des kürzesten, hypothetischen Kanalverlaufs. Dieser ist ähnlich wie das gebaute Original auffällig S-förmig and belegt die topographische Abhängigkeit des Kanalbaus vom Relief. Hydrogeographische Indizes, wie der topographic wetness index, wurden herangezogen um die Abweichungen des tatsächlichen Kanalverlaufs von der Ideallinie zu begründen. Es zeigt sich, dass die karolingischen Bauherren ein umfassendes Wissen zur naturräumlichen Ausstattung des Gebietes hatten. Sie vermieden gezielt feuchte Gebiete mit hoch anstehendem Grundwasser and organischen Sedimenten, welche zu instabil für die Erdarbeiten des Kanals gewesen wären.

Ansätze um verschiedene geoarchäologische and geophysikalische Methoden zu vereinen and sind nicht unbekannt, aber letztlich meistens qualitativer Natur. Die bisherigen Untersuchungen and Analysen am Karlsgraben konnten durch einen neuen integrativen and quantitativen Ansatz kombiniert werden. Es flossen geometrische and stratigraphische Ergebnisse aus 3 archäologischen Grabungsschnitten, 39 Rammkernbohrungen, sowie 2 Direct push sensing Transekten mit 105 colour logs ein. Das resultierende numerische 3D-Modell des maximalen Ausbauzustands des ca. 3 km langen Kanals ermöglicht zum ersten Mal eine Berechnung des Gesamtaushubvolumens. Knapp 300.000 m³ Material wurden für den Bau bewegt. Die heute noch sichtbaren Aushubwälle besitzen hingegen ein Gesamtvolumen von nur ca. 120.000 m³. Fast Zwei Drittel des gesamten Aushubs sind also heute nicht mehr in den Wällen konserviert, sondern seit dem Ende des Baus erodiert oder abgetragen. Das Aushubvolumen ist zudem nicht über den gesamten Kanalverlauf gleich verteilt, sondern der Hauptanteil liegt im Bereich der durchstochenen Europäischen Hauptwasserscheide. Das 3D-Modell ermöglicht zudem die Berechnung des Sohlniveaus des Kanallängsschnitts. Dabei konnte die Scheitelzone des getreppten Kanals ausgewiesen werden.

Die Basis der Verfüllung des Kanals bilden von den Aushubwällen zurückgerutschte Sedimente (sog. backfills). Stratigraphisch liegen diese direkt auf der Kanalsohle. Eine systematische Auswertung der geochemischen and stratigraphischen Befunde konnte zeigen, dass die backfills als initiale Kollapssedimente deklariert werden können. Radiokohlenstoffdatierungen von Makroresten datieren die backfills rund um die Bauzeit 792/793 AD. Dendrochronologische Analysen zeigen weiterhin, dass die verbauten Hölzer des Karlsgrabens in den Jahren 792 AD and 793 AD gefällt and verbaut wurden. Die räumliche Anaylse der Fällalter kann zudem einen Baufortschritt von Nord nach Süd nachweisen. Das Fehlen von Stillwassersedimenten auf der Kanalsohle bzw. den Bauhölzern dient als chronologischer Rahmen and Relativdatierung der backfill Sedimente. Die Anwendung des 3D-Modellierungsansatzes wurde auf die backfill-Sedimente übertragen. Die resultierende räumliche and quantitative Verbreitung der backfills entlang des gesamten Kanalverlaufs zeigt die Konzentration dieser Sedimente im zentralen Grabenbereich rund um die Europäische Hauptwasserscheide. Eine Sedimentbudgetierung, wie sie in der V

Geomorphologie angewandt wird, konnte nun hier adaptiv eingesetzt werden. Fast 35.000 m³ backfill- Sedimente sind an der Basis der Verfüllung allein im Zentralen and West-Ost Bereich zu finden. Dies entspricht ca. 15 % des Gesamtaushubvolumens in diesem Bereich. Daher kann diese Zone als Kollapszone ausgewiesen werden. Nach den zeitgenössischen Quellen wurde der Kanal aufgrund von instabilen Wällen and zurückgerutschtem Material aufgegeben. Die quantitativen Ergebnisse liefern nun zum ersten Mal einen naturwissenschaftlichen Beleg für diese These.

Zusammenfassend ist diese Arbeit ein Beispiel für die Anwendung für numerischen and quantitativen Methoden im Bereich der Geoarchäologie und zeigt den Mehrwert der Methoden im Verständnis für geoarchäologische Prozesse and deren Rekonstruktion.

Acknowledgements

I am grateful for the opportunity to work on the “Fossa Carolina project” with Prof. Dr. Christoph Zielhofer and the rest of his group at the Institute of Geography of Leipzig University. With his support, I undertook several field campaigns to Charlemagne’s canal and the surrounding landscape in Franconia. Furthermore, he encouraged me in many ways and he gave me the freedom to develop my ideas, share and discuss them with colleagues.

Lukas Werther always had an ear for my ideas and problems. His enduring constructive nature and open- minded thoughts helped me to orient myself in the scientific world. Thanks to our fruitful communication, we were able to ask new kinds of questions and go deeper into interdisciplinarity in the intersection of geosciences, environmental history and archaeology.

Johannes Rabiger-Völlmer as scientific phd-colleague and friend was a nearly daily interlocutor about the Fossa Carolina, the landscape and all our data. Furthermore, we had great field campagins together and there were always sweets in the car!

The large group of project members were always of benefit to the interdisciplinarity and after the years, I understood more and more of the languages of the disciplines and their aims and challenges. I would like to thank Stefanie Berg, Sven Linzen, Peter Ettel, Ulrike Werban, Peter Dietrich, Dennis Wilken, Annika Fediuk, Franz Herzig and Andreas Stele to name just a few.

A lot of laboratory work was done by Birgit Schneider and Katja Pöhlmann. Thank you for all your effort and time. Nevertheless, several students are thanked for their help in the lab as student assistances. I used a lot of already available data for my modelling approaches. Most of the stratigraphic data was produced by Eva Leitholdt. Thank you for sharing your results and all your metadata at the beginning of my work. This helped a lot.

Every day, I met interested colleagues from our institute and we shared the daily challenges. With your ready ears, my daily work routine was easier. Thank you. By the way, I learned a lot from your disciplines and methods; which has widened my mind.

I want to thank Jürgen Heinrich, as chairman of the phd-commission for his words of appreciation. Further, he supported me throughout my Bachelor and Master studies and taught me the basics of geomorphology and physical geography.

I thank also all colleagues from all over the world, whom I met during workshops and conferences. The discussions about scientific concepts, ideas and experiences were invaluable…

VII

The German Research Foundation (DFG) is thanked for their funding within the scope of the priority program 1630 “Häfen von der Römischen Kaiserzeit bis zum Mittelalter - Zur Archäologie and Geschichte regionaler and überregionaler Verkehrssysteme“.

Scientific results should be open and therefore I published my results in Open Access Journals. The Article processing charges were paid by Leipzig University within their program of Open Access publishing. Thank you for open science.

Of course, I don’t forget my friends and my family. Life is not only work. We had and have wonderful times together, where the dissertation didn’t play a role. Your support gave me the resilience required to work over the years on this topic.

I want to end with a quote by Ingo Eichfeld on a phd-meeting of the priority program 1630 Harbours of the german research foundation “Eine Promotion wird nicht zu Ende gebracht, Sie wird kontrolliert abgebrochen“ (“A dissertation will not be completed, it will be halted in a controlled manner”).

VIII

Preface

The present dissertation thesis was prepared and written within the project “Fossa Carolina – Bindeglied der Hafennetzwerke an Rhein and Donau. Studien zur Überwindung der europäischen Hauptwasserscheide“ (“Fossa Carolina – Connection of harbour networks between Rhine and Danube. Studies about bridging the Central European Watershed”). The project was part of the priority program 1630 “Häfen von der Römischen Kaiserzeit bis zum Mittelalter - Zur Archäologie and Geschichte regionaler and überregionaler Verkehrssysteme” (“Harbours from the Roman Period to the Middle Ages – Archaeology and History of regional and supra-regional traffic systems”)

During my scientific work in the project I published several studies either as first-author or as co-author. The present dissertation thesis includes the three main papers about modelling and quantification approaches at Charlemagne’s summit canal. Additionally, one co-authorship paper is attached in the supplementary material, because the results of this study are part of the basic data base for the modelling procedures. Further co-author publications, that accrued within the project about the Fossa Carolina are not claimed in this thesis. Instead, they are mentioned in the personal publication list in the appendix. Furthermore, each manuscript has its own reference section. The reference section at the end of the dissertation includes all references from the framework. The list of tables and figures includes only those of the framework.The supporting online material of the published papers can be found at each journal homepage or on the attached CD.

Table of contents

Abstract …… I Kurzzusammenfassung …… IV Acknowledgements …… VII Preface …… IX Table of contents …… XI List of Tables and Figures …… XIV

Chapter 1 - Introduction …… 1 1.1 General Information about Fossa Carolina …… 1 1.2 Study area …… 5 1.2.1 The landscape …… 5 1.2.2 The canal …… 6 1.3 State of the Art “Quantification and modelling approaches of geoarchaeological processes” …… 9 1.4 Material and methods …… 12 1.4.1 Data …… 13 1.4.2 Methodological approaches …… 16 1.5 Aims of the study …… 18

Chapter 2 - Shaping pre-modern digital terrain models: The former topography at Charlemagne’s canal construction site ……20 2.1 Abstract …… 21 2.2 Introduction …… 22 2.3 Study area …… 24 2.4 Material and methods …… 26 2.4.1 Basic data …… 26 2.4.2 Modelling approach …… 27 2.4.3 Local-relief model …… 28 2.4.4 Validation of the modelled pre-modern DTM …… 28 2.4.5 Least cost path analysis …… 29 2.4.6 Hydrogeographical analysis …… 30 2.5 Results …… 30 2.5.1 Interim results from stage 2 of the modelling approach …… 30 2.5.2 Results from stage 3 of the modelling approach …… 31 2.5.3 Model comparison …… 32 2.5.4 Validation of the modelled pre-modern DTM …… 33 2.5.5 The modelled pathway of Fossa Carolina …… 36 2.5.6 Hydrogeographic indices …… 37 2.6 Discussion …… 38 2.6.1 Palaeo-surface modelling approaches …… 38 2.6.2 Model performance …… 42

2.6.3 Evidence of excellent Carolingian knowledge in engineering …… 42 2.7 Conclusion …… 43 2.8 References …… 44

Chapter 3 - 3D-Modelling of Charlemagne’s Summit Canal (Southern Germany) – Merging Remote Sensing and Geoarchaeological Subsurface Data …… 50 3.1 Abstract …… 51 3.2 Introduction …… 52 3.3 Material and methods …… 55 3.3.1 Study area …… 55 3.3.2 Data acquisition …… 59 3.3.2.1 LiDAR Digital Terrain Model …… 59 3.3.2.2 Pre-modern Digital Terrain Model …… 59 3.3.2.3 Magnetic survey …… 59 3.3.2.4 Vibra-Coring …… 59 3.3.2.5 Direct push sensing …… 60 3.3.2.6 Archaeological excavations …… 60 3.3.3 Modelling routine …… 61 3.4 Results …… 63 3.4.1 Canal course …… 63 3.4.2 Cross-section reference geometries …… 64 3.4.3 Application of cross section reference geometries to vibra-coring and additive transects … 66 3.4.4 3D-Model …… 66 3.4.5 Volume calculation …… 66 3.5 Discussion …… 67 3.5.1 3D-modelling approach and quality …… 67 3.5.2 The scientific history of Fossa Carolina volume calculations …… 68 3.5.3 Where has all the material gone? …… 70 3.6 Conclusions …… 70 3.7 References …… 71

Chapter 4 - Sediment budgeting of short-term backfilling processes – the erosional collapse of a Carolingian canal construction …… 77 4.1 Abstract …… 78 4.2 Introduction …… 79 4.3 Fossa Carolina and its geographical setting …… 80 4.4 Material and Methods …… 84 4.4.1 Data acquisition …… 84 4.4.1.1 Geodata …… 84 4.4.1.1.1 LiDAR Digital Terrain Model …… 84 4.4.1.1.2 Pre-modern Digital Terrain Model …… 84 4.4.1.1.3 3D-Model of the Fossa Carolina …… 84 4.4.1.2 Drillings …… 85

XII

4.4.1.3 Sediment geochemistry …… 85 4.4.1.4 Radiocarbon dating …… 85 4.4.1.5 Direct push sensing …… 87 4.4.1.6 Archaeological excavations …… 87 4.4.1.7 Dendrochronological analysis …… 88 4.4.2 Modelling and sediment budgeting …… 89 4.5 Results …… 89 4.5.1 Backfill sediment identification …… 89 4.5.2 Radiocarbon results …… 90 4.5.3 Dendrochronological results …… 91 4.5.4 Backfill sediment storage …… 91 4.5.5 Backfill sediment budget …… 92 4.6 Discussion …… 93 4.6.1 Sediment budget approach …… 93 4.6.2 Spatial distribution of backfill sediments …… 94 4.6.3 Canal construction progress, abandonment and collapse …… 95 4.6.4 Large-scale control or local feature? …… 97 4.7 Conclusion …… 99 4.8 References …… 99

Chapter 5 – Discussion …… 108 5.1 Methodological approaches …… 108 5.1.1 Palaeo-surface modelling …… 108 5.1.2 3D-modelling …… 109 5.1.3 Geoarchaeological sediment budgeting …… 110 5.2 Key findings …… 111 5.2.1 Canal course …… 111 5.2.2 Canal trench volume …… 112 5.2.3 Summit level of the Fossa Carolina …… 114 5.2.4 Canal construction, collapse and backfill sediment budget …… 114

Chapter 6 – Conclusions …… 118

References …… 120

Appendix …… A A1 Minimalinvasive Direct Push Erkundung in der Feuchtboden(geo)archäologie am Beispiel des Karlsgrabens (Fossa Carolina) …… B A2 Short Curriculum vitae …… Y A3 Author publication list …… Z A4 Paper Contributions …… CC A5 Declaration of Originality …… EE A6 Bibliographic information …… FF XIII

XIV

List of Tables and Figures

Figure 1.1: European context of the position of the Fossa Carolina. Blue lines show major rivers. The dashed red line shows the Central European Watershed. The yellow star indicates the position the canal. …… 1

Figure 1.2: The earth work of the Fossa Carolina canal construction (Bl. 20r) from the Würzburger Bischofschronik des Lorenz Fries (M. ch. f. 760) of the Würzburg University library (modified). …… 2

Figure 1.3: Aerial image of the Fossa Carolina with shadow marks in the snow highlighting terrain differences. Black arrows show the canal course. (Bavarian State Department of Cultural Heritage BLfD 1985). …… 3

Figure 1.4: Fossa Carolina and the surrounding landscape with main villages on the sub-regional to local scale. …… 6

Figure 1.5: Impressions of the Fossa Carolina. a) Aerial image of the canal with shadow marks highlighting terrain differences, white arrows show the canal course (modified after Bavarian State Department of Cultural Heritage BLfD 1985), b) Pond at the southern edge of the canal in the village Graben (Photo: Lukas Werther), c) Central Section with afforested dams (Photo: Lukas Werther), d) West-East Section with drainage ditch, lateral dams (Photo: Eva Leitholdt), e) Northeastern Section with a flat present relief and agricultural land use as well as ongoing direct push sensing survey in the background (Photo: Johannes Schmidt), f) Canal profile in the Northeastern Section of the Fossa Carolina (2016–S1) during the excavation. The trench which has been excavated in 792/793 AD is clearly visible, the trench bottom is not yet excavated (Photo: Lukas Werther). …… 8

Figure 1.6: Flowchart of the input data and application in the different modelling and quantification approaches used in this thesis. …… 13

Figure 2.1: Geographical setting of the study area on different scales. a) Central European setting in relation to main drainage basins and the Central European Watershed. b) Regional setting of the Fossa Carolina in relation to tributaries of the Rhine-Main drainage system and the Danube drainage system. c) Local setting of the Fossa Carolina (hillshade) with five sub-divisions (Zielhofer et al. 2014) and ground-truth validation points. LiDAR data have been provided by the Bavarian State Office for Land Surveying. …… 25

Figure 2.2: Photographs of Charlemagne’s summit canal. a) Aerial image of the canal with shadow marks highlighting terrain differences (Bavarian State Department of Cultural Heritage BLfD 1985), b) Pond at the southern edge of the canal in Graben (Photo: Werther 2017), c) Central section with afforested ramparts (Photo: Werther 2017), d) West-East-section (Photo: Leitholdt 2014). …… 26

Figure 2.3: Flowchart of modelling approach in three steps. 1.) Base data acquisition (recent LiDAR DTM, aerial images, shapes of recent land use, historical map), 2.) Deriving specific data layers (* compound specific buffer widths), 3.) Creating a comprehensive buffer layer, removing the affected cells from the modern DTM, interpolating and filtering the residuals. …… 29

Figure 2.4: Interim results of the derived data from stage 2 of the modelling approach (see Fig. 3). a) Roads, b) railway lines, c) buildings, d) cadastre boundaries derived from present land use layer, e) cadastre boundaries derived from the 1st edition of cadastre sheets, f) additional anthropogenic structures derived from LiDAR-based DTM and aerial image, g) Fossa Carolina, h) comprehensive layer. …… 31

Figure 2.5: Pre-modern DTM. a) Pre-modern DTM illustrated with hillshade, b) enlarged section with present DTM and hillshade, c) enlarged section with the pre-modern DTM and hillshade. …… 32

Figure 2.6: Local Relief Model of the Fossa Carolina Central Section (a) and the Altmühl floodplain (b). Red indicates sediment aggradation in comparison with the modelled pre-modern DTM; blue indicates sediment removal in comparison with the modelled pre-modern DTM. …… 33

Figure 2.7: Validation at drilling transect in West-East section of the Fossa Carolina. Drilling data from Zielhofer et al. (Zielhofer et al. 2014); the modelled surface is derived from the pre-modern DTM. …… 34

Figure 2.8: Validation at archaeological excavation site (Werther et al. 2015) supplemented with drilling data (Zielhofer et al. 2014) in northern section of Fossa Carolina. The modelled surface is derived from the pre-modern DTM. …… 34

Figure 2.9: Validation plots of the modelled surface (black dots) and the present surface (red dots) against the observed surfaces (pre-modern height) derived from drillings and excavations. The RMSE (root- mean-square error) is shown in each legend box to estimate the error. a) Total set of all validation points; b) validation points (Kirchner et al. 2018) in the Altmuehl floodplain reflecting areas with less anthropogenic overprint (mainly quasi-natural accumulation of alluvial deposits); c) validation from previous studies (Werther et al. 2015) in the direct surroundings of Fossa Carolina reflecting areas with strong anthropogenic overprint (linear structures). …… 35

Figure 2.10: Pre-modern DTM with the Carolingian course of the canal (yellow line) and the modelled course (least cost path analysis). …… 37

Figure 2.11: Topographic wetness index with the present course of Fossa Carolina (yellow line) and the modelled course (brown line) based on the pre-modern DTM. Topographic Wetness Index (TWI) in greyscale shows potential wet areas (white colours). Grey arrows show slight deviations between both courses. Blue dotted lines show the present waterways. …… 38

Table 2.1: Basic data acquisition and quality. …… 27

XVI

Table 2.2: Compilation of compound-specific buffer widths. …… 28

Table 2.3: Comparison of Root Mean Square Errors (RMSE) between the Fossa Carolina and Altmühl validation point clusters and between the measured vs. modelled pre-modern surface and the measured pre-modern surface vs. the present DTM. …… 36

Table 2.4: Comparison of modelling approaches in palaeo-terrain research with geoarchaeological issues. …… 39

Figure 3.1: Geographical setting of the study area. a) Main Central European drainage basins and the Central European Watershed. b) Regional setting of the Fossa Carolina in relation to tributaries of the Rhine-Main drainage system and the Danube drainage system (modified after Zielhofer et al. 2014). …… 55

Figure 3.2: Local setting and course of the Fossa Carolina and its subdivision in I) Central Section, II) West-East Section, III) Northern Section, IV) North-Eastern Section. All input data for the subsequent modelling are shown, including drillings, modelled trench bottom transects and additive transects. Cross-section reference geometries a) “West-East Section”, b) “The Anomaly” base on direct push sensing data. Cross-section reference geometries c) “2013”, d) “2016 – S1”, e) “2016 – S2” base on archaeological excavations. LiDAR data have been provided by the Bavarian Land Surveying Office. …… 57

Figure 3.3: Impressions of the Fossa Carolina. a) Aerial image of the canal with shadow marks highlighting terrain differences, white arrows show the canal course (Bavarian State Department of Cultural Heritage BLfD 1985), b) 3D view of the present DTM derived from LiDAR-data, which shows the prominent dams and extent of the construction, c) Central Section with afforested dams (Photo: Lukas Werther), d) West-East Section with drainage ditch, lateral dams and ongoing direct push sensing (Photo: Johannes Völlmer), e) Canal profile in the North-Eastern Section of the Fossa Carolina (2016 – S1) during the excavation. The trench which has been excavated in 793 AD is clearly visible, the trench bottom is not yet excavated. Oak piles mark the Eastern edge of the fairway (Photo: Lukas Werther; see also Fig. 3.8b), f) Canal profile at the Northern end of the Fossa Carolina (2016 – S2) with least depth, width and excavated volume. The trench which has been excavated in 793 AD and later filled is clearly visible. North of the profile, oak stakes mark the edges of the fairway (Photo: Lukas Werther; see also Fig. 3.8c). …… 58

Figure 3.4: Flow chart of the modelling approach and subsequent calculation of volumes. a) Creation of cross-section reference geometries, b) Modelling of trench bottom cross-section geometries, c) Equidistant spacing of trench bottom cross-sections, d) 3D-model of the Fossa Carolina trench bottom, e) 3D-model of the Fossa Carolina and the surrounding pre-modern topography, f) Calculation of the volume of the excavated material and of the volume of the present dams. …… 62

Figure 3.5: Localisation of the Fossa Carolina course. a) Total study area, magnetic maps and DTM data, b) Central Section with canal course and dam ridges, c) West-East Section with canal course, dam ridges and Fluxgate magnetic map, d) North-Eastern Section with canal course and SQUID magnetic map, Greyscale: ± 10nT/m. …… 63

XVII

Figure 3.6: Cross-section reference geometry based on direct push sensing. The sketch of the “WE cross- section” transect shows present and pre-modern surface levels, trench geometries and maximum excavation depth. …… 64

Figure 3.7: Cross-section reference geometry based on direct push sensing. The sketch of the “The Anomaly” transect shows present and pre-modern surface levels, trench geometries and maximum excavation depth. …… 64

Figure 3.8: Archaeological excavation reference geometries. a) “excavation 2013”, b) “2016/S1” and c) “2016/S2”. …… 65

Figure 3.9: Exemplary transfer of the cross-section reference geometry “2016/S1” to the vibra-coring transect “Märzkampagne”. …… 66

Figure 3.10: 3D model of the Fossa Carolina with surrounding pre-modern topography and present remnants of the dams. 3D model is 15x exaggerated and its representation method is “perspective”. a) Entire canal course with a view from southeast, b) enlarged section of the prominent bend between the West-East and Northern Section, c) enlarged section of the Central Section. …… 67

Figure 3.11: Spatial distribution of excavated volume at the Fossa Carolina. Volume amounts are given in m³ per cell (0.25 m²). …… 69

Table 3.1: Depth accuracy, scale and stratigraphical resolution of trench geometry data; scale classification according to Zielhofer et al. (2018a). …… 54

Table 3.2: Sections of the Fossa Carolina and corresponding cross-sections reference geometries. …… 62

Table 3.3: Scientific history of Fossa Carolina volume estimations. …… 69

Table 3.4: Relative amounts of excavated volume at individual Fossa Carolina sections. …… 70

Figure 4.1: The geographical setting of the study area. a) Main Central European drainage basins and the Central European Watershed. b) Regional setting of the Fossa Carolina in relation to tributaries of the Rhine-Main drainage system and the Danube drainage system. …… 81

Figure 4.2: Local setting and course of the Fossa Carolina and its subdivisions in the I) Altmühl floodplain, II) Central Section, III) West-East Section, IV) Northern Section, and V) North-Eastern Section. Direct push sensing transects a) “WE-section”, b) “The Anomaly” and Archaeological excavations c) “2013 – Trench 1”, d) “2016 – Trench 2”, e) “2016 – Trench 3”. LiDAR data were provided by the Bavarian Land Surveying Office. …… 82

XVIII

Figure 4.3: Impressions of the Fossa Carolina, drillings, and archaeological excavation. a) Aerial image of the canal with shadow marks highlighting terrain differences, white arrows show the canal course (Bavarian State Department of Cultural Heritage BLfD 1985), b) 3D view of the present Digital Terrain Model derived from LiDAR-data, which show prominent dams and extent of the construction, c) Exemplary drilling results from the Central Section with stratigraphic description according to Zielhofer et al. (2014); the red line marks the proven trench bottom, d) Recovered timber from trench 2 (“d” in Fig. 4.2), e) Recovered timber from trench 1 (“c” in Fig. 4.2). …… 83

Figure 4.4: Summarised statistics of the a) sand content and b) TOC content of all samples from drillings of the Fossa Carolina of the sediment facies of the “natural, sterile parent material” and the excavated, eroded and re-accumulated backfills. …… 90

Figure 4.5: Calibrated radiocarbon age multiplot of the backfill sediments. Grey bars show 2-sigma ranges and black bars show 1-sigma ranges. Radiocarbon ages new calibrated using CALIB software (Stuiver et al. 2019) with the Intcal13 calibration curve (Reimer et al. 2013). The black line shows the construction of the Fossa Carolina in 792/793 AD. …… 91

Figure 4.6: Spatial sediment budgeting model of the backfill thickness of the Carolingian canal Fossa Carolina. The colour ramp of the backfill volume is displayed with a gamma stretch of 2. …… 92

Figure 4.7: Canal longitudinal section with the present surface (green line; derived from the present LiDAR DTM), pre-modern surface (grey line; derived from the pre-modern DTM from Schmidt et al. 2018), the maximum Carolingian excavation depth (black line, derived from the 3D-model of the Fossa Carolina from Schmidt et al. 2019) and the top level of the backfills (orange line; derived from the results presented in this study). The grey area shows the thickness of the backfills for the entire Fossa Carolina. The y-axis is 80-fold superelevated to show prominent information. …… 95

Figure 4.8: Spatial distribution of dendrochronological derived felling dates of timber from the three archaeological excavation along the Fossa Carolina. …… 97

Table 4.1: Recalibrated radiocarbon ages from Zielhofer et al. (2014) and Leitholdt et al. (2014) of the backfill sediments. KIA samples were conducted at the Kiel AMS facility, SUERC samples were conducted at the Glasgow AMS facility and MAMS samples were conducted at the Mannheim AMS facility. …… 86

Table 4.2: Dendrochronological results of recovered timber (oak) and wood fragments from three archaeological excavation trenches. …… 87

Table 4.3: Quantitative results of the sediment budgeting model of the backfill thickness and its spatial distribution for the entire canal and broken down to each canal section. a Maximum excavation volume (from Schmidt et al. 2019), b Canal trench volume without backfills (bottom edge of the organic fills). …… 93

the maximum Carolingian excavation depth (black line, derived from the 3D-model of the Fossa Carolina from Schmidt et al. 2019) and the top level of the backfills (orange line; Schmidt et al. 2020). The grey area shows the thickness of the backfills for the entire Fossa Carolina. The y-axis is 80-fold superelevated to show prominent information (modified after Schmidt et al. 2020). …… 110

Figure 5.2: Comparison of modelled (brown line) and present (yellow line) canal course of the Fossa Carolina. The topographical wetness index (TWI; greyscale layer) in the background shows potential wet areas (white colours). Arrows show the main deviations of both canal courses (modified after Schmidt et al. 2018). …… 112

Figure 5.3: Spatial distribution of the maximum excavation volume of the Fossa Carolina. Volume amounts given in m³ per cell (0.25m²) Background: Analytical hillshade of the LiDAR DTM in greyscale (modified after Schmidt et al. 2019). …… 113

Figure 5.4: Canal construction progress reconstruction. Spatial distribution of dendrochronological derived felling dates from recovered construction timber from three archaeological excavations along the Fossa Carolina canal course (modified after Schmidt et al. 2020). …… 115

Figure 5.5: Spatial distribution of backfill sediment thickness along the Fossa Carolina canal course (modified after Schmidt et al. 2020). …… 116

Table 5.1: Quantitative results of the sediment budgeting model of the backfill thickness and its spatial distribution for the entire canal and broken down to each canal section. a) Maximum excavation volume (from Schmidt et al. 2019), b) Canal trench volume without backfills (bottom edge of the organic fills). …… 117

1 Introduction

Chapter 1 - Introduction

1.1 General Information about Fossa Carolina

Charlemagne’s summit canal also known as Fossa Carolina is an Early Medieval hydro-engineering project to bridge the Central European watershed (Zielhofer et al. 2017; Werther et al. 2019); Fig. 1.1). The canal was built on the order of Charlemagne in 792/793 AD to connect the drainage basins of the Rhine-Main system and the Danube system to create a navigable waterway (Leitholdt et al. 2012; Werther et al. 2018; Preiser-Kapeller and Werther 2018). In the Carolingian period, inland navigation was very important for mobility, communication, military operations and economic exchange in the expanding Frankish empire (Werther et al. 2018; McCormick 2010; McKitterick 2008; Ehlers 2014; Squatriti 2002). After a long period without any large canal construction schemes from 3rd to 8th century AD, the Fossa Carolina is the first post-Roman large-scale canal North of the Alps (Werther et al. 2018; Bond 2007; Elmshäuser 1992; Squatriti 2002). It is the only pre-modern attempt to bridge the Central European watershed, which was at least partially finished – a Roman attempt in the 1st century AD did not go beyond the planning phase (Werther et al. 2018).

1 Introduction

Bridging this watershed was crucial, because it was a serious obstacle for mobility between different drainage systems (Werther et al. 2018; Preiser-Kapeller and Werther 2018; Westerdahl 2006). To transport cargo, passengers and sometimes also ships from one river catchment to another, terrestrial routes had to be used if there was no canal. These portages have been bottlenecks, because they caused an enormous additional effort due to transhipment (Preiser-Kapeller and Werther 2018; Bond 2007; Westerdahl 2006).

The canal construction is extraordinary. The extent and the hydrological engineering as well as the potential impact on inland navigation is more than odd and canal construction also received artistic interest and reception (Werther 2019); Fig. 1.2). The canal has a proven length of c. 3 km and is therefore one of the most significant artificial terrain modifications of the Early Middle Ages (Squatriti 2002). Further, this important hydro-technical construction in the Early Middle Ages is one of the spatially largest features of cultural heritage in Bavaria.

Figure 1.2: The earth work of the Fossa Carolina canal construction (Bl. 20r) from the Würzburger Bischofschronik des Lorenz Fries (M. ch. f. 760) of the Würzburg University library (modified).

The historical research on the Fossa Carolina is diverse. Hack (2014) gave a good overview of the written sources, either contemporary written or later. Also, Werther (2019) gives a comprehensive summary of the written sources. However, the verification and critical discussion of these sources in the context of their origin is crucial. The sources are in disagreement about the duration of the canal construction. Some define the beginning of the construction works in the year 792 AD other in 793 AD (Werther et al. 2020; Hack 2014). Further, there is a controversial in the written sources about the completion of the canal. Some sources report a finished canal, others disagree. The same pattern is apparent for the causes of the failure of the canal (if a failure is reported). The causes are manifold. The most concise hypothesis is the colluvial collapse. Due to strong rainfalls in autumn 793 AD, the excavated material was washed back into the construction pit, unless a re-excavation was hopeless (Hack 2014). A detailed analysis of the written sources, its criticism and the potential of integrating historical data with archaeological data can be found in Werther et al. (2020). 2

1 Introduction

The (geo)archaeological research on the canal reaches back to the beginning of the 20th century with a publication of Beck (1911). The time before was characterised by chronicles, oral traditions and descriptions (Berg-Hobohm 2014a). During the 19th and 20th century, several archaeological (also geoarchaeological) surveys were conducted (Koch 1993, 1996; Beck 1911; Berg-Hobohm 2014a). Unfortunately, well documented, trustworthy, information about the canal and its hydrotechnical concept is sparse. Solely, Berg-Hobohm and Kopecky-Hermanns (2012) could disprove the hypothesis of Hilgart (1999) who proposed an artificial Carolingian water reservoir, that could had been used for water regulation of the canal hydrology. Further information about the scientific history of Fossa Carolina research can be found in Berg-Hobohm (2014a) and Werther (2019). During the last years, the extended project group of produced several new data and results about the canal itself, but also about the landscape situation and landscape evolution during the last 1200 years. Analysis of aerial images (Fig. 1.3) and the analyses of Digital Terrain Models indicate a probable canal course (Leitholdt et al. 2012; Zielhofer et al. 2014).

Geophysical prospection techniques like SQUID magnetic (Superconducting Quantum Interference Device (Zielhofer et al. 2014; Schneider et al. 2013; Linzen et al. 2017; Linzen and Schneider 2014) and Fluxgate magnetic (Zielhofer et al. 2014; Stele et al. 2019; Stele 2017) surveys show the precise canal course, based on the sedimentary contrast of the canal fillings. Archaeological excavations and geoarchaeological drilling campaigns checked and proved this prospected canal course by stratigraphic findings (Zielhofer et al. 2014; Werther et al. 2015; Werther and Feiner 2014; Leitholdt et al. 2012; Leitholdt et al. 2014; Leitholdt 2014). Further, Kirchner et al. (2018) disproved canal remains in the Altmühl floodplain by more than 100 drillings

1 Introduction and combination of historical map and LiDAR DTM analysis as well as by geophysical prospection techniques (electrical resistivity tomography, ground penetrating radar, seismic refraction tomography and SQUID-magnetic). Direct push sensing surveys revealed precise archaeometric information about several canal cross-sections and the stratigraphic and sedimentary context (Völlmer et al. 2018; Rabiger-Völlmer et al. 2020; Hausmann et al. 2018). Köhn et al. (2019) used seismic refraction inversion techniques to improve the prospection accuracy of subsurface archaeological features.

Beside archaeological excavations, archaeological aerial surveys were conducted (Kirchner et al. 2018; Zielhofer et al. 2017) and also archaeological field surveys revealed information about the settlement development and cultural landscape (Werther 2014). As timber was used for the stabilization of the embankments of the canal, several specimens could be recovered during the excavations (Werther and Feiner 2014; Werther et al. 2015). Further information about the archaeological excavations can be found in Werther (2019). For the first time, Herzig and Werther (2014) concentrate on the recovered timbers in terms of high resolution scanning and revelation of processing traces. Subsequent, Werther (2016) focused on wood as critical resource of the construction of the canal. Further, timber can be used for dendrochronological analysis. Herzig (2018) dated the timbers and Werther et al. (2020) compared these dates with written sources in terms of the beginning of the construction of the Fossa Carolina. Moreover, the dendrochronological dates of more than 40 timbers and their spatial distribution were used to decipher the Early Medieval construction progress. Hence, Schmidt et al. (2020) could reveal a distinct construction progress from North to South. Dendroarchaeological analysis can also be used for climate reconstructions, due to climate dependent tree ring growth. Muigg et al. (2020) used the dendro-data from the recovered oak timbers to set up a hydroclimate reconstruction model of spring to summer conditions. It revealed relatively dry months during canal construction. The studies, presented in this thesis will be introduced more in detail posterior, but they have to be noticed here, because they are part of the project. They have the focus on modelling and quantification approaches of the Carolingian canal. By terrain modelling, the topographical situation of the canal could be explained and discussed (Schmidt et al. 2018). By multi- method integration of subsurface (geophysics, sediment data) information of the canal and LiDAR DTM and DTM derivate data a 3D-model of the canal in its maximum excavation state could be established (Schmidt et al. 2019). This revealed fort example a spatial differentiated picture of the excavated earth volumes. The same modelling procedure was used to determine the amount of backfills (sediments, which were eroded from the surrounding, excavated earthen dams of the canal) and their spatial distribution within the canal trench (Schmidt et al. 2020). The conceptual approach of sediment budgets in the specific case of the Fossa Carolina was used to decipher the amount of erosion of dams in comparison to the backfills.

1 Introduction

1.2 Study area

1.2.1 The landscape

The Central European watershed divides continental Europe in basins draining to the North, respectively to the Baltic Sea, North Sea and Atlantic Ocean and basins draining to the South, respectively to Mediterranean Sea and Black Sea. As mentioned above, the navigable waterway between both drainage directions is of highest geostrategic relevance, especially in the Middle Ages (Preiser-Kapeller and Werther 2018). At the Fossa Carolina, the watershed divides the Altmühl catchment (Danube drainage system) from the Swabian Rezat catchment (Rhine-Main drainage system). In the study area, the Altmühl floodplain with wide meander loops and a modern straightened watercourse are typical elements of the landscape (Kirchner et al. 2018). In contrast, the Swabian Rezat River has been straightened and moved from the natural riverbed in modern times, and it is difficult to precisely detect the natural riverbed via DTM or aerial images. The Altmühl river flow at a level of c. 408 m a.s.l. and the Swabian Rezat river flows at a level of c. 414 m a.s.l. Today, the shortest linear distance of both rivers is c. 2000 m.

The narrow landscape of the Fossa Carolina has slightly undulating relief. The valley and the valley watershed are built up by sandy to loamy, fluvial sediments of Pleistocene age with a slight Loess cover, especially at the lower slopes (Schmidt-Kaler 1993; Zielhofer and Kirchner 2014). The sediments are almost free of organic material and contains mainly fine sands (Leitholdt et al. 2014; Zielhofer et al. 2014). Below the quaternary sediments, Lias and Dogger (Jurassic; mud- and sandstones) formations are present. The surrounding steep slopes of the escarpment landscape are built up by limestones and marl of the Malm formation. The Northern and Northeastern Sections of the canal are located beside the Rezat fen area, a groundwater influenced depression (Fig. 1.4) (Koch and Leininger 1993). With additional interflow from the surrounding slopes, the fen area is characterised by peat sediments (up to 3 m depth) (Zielhofer and Kirchner 2014). According to the effective climate classification of Köppen/Geiger, the study area is related to the Cfb climate (Beck et al. 2018). The annual mean temperature is 8.2 °C with a mean maximum in July (17.6 °C) and a mean minimum in January (- 1.7 °C). The long-time annual precipitation is 726 mm with a maximum in June (90 mm) and a minimum in February (43 mm; Climate data 2020).

The sedimentary features, the proximity of both rivers, the small level differences, the hydrogeological situation with a karst spring and apparent interflow are appropriate as location for a canal construction with early medieval techniques (Leitholdt 2014; Koch and Leininger 1993).

1 Introduction

Figure 1.4: Fossa Carolina and the surrounding landscape with main villages on the sub-regional to local scale.

1.2.2 The canal

The Fossa Carolina has a proven length of c. 2.9 km (Zielhofer et al. 2014; Kirchner et al. 2018), that starts at the southern slope of the Central European watershed and passes the watershed in northern directions. There, the s-shape of the canal course is visible (Fig. 1.5a; Schmidt et al. 2018). The canal can be differentiated in 5 sub-divisions due to their geographical properties and geoarchaeological conditions; (I) 6

1 Introduction the Altmühl floodplain Section, (II) the Central Section, (III) the West-East Section, (IV) the Northern Section and (V) the Northeastern Section (Fig. 1.4).

(I) The Altmühl floodplain consists of late Pleistocene fluvial deposits, mainly gravels and sands. Holocene alluvial sediments are deposited on top and are stratigraphically interrupted by a mid-Holocene soil formation that indicates stable floodplain conditions with no evidence for late Holocene meander migration. Furthermore, Kirchner et al. (2018) concluded, that the Fossa Carolina was never built in this section and, therefore, the canal has never been finished. At the village Graben (named after the canal construction) a pond is still visible as result of the canal construction (Fig. 1.5b). This pond and two drillings a few metres to the south are the southernmost proven part of the canal. A photograph from 1932 indicates a tubing of a possible canal trench, that now drains in the direction of the Altmühl floodplain (Trögl 2004). Today, the main road follows the canal course (Koch 1993).

(II) The area of the watershed (Central Section) is sedimentary characterised by Pleistocene valley fills of sandy grain sizes (Leitholdt et al. 2012). The sediments are of reddish to greyish colours, due to different redox conditions (Zielhofer et al. 2014; Hausmann et al. 2018; Völlmer et al. 2018). Further, large dams edge the canal trench. Today, they are afforested and raise up to 13 m above the present trench level (Fig. 1.5c), which is raised up to 10 m above the Carolingian excavation level (Schmidt et al. 2019). The railway line (Nuremberg-Munich, finished in 1869) disturbs the Central Section as it cuts through the dams and the canal trench at its most eastern part of the Central Section.

(III) The West-East Section marks the transition from the watershed to the Swabian Rezat floodplain. The sandy parent material is similar to the Central Section (Völlmer et al. 2018; Hausmann et al. 2018; Zielhofer et al. 2014; Leitholdt et al. 2014). The dams are also still present in this section, but they are smaller and tend to stronger erosion (Fig. 1.5d; Schmidt et al. 2020). It is likely that the Central and West-East Sections of the canal were modified during the War of Spanish Succession, as the canal was part of the defensive line (Berg-Hobohm 2014b).

(IV) The Northern Section covers the canal course from the sharp bend of the transition from the West-East Section to the interface of the canal with the Treuchtlingen-Nuremberg railway track (Fig. 1.4). Along the course, sandy to loamy fluvial sediments dominate but close to the Rezat fen in the east, a half-bog soil is developed (Zielhofer et al. 2014). An archaeological excavation in this section shows a final stage of the Early Medieval construction progress with oak timber stabilising the embankments (Werther et al. 2015; Werther and Feiner 2014). The canal floor is located 2 to 4 m below the present surface. The dams are nearly levelled, and they are only slightly visible in aerial photographs and DTMs. The intense agricultural land use and accompanying drainage interventions caused the levelling at least since the beginning 19th century (Koch and Leininger 1993).

1 Introduction

Figure 1.5: Impressions of the Fossa Carolina. a) Aerial image of the canal with shadow marks highlighting terrain differences, white arrows show the canal course (modified after Bavarian State Department of Cultural Heritage BLfD 1985), b) Pond at the southern edge of the canal in the village Graben (Photo: Lukas Werther), c) Central Section with afforested dams (Photo: Lukas Werther), d) West- East Section with drainage ditch, lateral dams (Photo: Eva Leitholdt), e) Northeastern Section with a flat present relief and agricultural land use as well as ongoing direct push sensing survey in the background (Photo: Johannes Schmidt), f) Canal profile in the Northeastern Section of the Fossa Carolina (2016–S1) during the excavation. The trench which has been excavated in 792/793 AD is clearly visible, the trench bottom is not yet excavated (Photo: Lukas Werther).

(V) The Northeastern Section is similar to the Northern Section, with clastic valley fills but the influence the organic sediments in the northernmost part increases (Zielhofer et al. 2014). Two archaeological excavation trenches cutting the canal show also a final to almost final stage of construction (Werther et al. 2020; Werther 2019). The canal floor is located 1 to 2.5 m below the surface (Fig. 1.5f). Surrounding dams

1 Introduction can hardly been seen in aerial photographs or LiDAR DTMs (Fig. 1.5e). Agricultural land use issues and drainage construction led to a levelling of the dams at the latest than the beginning 19th century (Koch and Leininger 1993).

Nowadays, the canal trenches are at least partially refilled, the dams are partially eroded or levelled and the canal is only slightly visible in the Northern and Northeastern Sections (Berg-Hobohm and Werther 2014). However, the canal is clearly visible in the Central and West-East Sections, but the canal trench bottom is raised by sediment accumulation and formation of organic-rich peat layers up to 11 m above the Carolingian trench bottom level. By now, no artificial inflow was detected and proved and the hydroengineering approach remains unclear (Rabiger-Völlmer et al. 2020). Indeed, drilling surveys and geophysical prospections revealed a summit canal construction as hydrotechnical concept (Zielhofer et al. 2014; Werther et al. 2018), but the hydrological control is an unresolved issue.

The sediment deposits, that accumulated since the Carolingian excavation were studied by (Leitholdt 2014; Leitholdt et al. 2014; Leitholdt et al. 2012; Zielhofer et al. 2014; Völlmer et al. 2018; Werther et al. 2015; Werther and Feiner 2014). According to them the general stratigraphy of the canal fillings can be described as follows. The pre-Carolingian parent material consists of sandy to loamy fluvial sediments with almost no organic remains. The timber (oaks recovered during archaeological excavations, which were used to stabilise the embankments of the canal trench) represent the construction time. The initial trench fills feature abruptly redeposited sediments with less organic material. Subsequently, thick organic sediments cover the sandy fills. These organic sediments consist of peat and sapropel layers, representing open water bodies and former ponds. The youngest fills feature clastic sediments from mainly modern times. They indicate the ongoing erosion of the dams and an intensive land use with intentional levelling of the northern canal sections.

1.3 State of the Art “Quantification approaches of geoarchaeological processes”

Quantification and modelling approaches of geoarchaeological processes are defined as terrain- or sediment-regarded and in context of intended earthworks.

(i) Palaeo-terrain research and subsequent quantification (detailed information can be found in chapter 2 – Shaping pre-modern digital terrain models. The former topography at Charlemagne’s canal construction site)

Palaeo-terrain research can be divided in two major sub-disciplines; inductive studies, that deal with ground-truth data and a spatial linkage (e.g. interpolation) of the information. On the other hand, deductive studies reconstruct a former topography top-down, removing younger features from a Digital Terrain Model until the temporal state of interest is reached. Studies, that have access to DTM data and spatially well distributed ground truth data combine both subdisciplines.

1 Introduction

In the geoarchaeological context, often studies deal with ground-truth data. They use information from archaeological excavations and prospections, drilling surveys or drilling data bases or outcrops (Baubinienė et al. 2015; Grimm and Heinrich 2019; Vermeer et al. 2014; Pröschel and Lehmkuhl 2019). The cited studies use large datasets and deal with medium to large spatial scales. In contrast, Kirchner et al. (2018) used more than 100 drillings in small scale river floodplain to reconstruct the Carolingian floodplain surface. Another opportunity to collect ground-truth data (in the sense of remote sensing derived DTMs vs. on site data generation) are geophysical prospection techniques. Saey et al. (2008) used electromagnetic induction to reconstruct a former topography. Others use the combination of geophysical prospection techniques with information of archaeological excavations (Verhegge et al. 2017; Schneider et al. 2017). Smedt et al. (2013b) used electromagnetic induction surveys coupled with excavation information and could reveal the medieval wetland reclamation.

Deductive studies have the advantage of working on study areas, without or with less ground-truth information. Indeed, ground truth data is important to check the accuracy of the modelling procedure But in general, deductive approaches can be used as prospection tool in geoarchaeological issues. Hesse (2010) and Höfler et al. (2015) use non supervised filtering with a GIS environment to remove artificial and man- made structures. van Loon et al. (2009) could reconstruct a managed fan area in the Netherlands by means of peat subsidence ratios. In contrast, Werbrouck et al. (2011) supervised their feature elimination by a specific selection of features and specific buffer widths. Recently, van der Meulen et al. (2020) used a top- down approach for the Lower Rhine valley (incl. the Upper Delta area) and made a step forward in using such an approach on the large scale.

Combined studies tackle the challenge of “extrapolate” ground-truth data with a deductive DTM-based concept. Verhegge et al. (2016) use a combined approach to reconstruct the Late Mesolithic to Neolithic landscape by means of several geophysical prospection techniques (electromagnetic induction, electrical resistivity tomography, seismic survey) and drillings with a broader DTM context. Kirchner et al. (2020) quantified hollow ways in the Hildesheimer Wald Mountains with a small to medium scale GIS approach. They use a semi-automatic feature detection which bases on LiDAR-DTM data with coupled geomorphological field mapping results. In a marine/shore-related geoarchaeological context in remote Oceania, Carson (2014) used time series of sea level changes as input data. Vermeer et al. (2014) combine large geological databases with a deductive-derived purged DTM.

(ii) Sediment volume quantification (Detailed information can be found in chapter 3 – 3D-modelling of Charlemagne’s Summit Canal (Southern Germany) – Merging Remote Sensing and Geoarchaeological Subsurface Data)

Human impact in the landscape (in settlement or development context), often results in full form (positive or raised) or hollow form (negative or burrowed) features. Extensive consequences of human land use, like erosion, colluvial/alluvial deposition, reworking and transport through the whole sediment cascade are not part of this thesis. Hence, the thesis focusses on the settlement-related (in broad sense) features. 10

1 Introduction

Full form geoarchaeological features can be detected by remote sensing techniques (Luo et al. 2019; Agapiou and Lysandrou 2015; Campana 2017). In the last decades, LiDAR (Light Detection and Ranging) approaches with derived Digital Terrain Models (DTM) are subject of increasing studies published (Zielhofer et al. 2018; Grammer et al. 2017; Cowley et al. 2018). However, also terrestrial laser scanners can be used to detect and quantify earth mounds (Larsen et al. 2017). Once they are detected, the volume calculation can start. Pickett et al. (2016) and Lacquement (2010) used geometric simplifications of the feature to calculate feature volume. Hence, the amount of simplification and reduction of feature complexity are key settings. Magnani and Schroder (2015) compare geometric models and real 3D-models of earthen mounds in terms of accuracy. There, 3D-models have clear advantages. Mostly, geometric information are a result of archaeological excavations (Sherwood and Kidder 2011; Pickett et al. 2016). However, the quantification of full form features needs a reference level, from which the feature was risen (built up). This information can be derived from on-site data, like archaeological excavations or outcrops. But for larger features or a lack of on-site data, a palaeo-topography from a deductive modelling procedure is needed.

For this thesis, negative forms are only important if they are at least partially refilled. Non-filled features can be detected via remote sensing and related to palaeo-surfaces. The volume can be estimated or calculated, depending on the resolution of the data. Subsequent, refilled sediment quantification are often done by complete archaeological excavations (Andersen 1997). But this is only possible for small-scale features. Recently, the use of geophysical methods is striking in reconstructing refilled features (Smedt et al. 2011; Smedt et al. 2013b; Diamanti et al. 2005; Köhn et al. 2019; He et al. 2015). Drilling techniques can also derive stratigraphic data of filled features (Seeliger et al. 2018; Hadler et al. 2018; Zielhofer et al. 2014). A combination of geophysical prospection and drilling technique is the direct push sensing application, which delivers physically sensed information along a drilling stratigraphy (Hausmann et al. 2018; Völlmer et al. 2018; Rabiger-Völlmer et al. 2020; Missiaen et al. 2015; Fischer et al. 2016). Further, the combination of specific archaeological excavation data and geoarchaeological data is becoming popular (Canti and Huisman 2015; Beuzen-Waller et al. 2018). Ground-truth data of the filled features have to be related to the respective surrounding topography (palaeo-topography) to determine the feature volume.

(iii) Geoarchaeological sediment budgeting (Detailed information can be found in chapter 4 – Sediment budgeting of short-term backfilling processes – the colluvial collapse of a Carolingian canal construction)

The concept of sediment budgeting is widespread in geomorphology (Brown et al. 2009; Hinderer 2012; Hoffmann 2006). The main aim is to quantify the sediment storage of a specific landscape or archive and link it to erosion processes and quantities. Several studies show that the spatial and temporal scales vary widely from large catchment scale studies to small slope-based studies (Hoffmann et al. 2007; Kesel et al. 1992; Rascher et al. 2018; Bussmann et al. 2014). Also, the temporal scale of erosion-accumulation processes varies from orogenetic fluxes to short-term precipitation events (Chen et al. 2018; Hinderer and Einsele 2001). Further, the sediment budgeting concept is used in pristine and anthropogenic influenced catchments (Voiculescu et al. 2019; Förster and Wunderlich 2009). However, there is a lack of sediment

1 Introduction budgeting approaches in geoarchaeological issues. There are studies dealing with sediment quantification in geoarchaeological contexts (Pickett et al. 2016; Lacquement 2010), but studies dealing with sediment budget are rarely existing (Bork et al. 2003).

The first major task is the quantification of the accumulated sediments (accumulation position). Recent and sub-recent sediment storages can be studied by instrumental methods or monitoring approaches (Gellis et al. 2017; Griffiths and Topping 2017). Sedimentary archives can be investigated by drillings, outcrops, trenches etc (Zolitschka 1998; Erkens et al. 2006; Förster and Wunderlich 2009; Bork et al. 2003). However, the spatial interpolation of the sediment layer is crucial. The larger the study area, the more difficult is the establishment of a dense subsurface data network (Hoffmann et al. 2007). Hence, some authors use geostatistical interpolations (Bussmann et al. 2014; Rommens et al. 2005), or assume geometrical forms of sediment bodies (Suchodoletz et al. 2009) or other use geophysical techniques to estimate the spatial distribution and quantification of the sediment layers (Guillocheau et al. 2012). The second task is the calculation or measurement of the lack of sediment of the “parent material” (erosion position). In most cases, this the result of the budgeting approach, whereas only few studies achieve this information directly (Bork et al. 2003). Finally, the comparison of the sediment storages reflects the budgeting of the sediment flux.

1.4 Material and methods

The thesis has two major thematic strands. The first focuses on open question in the distinct Fossa Carolina research. The second focuses on the development and application of quantification approaches and modelling procedures in geoarchaeological contexts. The aims of the study will be presented in the next chapter. The tackle both challenging issues, I will briefly present the main data input and the main methodological approaches (Fig. 1.6). Detailed information about the data and methods and their accuracy and resolution can be found in each paper (Chapter 2 to 4).

1 Introduction

Figure 1.6: Flowchart of the input data and application in the different modelling and quantification approaches used in this thesis.

1.4.1 Data

LiDAR DTM

The LiDAR-based (Light Detection and Ranging) Digital Terrain Model reflects the bare-earth surface of the study area. It has a spatial resolution of 1 x 1 m and height accuracy of c. 0.2 m. It was provided by the Bavarian land surveying Office (2018b) and produced by airborne laser scanning. The basic data was accessed at 2012-11-06 and 2013-08-08. The LiDAR DTM was used in all three papers (Schmidt et al. 2018; Schmidt et al. 2020; Schmidt et al. 2019) presented in this thesis (chapter 2 to 4).

Aerial image (high-resolution orthophoto)

The high-resolution orthophoto of the study area was provided by the Bavarian land surveying Office (2018d). It was accessed at 2012-11-06 and has a spatial resolution of 0.2 m. The orthophoto was used in the first paper (Schmidt et al. 2018) presented in this thesis (chapter 2).

Land-use vector layers

Two data sets of the present land use were used in the first paper (Schmidt et al. 2018) presented in this thesis (chapter 2). On the one hand, the official land use vector layer from Bavarian land surveying Office (2018e) and the present cadastre boundaries (Bavarian land surveying Office 2018a). It was accessed at 2012-11-06 and reflects the land use (buildings, roads, railway lines, etc.) with present cadastre boundaries

1 Introduction at a scale of 1:1000. On the other hand, I used the vector layers of the OpenStreetMap data base (Geofabrik 2016). It does not have a homogenous scale, but is on an almost daily basis. It was used for detailed checks and was first accesses at 2016-05-31.

Historical cadastre map

To achieve spatial historical data, I used the historical cadastre sheet recorded from 1820 to 1822. It is well georeferenced and reflects the cadastre boundaries and roads before the land consolidation. It was provided by the Bavarian land surveying Office (2018c). The map was provided (accesses at 2012-11-06) as georeferenced digital image and was vectorized by hand to include the information in the modelling.

Pre-modern DTM

The pre-modern DTM is the main result of paper 1 (Schmidt et al. 2018) presented in this thesis. It shows the topography of the study area before the massive intentional topographical changes and is free of anthropogenic structures. The pre-modern DTM has a spatial resolution of 1 x 1 m, because it bases mainly on the LiDAR DTM. It was used as palaeo-landscape for calculation of sediment application and removal and as base for the Carolingian canal trench excavation assumption in paper 2 and 3 (Schmidt et al. 2020; Schmidt et al. 2019) presented in this thesis (chapter 3 and 4).

3D-model of the Fossa Carolina

The 3D-model of the Fossa Carolina is main result of paper 2 (Schmidt et al. 2019) presented in this thesis. It shows the Carolingian canal in its maximum state of construction. Due to the high-resolution input data (archaeological excavations and direct push sensing transects, 50 cm up to 12.5 cm spacing), the 3D-model has a spatial resolution of 0.5 x 0.5 m. It was used in paper 3 (Schmidt et al. 2020) as GIS raster base layer for the calculation of the initial backfills.

Magnetic surveys

For the exact localization of the Carolingian canal magnetic surveys were conducted and used in the studies Schmidt et al. (2019) and Schmidt et al. (2020) of this thesis (chapter 3 and 4). On the one hand, I used magnetic maps measured with a Superconducting Quantum Interference Device (SQUID) System (Linzen and Schneider 2014; Linzen et al. 2017; Linzen et al. 2009). On the other hand, I used Fluxgate magnetic (Bartington Grad601 and Geoplot v3.0) survey maps (Stele 2017). Both methods detect magnetic subsurface anomalies, but range in different resolutions (SQUID on a centimetre scale, Fluxgate with a pixel size of 0.25 x 0.5 m) The magnetic maps of the Fossa Carolina and its surrounding landscape were 14

1 Introduction published before (Zielhofer et al. 2014; Kirchner et al. 2018; Stele et al. 2019). Nevertheless, due to topographical conditions of the area (and the soil moisture) both methods have their advantages.

Vibra-coring and subsequent sedimentological analysis

For sedimentary and stratigraphic data of the canal trench filling and the palaeosurface (fossil topsoils), I used several dozen vibra-corings (with Atlas Copco hammer and 60mm open corer) along the canal and in its vicinity. The verification of the pre-modern DTM modelling procedure (used in Schmidt et al. (2018) – chapter 2) was done by comparison with ground-truth coring data (levels of fossil topsoils) in the Altmühl floodplain (Kirchner et al. 2018) and Rezat fen area and floodplain (Zielhofer et al. 2014). More than 30 vibra-corings along the canal trench were used for 2 papers (Schmidt et al. 2020; Schmidt et al. 2019) of the presented thesis (chapter 3 and 4). The trench bottom and the depth of the initial backfills at its base could show by macroscopic sediment-stratigraphic data, but also geochemical parameters (total organic carbon, grain size distribution), were used. Furthermore, radiocarbon dates of macro remains were used to compile a chronological framework of the initial backfilling process. The analysis results were published before (Kirchner et al. 2018; Leitholdt et al. 2014; Leitholdt et al. 2012; Zielhofer et al. 2014), but reassessed in terms of re-calibration and synthesis.

Direct push sensing

Direct push sensing is a fast, minimal-invasive and depth accurate tool for in-situ characterisation of sediment stratigraphies (Dietrich and Leven 2009; Leven et al. 2011). Steel rods with a small diameter (38 mm) and different probes are pushed in the unconsolidated sediments. The data I used in the papers 2 and 3 (Schmidt et al. 2020; Schmidt et al. 2019), in the presented thesis (chapter 3 and 4) were measured with the colour logging tool (SCOSTTM , Dakota Technologies, Fargo, USA) to describe different sediment layers and their colour-dependent properties like organic content or redox characteristics (Hausmann et al. 2016; Hausmann 2013; Rabiger-Völlmer et al. 2020). Moreover, results of electrical conductivity logging (SC-500, Keijr Engineering Inc. – Geoprobe Systems, USA) were used to achieve information about the grain size distribution (Völlmer et al. 2018; Butler et al. 1999; Schulmeister et al. 2003; Fischer et al. 2016). In paper 2 (Schmidt et al. 2019) and 3 (Schmidt et al. 2020), two direct push sensing transects were used to create a 2D geometry of the Fossa Carolina. Transect “WE-Section” was already published before (Völlmer et al. 2018; Hausmann et al. 2018), whereas “TheAnomaly” was novel and unpublished. The pushes could be conducted with a spacing of up to 12.5 cm in a transect and are therefore appropriate for archaeo- geometrical analysis of the archaeological feature.

1 Introduction

Archaeological excavations and dendroarchaeology

For paper 2 (Schmidt et al. 2019) and 3 (Schmidt et al. 2020) of the presented thesis, distinct archaeological information about the Fossa Carolina were used. In two excavation phases (2013 and 2016) three archaeological excavation trenches were opened, cutting the canal rectangular to the embankments (Werther and Feiner 2014; Werther 2017, 2019; Werther et al. 2015). These trenches are located in the Northern and Northeastern Sections of the canal. The edges of the Carolingian trench bottom at both banks could be identified precisely with cm-accuracy, because the well-preserved timber revetments show clear signs of decay in the upper part, which was exposed to the water (Werther et al. 2015). The trench bottom between both banks and the excavation level of the accompanying slopes have been identified based on initial infills such as sapropel and re-located sandy material with higher organic contents compared to the Pleistocene parent material. Further, a large group of timbers and wood waste was recovered (Werther 2019; Werther et al. 2015; Werther and Feiner 2014) and in total 44 samples offered a reliable basis for a chronology (Werther et al. 2020). Technically, the timber and wood waste were carefully cleaned and prepared to preserve all processing traces. Subsequently, tree ring widths were measured with an accuracy of 1/100 mm using a stereoscopic microscope (Herzig 2018).

1.4.2 Methodological approaches

Palaeotopography modelling and GIS-based derivates

As mentioned in chapter 1.3, there are two possibilities to reconstruct a former topography. On sub-regional scale an inductive approach isn’t applicable, due to a lack of a dense subsurface data set. Hence, a deductive approach was used. After separation of the different land use thematic shape layers, we developed specific buffer width (by manual measurement of topographic impact), which cover the spatial impact of each type. For example, railway lines have a broad spatial impact along their course because they have wide ballast beds, which disrupt the present topography. The orthophoto was used to map additional anthropogenic structures, that are not embedded in the land use layers (such as archaeological and industrial features). Further, the digitized historical cadastre boundaries from the first edition of cadastre sheets (1820 to 1822) at a scale of 1:5,000 was equipped with a specific buffer width. Subsequently, we merged all buffer layers into one comprehensive layer. This comprehensive layer is used as a template to remove all detected cells with an anthropogenic impact from the LiDAR-based DTM. This procedure creates a perforated data layer. We interpolated the resulting residual points via a multilevel B-spline. This polynomial function allows for the creation of a continuous and consistent topography and is suitable for unregularly spaced points (Lee et al. 1997). The purged DTM with a spatial resolution of 1x1 m no longer contains any larger anthropogenic surface structures and represents the pre-modern topography. The verification of the modelling procedure was done by statistical comparison of modelled and ground-truth data (fossil topsoils) and given in a RMSE (root mean square error). Afterwards, the pre-modern DTM allows to calculate the favoured canal course 16

1 Introduction between the Altmühl and Swabian Rezat rivers. The so-called Least Cost Path analysis is widespread in supra-regional (geo)archaeological studies (Siart et al. 2008; Supernant 2017; Verbrugghe et al. 2017). The quantitative comparison of the present LiDAR DTM and the pre-modern DTM can be done by a simple subtraction of both layers. This is called local relief model (Hesse 2010) and the resulting raster layer shows changes in elevation, respectively erosion, sediment removal, sediment application or accumulation (Kokalj and Hesse 2017). Because of its hydrographic correct topography, the topographic wetness index could be calculated. It is based on the slopes and upstream catchment area and indicates potentially wet or dry areas (Conrad et al. 2015), which is an important information to interpret the present and modelled canal course of the Fossa Carolina.

Data integration and 3D-Modelling

The main methodological challenge of paper 2 (Schmidt et al. 2019) was the combination geoarchaeological subsurface data and the integration of in the pre-modern DTM, in order to create a 3D- model of the Fossa Carolina. Therefore, high-resolution cross-section were used as precise geometry markers of the canal trench. Due to the quality and resolution, archaeological excavations and direct push sensing transects serve as reference cross-sections. The derived canal trench geometry of each cross-section was transferred to maximum canal trench depth information from vibra-corings along the canal course and equidistant spaced additional transects (synthetic). Neither the archaeological excavations, nor the direct push sensing transects are distributed equally along the canal. This is mainly due to high effort of archaeological excavations (Werther and Feiner 2014) and the impassibility for the direct push caterpillar in some sections (Hausmann et al. 2018). Also, the vibra-coring position are not equally distributed along the canal course. On this account, we developed a modelling approach, which suits this challenge reflecting four steps to create a trustworthy 3D-model. The first step summarizes the selection of reference cross-sections due to their specific 2D imaging quality. In a second step the derived canal trench geometries were transferred to the vibra-coring positions using their canal trench depth information. The specific reference geometry was adjusted to the level of the trench bottom. In a third step, we created additional transects, that have an equidistant spacing of c. 50m along the canal course. The trench bottom level at each additional transect was linear interpolated from neighbouring ground-truth information (vibra-corings, direct push sensing or archaeological excavations). Subsequent, specific reference cross-sections were transferred and adjusted to the additional transects and their interpolated canal trench level. In the last step, a spatial interpolation of all transects by triangulation created the 3D-model of the Fossa Carolina. Due to the high-resolution of the reference cross-sections, the resulting model has a spatial resolution of 0.5 x 0.5 m. The same procedure was performed in the third paper (Schmidt et al. 2020) presented in this thesis. However, the top level of the initial sedimentary backfills was used as depth information. Thus, the quantitative comparison of the first 3D-model (in maximum excavation state) and the second 3D-model (after the backfills accumulated), could be used for sediment quantification.

1 Introduction

Sediment storage quantification and sediment budgeting

Sediment storages of the residual remnants of the Carolingian dam, the maximum excavation volume of the total canal trench and the sediment storage of the initial backfills were quantified using the above mentioned modelling results in paper 2 (Schmidt et al. 2019) and paper 3 (Schmidt et al. 2020). Therefore, the main results of the modelling procedures were compared to each other. The dam volume was quantified by subtracting the pre-modern DTM from the present LiDAR DTM. The maximum excavation volume was quantified by subtracting the 3D-model raster layer from the pre-modern DTM. The sedimentary backfill volume was quantified by subtracting the “second” 3D-model (after the backfills accumulated) from the first 3D-model (maximum excavation state). All sediment quantities were broken down to the sections of the canal, so that canal section specific sediment storage and budgeting questions can be answered. The sediment budget is a complied analysis, of the information on maximum excavation volume, the residual dam volume and the volume of the backfill sediments. Separated by the canal sections, we can derive spatial differentiated ratios of dam erosion and backfill accumulation.

1.5 Aims of the thesis

The present thesis combines GIS-based modelling approaches using spatial data and geoarchaeological subsurface information. For the first time, quantities of geoarchaeological processes of the Fossa Carolina will be tackled. According to the state of the Art of research on the Fossa Carolina, a detailed and the previous methodological approaches in quantification and modelling of geoarchaeological processes, I identified the following issues/aims:

1. The development of a specific palaeo-surface modelling approach is necessary to answer the issues of former spatial dependencies. Only a purged Digital Terrain Model without anthropogenic structure disturbing the relief is suitable for such analysis. The development of an approach that is not only applicable to my study, but rather broadly applicable will be of merit. 2. Where would the “best” (shortest connection between the anchor points at the Altmühl and Swabian Rezat rivers, respectively the least earth volume to excavate) canal course be? Does the Carolingian hydro engineers have took this path? If not, are there severe reason to avoid this “best canal course”? Which reasons could be mentioned in discussing the canal course deviation? 3. A detailed 3D-model of the maximum excavation state of the canal is crucial to quantify the canal excavation processes. There is no published approach dealing with such a large archaeological feature and its 3D reconstruction. To challenge the extension of the canal, I create a reproducible and applicable modelling approach by integrating different geoarchaeological subsurface data within a purged DTM. The approach will be unique, but transparent and useful for other geoarchaeological studies, dealing with volumes of refilled archaeological features. 18

1 Introduction

4. In the Fossa Carolina research, the construction’s excavated earth volume remains unclear. Several authors estimated values, but with a doubtful or no data base. The exact quantities of moved earth material and its spatial distribution along the canal trench can help to understand the Carolingian efforts building this waterway. Further, the model can be used to define the summit level of the canal. 5. As authors suggested before, the canal is collapsed (damaging erosion of the dams and accumulation of backfill material within the construction pits) directly after constructions site abandonment or maybe as cause for the relinquishment. How large is the volume of the backfills? How are these backfills distributed along the canal? Is it possible to define a “collapse-zone”? 6. The creation of sediment budgeting approach can help to understand the ratios of excavated material, earth volume remains in the dams and the volume of backfills deposited at the base of the canal trench. The assessment of the sediment budget can show the post-Carolingian terrain modifications and earth volume changes.

2 Shaping pre-modern Digital Terrain Models

Chapter 2 - Shaping pre-modern digital terrain models. The former topography at Charlemagne’s canal construction site

Schmidt, J.; Werther, L.; Zielhofer, C. (2018): Shaping pre-modern digital terrain models. The former topography at Charlemagne's canal construction site. In: PloS one 13 (7), e0200167. DOI: 10.1371/journal.pone.0200167.

2 Shaping pre-modern Digital Terrain Models

Shaping pre-modern digital terrain models: the former topography at Charlemagne’s canal construction site

Johannes Schmidt1*, Lukas Werther², Christoph Zielhofer1

1Institute of Geography, Leipzig University, D-04103 Leipzig, Germany

²Seminar of the Archaeology of Prehistory to the Early Middle Ages, Friedrich-Schiller University, D- 07743 Jena, Germany

*corresponding author: [email protected]

2.1 Abstract

The use of remote sensing techniques to identify (geo)archaeological features is wide spread. For archaeological prospection and geomorphological mapping, Digital Terrain Models (DTMs) on based LiDAR (Light Detection And Ranging) are mainly used to detect surface and subsurface features. LiDAR is a remote sensing tool that scans the surface with high spatial resolution and allows for the removal of vegetation cover with special data filters. Archaeological publications with LiDAR data in issues have been rising exponentially since the mid-2000s. The methodology of DTM analyses within geoarchaeological contexts is usually based on “bare-earth” LiDAR data, although the terrain is often significantly affected by human activities. However, “bare-earth” LiDAR data analyses are very restricted in the case of historic hydro-engineering such as irrigation systems, mills, or canals because modern roads, railway tracks, buildings, and earth lynchets influence surface water flows and may dissect the terrain. Consequently, a "natural" pre-modern DTM with high depth accuracy is required for palaeohydrological analyses. In this study, we present a GIS-based modelling approach to generate a pre-modern and topographically purged DTM. The case study focuses on the landscape around the Early Medieval Fossa Carolina, a canal constructed by Charlemagne and one of the major medieval engineering projects in Europe. Our aim is to reconstruct the pre-modern relief around the Fossa Carolina for a better understanding and interpretation of the alignment of the Carolingian canal. Our input data are LiDAR-derived DTMs and a comprehensive vector layer of anthropogenic structures that affect the modern relief. We interpolated the residual points with a spline algorithm and smoothed the result with a low pass filter. The purged DTM reflects the pre- modern shape of the landscape. To validate and ground-truth the model, we used the levels of recovered pre-modern soils and surfaces that have been buried by floodplain deposits, colluvial layers, or dam material of the Carolingian canal. We compared pre-modern soil and surface levels with the modelled pre-modern terrain levels and calculated the overall error. The modelled pre-modern surface fits with the levels of the buried soils and surfaces. Furthermore, the pre-modern DTM allows us to model the most favourable course of the canal with minimal earth volume to dig out. This modelled pathway corresponds significantly with

2 Shaping pre-modern Digital Terrain Models the alignment of the Carolingian canal. Our method offers various new opportunities for geoarchaeological terrain analysis, for which an undisturbed high-precision pre-modern surface is crucial.

2.2 Introduction

It is obvious that the present relief does not match the pre-modern topography. The present relief is often detected on a large scale via LiDAR (Light Detection and Ranging), whereas the pre-modern relief needs to be reconstructed. LiDAR is used to scan the surface from the air by measuring the optical distances and velocities of laser beams. Airborne Laser-Scanning (ALS) LiDAR provides direct measurements of vegetation cover (first pulse) and “bare earth” (last pulse), resulting in a 3D point cloud. Due to the penetration of the light signal through vegetation cover, it is possible to detect the topography and archaeological features under light forest canopy (Schindling and Gibbes 2014; Doneus et al. 2008). By filtering the data, vegetation cover can be removed so that the ground surface is displayed in the terrain model (Schindling and Gibbes 2014; Chase et al. 2012).

According to Web of Science (Web of Science 2018), the number of publications referring to LiDAR data in archaeological and environmental science has been growing exponentially since the mid-2000s, and the availability and successful application of these datasets have increased. From a geoarchaeological point of view, LiDAR offers a fast, non-destructive tool for remote sensing and large-scale prospection that provides valuable information about the location and extent of anthropogenic surface structures (Lasaponara and Masini 2011; Freeland et al. 2016; Bewley 2003; Doneus et al. 2008; Johnson and Ouimet 2014). Nevertheless, the method can only document the present relief, which is usually profoundly modified by human activity and does not match with the pre-modern relief. Furthermore, the relief has been modified significantly by extensive linear and non-linear structures such as settlements and infrastructure (roads, railway lines, ditches, etc.), especially in the modern era. All of these structures alter the natural landscape surface, especially with regard to palaeohydrological interpretability.

In recent years, there have been several approaches to reconstruct pre-modern terrain based on the interpolation of large datasets from drillings, archaeological excavations, and outcrops (Baubinienė et al. 2015; Vermeer et al. 2014). Other studies combine archaeological excavation data and geophysical data to interpolate detected pre-modern surface heights (Schneider et al. 2017; Verhegge et al. 2017; Verhegge et al. 2016). Both approaches are inductive methods based on field data interpolation and elaborate (geo- )archaeological and geophysical fieldwork, and post-processing. There are also deductive approaches to model and reconstruct the palaeo-terrain via the digital deconstruction of present LiDAR-derived Digital Terrain Models (DTMs) (Hesse 2010; Höfler et al. 2015; Werbrouck et al. 2011; Zwertvaegher et al. 2010).

Our case study examines Charlemagne’s summit canal, or Fossa Carolina, an Early Medieval hydro- engineering project to bridge the Central European watershed (see Fig. 2.1 and Fig. S1). The canal was built on the order of Charlemagne in 793 AD to connect the drainage basins of the Rhine-Main system and 22

2 Shaping pre-modern Digital Terrain Models the Danube system to create a navigable waterway (Leitholdt et al. 2012; Werther et al. 2018; Preiser- Kapeller and Werther 2018).

In the Carolingian period, inland navigation was very important for mobility, communication, military operations and economic exchange in the expanding Frankish empire (Werther et al. 2018; McCormick 2010; McKitterick 2008; Ehlers 2014; Squatriti 2002). After a long period without any large canal construction schemes from 3rd to 8th century AD, the Fossa Carolina is the first post-Roman large-scale canal North of the Alps (Werther et al. 2018; Bond 2007; Elmshäuser 1992; Squatriti 2002).

It is the only pre-modern attempt to bridge the Central European watershed, which was at least partially finished – a Roman attempt in the 1st century AD did not go beyond the planning phase (Werther et al. 2018). Bridging this watershed was crucial, because it was a serious obstacle for mobility between different drainage systems (Werther et al. 2018; Preiser-Kapeller and Werther 2018; Westerdahl 2006). To transport cargo, passengers and sometimes also ships from one river catchment to another, terrestrial routes had to be used if there was no canal. These portages have been bottlenecks, because they caused an enormous additional effort due to transhipment (Preiser-Kapeller and Werther 2018; Bond 2007; Westerdahl 2006).

Furthermore, the Fossa Carolina is also one of the most significant artificial terrain modifications of the Early Middle Ages (Squatriti 2002). It was planned as a summit canal due to the different levels of both tributaries (Zielhofer et al. 2014; Werther et al. 2018). Nevertheless, the hydrological concept and the reasons for the implemented pathway remain unclear due to the lack of a high-precision model of the topography from the time of construction.

We present a GIS-based modelling approach to improve and purge a present DTM on a sub-landscape scale in the surroundings of the Fossa Carolina. The DTM is based on LiDAR data with a spatial resolution of 1x1 m. We revised the model by eliminating all kind of detectable disturbing factors such as roads, railway lines, buildings, Carolingian features, present and historic cadastre boundaries, and other human features such as sewage plants, rain retention basins, etc. Furthermore, we use levels from geoarchaeological drillings and archaeological excavations to ground-truth the model and to validate the reconstructed pre- modern surface.

2 Shaping pre-modern Digital Terrain Models

2.3 Study area

The 12.5 km² study area is located in the range of the Southern Franconian Jura foothills in Bavaria, Southwest Germany (Fig. 2.1). The bedrock of the escarpment consists of bedded Upper Jurassic limestone, whereas the parent material of the foothills consists of Middle Jurassic claystone and Upper Pleistocene sandy valley fills (Zielhofer and Kirchner 2014). The study area is part of the Central European watershed and locally features two sub-drainage systems (Fig. 2.1b).

The Altmühl River is a tributary of the Danube and drains towards the Black Sea. In the study area, the Altmühl floodplain with wide meander loops and a modern straightened watercourse are typical elements of the landscape (Kirchner et al. 2018). In contrast, the Swabian Rezat River (Fig. 2.1b) is part of the Rhine- Main catchment and drains towards the North Sea. The Rezat fen is located along the upper course of the Swabian Rezat River and consists of thick organic sediments. In modern times, the Swabian Rezat has been straightened and moved from the natural riverbed, and it is difficult to precisely detect the natural riverbed via DTM or aerial images. The European watershed divides both sub-drainage systems and trends as a shallow valley ridge that mainly consists of sandy to loamy fluvial deposits from the Late Pleistocene age.

2 Shaping pre-modern Digital Terrain Models

The course of Charlemagne’s summit canal could be divided into five different sections, according to their geographical and geoarchaeological conditions (Zielhofer et al. 2014) (Fig. 2.1c and Fig. 2.2). The southernmost canal remains are located in the village Graben and are still visible as a pond (Fig. 2.2b). The central section is characterized by afforested ramparts (up to 13m above present pond level; Fig. 2.2c). In the West-East-section, the ramparts are lower and the former canal is silted (Fig. 2.2d). The Northern and North-Eastern section are marked by relatively flat ramparts and only hardly visible at the surface.

2.4 Material and methods

2.4.1 Basic data

The basic digital data for our approach are a LiDAR DTM (1x1 m spatial resolution, height accuracy of ±0.2 m); a contemporary high-resolution orthophoto (0.2-m spatial resolution); a contemporary high- resolution land use vector layer with buildings, roads, railway lines, and present cadastre boundaries (Bavarian State Office for Land Surveying and Open Street Map); and the first high-resolution historical cadastre map from 1820-1822 AD, which includes cadastre boundaries and roads from before modern land consolidation (provided by the Bavarian State Conservation Office, Tab. 2.1).

2 Shaping pre-modern Digital Terrain Models

Table 2.1: Basic data acquisition and quality.

Present Historical cadastre Land use layer LiDAR DTM Digital Orthophoto cadastre and sheet Open Street Map (DOP 20) land use layers Metadata (Bavarian land (Bavarian land (Geofabrik 2016) (Bavarian land (Bavarian land surveying Office surveying Office surveying Office surveying Office 2018a, 2018e) 2018c) 2018b) 2018d) Source/version 2012-11-06, provided 2012-11-206 2016-05-31, 2012-11-06, 2013-08- 2012-11-06, provided by LDBV Bayern provided by LDBV Download 08, provided by by LDBV Bayern Bayern; Original LDBV Bayern maps 1820-1822 Resolution/scale 1:1000 1:5000 not homogenous 1m 0.2m

Format shapefile georeferenced tif shapefile ascii file, transferred in georeferenced tif a DTM raster file Accuracy official - high official - medium open data -medium official - high official - high (accuracy verfied based on high- accuracy LiDAR DTM and Digital Orthophotos) Percentage of c. 75 % c. 20% c. 5 % the comprehensive layer of anthropogenic structures

2.4.2 Modelling approach

We separated the land use vector layer from the Bavarian State Office for Land Surveying into different thematic shape layers (buildings, roads, railway lines, and present cadastre boundaries). Different types of anthropogenic structures reveal specific spatial impacts on the topography. For example, railway lines have a broad spatial impact along their course because they have wide ballast beds, which disrupt the present topography.

Aerial images, the land use vector layer from Open Street Map, and LiDAR-based DTM were used to map additional structures such as archaeological features (e.g. Fossa Carolina) and industrial features (e.g. sewage plants and quarries) as polygons. Furthermore, we manually digitized the historical cadastre boundaries from the first edition of cadastre sheets at a scale of 1:5,000. Hence, we created structure- specific buffers based on empirical knowledge (manual measurement of topographic impacts of each structure type; Tab. 2.2, Fig. 2.3). Subsequently, we merged all buffer layers into one comprehensive layer (Fig. S2). This comprehensive layer is used as a template to remove all detected cells with an anthropogenic impact from the LiDAR-based DTM. This procedure creates a perforated data layer. We interpolated the resulting residual points via a multilevel B-spline. This polynomial function allows for the creation of a continuous and consistent topography and is suitable for unregularly spaced points (Lee et al. 1997).

2 Shaping pre-modern Digital Terrain Models

Table 2.2: Compilation of compound-specific buffer widths.

Railway lines Roads Buildings Present cadastre 19th century cadastre boundaries boundaries Buffer widths 20 m 15 m 10 m 7 m 2 m

Finally, we smoothed the model with a low-pass filter. This procedure adjusts small residuals from the interpolation. The purged DTM with a spatial resolution of 1x1 m no longer contains any larger anthropogenic surface structures and represents the pre-modern topography. Fig 3 shows the stepwise procedure of the modelling approach. The data layer derived from the cadastre sheets is illustrated in the results section (see Fig. 2.4).

2.4.3 Local-relief model

We compared the present and the modelled pre-modern DTM via subtraction of both layers. This tool is known as a Local-Relief Model (LRM) (Hesse 2010). The resulting raster layer in a two-colour stretch visualises if there is a positive or a negative change in elevation between both datasets. This tool can be used to quantify the erosion, removal, accumulation, and aggradation.

2.4.4 Validation of the modelled pre-modern DTM

Validation of the modelled pre-modern DTM is crucial for interpretation of the results. Our quantitative approach is based on the comparison of the modelled pre-modern DTM with measured levels of buried soils and surfaces that have been recovered by geoarchaeological drilling campaigns (Zielhofer et al. 2014; Kirchner et al. 2018) and archaeological excavations (Werther et al. 2015) (Fig. 2.1c). The deviations are computed as the Root Mean Square Error (RMSE) and show the overall error.

2 Shaping pre-modern Digital Terrain Models

2.4.5 Least cost path analysis

The Least Cost Path Analysis (LCPA) is a common method for predicting best pathways (with the least cost) between different points. LCPAs are used in archaeology for reconstructing favoured pathways through present and past landscapes (Siart et al. 2008; Supernant 2017). At the early stage of the analysis, a criterion has to be selected for the calculation. We have chosen the altitude as a single criterion because we want to predict the most favourable course of the Carolingian canal in consideration of the minimum volume of excavation material required.

We fixed the assumed connection of the Fossa Carolina with the present Altmühl and Swabian Rezat Rivers as source and destination points. In this context, we selected the closest position of the Altmühl River to the Fossa Carolina because there is no geoarchaeological proof for a link between the Altmühl River and the Carolingian canal (Kirchner et al. 2018). The LCPA results in a line shape, which angular shaped due to raster cells. Thus, we smoothed the resulting line shape for illustrative purposes via iterative averaging (Bodansky et al. 2002; Mansouryar and Hedayati 2012).

2 Shaping pre-modern Digital Terrain Models

2.4.6 Hydrogeographical analysis

It is likely that the engineering concept and the course of the Fossa Carolina depend on the topographic and hydrogeographic features of the landscape. Therefore, we used hydrogeographic indices for a reliable interpretation of the course of the canal. The Topographic Wetness Index (TWI) indicates potentially wet or dry areas. The calculation is based on the slopes and upstream catchment area. This tool was implemented in SAGA GIS, which provides further descriptions for handling (Conrad et al. 2015).

2.5 Results

2.5.1 Interim results from stage 2 of the modelling approach

The interim results represent the extracted data from stage 2 (Fig. 2.3) of the modelling approach. The sample extract (Fig. 2.4) shows all layers with their compound-specific buffers. It is obvious that using the comprehensive layer as a mask (Fig. 2.4h) eliminates all larger anthropogenic structures.

2 Shaping pre-modern Digital Terrain Models

2.5.2 Results from stage 3 of the modelling approach

The modelling procedure results in the pre-modern DTM (Fig. 2.3, Stage 3). The modelled DTM no longer includes larger distinct anthropogenic structures in the study area (Fig. 2.5a, c and Fig. S4). The zoomed area in Figs. 2.5b and Fig. 2.5c offers a before-and-after comparison. While the terrain in Fig. 2.5b is dominated by many different anthropogenic structures like buildings, roads, and archaeological features, the terrain in Fig. 2.5c is smooth.

2 Shaping pre-modern Digital Terrain Models

Figure 2.5: Pre-modern DTM. a) Pre-modern DTM illustrated with hillshade, b) enlarged section with present DTM and hillshade, c) enlarged section with the pre-modern DTM and hillshade.

2.5.3 Model comparison

The subtraction of the present DTM from the modelled pre-modern DTM results in the LRMs (Fig. S3, Fig. 2.6a and 2.6b). Both LRMs feature the same colour scale, with red representing aggradation and blue representing the removal of sediment. The LRMs do not differentiate between direct human impact (e.g. buildings, roads) and indirect human impact (e.g. soil-erosion). Figure 2.6a clearly shows many anthropogenic structures like roads and buildings, the noticeable Fossa Carolina dams, and the high-impact railway line. Additionally, many fuzzy structures represent cadastre boundaries or land use boundaries. In Figure 2.6b, building activities in the area of Graben village and the thick red railway line dominate the eastern and northern parts of the map. The Altmühl floodplain seems to be less affected by human impact, and it is obvious that there are no visible Carolingian canal residuals. Remarkably, distinct levees from the modern age are clearly visible along the Altmühl River.

2 Shaping pre-modern Digital Terrain Models

2.5.4 Validation of the modelled pre-modern DTM

Every model has its uncertainties. To determine them precisely, we compared the levels of the modelled pre-modern DTM with measured levels of buried soil surfaces that have been recovered during former drilling campaigns or archaeological excavations. Our validation approach is semi-quantitative because an equal distribution of validation points would be required for a quantitative validation.

The first drilling transect (Zielhofer et al. 2014) is located in the west-east section of the Fossa Carolina (Fig. 2.1c). The drillings gave proof of buried soils under the Carolingian excavation material (Fig. 2.7). The grey dashed line represents the interpolated level of the buried A-horizons, and the red dashed line indicates the surface of the modelled pre-modern DTM. The two lines run parallel but with a slight vertical offset between 35 and 90 cm.

2 Shaping pre-modern Digital Terrain Models

The second transect corresponds with an archaeological excavation in the northern section of the Fossa Carolina (Fig. 2.1c). The stratigraphy of the archaeological excavation (Werther et al. 2015; Werther and Feiner 2014) and parallel drillings (Zielhofer et al. 2014) reveal a buried A-horizon in the Rezat fen floodplain. Carolingian excavation material and younger flood loam deposits of around 50 cm cover the pre-modern A-horizon. Here, the surface of the modelled pre-modern DTM has almost no deviation from the buried pre-modern surface (Fig. 2.8).

2 Shaping pre-modern Digital Terrain Models

In order to obtain a quantitative estimation of the error of the modelled pre-modern DTM, we calculated the RMSE (Tab. 2.3 and Fig. 2.9). This error is given in metres (Chai and Draxler 2014) and represents the mean deviations between the modelled pre-modern DTM and manifold measured levels of pre-modern buried surfaces derived from drillings (Zielhofer et al. 2014; Kirchner et al. 2018) and archaeological excavations (Werther et al. 2015). We manually clustered the spatial data in two subsets to estimate the error of different landscape types and associated canal sections. The first cluster includes all data in the direct surroundings of Fossa Carolina, reflecting a zone with an intense anthropogenic impact due to the Carolingian canal and railway lines. The second cluster in the Altmühl floodplain reflects an area with generally lower anthropogenic impact on the terrain.

Figure 2.9: Validation plots of the modelled surface (black dots) and the present surface (red dots) against the observed surfaces (pre-modern height) derived from drillings and excavations. The RMSE (root-mean-square error) is shown in each legend box to estimate the error. a) Total set of all validation points; b) validation points (Kirchner et al. 2018) in the Altmuehl floodplain reflecting areas with less anthropogenic overprint (mainly quasi-natural accumulation of alluvial deposits); c) validation from previous studies (Werther et al. 2015) in the direct surroundings of Fossa Carolina reflecting areas with strong anthropogenic overprint (linear structures).

We calculated the RMSE between the modelled pre-modern DTM and the measured pre-modern surface, as well as between the present DTM and the measured pre-modern surface to estimate the improvement of the modelled pre-modern DTM against the present DTM (Tab. 2.3). Generally, there is an improvement of the modelled pre-modern DTM in all areas (0.18 m, Tab. 2.3). However, there are noticeable differences between the subsets. The Fossa Carolina subset shows a mean improvement of 0.51 m, and the Altmühl subset reveals a mean improvement of only 0.05 m (Tab. 2.3). Consequently, there are higher improvements in zones of stronger and more direct human impact on the former topography. 35

2 Shaping pre-modern Digital Terrain Models

Total Fossa Carolina Altmühl floodplain

Measured (buried) pre-modern surface vs. modelled 0.69 0.62 0.71 pre-modern DTM Measured (buried) pre-modern surface vs. present 0.87 1.13 0.76 LiDAR-based DTM Improvement [m] 0.18 0.51 0.05

2.5.5 The modelled pathway of Fossa Carolina

The LCPA computes the most cost-efficient canal pathway and therefore the minimum earthmoving (Fig. 2.10). In general, it is striking that the modelled canal course follows the real Carolingian canal course quite well (Fig. 2.10). Both courses are S-shaped and indicate an almost identical point for crossing the Central European Watershed. The modelled course has some minor deviations from the real canal course. The first deviation is visible in the west-east section, where the modelled pathway is located more towards the centre of the depression line in the North. The second deviation is detectable in the northern section, where the Carolingian canal course is located slightly more in the west of the modelled pathway.

2 Shaping pre-modern Digital Terrain Models

Figure 2.10: Pre-modern DTM with the Carolingian course of the canal (yellow line) and the modelled course (least cost path analysis).

2.5.6 Hydrogeographic indices

The modelled pre-modern DTM allows the calculation of different hydrogeographic indices such as the Topographic Wetness Index (TWI), as disturbing anthropogenic structures that alter the surface runoff have been removed. The modelled wet areas are located in the Altmühl floodplain and in the area of the Rezat fen (Fig. 2.11, Fig. S5). The real course of the Fossa Carolina runs at the southern and western margins of the wet areas, whereas the modelled canal pathway runs directly through the wet depression line.

2 Shaping pre-modern Digital Terrain Models

2.6 Discussion

2.6.1 Palaeo-surface modelling approaches

The modelling of the palaeo-surface is an important tool for archaeological evaluation and geoarchaeological site interpretation (Carey et al. 2017). A palaeo-surface can be reconstructed by deductive, inductive, or combined approaches. To evaluate our study in terms of effort, accuracy, applicability, and validation, we compared it with available studies with a terrain reconstruction approach that result in a palaeo-DTM (see Tab. 2.4).

2 Shaping pre-modern Digital Terrain Models

Table 2.4: Comparison of modelling approaches in palaeo-terrain research with geoarchaeological issues.

Study area Dating Approach Applicability Low Accuracy Validation Data Spatial Reference effort scale general study - deductive ++ ++ - no LiDAR DTM none Hesse et al. 2010 Karlburg, Germany - deductive ++ ++ - no LiDAR DTM c. 1 km² Höfler et al. 2015

Sandy Flanders, - deductive + + + yes LiDAR DTM, topographical 1400 km² Werbrouk et al. 2011 Belgium vector data, historic map (1863, 1909)

Gooi and 5 time slices (0 AD, deductive + + - no present DTM 200 km² van Loon et al. 2014 Vechtsdreek area, 800 AD, 1,350 AD, Netherlands 1,885 AD, 2,000 AD)

Fossa Carolina, pre-modern deductive ++ + + yes LiDAR-DTM, present land 12.5 Km² This study Germany use data, historic map, aerial image

Scheldt Polders, Final Mesolithic, early inductive - - + yes EMI data, sediment-drilling data, c. 0.25 km² Verhegge et al. 2017 Belgium Neolithic CPT data, present DTM

Vilnius, Lithuania - inductive - - - no archival material, historical 2.6 km² Baubinienne et al. sources, cartographic and visual 2015 material, geological borehole data, geophysical data

Sandy Flanders, Medieval inductive + + ++ yes EMI data, archaeological 0.2 km² De Smedt et al. 2013 Belgium excavations

Altmühl, Germany Mid-Holocene inductive - -- + no sediment drilling data c. 0.2 km² Kirchner et al. 2017

2 Shaping pre-modern Digital Terrain Models

Table 2.4: continued.

Study area Dating Approach Applicability Low Accuracy Validation Data Spatial Reference effort scale Lausitz, Germany Mesolithic inductive -- -- + no archaeological excavation, GPR c. 0.1 km² Schneider et al. 2017 data, Drone-DTM Remote Oceania 1,500-1,000 BC combined + - + (yes) present DTM, sea level data few km² Carson 2014 (time series), archaeological data

Pederneira lowland, Pre-Holocene combined o o - no present DTM, geological c. 16 km² Lopez et al. 2013 Portugal borehole data, TEM data

Scheldt Polders, Final Mesolithic, early combined o - o no archaeological drilling data, EMI c. 0.5 km² Verhegge et al. 2016 Belgium Neolithic data, Seismic shear, Electrical resistivity imaging, CPTs

Sandy Flanders, 10,000 BC combined -- -- +(+) yes Filtered DEM from Werbrouck 584 km² Vermeer et al. 2014 Belgium et al. 2011, Holocene sediment data from literature and own field studies, c. 4000 drillings from geological database

Sandy Flanders, - combined -- -- + no for temporal DEM -> present 584 km² Zwertvaegher et al. Belgium DTM 2010

2 Shaping pre-modern Digital Terrain Models

Inductive approaches (Tab. 2.4) usually have an advantage in that the specific data points used for the interpolation are dated and have a stratigraphic context (Kirchner et al. 2018; Corrò and Mozzi 2017). The spatial distribution of data points is critical. Geophysical methods can generate many spatially well- distributed data points (Schneider et al. 2017; Faïsse et al. 2018; Baubinienė et al. 2015). The resolution and precision of the data depend on the sedimentary contrasts of the palaeo-surface and the overlaying sediments (Smedt et al. 2013). The amount of required fieldwork for obtaining and post-processing the data is time-consuming. In addition, the research area must be accessible or already explored, and archaeological, geological, or geoarchaeological data must be available.

Deductive approaches (Tab. 2.4) provide an advantage of handling the study area without laborious fieldwork and allow us to work with spatially well-distributed input data (DTMs, aerial images, etc.) that are relatively easy to obtain and well-achievable for large areas (Werbrouck et al. 2011; Vermeer et al. 2014; Zwertvaegher et al. 2010). Usually, a present landscape is deconstructed via eliminating different structures with different techniques until the necessary state has been reached (Vermeer et al. 2014). However, for a deductive approach, the temporal uncertainty of the modelled terrain and the validation of the result are critical. Van Loon et al. (2009) used a deductive method and gave their model a chronological frame. This exception is due to their approach using peat subsistence ratios to model the altitude in five time slices. In the present study, we eliminated modern and specific historic features during our modelling approach. Therefore, we are able to provide at least a pre-modern stage as a chronological frame.

A combination of deductive and inductive approaches (Tab. 2.4) might improve the performance of the model. Usually, there is an approximation of the palaeo-surface with a deductive approach, which is subsequently supported by a dense dataset of archaeological or geoscientific ground-truth data (Zwertvaegher et al. 2010; Carson 2014; Verhegge et al. 2016). Zwertvaegher et al. (2010) developed an integrative process model to challenge the complexity of landscape evolution by combining a deductive approach to obtain a purged base layer and an inductive approach for the predictive modelling of archaeological finds within the landscape. The amount of data, their quality, and the effort are enormous and require a holistic understanding of the landscape. On the other hand, such integrative approaches provide a more accurate result than a deductive or inductive approach alone.

Our approach is deductive and characterized by easily obtained data (LiDAR-DTM, land use shapefiles, historical cadastre sheets, areal images), and therefore, it is well applicable at different scales. Unlike many other studies with a deductive approach (Hesse 2010; Höfler et al. 2015; van Loon et al. 2009), we are able to give an accuracy estimation of our model. We use a ground-truth validation approach based on discrete palaeo-surface data from drillings and excavations to provide values for model accuracy. The qualitative 2D validation documents the specific offset of the model at two specific points of our study area. Our semi- quantitative validation approach compares point subsets from Zielhofer et al. (2014), Werther et al. (2015), and Kirchner et al. (2018) with the modelled terrain height. In addition, we measured the error between the different models, today's DTM, and the modelled DTM and developed an improvement value. The

2 Shaping pre-modern Digital Terrain Models validation results of our study show a spatial unequal distribution of the model error. In the Altmühl floodplain, which has a low anthropogenic impact (mainly planar accumulation of overbank fines), the improvement compared to the present DTM is negligible (see Tab. 2.2). On the other hand, in the proximal area of the Fossa Carolina construction site, the modelling of the pre-modern DTM significantly improves the accuracy.

2.6.2 Model performance

The performance of the model is crucial for the reliability of the results and their interpretation. For the qualitative validation, drilling and model data were connected in a 2D plot, and model inaccuracies were estimated (see Fig. 2.7 and 2.8). The offset of the modelled altitude and the detected palaeosols is low to moderate. However, if we consider the absence of a buried A-horizon and the thickness of the buried B- horizons (IUSS Working Group WRB 2015), it is clear that these soils were truncated before the Carolingian construction phase (Bussmann et al. 2014). Therefore, the “natural” offset should be lower since the original surface was higher. Furthermore, the height accuracy of the LiDAR-DTM with ±0.2 m is taken into account, and the semi-quantitative validation results are discussed.

No other deductive modelling approach has been validated with palaeo-pedological and stratigraphic field data (see Tab. 2.4). Only Werbrouck et al. (2011) validated their model at all, but only based on a historic map from 1909 to correlate drawn and modelled ditches qualitatively. However, the detection of palaeosols is not flawless. Among others, one part of the described offsets may result from sediment loss and sediment consolidation during canal construction. Additionally, the region between Weissenburg and Treuchtlingen was used intensively since at least the Iron Age, resulting in thick pre-medieval colluvial and fluvial deposits around the Fossa Carolina (Berg-Hobohm and Kopecky-Hermanns 2012).

Our approach only allows the removal of linear or punctual features of the model, so the main problem is erosion and accumulation with a wider spatial impact. This large-scale impact on the terrain is not clearly detectable with our modelling approach. On the other hand, our semi-quantitative validation does only show minor offsets. Furthermore, general relief characteristics remain undisturbed and suitable for hydrogeographic modelling approaches.

2.6.3 Evidence of excellent Carolingian knowledge in engineering

The LCPA reveals the most cost-efficient course of Fossa Carolina and allows for a comparison of the modelled and the real Carolingian canal course. The overall shapes of both courses are nearly identical. The sharp bend framed by the west-east section and the northern section is thus mainly induced by the topography and not by geological conditions, as postulated by Koch (1993). The topography and the predicted earth volume must therefore have been the crucial factors in the process of decision making 42

2 Shaping pre-modern Digital Terrain Models concerning the course of Fossa Carolina. Other factors may also have played a role but were not necessarily key, such as the possibility of connecting the Swabian Rezat with the canal in order to supply the summit section with water (as suggested by Zielhofer et al. (2014)).

As illustrated in Figure 2.11, there is a small offset between the modelled course and the real Carolingian course in several sections. We assume that these shifts are related to the hydro-engineering concept and practical reasons of organisation of the construction site. The course of the canal within the depression is less laborious (with minimal excavation material), but it has to start under wet conditions from the very beginning. Accessibility and the transport of men, building material, and excavated material would have been much more difficult there. Furthermore, keeping the construction site drained and as “water-free” as possible was most likely a major task. In the depression line, Carolingian hydro-engineers would have been confronted with interflow and inflowing groundwater. Furthermore, the sediments in the wet depression are organic-rich loams and peats (Zielhofer and Kirchner 2014). The stability of embankments of those sediments for canal construction is much more challenging.

Without using modern data and survey techniques, Carolingian engineers traced out the canal along the most effective route with an impressive level of precision. The deviations from the ideal line further underline that the people in charge had a deep understanding of the local topography and hydrology, as well as technical means to apply all that knowledge in a perfectly surveyed canal course. The course is thus a carefully chosen compromise between the minimum earthmoving and the maximum geotechnical stability and site accessibility.

2.7 Conclusion

We have provided a general approach for the revision of high-resolution DTMs. The concept of the pre- modern DTMs is a well-reproducible prospecting method for geoarchaeological, historic-geographical, geomorphological, and palaeohydrological issues. The high spatial resolution offers the possibility of making even small-scale changes of the terrain visible. Since most of the input data is widely available, the transferability to other study sites is very high. The subtraction of the pre-modern model from the modern DTM generates a highly significant raster layer that visualises the local human impact on the relief. This dataset has high potential as a prospection and visualization tool for geoarchaeological issues.

For the first time, we presented a deductive modelling approach with a two-way validation using palaeo- pedological data to estimate the error of our model. Based on the reliable pre-modern DTM, we modelled the most favourable course of Fossa Carolina. The general modelled course is nearly identical to the Carolingian course. Slight deviations of the predicted course document a carefully chosen compromise between the minimum earthmoving and the maximum geotechnical stability and site accessibility. This suggests that the Carolingian engineers had an impressive understanding of the landscape, hydro- engineering, large-scale construction site organisation, and surveying. 43

2 Shaping pre-modern Digital Terrain Models

2.8 References

Baubinienė, A.; Morkūnaitė, R.; Bauža, D.; Vaitkevičius, G.; Petrošius, R. (2015): Aspects and methods in reconstructing the medieval terrain and deposits in Vilnius. In: Quaternary International 386, S. 83–88. DOI: 10.1016/j.quaint.2014.09.068.

Bavarian Land Surveying Office (2018a): Flurkarte. https://www.ldbv.bayern.de/produkte/kataster/ flurkarte.html, accessed 11.04.2018.

Bavarian Land Surveying Office (2018b): Geländemodell. https://www.ldbv.bayern.de/produkte/ 3dprodukte/ gelaende.html, accessed 11.04.2018.

Bavarian Land Surveying Office (2018c): Historische Flurkarten. https://www.ldbv.bayern.de/produkte/ historisch/hist-flurkarten.html, accessed 11.04.2018.

Bavarian Land Surveying Office (2018d): Orthofotos. https://www.ldbv.bayern.de/produkte/luftbild/ orthophotos.html, accessed 11.04.2018.

Bavarian Land Surveying Office (2018e): Tatsächliche Nutzung. https://www.ldbv.bayern.de/produkte/ kataster/tat_nutzung.html, accessed 11.04.2018.

Bavarian State Department of Cultural Heritage BLfD (1985): Luftbildarchiv, Archivnummer: 7130_027, Filmnummer: 3840B, Bild 12, 19.02.1985. O. Braasch (Regie).

Berg-Hobohm, S.; Kopecky-Hermanns, B. (2012): Naturwissenschaftliche Untersuchungen in der Umgebung des Karlsgrabens (Fossa Carolina). In: Berichte der bayerischen Bodendenkmalpflege 52, S. 403–418.

Bewley, R. H. (2003): Aerial survey for archaeology. In: Photogrammetric Record 18 (104), S. 273–292. DOI: 10.1046/j.0031-868X.2003.00023.x.

Bodansky, E.; Gribov, A.; Pilouk, M. (2002): Smoothing and Compression of Lines Obtained by Raster- to-Vector Conversion. In: Goos, G.; Hartmanis, J.; van Leeuwen, J.; Blostein; D. and Kwon, Y.-B. (Hg.): Graphics Recognition Algorithms and Applications, Bd. 2390. Berlin, Heidelberg: Springer Berlin Heidelberg (Lecture Notes in Computer Science), S. 256–265.

Bond, J. (2007): Canal Construction in the Early Middle Ages. An Introductory Review. In: Blair, J. (Hg.): Waterways and Canal-Building in Medieval England. Oxford: Oxford University Press (Medieval History and Archaeology series).

Bussmann, J.; Stele, A.; Härtling, J. W.; Zielhofer, C.; Fuchs, M. C. (2014): Holocene Sediment Dynamics in the Vicinity of a Roman battlefield near Osnabrück (NW-Germany). In: Zeitschrift für Geomorphologie, Supplementary Issues 58 (1), S. 97–117. DOI: 10.1127/0372-8854/2013/S-00133. 44

2 Shaping pre-modern Digital Terrain Models

Carey, C.; Howard, A. J.; Jackson, R.; Brown, A. (2017): Using geoarchaeological deposit modelling as a framework for archaeological evaluation and mitigation in alluvial environments. In: Journal of Archaeological Science: Reports 11, S. 658–673. DOI: 10.1016/j.jasrep.2017.01.013.

Carson, M. T. (2014): Paleo-Terrain Research. Finding the First Settlement Sites of Remote Oceania. In: Geoarchaeology 29 (3), S. 268–275. DOI: 10.1002/gea.21457.

Chai, T.; Draxler, R. R. (2014): Root mean square error (RMSE) or mean absolute error (MAE)? – Arguments against avoiding RMSE in the literature. In: Geosci. Model Dev. 7 (3), S. 1247–1250. DOI: 10.5194/gmd-7-1247-2014.

Chase, A. F.; Chase, D. Z.; Fisher, C. T.; Leisz, S. J.; Weishampel, J. F. (2012): Geospatial revolution and remote sensing LiDAR in Mesoamerican archaeology. In: Proceedings of the National Academy of Sciences of the United States of America 109 (32), S. 12916–12921. DOI: 10.1073/pnas.1205198109.

Conrad, O.; Bechtel, B.; Bock, M.; Dietrich, H.; Fischer, E.; Gerlitz, L.; Wehberg, J.; Wichmann, V.; Böhner, J. (2015): System for Automated Geoscientific Analyses (SAGA) v. 2.1.4. In: Geosci. Model Dev. 8 (7), S. 1991–2007. DOI: 10.5194/gmd-8-1991-2015.

Corrò, E.; Mozzi, P. (2017): Water matters. Geoarchaeology of the city of Adria and palaeohydrographic variations (Po Delta, Northern Italy). In: Journal of Archaeological Science: Reports 15, S. 482–491. DOI: 10.1016/j.jasrep.2016.08.001.

Doneus, M.; Briese, C.; Fera, M.; Janner, M. (2008): Archaeological prospection of forested areas using full-waveform airborne laser scanning. In: Journal of Archaeological Science 35 (4), S. 882–893. DOI: 10.1016/j.jas.2007.06.013.

Ehlers, C. (2014): Der reisende Herrscher: Organisation and Kommunikation. In: Pohle, K (Hg.): Karl der Große. Charlemagne. Orte der Macht. Essays. Dresden: Sandstein Kommunikation, S. 40–47.

Elmshäuser, K. (1992): Kanalbau and technische Wasserführung im Mittelalter. In: Technikgeschichte 59 (1), S. 1–26.

Faïsse, C.; Mathé, V.; Bruniaux, G.; Labussière, J.; Cavero, J.; Jézégou, M.-P.; Lefèvre, D.; Sanchez, C. (2018): Palaeoenvironmental and archaeological records for the reconstruction of the ancient landscape of the Roman harbour of Narbonne (Aude, France). In: Quaternary International 463, S. 124–139. DOI: 10.1016/j.quaint.2017.03.072.

Freeland, T.; Heung, B.; Burley, D. V.; Clark, G.; Knudby, A. (2016): Automated feature extraction for prospection and analysis of monumental earthworks from aerial LiDAR in the Kingdom of Tonga. In: Journal of Archaeological Science 69, S. 64–74. DOI: 10.1016/j.jas.2016.04.011.

Geofabrik (2016): Open Street Map data: Mittelfranken. https://download.geofabrik.de/europe/germany/bayern/mittelfranken.html, accessed 11.04.2018. 45

2 Shaping pre-modern Digital Terrain Models

Hesse, R. (2010): LiDAR-derived Local Relief Models - a new tool for archaeological prospection. In: Archaeol. Prospect. 79 (3), n/a-n/a. DOI: 10.1002/arp.374.

Höfler, V.; Wessollek, C.; Karrasch, P. (2015): Modelling prehistoric terrain Models using LiDAR-data. A geomorphological approach. In: Proceedings SPIE Remote Sensing, 96440B. DOI: 10.1117/12.2194290.

IUSS Working Group WRB (2015): World reference base for soil resources 2014, update 2015. International soil classification system for naming soils and creating legends for soil maps. Rome: World Soil Ressources Reports (106).

Johnson, K. M.; Ouimet, W. B. (2014): Rediscovering the lost archaeological landscape of southern New England using airborne light detection and ranging (LiDAR). In: Journal of Archaeological Science 43, S. 9–20. DOI: 10.1016/j.jas.2013.12.004.

Kirchner, A.; Zielhofer, C.; Werther, L.; Schneider, M.; Linzen, S.; Wilken, D.; Wunderlich, T.; Rabbel, W.; Meyer, C.; Schmidt, J.; Schneider, B.; Berg-Hobohm, S.; Ettel, P. (2018): A multidisciplinary approach in wetland geoarchaeology: Survey of the missing southern canal connection of the Fossa Carolina (SW Germany). In: Quaternary International 473, S. 3–20. DOI: 10.1016/j.quaint.2017.12.021.

Koch, R. (1993): Fossa Carolina - 1200 Jahre Karlsgraben. München: Denkmalpflege Informationen.

Lasaponara, R.; Masini, N. (2011): Satellite remote sensing in archaeology. Past, present and future perspectives. In: Journal of Archaeological Science 38 (9), S. 1995–2002. DOI: 10.1016/j.jas.2011.02.002.

Lee, S.; Wolberg, G.; Shin, S. Y. (1997): Scattered data interpolation with multilevel B-splines. In: IEEE Trans. Visual. Comput. Graphics 3 (3), S. 228–244. DOI: 10.1109/2945.620490.

Leitholdt, E.; Zielhofer, C.; Berg-Hobohm, S.; Schnabl, K.; Kopecky-Hermanns, B.; Bussmann, J.; Härtling, J. W.; Reicherter, K.; Unger, K. (2012): Fossa Carolina. The First Attempt to Bridge the Central European Watershed-A Review, New Findings, and Geoarchaeological Challenges. In: Geoarchaeology 27 (1), S. 88–104. DOI: 10.1002/gea.21386.

Mansouryar, M.; Hedayati, A. (2012): Smoothing Via Iterative Averaging (SIA)A Basic Technique for Line Smoothing. In: IJCEE, S. 307–311. DOI: 10.7763/IJCEE.2012.V4.501.

McCormick, M. (2010): The origin of the European economy. Communications and commerce A.D. 300 - 900. Cambridge: Cambridge Univ. Press.

McKitterick, R. (2008): Charlemagne. The formation of a European identity. Cambridge: Cambridge Univ. Press.

2 Shaping pre-modern Digital Terrain Models

Preiser-Kapeller, J.; Werther, L. (2018): Connecting Harbours. A Comparison of Traffic Networks across Ancient and Medieval Europe. In: von Carnap-Bornheim, C.; Daim, F.; Ettel, P. and Warnke, U. (Hg.): Harbours as object of interdisciplinary research. Archaeology + history + geosciences. Mainz: Verl. des RGZM (RGZM Tagungen, 34), S. 7–31.

Schindling, J.; Gibbes, C. (2014): LiDAR as a tool for archaeological research. A case study. In: Archaeol Anthropol Sci 6 (4), S. 411–423. DOI: 10.1007/s12520-014-0178-3.

Schneider, A.; Hirsch, F.; Wechler, K.-P.; Raab, A.; Raab, T. (2017): Reconstruction of a Palaeosurface and Archaeological Site Location in an Anthropogenic Drift Sand Area. In: Archaeol. Prospect. 19 (8), S. 711. DOI: 10.1002/arp.1571.

Siart, C.; Eitel, B.; Panagiotopoulos, D. (2008): Investigation of past archaeological landscapes using remote sensing and GIS. A multi-method case study from Mount Ida, Crete. In: Journal of Archaeological Science 35 (11), S. 2918–2926. DOI: 10.1016/j.jas.2008.06.006.

Smedt, P. de; van Meirvenne, M.; Herremans, D.; Reu, J. de; Saey, T.; Meerschman, E.; Crombé, P.; Clercq, W. de (2013): The 3-D reconstruction of medieval wetland reclamation through electromagnetic induction survey. In: Scientific reports 3, S. 1517. DOI: 10.1038/srep01517.

Squatriti, P. (2002): Digging Ditches in Early Medieval Europe. In: Past & Present 176 (1), S. 11–65. DOI: 10.1093/past/176.1.11.

Supernant, K. (2017): Modeling Métis mobility? Evaluating least cost paths and indigenous landscapes in the Canadian west. In: Journal of Archaeological Science 84, S. 63–73. DOI: 10.1016/j.jas.2017.05.006. van Loon, A. H.; Schot, P. P.; Griffioen, J.; Bierkens, M.F.P.; Wassen, M. J. (2009): Palaeo-hydrological reconstruction of a managed fen area in The Netherlands. In: Journal of Hydrology 378 (3-4), S. 205– 217. DOI: 10.1016/j.jhydrol.2009.09.014.

Verhegge, J.; Missiaen, T.; Crombé, P. (2016): Exploring Integrated Geophysics and Geotechnics as a Paleolandscape Reconstruction Tool. Archaeological Prospection of (Prehistoric) Sites Buried Deeply below the Scheldt Polders (NW Belgium). In: Archaeol. Prospect. 23 (2), S. 125–145. DOI: 10.1002/arp.1533.

Verhegge, J.; Saey, T.; van Meirvenne, M.; Missiaen, T.; Crombé, P. (2017): Reconstructing Early Neolithic Paleogeography. EMI-Based Subsurface Modeling and Chronological Modeling of Holocene Peat below the Lower Scheldt Floodplain in NW Belgium. In: Geoarchaeology 32 (2), S. 159–176. DOI: 10.1002/gea.21581.

Vermeer, J. A. M.; Finke, P. A.; Zwertvaegher, A.; Gelorini, V.; Bats, M.; Antrop, M.; Verniers, J.; Crombé, P. (2014): Reconstructing a prehistoric topography using legacy point data in a depositional environment. In: Earth Surf. Process. Landforms 39 (5), S. 632–645. DOI: 10.1002/esp.3472.

2 Shaping pre-modern Digital Terrain Models

Web of Science (2018): Citation Report. http://apps.webofknowledge.com/CitationReport.do?product =WOS&search_mode=CitationReport&SID=E1LRaPBj7fw59Je5XkQ&page=1&cr_pqid=2&viewTy pe=summary&colName=WOS, accessed 11.04.2018.

Werbrouck, I.; Antrop, M.; van Eetvelde, V.; Stal, C.; Maeyer, Ph. de; Bats, M.; Bourgeois, J.; Court-Picon, M.; Crombé, Ph.; Reu, J. de; Smedt, Ph. de; Finke, P. A.; van Meirvenne, M.; Verniers, J.; Zwertvaegher, A. (2011): Digital Elevation Model generation for historical landscape analysis based on LiDAR data, a case study in Flanders (Belgium). In: Expert Systems with Applications 38 (7), S. 8178–8185. DOI: 10.1016/j.eswa.2010.12.162.

Werther, L.; Feiner, D. (2014): Der Karlsgraben im Fokus der Archäologie. In: Ettel, P.; Daim, F.; Berg- Hobohm, S.; Werther, L.; and Zielhofer, C. (Hg.): Großbaustelle 793. Das Kanalprojekt Karls des Großen zwischen Rhein and Donau. Mainz: Verl. des Römisch-Germanischen Zentralmuseums (Mosaiksteine, 11), S. 33–40.

Werther, L.; Kröger, L.; Kirchner, A.; Zielhofer, C.; Leitholdt, E.; Schneider, M.; Linzen, S.; Berg- Hobohm, S.; Ettel, P. (2018): Fossata Magna - a Canal Contribution to Harbour Construction in the 1st Millenium AD. In: von Carnap-Bornheim, C.; Daim, F.; Ettel, P.; and Warnke, U.; (Hg.): Harbours as object of interdisciplinary research. Archaeology + history + geosciences. Mainz: Verl. des RGZM (RGZM Tagungen, 34), S. 355–372.

Werther, L.; Zielhofer, C.; Herzig, F.; Leitholdt, E.; Schneider, M.; Linzen, S.; Berg-Hobohm, S.; Ettel, P.; Kirchner, A.; Dunkel, S. (2015): Häfen verbinden. Neue Befunde zu Verlauf, wasserbaulichem Konzept and Verlandung des Karlsgrabens. In: Schmidts, T. and Vučetić, M. M. (Hg.): Häfen im 1. Millennium AD. Bauliche Konzepte, herrschaftliche and religiöse Einflüsse. 1. Aufl. Regensburg, Mainz: Schnell & Steiner; Verl. des RGZM (Interdisziplinäre Forschungen zu den Häfen von der Römischen Kaiserzeit bis zum Mittelalter in Europa, 1), S. 151–185.

Westerdahl, C. (2006): On the significance of Portages. A survey of a new research theme. In: Westerdahl, C. (Hg.): The significance of Portages. Oxford: Archaeopress (BAR International series, 1499), S. 15– 51.

Zielhofer, C.; Kirchner, A. (2014): Naturräumliche Gunstlage der Fossa Carolina. In: Ettel, P.; Daim, F.; Berg-Hobohm, S.; Werther, L. and Zielhofer, C. (Hg.): Großbaustelle 793. Das Kanalprojekt Karls des Großen zwischen Rhein and Donau. Mainz: Verl. des Römisch-Germanischen Zentralmuseums (Mosaiksteine, 11), S. 5–8.

Zielhofer, C.; Leitholdt, E.; Werther, L.; Stele, A.; Bussmann, J.; Linzen, S.; Schneider, M.; Meyer, C.; Berg-Hobohm, S.; Ettel, P. (2014): Charlemagne's summit canal. An early medieval hydro-engineering project for passing the Central European Watershed. In: PloS one 9 (9), e108194. DOI: 10.1371/journal.pone.0108194.

2 Shaping pre-modern Digital Terrain Models

Zwertvaegher, A.; Werbrouck, I.; Finke, P. A.; Reu, J. de; Crombé, P.; Bats, M.; Antrop, M.; Bourgeois, J.; Court-Picon, M.; Maeyer, P. de; Smedt, P. de; Sergant, J.; van Meirvenne, M.; Verniers, J. (2010): On the use of integrated process models to reconstruct prehistoric occupation, with examples from Sandy Flanders, Belgium. In: Geoarchaeology 25 (6), S. 784–814. DOI: 10.1002/gea.20332.

3 3D-modelling of Charlemagne’s Summit Canal

Chapter 3 - 3D-modelling of Charlemagne’s Summit Canal (Southern Germany) – Merging Remote Sensing and Geoarchaeological Subsurface Data

Schmidt, J.; Rabiger-Völlmer, J.; Werther, L.; Werban, U.; Dietrich, P.; Berg, S.; Ettel, P.; Linzen, S.; Stele, A.; Schneider, B.; Zielhofer, C. (2019): 3D-Modelling of Charlemagne’s Summit Canal (Southern Germany) — Merging Remote Sensing and Geoarchaeological Subsurface Data. In: Remote Sensing 11 (9), S. 1111. DOI: 10.3390/rs11091111.

3 3D-modelling of Charlemagne’s Summit Canal

3D modelling of Charlemagne’s summit canal (Southern Germany) – merging remote sensing and geoarchaeological subsurface data

Johannes Schmidt1*, Johannes Rabiger-Völlmer1, Lukas Werther², Ulrike Werban³, Peter Dietrich³, Stefanie Berg4, Peter Ettel², Sven Linzen5, Andreas Stele6, Birgit Schneider1, Christoph Zielhofer1

1Chair of Physical Geography, Leipzig University, D-04103 Leipzig, Germany 2Chair of Prehistory and Early History, Friedrich-Schiller University, D-07743 Jena, Germany 3Helmholtz Centre for Environmental Research UFZ, Department Monitoring and Exploration Technologies, D-04318 Leipzig, Germany 4Bavarian State Department of Cultural Heritage BLfD, D-80539 Munich, Germany 5Leibniz Institute of Photonic Technology IPHT, D-07745 Jena, Germany 6Institute of Geography, Osnabruck University, D-49074 Osnabruck, Germany

*Correspondence: [email protected];

3.1 Abstract

The Early Medieval Fossa Carolina is the first hydro-engineering construction that bridges the Central European Watershed. The canal was built in 792/793 AD on order of Charlemagne and should connect the drainage systems of the Rhine-Main catchment and the Danube catchment. In this study, we show for the first time, the integration of Airborne LiDAR and (geo)archaeological subsurface datasets with the aim to create a 3D-model of Charlemagne’s summit canal. We used a purged Digital Terrain Model that reflects the pre-modern topography. The geometries of buried canal cross-sections are derived from three archaeological excavations and four high-resolution direct push sensing transects. By means of extensive core data, we interpolate the trench bottom and adjacent edges along the entire canal course. As a result, we are able to create a 3D-model that reflects the maximum construction depth of the Carolingian canal and calculate an excavation volume of approx. 297,000 m³. Additionally, we compute the volume of the present dam remnants by Airborne LiDAR data. Surprisingly, the volume of the dam remnants reveals only 120,000 m3 and is much smaller than the computed Carolingian excavation volume. The difference reflects the erosion and anthropogenic overprint since the 8th century AD.

3 3D-modelling of Charlemagne’s Summit Canal

3.2 Introduction

The Central European Watershed divides the Rhine-Main catchment and the Danube catchment. During the Early Middle Ages, navigable waterways were one of the most important transportation, communication and economic exchange routes in Central Europe (McCormick 2010; Squatriti 2002). In the early 790s AD, Charlemagne decided to link the Danube and Rhine-Main catchment by the construction of a canal between Altmühl and Swabian Rezat rivers. Thus, he wanted to create an important node within the waterway network in the expanding Franconian empire (Leitholdt et al. 2012; Werther et al. 2018). This important hydro-technical construction in the Early Middle Ages is one of the spatially largest features of cultural heritage in Bavaria.

The archaeological research on the canal reaches back at the beginning of the 20th century with a publication of Beck (1911). The time before was characterised by chronicles, oral traditions and descriptions (Berg-Hobohm 2014). During the 20th century, several archaeological surveys were conducted. There is a long-lasting debate on the volume of the moved material for the construction and subsequently on the number of workers that were needed (Beck 1911; Birzer 1958; Koch 1993, 1996). Former studies estimated the volume just on the base of few recovered cores (Koch 1993, 1996) or solely simple observations (Beck 1911; Birzer 1958). Thus, none of these studies came to a satisfying result, mostly because of a lack of precise data of the canal extent and geometry as well as its hydro-technical concept. Therefore, a 3D-modelling approach is essential for answering these specific quantitative questions, like excavation volumes and its spatial distribution.

So far, three-dimensional approaches in archaeology are mainly image-based (photogrammetric) studies describing and preserving architecture, objects and features (Bruno et al. 2010; Ducke et al. 2011; Reu et al. 2014). Richards-Rissetto (2017) sees the main advantage of 3D-views in the change from the bird eye’s view in GIS to a more human perspective. These approaches are sometimes used in terms of cyber- archaeology and virtual reality for museums and education (Koutsoudis et al. 2007; Forte 2014).

In this study, we understand 3D-modelling as a high precision reconstruction technique in terms of ruined, buried or overprinted features. Archaeological remains, which are detectable as positive landforms can be reconstructed by means of geometric simplification of the structure (Pickett et al. 2016). Hence, it is possible to calculate the volume. Thereby, the type of simplification and the amount of complexity is crucial for a reliable result (Lacquement 2010; Pickett et al. 2016). However, full form archaeological features are often reconstructed by means of archaeological excavation (Sherwood and Kidder 2011; Pickett et al. 2016). Mapping of these structures is mostly done by LiDAR (Light Detection and Ranging), unmanned airborne vehicles (UAV, drones; equipped with laser scanning or photogrammetric devices) or ground based differential GPS survey (Grammer et al. 2017; Cowley et al. 2018; Zielhofer et al. 2018b).

Covered and filled archaeological features like ditches, trenches and pits are often reconstructed (with the aim of volume calculation) with methods of archaeological excavation (Andersen 1997). Currently, the use

3 3D-modelling of Charlemagne’s Summit Canal of geophysical methods (Electrical resistivity tomography, electromagnetic induction, seismics) for the reconstruction of archaeological features become also more common (Diamanti et al. 2005; Smedt et al. 2011; Smedt et al. 2013; Hausmann et al. 2018). These methods provide fast spatial approaches of mapping but lacking of detailed stratigraphic data. Direct push sensing makes an exception. This geophysical method produces high resolution stratigraphic data (Hausmann et al. 2018). Further, vibra-coring, subsequent sediment sampling and geochemical laboratory analysis are useful and widespread in geoarchaeology (Hadler et al. 2018; Seeliger et al. 2018). Furthermore, combination of archaeological excavation data with geoarchaeological data become more common (Canti and Huisman 2015; Beuzen-Waller et al. 2018), but there are no studies, which tend to model or reconstruct a whole feature by integration of archaeological, remote sensing, geophysical and sediment core data.

Generally, 3D reconstructions depend on the spatial extent of a feature. The Fossa Carolina has a proven length of approx. 2.9km and a maximum depth of approx. 11m below the present surface (Zielhofer et al. 2014). The whole canal trench is accompanied by dams, originating from Early Medieval excavations. Therefore, a complete archaeological survey is not feasible. In addition, the deeper the structure is buried the more difficult is an excavation due to inflowing groundwater and stabilisation of the trench embankments (Werther and Feiner 2014; Hausmann et al. 2018). Therefore, vibra-coring, direct push sensing or other geoarchaeological techniques are essential. Depending on its accuracy, the data must be integrated to generate meaningful and reliable results.

In this study, we summarise trustworthy and high precision data of the geometry of the Fossa Carolina. We merge levels from driving cores, direct push sensing and archaeological excavations. We combine these datasets with a LiDAR-based, pre-modern DTM (Schmidt et al. 2018) with the aim of creating a high- resolution 3D-model of Charlemagne’s summit canal. We present and use new direct push sensing transects, profiles of archaeological excavations and multiple driving cores. For the first time, we use LiDAR data for modelling the present dam remnants but also refer to published canal geometry data (see Tab. 3.1).

3 3D-modelling of Charlemagne’s Summit Canal

Table 3.1: Depth accuracy, scale and stratigraphical resolution of trench geometry data; scale classification according to Zielhofer et al. (2018a).

Technique Number Name (Label References Depth Lateral Scale Resolution Pace in Figure 2) accuracy distances of stratigraphy Excavation 3 trenches “2013” (c) Werther and ++ cm-scale micro to small +++ - “2016 – S1” (d) Feiner 2014; “2016 – S2” (e) Werther et al. 2015; Werther 2017 Direct-push 2 transects “WE-Section” Völlmer et al. ++ 12.5 to 50 cm micro to medium ++ ++ (a) 2018; “TheAnomaly” This study (b)

Drilling 26 transects - Leitholdt et al. ○ Up to 1 m small to medium + + 2012; Leitholdt et al. 2014; Zielhofer et al. 2014; Kirchner et al. 2018; This study

At a first step, we present standard cross-section reference geometries derived from archaeological excavations and direct push sensing transects. Subsequently, we conduct our 3D-modelling approach by the transfer of the cross-section reference geometries to extensive vibra-coring positions along the entire canal course. Further, we integrate the ground truth data within the pre-modern topography to establish the 3D-model. Finally, we calculate the volume of the present dam remnants and compare both volume results in the context of the construction and decay of the Fossa Carolina.

The main objectives of our study are:

1. Integrating different geoarchaeological datasets and creating a standard routine for high precision 3D-modelling of the Carolingian excavation depth and volume.

2. Calculation of the earth excavation volume based on the Fossa Carolina 3D-model and the pre- modern DTM.

3. Calculation of the dam volume based on the present LiDAR DTM and the pre-modern DTM.

4. Comparison of both volumes from the canal trench and dams, describing the differences and its implications for the Carolingian and post-Carolingian history of the canal.

3 3D-modelling of Charlemagne’s Summit Canal

3.3 Materials and Methods

3.3.1 Study area

Charlemagne’s summit canal is located on a valley watershed between Altmühl and Swabian Rezat rivers in Middle Franconia, Bavaria, Southern Germany (Fig. 3.1). The valley fills consist of Pleistocene sandy and loamy fluvial sediments building up the valley watershed (Schmidt-Kaler 1993) as part of the Central European Watershed (approx. 420 m a.s.l.) (Zielhofer and Kirchner 2014). In the South, the Altmühl River level corresponds with an altitude of 408.3 m a.sl. The Altmühl is a tributary of the Danube River. In the North of the Fossa Carolina, the Swabian Rezat River, as part of the Rhine-Main catchment, corresponds with an altitude of 413.5 m a.s.l.

The canal course can be divided into five sections in relation to surface structures and trench bottom depths (see Fig. 3.2).

(a) The Altmühl floodplain consists of late Pleistocene fluvial deposits, mainly gravels and sands. Holocene alluvial sediments accumulated subsequently with an intercalated mid-Holocene soil indicating neglectable meander migration during the late Holocene (Kirchner et al. 2018). Further, Kirchner et al. (2018) deduced, that the Fossa Carolina was never built in the section and, therefore, the canal has never been finished.

(b) The area of the watershed (Central Section) is sedimentary characterised by Pleistocene valley fills of sandy grain sizes (Leitholdt et al. 2012). The sediments are of reddish to greyish colours, due to different

3 3D-modelling of Charlemagne’s Summit Canal redox conditions (Zielhofer et al. 2014; Hausmann et al. 2018; Völlmer et al. 2018). Here, large lateral dams are still present in this section. These reach altitudes up to 13 m above present pond level.

(c) The West-East Section marks the transition from the watershed to the Swabian Rezat floodplain. The sandy parent material is similar to the Central Section (Hausmann et al. 2018; Völlmer et al. 2018). The lateral dams are smaller compared to the Central Section, but still prominent with altitudes up to 5 m above the inner trench levels.

(d) The Northern Section represents the South-North canal course parallel to the Rezat fen in the East. Sandy to loamy fluvial sediments dominate and close to the Rezat fen a half-bog soil is developed (Zielhofer et al. 2014). Here, the dams are almost not visible in the field, but noticeably identifiable in the LiDAR DTM.

(e) The Northeast Section is similar to the Northern Section, with clastic valley fills but the influence the organic sediments in the northernmost part increases. The Swabian Rezat floodplain (Rezat fen) is characterised by organic sediments that reach up to 3m thickness (Zielhofer and Kirchner 2014). Similar to the Northern Section, the dams are almost not visible in the field but can be clearly detected in the DTM.

The Fossa Carolina has a length approx. 2.9 km (Zielhofer et al. 2014) and the prove of the Early Medieval summit canal was done by vibra-coring (Leitholdt et al. 2012; Leitholdt et al. 2014; Zielhofer et al. 2014; Kirchner et al. 2018), direct push sensing (Hausmann et al. 2018; Völlmer et al. 2018) and archaeological excavations (Werther and Feiner 2014; Werther et al. 2015). The canal course starts on the southern slope of the valley watershed and runs with a noticeable s-shape in the northern direction (Fig. 3.2). The s-shape is a result of the impressive knowledge of the Carolingian constructors to make the best alignment of the canal course in relation to a minimal excavation workload (Schmidt et al. 2018).

3 3D-modelling of Charlemagne’s Summit Canal

Figure 3.2: Local setting and course of the Fossa Carolina and its subdivision in I) Central Section, II) West-East Section, III) Northern Section, IV) North-Eastern Section. All input data for the subsequent modelling are shown, including drillings, modelled trench bottom transects and additive transects. Cross- section reference geometries a) “West-East Section”, b) “The Anomaly” base on direct push sensing data. Cross-section reference geometries c) “2013”, d) “2016 – S1”, e) “2016 – S2” base on archaeological excavations. LiDAR data have been provided by the Bavarian Land Surveying Office.

According to Werther and Feiner (2014) and Völlmer et al. (2018) the general stratigraphy of the canal fills can be summarised: the pre-Carolingian parent material consists of sandy to loamy fluvial sediments with almost no organic remains. The timber (oaks recovered during archaeological excavations, which were used to stabilise the embankments of the Carolingian canal trench) documents the construction time. The initial trench fills feature abruptly redeposited sediments with less organic material. Subsequently, thick organic sediments cover the initial, sandy trench fills. These organic sediments consist of peat and sapropel layers, representing open water bodies and former ponds. The youngest fills feature clastic sediments from mainly 57

3 3D-modelling of Charlemagne’s Summit Canal modern times. They indicate the ongoing erosion of the dams and an intensive land use with intentional levelling of the northern canal sections. Today, parts of the canal structures are fully eroded, refilled and under agricultural use. However, in the Central and West-East Section, massive dam remnants and the course of the canal are still visible (Fig. 3.3).

3 3D-modelling of Charlemagne’s Summit Canal

3.3.2 Data acquisition

3.3.2.1 LiDAR Digital Terrain Model

High-resolution airborne laser scanning data were provided by the Bavarian Land Surveying Office (provided 2012-11-06 and 2013-08-08) (Bavarian land surveying Office 2018a, 2018b). We used the derived 1 x 1 m spatial resolved DTM for the localisation of the canal course and the determination of the midway position between both dam ridges. Furthermore, we use the present LiDAR DTM for the modelling and volume calculation of the present remnants of the dams.

3.3.2.2 Pre-modern Digital Terrain Model

For the modelling of the canal geometry and its integration in the former landscape, the availability of a high- resolution pre-modern DTM is important. In this study, we use the pre-modern DTM with a spatial resolution of 1 x 1 m from Schmidt et al. (2018). This model bases on the LiDAR DTM mentioned above. The DTM is purged of all modern anthropogenic structures, by removing all grid cells, which are affected by anthropogenic disturbance. Finally, remaining cells were interpolated to create a smoothed pre-modern terrain.

3.3.2.3 Magnetic survey

We prospected the large areas of the Northern and North-Eastern Section magnetically with a fast, motorised measurement system developed for geoarchaeological issues by Linzen et al. (2009). The system bases on a set of SQUID (Superconducting Quantum Interference Devise) which provide very high resolution (local centimetre resolution; sensor sampling rate of 1 kHz) and a maximum of magnetic information (Schneider et al. 2013; Linzen et al. 2017). Thus, we detected buried canal remains and localised its course precisely over a distance of more than 1.2 km (Linzen and Schneider 2014).

Within the West-East Section of the canal, we conducted a manually operated Fluxgate magnetic survey to detect the canal course between the present dams. We used a Bartington Grad601 fluxgate magnetometer and analysed and visualised the data with the software Geoplot 3.0 (Zielhofer et al. 2014). We mapped the precise canal course with the aid of the georeferenced magnetic maps.

3.3.2.4 Vibra-Coring

The basic data for the 3D-modelling approach represent sedimentary stratigraphic data recovered from vibra- coring, direct push sensing and archaeological excavations (Tab. 3.1). For vibra-coring, we use an Atlas Copco Cobra Pro hammer and 60 mm open corer. In total, we drilled 39 cores within the trench fills with

3 3D-modelling of Charlemagne’s Summit Canal coring depths between 200 and 800 cm). Mostly, we could identify the trench bottom macroscopically but we verified it by geochemical analysis. The main contrast of the natural sediments and the first anthropogenic backfills is the organic carbon content, because the Pleistocene sediments are sterile and backfills are TOC enriched (Zielhofer et al. 2014).

3.3.2.5 Direct push sensing

Direct push sensing is a fast, minimal-invasive and depth accurate tool for in-situ characterisation of sediment stratigraphies (Dietrich and Leven 2009; Leven et al. 2011). Steel rods with a small diameter (38 mm) and different probes were pushed in the unconsolidated sediments. For the data acquisition we used the colour logging tool (SCOST™, Dakota Technologies, Fargo, USA) to describe different sediment layers and their colour-dependent properties like organic content or redox characteristics (Hausmann et al. 2016; Hausmann et al. 2018). An appropriate pace of measuring (2 cm/s) and an integration time of 300 ms result in 3 values per 2 cm. This high vertical resolution is accompanied by horizontal spacing of up to 12.5 cm. Usually, we used 50 cm spacing. Additionally, an electrical conductivity probe (SC-500, Keijr Engineering Inc. – Geoprobe Systems, USA) provides evidence for grain size changes (Butler et al. 1999; Schulmeister et al. 2003). A Geoprobe 6610DT caterpillar drives the system. For this study, we used 105 direct push colour logs divided in two transects (Tab. 3.1). The sensing depths range between 400 and 800 cm, depending on the specific depth of the canal.

3.3.2.6 Archaeological excavations Based on our detailed (geo-)archaeological and geophysical survey, we conducted three archaeological excavations in 2013 and 2016, cutting the canal rectangular to the embankments (Werther and Feiner 2014; Werther et al. 2015; Werther 2017). The stratigraphy has been documented in detail in the field and validated afterwards by sedimentological analysis (grain size, organic carbon content) as well as archaeobotanical samples (Werther et al. 2015). We measured all points of the local network with a Topcon HiPER II D-GPS. We took the absolute position official measuring points. Additionally, we proved the reference heights with an analogue levelling tool. The edges of the Carolingian trench bottom at both banks could be identified precisely with cm-accuracy, because the well-preserved timber revetments show clear signs of decay in the upper part, which was exposed to the water (Werther et al. 2015). The trench bottom between both banks has been identified based on initial infills such as sapropel and re-located sandy material with higher organic contents compared to the Pleistocene parent material. In 2013, a decimetre-uncertainty of the depth in some parts of the cross-section has been inevitable due to excavation conditions (Werther et al. 2015). In 2016, depth accuracy is on a cm-scale in all parts of the cross-section.

3 3D-modelling of Charlemagne’s Summit Canal

3.3.3 Modelling routine

The main challenge of our study is the combination of geoarchaeological datasets (vibra-coring data, direct push sensing data, archaeological excavation data) and the subsequent integration of the combined data within the pre-modern topography. High-resolution cross-sections of the archaeological excavations and direct push sensing transects are not equally distributed along the canal course. This is mainly due to the high effort of archaeological excavations (Werther and Feiner 2014) and the impassibility for the direct push caterpillar in the Central Section. Furthermore, the vibra-coring positions are neither equally distributed throughout the canal. Thus, we developed a modelling approach, which suits this challenge reflecting four steps to create a 3D-model (Fig. 3.4a-d) and two steps to calculate the excavation and dam volumes (Fig. 3.4e-f).

a) First of all, we create cross section reference geometries that base on high resolution transects. Here, archaeological excavations provide the most precise data in terms of identifying the trench bottom geometry (Tab. 3.1). The spatial and vertical resolution of the direct push transects are also very high. The horizontal

3 3D-modelling of Charlemagne’s Summit Canal spacing is usually 50 cm but can reach up to 12.5 cm. The vertical resolution is at centimetre scale. Hence direct push transects provide significant data for cross-section reference geometries (Fig. 3.4a). In total, we compiled cross-section reference geometries for five cross-sections (a-e in Fig. 3.2). These geometries are also representative for their adjacent sections. b) In a second step, we transferred the cross-section reference geometries to the respective vibra-coring positions (Tab. 3.2). Here, we used trench bottom levels from recovered core stratigraphies. The specific reference geometry was adjusted to the level of the trench bottom, inferred from individual cores (Fig. 3.4b). In the case that the trench bottom is deeper than the cross-section reference geometry, we extended the embankments with its specific slope angle until they reach the pre- modern DTM surface. In the case that the trench bottom is above the depth of the cross-section reference geometry, we cut the supernatant embankments at the pre-modern DTM surface.

Table 3.2: Sections of the Fossa Carolina and corresponding cross-sections reference geometries.

Fossa Carolina Section Length [m] Cross-Section Reference Type Transferred to n core Geometry positions Central Section 803 "WE cross-section" direct push sensing 10 WE Section 494 "WE cross-section" direct push sensing 16 Northern Section I (S) 368 "The Anomaly" direct push sensing 1 Northern Section II (N) 370 "2013" archaeological 6 excavation North-Eastern Section I (S) 476 "2013" archaeological 0 excavation North-Eastern Section II 198 "2016/S1" archaeological 2 (M) excavation North-Eastern Section III 120 "2016/S2" archaeological 4 (N) excavation

c) Subsequently, we created additional transects, which are not based on vibra-coring. These additional transects are important for the interpolation of all transects in the next step. The larger the distance between two transects, the bigger are potential interpolation errors and disturbances (Davis and Herzfeld 1993). For this reason, we added additional transects equidistantly (~50 m spacing) (Fig. 3.2). The depths were transferred from neighbouring vibra-coring trench bottom levels, direct push sensing transects or archaeological excavations (Fig. 3.4c). d) Once all transects have been calculated, we spatially interpolated neighbouring transects via triangulation with an output raster of 0.5 m resolution (Fig. 3.4d). Altogether, we interpolated 73 single segments. Finally, we merged all single segments into one raster dataset and used a low-pass filter (5m radius) to slightly smooth the raster data. The result is a 3D digital terrain model of the Fossa Carolina in its maximum construction depth in a 0.5 x 0.5 m resolution. e+f) To answer the question of the volumes, we integrated the 3D-model in the pre-modern landscape (Fig. 3.4e) provided by Schmidt et al. (2018). We determined the excavation volume of the canal trench by calculating the difference between the 3D-model and the pre-modern DTM (Fig. 3.4f). The volume of the Fossa Carolina dam remnants is computed as the difference between the present LiDAR DTM and the pre-modern DTM.

3 3D-modelling of Charlemagne’s Summit Canal

3.4 Results

3.4.1 Canal course

We localised the canal course with different prospection methods (SQUID magnetic and Fluxgate magnetic prospection as well as DTM analyses), depending on the data availability. In the Central Section, no geophysical data are available due to wet and barley passable ground conditions. Therefore, we used the middle of the dam ridges as alignment of the canal centre (Fig. 3.5b). In the West-East Section, we used a combination of dam ridge positions and a Fluxgate magnetic map to explore the canal course (Fig. 3.5c). Low relief changes and nearly no visible dams characterise the Northern and North-Eastern Sections. Here, we used SQUID magnetic prospection maps for precisely reconstructing the canal course (Fig. 3.5d).

3 3D-modelling of Charlemagne’s Summit Canal

3.4.2 Cross-section reference geometries

We established five cross-section reference geometries based on two direct push sensing transects and three archaeological excavations. These geometries are the first interim results of our study.

The “WE cross-section” is situated in the western part of the West-East Section of the Fossa Carolina (Völlmer et al. 2018) (a in Fig. 3.2; Fig. 3.6). The second direct push sensing cross-section (“The Anomaly”) is situated in the Northern Section (b in Fig. 3.2). We detected the deepest level of the trench bottom at 5.5 m depth below modern surface (Fig. 3.7).

3 3D-modelling of Charlemagne’s Summit Canal

About 300 m further north, we derived a cross-section reference geometry from an archaeological excavation that took place in 2013 (c in Fig. 3.2). The excavation revealed a trench bottom of approx. 3 m below surface (Fig. 3.8a). Another excavation took place in 2016 that provided two cross-section reference geometries in the northernmost part of the canal. At excavation “2016/S1” (d in Fig. 3.2) the maximum depth of the trench bottom is approx. 2 m below surface (Fig. 3.8b) and at excavation “2016/S2” (e in Fig. 3.2) the trench bottom was recovered at approx. 1.2 m below surface (Fig. 3.8c). The general geometry over all cross-sections looks similar. Both cross-sections “2013” and “2016 - S1” show fairway width of approx. 5m and an almost finished canal construction. In contrast, the cross-section “2016 – S2” reveals a fairway width of just approx. 2.5 m.

Figure 3.8: Archaeological excavation reference geometries. a) “excavation 2013”, b) “2016/S1” and c) “2016/S2”.

3 3D-modelling of Charlemagne’s Summit Canal

3.4.3 Application of cross section reference geometries to vibra-coring and additive transects

The application of the cross-section reference geometries to the respective vibra-coring positions and additional trench bottom transects result in an interim data sets of our study (Fig. 3.9). In total, we created 26 transects based on vibra-coring positions and 42 additional trench bottom transects.

Figure 3.9: Exemplary transfer of the cross-section reference geometry “2016/S1” to the vibra-coring transect “Märzkampagne”.

3.4.4 3D-Model

The major step of our study is the development of a 3D-model of the entire Fossa Carolina trench bottom (Fig. 3.10). The interpolation of the all geometry transects resulted in the smooth integration of the canal geometry in the pre-modern landscape. This model is a raster layer and can be used like a DTM within a GIS environment. It has a spatial resolution of 0.5 x 0.5 m.

3.4.5 Volume calculation

The result of the volume calculation can be separated in two data sets. First, the excavation volume calculation bases on the 3D-model of the Fossa Carolina and pre-modern DTM. The material that was

3 3D-modelling of Charlemagne’s Summit Canal moved during the construction has a volume of approx. 297,000 m³. Second, the present remnants of the dams have a volume of approx. 120,000 m³. This calculation bases on the pre-modern DTM and the present LiDAR DTM.

3.5 Discussion

3.5.1 3D-modelling approach and quality

Applications of GIS in geoarchaeological issues are recently common (Carey et al. 2018; Earley-Spadoni 2017). The integration of geoarchaeological data with GIS tools offer various possibilities of data management and analysis (Chapman et al. 2009; McCoy 2017). However, modelling approaches that aim to the reconstruction of archaeological features are rare. Here, reconstructions are mainly on a small scale and include available spatial geometry information such as geophysics or excavation data. Smedt et al. (2013) used a geophysical survey (electromagnetic induction) to reconstruct a medieval wetland reclamation. Further, Diamanti et al. (2005) used geophysical survey data from electrical resistivity tomography to reconstruct buried city ruins. In contrast, Pickett et al. (2016) reconstructed a medieval burial

3 3D-modelling of Charlemagne’s Summit Canal mound by means of data from archaeological excavations. Studies with a combination of geoarchaeological techniques are lacking.

For the first time, we developed an approach for the integration of different geoarchaeological data for a large-scale feature of 2.9 km length. We think that we produced reliable results but we have to discuss the quality and potential sources of uncertainty. The majority of our input data consists of published geoarchaeological cross-sections, archaeological excavations and a pre-modern DTM (Tab. 3.1). Direct push sensing data and geometric information derived from archaeological excavations provide excellent depth-accuracies. Vibra-coring data may have a coarser vertical resolution and uncertainties in depth- accuracy due to the compaction of organic sediments (Hausmann et al. 2018). The overall Root Mean Square Error (RMSE) of the pre-modern DTM (Schmidt et al. 2018) was calculated using vibra-coring, direct push sensing and archaeological excavation data that recovered a buried paleosol and, therefore, the former level of the pre-modern surface. The RMSE of 0.69 shows a general overestimation of the modelled pre-modern DTM levels (69 cm). Hence, the computed volume of the Fossa Carolina excavation material can be slightly lower in contrast to the calculated volume of the dams that would increase.

3.5.2 The scientific history of Fossa Carolina volume calculations

First of all, Birzer (1958) assumed an excavation volume only for the Central and West-East Section of approx. 80,000 m³. As the author had no reliable information about the total length of the canal, he estimated in the next step a 4.5 km long canal course with a constant trench bottom level. As a result, he estimated a total volume of approx. 450,000 m³ for the entire canal course (Tab. 3.3). Hofmann (1976) works in detail with the building energetics. He hypothesises a single trench bottom level and no summit concept. With a length of 1.4 km, 30 m width and an assumed depth of 6 m, he calculated a volume of approx. 130,000 m³. The most recent volume estimation is from Koch (1993). He drilled several cores and recovered that the detected trench bottom levels do not reach the level of the Altmühl River. Therefore, he concluded that the canal was presumable constructed as a summit canal and carefully assumed roughly several 100,000 m³ of excavation material (Tab. 3.3).

Our study presents for the first time a 3D-modelling approach that base on precise excavation, vibra-coring and direct push sensing data sets. The resulting volume of approx. 297,000 m³ represents an improved calculation in comparison with former estimations. With respect to the summit concept, no other authors assumed a volume as large as we calculated. We can summarise that the former lack of geoarchaeological data led to uncertain conclusions. Furthermore, this precise volume calculation is a key for reliable future modelling of building energetics of this unique canal construction.

3 3D-modelling of Charlemagne’s Summit Canal

Table 3.3: Scientific history of Fossa Carolina volume estimations.

No. Study Volume Object Method Comments 1 Birzer 1958 80,000 m³ canal trench estimation/calculation minimum; only Central and WE-Section

3 Birzer 1958 450,000 m³ canal trench estimation/calculation Assumed canal length of 4.5 km and constant trench bottom level 6 Hofmann 1976 130,000 m³ canal trench calculation Assumed canal length 1.4 km, width 30 m, depth 6 m 5 Koch 1993 several 100,000 canal trench estimation Assumed canal length 5-7 km m³ 2 This study 297,667 m³ canal trench calculation Integrative approach

7 This study 119,681 m³ dams calculation Calculated based on present dams in comparison to pre-modern DTM

The spatial distribution of the excavated volume shows that the majority (54%) was excavated in the Central Section (Fig. 3.11). Further, the West-East Section has a proportion of roughly 32% of the volume and only 14% corresponds with the Northern and North-Eastern Sections (Tab. 3.4).

Figure 3.11: Spatial distribution of excavated volume at the Fossa Carolina. Volume amounts are given in m³ per cell (0.25 m²).

3 3D-modelling of Charlemagne’s Summit Canal

Table 3.4: Relative amounts of excavated volume at individual Fossa Carolina sections.

Section Section Length Trench Trench volume Dam Ratio length proportion volume proportion volume dams/trench Total 2829 m 100% 297,667 m³ 100% 119,681 m³ 40% Central Section 803 m 28% 160,815 m³ 54% 83,826 m³ 52% WE Section 494 m 17% 96,496 m³ 32% 20,449 m³ 21% Northern Section 738 m 27% 26,267 m³ 9% 10,864 m³ 41% North-Eastern 794 m 27% 14,088 m³ 5% 4,558 m³ 32% Section

3.5.3 Where has all the material gone?

It is obvious, that more than half of the excavated volume is not stored in the remaining dams. Vibra-coring (Kirchner et al. 2018), direct push sensing (Völlmer et al. 2018) and archaeological excavation data (Werther and Feiner 2014) show that significant amounts of excavated material was already washed back into the open trench shortly after the construction site was abandoned. Upper trench fills reflect relocated material originating from the adjacent dams (Zielhofer et al. 2014). The canal and the corresponding dams were fundamentally modified even during modern times. Especially the northern canal sections were levelled for agricultural purposes (Berg-Hobohm and Werther 2014) and missing dam volumes in the central canal sections result from massive sand and loam mining activities for modern infrastructure and buildings (Beck 1911; Schmidt et al. 2018).

3.6 Conclusions

Our 3D modelling approach of the Early Medieval Fossa Carolina integrates archaeological excavations, direct push sensing and vibra-coring techniques, as well as the present LiDAR DTM, Fluxgate and SQUID magnetic surveys and a palaeo-surface (pre-modern DTM) of the study area. We identified the buried canal trench by LiDAR DTM analysis of the present dam remnants and interpretation of magnetic survey maps. We transferred cross-section reference geometries (derived from archaeological excavations and direct push sensing transects) to vibra-coring positions. Finally, we interpolated the canal trench geometry to create the 3D-model of the Carolingian canal. The spatial and vertical accuracy of the model depends on the quality of its input data. The reference cross-sections are of high quality as well as the LiDAR DTM and pre-modern DTM. Because of the large spatial extent of the canal, we included several vibra-corings, which have at least dm depth-accuracy. Our modelling routine minimises the uncertainties by creating cross-section reference geometries.

For the first time, the 3D-model provide a data-based calculation of the amount of material moved during the Early Medieval construction. The calculation of the earth volume was done by calculating the difference between the 3D-model and the pre-modern digital terrain model. The pre-modern DTM reflects a 70

3 3D-modelling of Charlemagne’s Summit Canal deconstructed landscape nearly free of human induced terrain changes. Altogether, approx. 297,000 m³ of material was excavated during the Early Medieval construction time.

In comparison, we calculated the volume of the preserved dams. This was also done by subtracting the present shape from the pre-modern DTM. Approximately 120,000 m³ material is still remaining in the dams. Nevertheless, more than the half of the excavation material was eroded, redistributed, or backfilled in the canal.

3.7 References

Andersen, N. H. (1997): The Sarup Enclosures. The funnel Beaker Culture of the Sarup site including two causewaysed camps compared to the contemporary settlements in the area and other European enclosures. Aarhus: Aarhus University Press (Jutland Archaeological Society Publications, XXXIII).

Bavarian Land Surveying Office (2018a): Airborne Laserscanning. https://www.ldbv.bayern.de/ produkte/3dprodukte/laser.html, accessed 11.04.2018.

Bavarian Land Surveying Office (2018b): Geländemodell. https://www.ldbv.bayern.de/produkte/ 3dprodukte/gelaende.html, accessed 11.04.2018.

Bavarian State Department of Cultural Heritage BLfD (1985): Luftbildarchiv, Archivnummer: 7130_027, Filmnummer: 3840B, Bild 12, 19.02.1985. O. Braasch (Regie).

Beck, F. (1911): Der Karlsgraben. Eine historische, topographische and kritische Abhandlung. Nürnberg: Verlag der Friedrich Kornschen Buchhandlung.

Berg-Hobohm, S. (2014): Archäologische Forschungsgeschichte der Fossa Carolina. In: Ettel, P.; Daim, F.; Berg-Hobohm, S.; Werther, L.; and Zielhofer, C. (Hg.): Großbaustelle 793. Das Kanalprojekt Karls des Großen zwischen Rhein and Donau. Mainz: Verl. des Römisch-Germanischen Zentralmuseums (Mosaiksteine, 11), S. 1–4.

Berg-Hobohm, S.; Werther, L. (2014): Das rezente Erscheinungsbild des Karlsgrabens. In: Ettel, P.; Daim, F.; Berg-Hobohm, S.; Werther, L. and Zielhofer, C. (Hg.): Großbaustelle 793. Das Kanalprojekt Karls des Großen zwischen Rhein and Donau. Mainz: Verl. des Römisch-Germanischen Zentralmuseums (Mosaiksteine, 11), S. 9–12.

Beuzen-Waller, T. Stock, F.; Kondo, Y. (2018): Geoarchaeology: A toolbox for revealing latent data in sedimentological and archaeological records. In: Quaternary International 483, S. 1–4. DOI: 10.1016/j.quaint.2018.05.029.

Birzer, F. (1958): Der Kanalbauversuch Karls des Großen. In: Geologische Blätter für Nordost-Bayern and angrenzende Gebiete 8, S. 171–178.

3 3D-modelling of Charlemagne’s Summit Canal

Bruno, F.; Bruno, S.; Sensi, G. de; Luchi, M.-L.; Mancuso, S.; Muzzupappa, M. (2010): From 3D reconstruction to virtual reality: A complete methodology for digital archaeological exhibition. In: Journal of Cultural Heritage 11 (1), S. 42–49. DOI: 10.1016/j.culher.2009.02.006.

Butler, J. J.; Healey, J. M.; Zheng, L.; McCall, G. W.; Schulmeister, M. K. (1999): Hydrostratigraphic characterization of unconsolidated alluvial deposits with direct-push sensor technology. In: Kansas Geological Survey Open-File report 99-40. http://www.kgs.ku.edu/Hydro/Publications/OFR99_40/.

Canti, M.; Huisman, D. J. (2015): Scientific advances in geoarchaeology during the last twenty years. In: Journal of Archaeological Science 56, S. 96–108. DOI: 10.1016/j.jas.2015.02.024.

Carey, C.; Howard, A. J.; Knight, D.; Corcoran, J.; Heathcote, J. (2018): Deposit Modelling and Archaeology. Brighton: University of Brighton.

Chapman, H.; Adcock, J.; Gater, J. (2009): An approach to mapping buried prehistoric palaeosols of the Atlantic seaboard in Northwest Europe using GPR, geoarchaeology and GIS and the implications for heritage management. In: Journal of Archaeological Science 36 (10), S. 2308–2313. DOI: 10.1016/j.jas.2009.06.015.

Cowley, D.; Moriarty, C.; Geddes, G.; Brown, G.; Wade, T.; Nichol, C. (2018): UAVs in Context: Archaeological Airborne Recording in a National Body of Survey and Record. In: Drones 2 (1), S. 2. DOI: 10.3390/drones2010002.

Davis, J. C.; Herzfeld, U. C. (1993): Computers in geology. 25 years of progress. New York: Oxford University Press (International Association for Mathematical Geology studies in mathematical geology, 5). http://site.ebrary.com/lib/academiccompletetitles/home.action.

Diamanti, N. G.; Tsokas, G. N.; Tsourlos, P. I.; Vafidis, A. (2005): Integrated interpretation of geophysical data in the archaeological site of Europos (northern Greece). In: Archaeol. Prospect. 12 (2), S. 79–91. DOI: 10.1002/arp.249.

Dietrich, P.; Leven, C. (2009): Direct Push-Technologies. In: Kirsch, R. (Hg.): Groundwater Geophysics. 2. Aufl. Berlin: Springer, S. 347–366.

Ducke, B.; Score, D.; Reeves, J. (2011): Multiview 3D reconstruction of the archaeological site at Weymouth from image series. In: Computers & Graphics 35 (2), S. 375–382. DOI: 10.1016/j.cag.2011.01.006.

Earley-Spadoni, T. (2017): Spatial History, deep mapping and digital storytelling. Archaeology’s future imagined through an engagement with the Digital Humanities. In: Journal of Archaeological Science 84, S. 95–102. DOI: 10.1016/j.jas.2017.05.003.

3 3D-modelling of Charlemagne’s Summit Canal

Forte, M. (2014): Virtual Reality, Cyberarchaeoloogy, teleimmersive Archaeology. In: Remondino, F. and Campana, S. (Hg.): 3D Recording and Modelling in Archaeology and Cultural Heritage. Theory and best practises. Oxford: Archaeopress (BAR International series, 2598), S. 113–127.

Grammer, B.; Draganits, E.; Gretscher, M.; Muss, U. (2017): LiDAR-guided Archaeological Survey of a Mediterranean Landscape. Lessons from the Ancient Greek Polis of Kolophon (Ionia, Western Anatolia). In: Archaeol. Prospect. 54 (2-3), S. 64. DOI: 10.1002/arp.1572.

Hadler, H.; Vött, A.; Newig, J.; Emde, K.; Finkler, C.; Fischer, P.; Willershäuser, T. (2018): Geoarchaeological evidence of marshland destruction in the area of Rungholt, present-day Wadden Sea around Hallig Südfall (North Frisia, Germany), by the Grote Mandrenke in 1362 AD. In: Quaternary International 473, S. 37–54. DOI: 10.1016/j.quaint.2017.09.013.

Hausmann, J.; Dietrich, P.; Vienken, T.; Werban, U. (2016): Technique, analysis routines, and application of direct push-driven in situ color logging. In: Environ Earth Sci 75 (11), S. 659. DOI: 10.1007/s12665- 016-5515-7.

Hausmann, J.; Zielhofer, C.; Werther, L.; Berg-Hobohm, S.; Dietrich, P.; Heymann, R.; Werban, U. (2018): Direct push sensing in wetland (geo)archaeology. High-resolution reconstruction of buried canal structures (Fossa Carolina, Germany). In: Quaternary International (473), S. 21–36. DOI: 10.1016/j.quaint.2017.02.008.

Hofmann, H. H. (1976): Kaiser Karls Kanalbau. “Wie Künig Carl der Grosse unterstünde die Donaw vnd den Rhein zusam̄ enzugraben”. 2. Aufl. Sigmaringen: J. Thorbecke (Kulturgeschichtliche Miniaturen).

Koch, R. (1993): Fossa Carolina - 1200 Jahre Karlsgraben. München: Denkmalpflege Informationen.

Koch, R. (1996): Neue Beobachtungen and Forschungen zum Karlsgraben. In: Jahrbuch des Historischen Vereins für Mittelfranken 97, S. 1–16.

Koutsoudis, A.; Arnaoutoglou, F.; Chamzas, C. (2007): On 3D reconstruction of the old city of Xanthi. A minimum budget approach to virtual touring based on photogrammetry. In: Journal of Cultural Heritage 8 (1), S. 26–31. DOI: 10.1016/j.culher.2006.08.003.

Lacquement, C. H. (2010): Recalculating mound volume at moundville. In: Southeastern Archaeology 29 (2), S. 341–354.

3 3D-modelling of Charlemagne’s Summit Canal

Leitholdt, E.; Krüger, A.; Zielhofer, C. (2014): The medieval peat layer of the Fossa Carolina – Evidence for bridging the Central European Watershed or climate control? In: Zeit fur Geo Supp 58 (1), S. 189– 209. DOI: 10.1127/0372-8854/2012/S-00113.

Leven, C.; Weiß, H.; Vienken, T.; Dietrich, P. (2011): Direct-Push-Technologien – Effiziente Untersuchungsmethoden für die Untergrunderkundung. In: Grundwasser 16 (4), S. 221–234. DOI: 10.1007/s00767-011-0175-8.

Linzen, S.; Schneider, M. (2014): Der Karlsgraben im Fokus der Geophysik. In: Ettel, P.; Daim, F.; Berg- Hobohm, S.; Werther L. and Zielhofer, C. (Hg.): Großbaustelle 793. Das Kanalprojekt Karls des Großen zwischen Rhein and Donau. Mainz: Verl. des Römisch-Germanischen Zentralmuseums (Mosaiksteine, 11), S. 29–32.

Linzen, S.; Schneider, M.; Berg-Hobohm, S.; Werther, L.; Ettel, P.; Zielhofer, C.; Schmidt, J.; Fassbinder, J. W. E.; Wilken, D.; Fediuk, A.; Dunkel, S.; Stolz, R.; Meyer, H.-G.; Sommer, C. S. (2017): From magnetic SQUID prospection to excavation - investigations at Fossa Carolina, Germany. In: Jennings, B.; Gaffney, C.; Sparrow T. and Gaffney, S. (Hg.): 12th International Conference of Archaeoloigcal Prospection. Bradford. The University of Bradford: Archaeopress, S. 144–145.

Linzen, S.; Schultze, V.; Chwala, A.; Schüler, T.; Schulz, M.; Stolz, R.; Meyer, H.-G. (2009): Quantum Detection Meets Archaeology – Magnetic Prospection with SQUIDs, Highly Sensitive and Fast. In: Reindel, M. and Wagner, G. A. (Hg.): New Technologies for Archaeology. Berlin, Heidelberg: Springer, S. 71–85.

McCormick, M. (2010): The origin of the European economy. Communications and commerce A.D. 300 - 900. Cambridge: Cambridge Univ. Press.

McCoy, M. D. (2017): Geospatial Big Data and archaeology. Prospects and problems too great to ignore. In: Journal of Archaeological Science 84, S. 74–94. DOI: 10.1016/j.jas.2017.06.003.

Pickett, J.; Schreck, J. S.; Holod, R.; Rassamakin, Y.; Halenko, O.; Woodfin, W. (2016): Architectural energetics for tumuli construction. The case of the medieval Chungul Kurgan on the Eurasian steppe. In: Journal of Archaeological Science 75, S. 101–114. DOI: 10.1016/j.jas.2016.09.006.

Reu, J. de; Smedt, P. de; Herremans, D.; van Meirvenne, M.; Laloo, P.; Clercq, W. de (2014): On introducing an image-based 3D reconstruction method in archaeological excavation practice. In: Journal of Archaeological Science 41, S. 251–262. DOI: 10.1016/j.jas.2013.08.020.

3 3D-modelling of Charlemagne’s Summit Canal

Richards-Rissetto, H. (2017): What can GIS + 3D mean for landscape archaeology? In: Journal of Archaeological Science 84, S. 10–21. DOI: 10.1016/j.jas.2017.05.005.

Schmidt, J.; Werther, L.; Zielhofer, C. (2018): Shaping pre-modern digital terrain models. The former topography at Charlemagne’s canal construction site. In: PloS one 13 (7), e0200167. DOI: 10.1371/journal.pone.0200167.

Schmidt-Kaler, H. (1993): Geologie and Landschaftsentwicklung im Rezat-Altmühl Bereich. In: Bau intern Special Issue, S. 8–10.

Schneider, M.; Stolz, R.; Linzen, S.; Schiffler, M.; Chwala, A.; Schulz, M.; Dunkel, S.; Meyer, H.-G. (2013): Inversion of geo-magnetic full-tensor gradiometer data. In: Journal of Applied Geophysics 92, S. 57–67. DOI: 10.1016/j.jappgeo.2013.02.007.

Schulmeister, M. K.; Butler, J. J.; Healey, J. M.; Zheng, L.; Wysocki, D. A.; McCall, G. W. (2003): Direct- Push Electrical Conductivity Logging for High-Resolution Hydrostratigraphic Characterization. In: Groundwater Monitoring & Remediation 23 (3), S. 52–62. DOI: 10.1111/j.1745-6592.2003.tb00683.x.

Seeliger, M.; Pint, A.; Frenzel, P.; Weisenseel, P.; Erkul, E.; Wilken, D.; Wunderlich, T.; Başaran, S.; Bücherl, H.; Herbrecht, M.; Rabbel, W.; Schmidts, T.; Szemkus, N.; Brückner, H. (2018): Using a Multi- Proxy Approach to Detect and Date a Buried part of the Hellenistic City Wall of Ainos (NW Turkey). In: Geosciences 8 (10), S. 357. DOI: 10.3390/geosciences8100357.

Sherwood, S. C.; Kidder, T. R. (2011): The DaVincis of dirt. Geoarchaeological perspectives on Native American mound building in the Mississippi River basin. In: Journal of Anthropological Archaeology 30 (1), S. 69–87. DOI: 10.1016/j.jaa.2010.11.001.

Smedt, P. de; van Meirvenne, M.; Meerschman, E.; Saey, T.; Bats, M.; Court-Picon, M.; Reu, J. de; Zwertvaegher, A.; Antrop, M.; Bourgeois, J.; Maeyer, P. de; Finke, P. A.; Verniers, J.; Crombé, P. (2011): Reconstructing palaeochannel morphology with a mobile multicoil electromagnetic induction sensor. In: Geomorphology 130 (3-4), S. 136–141. DOI: 10.1016/j.geomorph.2011.03.009.

Squatriti, P. (2002): Digging Ditches in Early Medieval Europe. In: Past & Present 176 (1), S. 11–65. DOI: 10.1093/past/176.1.11.

Völlmer, J.; Zielhofer, C.; Hausmann, J.; Dietrich, P.; Werban, U.; Schmidt, J.; Werther, L.; Berg, S. (2018): Minimalinvasive Direct-push Erkundung in der Feuchtboden(geo)archäologie am Beispiel des Karlsgrabens (Fossa Carolina). In: Archäologisches Korrespondenzblatt 48 (4), S. 577–608.

3 3D-modelling of Charlemagne’s Summit Canal

Werther, L. (2017): Karlsgraben doch schiffbar? Aktuelles aus der Landesarchäologie. In: Archäologie in Deutschland (5), S. 41–42.

Werther, L.; Feiner, D. (2014): Der Karlsgraben im Fokus der Archäologie. In: Ettel, P.; Daim, F.; Berg- Hobohm, S.; Werther, L. and Zielhofer, C. (Hg.): Großbaustelle 793. Das Kanalprojekt Karls des Großen zwischen Rhein and Donau. Mainz: Verl. des Römisch-Germanischen Zentralmuseums (Mosaiksteine, 11), S. 33–40.

Werther, L.; Kröger, L.; Kirchner, A.; Zielhofer, C.; Leitholdt, E.; Schneider, M.; Linzen, S.; Berg- Hobohm, S.; Ettel, P. (2018): Fossata Magna - a Canal Contribution to Harbour Construction in the 1st Millenium AD. In: von Carnap-Bornheim, C.; Daim, F.; Ettel, P. and Warnke, U. (Hg.): Harbours as object of interdisciplinary research. Archaeology + history + geosciences. Mainz: Verl. des RGZM (RGZM Tagungen, 34), S. 355–372.

Zielhofer, C.; Leitholdt, E.; Werther, L.; Stele, A.; Bussmann, J.; Linzen, S.; Schneider, M.; Meyer, C.; Berg-Hobohm, S.; Ettel, P. (2014): Charlemagne’s summit canal. An early medieval hydro-engineering project for passing the Central European Watershed. In: PloS one 9 (9), e108194. DOI: 10.1371/journal.pone.0108194.

Zielhofer, C.; Rabbel, W.; Wunderlich, T.; Vött, A.; Berg, S. (2018a): Integrated geophysical and (geo)archaeological explorations in wetlands. In: Quaternary International 473, S. 1–2. DOI: 10.1016/j.quaint.2018.04.008.

Zielhofer, C.; Wellbrock, K.; al-Souliman, A. S.; Grafenstein, M. von; Schneider, B.; Fitzsimmons, K.; Stele, A.; Lauer, T.; Suchodoletz, H. von; Grottker, M.; Gebel, H. G. K. (2018b): Climate forcing and shifts in water management on the Northwest Arabian Peninsula (mid-Holocene Rasif wetlands, Saudi Arabia). In: Quaternary International 473, S. 120–140. DOI: 10.1016/j.quaint.2018.03.001.

4 Sediment budgeting of short-term backfilling processes

Chapter 4 - Sediment budgeting of short-term backfilling processes – the colluvial collapse of a Carolingian canal construction

Schmidt, J.; Werther, L. Rabiger-Völlmer, J.; Herzig, F.; Schneider, B.; Werban, U.; Dietrich, P.; Berg, S.; Linzen, S.; Ettel, P.; Zielhofer, C. (2020): Sediment budgeting of short-term backfilling processes - the colluvial collapse of a Carolingian canal construction. accepted. In: Earth Surf. Process. Landforms.

4 Sediment budgeting of short-term backfilling processes

Sediment budgeting of short-term backfilling processes - the erosional collapse of a Carolingian canal construction

Johannes Schmidt1*, Lukas Werther2, Johannes Rabiger-Völlmer1, Franz Herzig3, Birgit Schneider1, Ulrike Werban4, Peter Dietrich4, Stefanie Berg5, Sven Linzen6, Peter Ettel7, Christoph Zielhofer1

1Institute of Geography, Leipzig University, D-04103 Leipzig, Germany 2Department for Medieval Archaeology, University of Tübingen, D-72070 Tübingen, Germany 3Bavarian State Department for Cultural Heritage BLfD, D-86672 Thierhaupten, Germany 4Helmholtz Centre for Environmental Research UFZ, Department Monitoring and Exploration Technologies, D-04318 Leipzig, Germany 5Bavarian State Department of Cultural Heritage BLfD, D-80539 Munich, Germany 6Leibniz Institute of Photonic Technologies IPHT, D-07745 Jena, Germany 7Prehistory and Early History, Friedrich-Schiller University, D-07743 Jena, Germany

*corresponding author: [email protected]

4.1 Abstract

Sediment budgeting concepts serve as quantification tools to decipher the erosion and accumulation processes within a catchment and help to understand these relocation processes through time. While sediment budgets are widely used in geomorphological catchment-based studies, such quantification approaches are rarely applied in geoarchaeological studies. The case of Charlemagne’s summit canal (also known as Fossa Carolina) and its erosional collapse provides an example for which we can use this geomorphological concept and understand the abandonment of the Carolingian construction site. The Fossa Carolina is one of the largest hydro-engineering projects in Medieval Europe. It is situated in Southern Franconia (48.9876 N, 10.9267 E; Bavaria, Southern Germany) between the Altmühl and Swabian Rezat rivers. It should have bridged the Central European watershed and connected the Rhine-Main and Danube river systems. According to our dendrochronological analyses and historical sources, the excavation and construction of the Carolingian canal took place in 792 and 793 AD. Contemporary written sources describe an intense backfill of excavated sediment in autumn 793 AD. This short-term erosion event has been proposed as the principle reason for the collapse and abandonment of the hydro-engineering project. We use subsurface data (drillings, archaeological excavations and direct-push sensing) and geospatial data [a LiDAR Digital Terrain Model (DTM), a pre-modern DTM and a 3D-model of the Fossa Carolina] for the identification and sediment budgeting of the backfills. Dendrochronological findings and radiocarbon ages of macro-remains within the backfills give clear evidence for the erosional collapse of the canal project 78

4 Sediment budgeting of short-term backfilling processes

during or directly after the construction period. Moreover, our quantification approach allows the detection of the major sedimentary collapse zone. The exceedance of the manpower tipping point may has caused the abandonment of the entire construction site. The spatial distribution of the dendrochronological results indicates a north-to-south direction of the early medieval construction progress.

4.2 Introduction

Quantification of sedimentary processes is a hot topic in geomorphology and geoarchaeology (Brown et al. 2009; Hinderer 2012). One of the main challenges of sediment budgeting is to quantify the sediment storage and the amount of eroded sediment (Brown et al. 2009). The spatial scale of sediment budget studies can be large [e.g., the Rhine (Hoffmann et al. 2007) or Mississippi river catchments (Kesel et al. 1992)], medium [e.g., small sub-catchments (Weber and Pasternack 2017; Rascher et al. 2018)], or small [e.g., gully systems (Dotterweich et al. 2003) or hillslope processes (Smetanová et al. 2017; Bussmann et al. 2014)]. The temporal scales vary from orogenetic sediment fluxes of up to several millions of years (Hinderer and Einsele 2001; Bhattacharya et al. 2016) to short-term rainfall events (Chen et al. 2018). The terms and concepts of sediment budgeting are used in both pristine (Voiculescu et al. 2019) and strongly anthropogenic influenced (Gellis et al. 2017) catchments, or among them (Förster and Wunderlich 2009). Sediment budgets can be calculated by using different data measurement approaches. Recent and sub-recent sedimentary budgets can be investigated by instrumental data and monitoring approaches (Gellis et al. 2017; Griffiths and Topping 2017). Sedimentary archives like marine deposits (Qiao et al. 2017), lakes (Breuer et al. 2013; Zolitschka 1998), fluvial/alluvial sequences (Hoffmann et al. 2007), deltas (Kondolf et al. 2018; Erkens et al. 2006), colluvial and slope deposits (Förster and Wunderlich 2009) are used to calculate sediment budgets. However, sediment budgeting approaches are not typically applied in geoarchaeological research. Though there are studies that challenge sediment quantification in archaeological contexts (Pickett et al. 2016; Schmidt et al. 2019; Lacquement 2010; Sherwood and Kidder 2011), sediment budgets dealing with quantities of erosion and accumulation are sparse (Bork et al. 2003).

Our study comprises sedimentary and dendroarchaeological data using a modelling approach in the context of a very well-dated archaeological site in Southern Germany. The object of our study is the Fossa Carolina, an Early Medieval canal that should have bridged the Central European Watershed (Leitholdt et al. 2012). Recently, extensive research has been done on the sediments within the canal trench (Völlmer et al. 2018; Zielhofer et al. 2014; Kirchner et al. 2018). Stratigraphic data show initial erosion of the dams, and nearly all drillings show initial sedimentary backfills at the trench bottom (Zielhofer et al. 2014). To our knowledge, no published work has focused on the backfills or their value for understanding canal construction and site abandonment. Furthermore, to assess the linkage with wider landscape dynamics, we will compare the backfilling collapse with already published phases of soil erosion elsewhere (Dotterweich 2008; Dreibrodt et al. 2010) and climate reconstructions (Muigg et al. 2020). Therefore, in this study, we

4 Sediment budgeting of short-term backfilling processes

present stratigraphic and geochemical results to identify these backfills. Three archaeological excavations in the Northern and North-Eastern Sections have revealed the status of construction at the moment of abandonment (Werther and Feiner 2014; Werther et al. 2015) and several oak timbers were recovered (Werther et al. 2020). Furthermore, we compile the radiocarbon ages of the backfills, and we determine the spatial distribution of dendrochronological data from the Early Medieval canal construction. Following the modelling approach of Schmidt et al. (2019), we use the spatial data of the backfills and set up a quantitative model of the backfill sediment budget and, subsequently, of the collapse of the Early Medieval canal construction.

The main objectives of this study were:

the identification of the backfill sediments and its spatial distribution; the development of a sediment budgeting model for the short-term backfilling processes; the creation of a spatiotemporal model of the canal construction and abandonment progress.

4.3 Fossa Carolina and its geographical setting

The Central European watershed divides the Rhine-Main catchment and the Danube catchment. At the Fossa Carolina, it divides the Altmühl catchment (Danube drainage system) from the Swabian Rezat catchment (Rhine-Main drainage system). According to historical sources, the Fossa Carolina was built in 792/793 AD on a valley watershed (Fig. 4.1) and according to the revised version of the Royal Frankish Annals its termination ought to have been caused, among other things, by heavy rainfalls, which induced intense erosional processes of the construction earthwork (Werther et al. 2020; Hack 2014; Nelson 2015). “For because of the continual rain and the bogginess of the land which was in the nature of things completely waterlogged, the work that was being done could not hold firm, given the excessive wetness, and as much of the earth as was excavated by the diggers during the day slid back again and sank into the soil during the night” (Translation of the revised version to the year 793, after Nelson 2015).

The surrounding escarpment landscape is built up by Middle to Upper Jurassic rocks (mudstones, sandstones, marl, and limestones). The valley and the valley watershed are built up by sandy to loamy, fluvial sediments of Pleistocene age with a slight Loess cover, especially at the lower slopes (Schmidt- Kaler 1993; Zielhofer and Kirchner 2014). The sediments are almost free of organic material and contain mainly fine sands (Leitholdt et al. 2014; Zielhofer et al. 2014). The current level of the Swabian Rezat is at 413.5 m a.s.l. and of the Altmühl River is at 408.3 m a.s.l. Due to the difference of both river levels and to excavate as little material as possible, the canal was set up as a summit canal (Zielhofer et al. 2014). By now, no artificial inflow was detected yet, and the hydro-engineering approach remains unclear (Rabiger- Völlmer et al. 2020). According Kirchner et al. (2018) the canal construction was never finished, because canal structures are missing in the Altmühl floodplain.

4 Sediment budgeting of short-term backfilling processes

Furthermore, the canal course was documented by geophysical prospections [Superconducting Quantum Interference Device (SQUID) and Fluxgate magnetic, seismic profiles and geoelectric survey), aerial photography, and analysis of LiDAR (Light Detection and Ranging) data (Linzen et al. 2017; Köhn et al. 2019; Zielhofer et al. 2014; Stele et al. 2019). The canal course can be divided into five sections (Fig. 4.2) based on their specific geomorphological characteristics. The canal has a length of c. 2.9 km, and the proof was done by drillings with subsequent sediment analyses (Leitholdt et al. 2014; Zielhofer et al. 2014; Kirchner et al. 2018), direct push sensing (Hausmann et al. 2018; Völlmer et al. 2018) and archaeological excavations (Werther et al. 2015). The canal course has an apparent S-shape (Fig. 4.2), which is the best alignment in relation to the minimal excavation workload and avoidance of with unfavourable site conditions, such as wet areas and unstable organic-rich sediments (Schmidt et al. 2018). Therefore, the S- shape (Fig. 4.3a and b) is a result of the impressive knowledge of the Carolingian constructors. In total, the Carolingian workers excavated almost 300,000 m³ of material and raised two surrounding dams (Schmidt et al. 2019).

4 Sediment budgeting of short-term backfilling processes

The canal trench has now been refilled and the canal is only slightly visible in the Northern and North- Eastern Sections. However, the canal is clearly visible in the Central and West-East Sections (due to the remains of the dams taller than 10 m height), but the canal trench bottom has been raised by sediment accumulation and formation of organic-rich peat layers up to 11 m above the Carolingian trench bottom level. Furthermore, the dams are partly eroded, and only c. 30% of the dam material is still remaining (Schmidt et al. 2019). 82

4 Sediment budgeting of short-term backfilling processes

4.4 Material and Methods

The chapter will give an overview of the used data, the accuracy and the modelling and budgeting approach.

4.4.1 Data acquisition

Data are essential for modelling approaches. In the following subchapter, we will therefore describe our data acquisition and data processing.

4.4.1.1 Geodata

The largest part of the input data for the present study are geodata, processed within a GIS environment (SAGA GIS; Conrad et al. 2015).

4.4.1.1.1 LiDAR Digital Terrain Model

We used the 1 × 1 m spatial resolved Digital Terrain Model (DTM), which was provided by the Bavarian Land Surveying Office (Bavarian land surveying Office 2018). It was created using high-resolution airborne laser scanning data. The LiDAR DTM is a basic part of the sediment budgeting procedure.

4.4.1.1.2 Pre-modern Digital Terrain Model

The present landscape is often disturbed by anthropogenic structures, such as roads, buildings, and railway tracks. We use a pre-modern DTM of the study area, where nearly all anthropogenic structures are filtered. Schmidt et al. (2018) removed all grid cells that are affected by anthropogenic disturbance and interpolated the residual cells to create a smooth pre-modern terrain of the study area around the Fossa Carolina. This pre-modern DTM is based on the LiDAR DTM mentioned above, and has a spatial resolution of 1 × 1 m.

4.4.1.1.3 3D-Model of the Fossa Carolina

Schmidt et al. (2019) developed a quantitative approach to creating a 3D-Model of the Fossa Carolina in its maximum state of construction. The modelling procedure was based on five reference-cross sections (archaeological excavations and direct push sensing transects) and 39 drillings. By data integration, these authors created a dense network of depth information along the canal. Due to the high resolution of the reference cross-sections (50 cm up to 12.5 cm), the resulting “backfill model” has a spatial resolution of

4 Sediment budgeting of short-term backfilling processes

0.5 × 0.5 m. The 3D-Model shows the maximum excavation depth of the Fossa Carolina. In this study, we use this information to calculate the volume of the backfill sediments and create a sediment budget.

4.4.1.2 Drillings

The basic stratigraphical data for the modelling of the initial backfills are the results taken from several cores (Fig. 4.3c). The drillings were conducted during the last decade in several field campaigns. We used an Atlas Copco Cobra Pro hammer with a 60 mm open corer. For this study, we used 39 drilling cores that are situated within the trench fills (Fig. 4.2). Subsequently, we sampled and analysed selected cores (Leitholdt et al. 2014; Zielhofer et al. 2014; Kirchner et al. 2018).

4.4.1.3 Sediment geochemistry

We performed grain size analysis by X-ray granulometry using a SediGraph III 5120 (Micrometrics). To remove the organic matter content, we dissolved the sample in 50 ml 35 % hydrogen peroxide, left it overnight, and heated the sample the next day. We added the oxidising agent at a high temperature until all organic matter was removed. Next, we dispersed the sample using 10 ml 0.4 N sodium pyrophosphate solution and ultrasonic treatment for 45 min. By wet sieving with a 63 µm sieve, the samples were separated into the sand fraction and the silt / clay fraction (<63 µm). The silt / clay fraction was analysed by means of X-ray absorption. To analyse the sand fraction, we used the dry-sieving technique. Furthermore, we calculated the organic content by measuring the total carbon content (TC) using a CNS analyser Vario EL cube (Elementar). Additionally, we determined the total inorganic carbon content (TIC) using the calcimeter technique (Scheibler method, Eijkelkamp). By subtraction, we calculated the total organic carbon content (TOC).

4.4.1.4 Radiocarbon dating

Chronological information for the backfill sediments were obtained using wood remains or charcoal material (Tab. 4.1, stratigraphic positions of radiocarbon age samples can be found in Supplemental Fig. S1). 14C dating was processed using accelerator mass spectrometry (AMS). We use published radiocarbon ages from Leitholdt et al. (2012) and Zielhofer et al. (2014). However, we newly calibrated all radiocarbon ages using CALIB 7.1 (Stuiver et al. 2019) with Intcal13 calibration curve (Reimer et al. 2013).

4 Sediment budgeting of short-term backfilling processes

No. Material Location Profile/Core Depth (level in m Lab-No. 14C-Age BP Intcal13-Calibration Intcal13-Calibration d13C (‰) Reference ID asl) (1sigma) (1sigma) (2sigma)

1 charcoal Central Section A/A26 410.9 KIA36404 1267 +- 27 689-727 666-777 -23.67 +- 0.09 Leitholdt et al. 2012 737-753 793-801 757-768 844-854 2 charcoal Central Section A/A35 409.4 KIA36406 1269 +- 27 689-725 665-776 -24.50 +- 0.14 Leitholdt et al. 2012 738-768 793-800 848-851 3 wood Central Section K/K3 413.5 SUERC-42075 1253 +- 26 691-749 675-778 -29.40 Zielhofer et al. 2014 761-772 791-828 838-864 4 wood Central Section K/K4 413.3 SUERC-42076 1163 +- 26 778-791 774-902 -27.90 Zielhofer et al. 2014 804-842 919-963 860-896 927-941 5 wood Central Section K/K5 412.9 SUERC-42140 1170 +- 34 777-793 770-907 -28.70 Zielhofer et al. 2014 801-893 914-968 6 wood Central Section L/L3 412.6 SUERC-42143 1110 +- 37 894-931 778-790 -27.00 Zielhofer et al. 2014 937-980 809-813 826-841 863-1017 7 wood Central Section M/M2 412.9 SUERC-42148 1223 +- 37 721-740 688-888 -27.00 Zielhofer et al. 2014 767-779 788-873 8 wood West-East Section Q/Q1 412.0 SUERC-44082 1271 +- 29 687-725 663-777 -26.40 Zielhofer et al. 2014 738-768 793-802 844-855 9 wood West-East Section S/S2 412.1 MAMS 17461 1312 +- 17 664-689 660-713 -22.50 Zielhofer et al. 2014 752-760 744-765

4 Sediment budgeting of short-term backfilling processes

4.4.1.5 Direct push sensing

Direct push sensing is a minimally invasive and depth-accurate technique for in-situ characterisation of sediment stratigraphies (Dietrich and Leven 2009; Leven et al. 2011). A caterpillar pushes steel rods with a small diameter into the ground using various probes. We used the colour logging tool (SCOST™, Dakota Technologies, Fargo, USA) to differentiate the sediment stratigraphy of the canal trench (Hausmann et al. 2018; Völlmer et al. 2018). By an appropriate measuring pace (2 cm/s) and an integration time of 300 ms, the direct push sensing results in a vertical resolution of 3 values per 2 cm. This high vertical resolution is accompanied by a horizontal spacing of up to 12.5 cm. To obtain reference cross-sections for the sediment budgeting process, we used two direct push sensing transects with 105 direct push colour logs (Fig. 4.2).

4.4.1.6 Archaeological excavations

We conducted three archaeological excavations in 2013 and 2016 in the Northern and North-Eastern Sections (Fig. 4.2), cutting the canal rectangular to the embankments (Werther and Feiner 2014; Werther et al. 2015; Werther 2017; Werther et al. 2020). The localisation of the excavation trenches was based on geoarchaeological and geophysical surveys (Zielhofer et al. 2014; Köhn et al. 2019; Linzen and Schneider 2014). We recovered different kinds of timber and wood waste (Tab. 4.2). The timbers are situated alongside the trench edges (Fig. 4.3d and e) and stabilised the trench revetments. In this study, we use for the first time high-resolution canal trench geometries of all three trenches as cross-section reference geometries for the sediment budget calculation. Furthermore, the archaeological excavations give the chronostratigraphic context of the recovered timbers.

Table 4.2: Dendrochronological results of recovered timber (oak) and wood fragments from three archaeological excavation trenches. No. Location Excavation year Find label Timber type Felling date 1 Trench 1 2013 77 Plank 793, season unspecified 2 Trench 1 2013 81 Pile 793, c. August - September 3 Trench 1 2013 82 Pile 793, season unspecified 4 Trench 1 2013 85 Pile 793, c. August - September 5 Trench 1 2013 89 Pile 793, c. August - September 6 Trench 1 2013 92 Pile 793, c. August - September 7 Trench 1 2013 93 Pile 793, c. August - September 8 Trench 1 2013 94 Pile 793, c. August - September 9 Trench 2 2016 11 Pile 793, c. April - May 10 Trench 2 2016 22 Pile 793, c. April - May 11 Trench 2 2016 21/25 Splinter 793, c. April - May 12 Trench 2 2016 25 Splinter 793, season unspecified 13 Trench 2 2016 25 Splinter 793, c. April - May 14 Trench 2 2016 25 Splinter 792, season unspecified

4 Sediment budgeting of short-term backfilling processes

Table 4.2: continued.

No. Location Excavation year Find label Timber type Felling date 15 Trench 2 2016 25 Splinter 793, season unspecified 16 Trench 2 2016 28 Plank 793, c. April - May 17 Trench 2 2016 28/29 Splinter 792, season unspecified 18 Trench 2 2016 29 Splinter 793, season unspecified 19 Trench 2 2016 32 Pile 793, c. April - May 20 Trench 2 2016 33 Pile 793, c. April - May 21 Trench 2 2016 35 Pile 793, c. April - May 22 Trench 2 2016 37 Pile 793, c. April - May 23 Trench 2 2016 38 Pile 793, c. April - May 24 Trench 2 2016 39 Pile 793, c. April - May 25 Trench 2 2016 40 Pile 793, c. April - May 26 Trench 2 2016 41 Pile 793, c. April - May 27 Trench 2 2016 56 Plank 792, season unspecified 28 Trench 2 2016 80 Splinter 793, c. April - May 29 Trench 2 2016 82 Splinter 793, c. April - May 30 Trench 2 2016 89 Splinter 793, season unspecified 31 Trench 2 2016 88/91 Splinter 793, c. April - May 32 Trench 2 2016 200 Forked wood 793, season unspecified 33 Trench 2 2016 203 Plank 792, c. October - 793, c. March 34 Trench 3 2016 122 Splinter 792, season unspecified 35 Trench 3 2016 126 Wood fragment 792, season unspecified 36 Trench 3 2016 126 Pile 792, season unspecified 37 Trench 3 2016 127 Pile 793, c. April - May 38 Trench 3 2016 128 Pile 792, c. May - June 39 Trench 3 2016 134 Pile 792, c. October - 793, c. March 40 Trench 3 2016 135 Plank 792, c. October - 793, c. March 41 Trench 3 2016 140 Pile 793, c. April - May 42 Trench 3 2016 141 Pile 792, c. October - 793, c. March 43 Trench 3 2016 142 Pile 792, c. October - 793, c. March 44 Trench 3 2016 143 Pile 792, c. October - 793, c. March

4.4.1.7 Dendrochronological analysis

We have used a large group of timbers and wood waste to tackle chronological questions of the canal construction. In total, 44 samples offer a reliable basis for chronological analysis (Tab. 4.2). We used 30 timbers and 14 samples of wood waste with preserved terminal tree rings. Naturally, a tree forms a tree- ring every year. Tree-ring growth starts in spring and ends in autumn. Within this vegetation period (c. April to September), the tree ring grows constantly. Tree growth ends with the felling, and the terminal ring dates the felling date (Haneca et al. 2009). Due to the seasonal information of the felling date, Werther et al. (2020) could characterise the dendrochronological dates and discuss the construction progress in high chronological resolution. Besides the chronological information, the timbers might have marks of decay (unless they were stored in anaerobic conditions, which stops further decay) (Schweingruber 1988). Technically, timber and wood waste were carefully cleaned and prepared to preserve all processing traces. 88

4 Sediment budgeting of short-term backfilling processes

Subsequently, tree ring widths were measured with an accuracy of 1/100 mm using a stereoscopic microscope (Herzig 2018).

4.4.2 Modelling and sediment budgeting

The first challenge is the integration of all datasets to create a dense spatial network of backfill information. According to Schmidt et al. (2019), our modelling approach combines the drilling results with the high- resolution cross-section reference geometries (three archaeological excavation trenches, two direct push sensing transects). For the first time, we interpolate the upper and lower limit of the backfills derived from the drillings according to the respective canal geometry. To create a dense network of depth information, we add further synthetic transects along the canal course with an equidistant spacing of c. 50 m. We derived the backfill depth information from the neighbouring drillings, archaeological excavation, or direct push sensing transects. We also use the respective cross-section reference geometry to interpolate the depth information on the transect. Finally, we are able to spatially interpolate all transects to generate a 3D-model of the backfill top level. Similar to the 3D-model with the maximum excavation depth, due to the high- resolution reference cross-sections, the raster layer of the backfill sediment storage has a spatial resolution of 0.5 × 0.5 m.

Schmidt et al. (2019) recently produced a 3D-model of the backfill lower level, and estimated the maximum excavation depth with the same modelling approach. Quantitatively, we applied the sediment storage of the backfills through the subtraction of both 3D-models. Thus, we generate a map of the spatial distribution of the backfill thickness for the entire canal. Additional separations of the backfill volumes by the main sections of the canal helps to understand the spatial distribution of the short-term erosional process. The sediment budget is a complied analysis, of the information on maximum excavation volume, the residual dam volume, and the volume of the backfill sediments. Separated by the canal sections, we can derive spatially differentiated ratios of dam erosion and backfill accumulation.

4.5 Results

4.5.1 Backfill sediment identification

We identified the trench bottom (lower limit of the backfills) in the drillings mainly macroscopically, because the underlying Pleistocene valley fills (Schmidt-Kaler 1993) are almost free of organic content. However, the sedimentary material is the same in both categories. The grain-size distributions of both facies are similar (e.g., the sand content; Fig. 4.4a). For a detailed sedimentary description, see Zielhofer et al. (2014). The TOC content is raised in the backfill sediments compared to the “sterile”, sandy to loamy valley fills (Fig. 4.4b). The identification of the lower limit in the archaeological excavation trenches was

4 Sediment budgeting of short-term backfilling processes

performed macroscopically while seeing the total cross-section with the underlying “natural” material and the timber. The backfills were stratigraphically identified. The high-resolution direct push sensing transects have the advantage of also presenting the total cross-section of the canal, including the trench bottom and the revetments, but with fewer costs and effort compared to an archaeological excavation. The upper limit of the backfills is characterised by autochthone organic-rich sediments, mainly peat and sapropel. Therefore, the sedimentary contrast between the backfills and the overlaying organic facies is very good, and detectable macroscopically in the drillings and archaeological excavations, as well as in the direct push sensing transects.

4.5.2. Radiocarbon results

The radiocarbon results of the backfill sediment show consistent ages (Tab. 4.1, stratigraphic positions of radiocarbon samples can be found in Supplemental Fig. S1). All samples are macro-remains (charcoal and wood), which show, at least within their 2-sigma range ages that fit in the time of the Fossa Carolina construction 792/793 AD (Fig. 4.5). Within their 1-sigma range, the samples date to the construction time or before. Only sample “S2” shows a slightly older age that does not fit in the construction time, neither within the 1-sigma nor within the 2-sigma range.

4 Sediment budgeting of short-term backfilling processes

4.5.3 Dendrochronological results

All 44 samples, which have the terminal tree ring preserved, date the tree cut-off to 792 or 793 AD (Tab. 4.2). All timber in trench 1 date to 793 AD, more specifically to summer/autumn 793 AD. The results from trench 2 show mainly felling dates in spring 793 AD. In trench 3, the northernmost archaeological excavation trench, the results reveal felling dates from summer 792 AD or between the growing seasons of 792 AD and 793 AD. Most of the piles have been found in situ, rammed into the ground as bank revetments along the canal course. Furthermore, some planks have been found in a semi-finished condition, together with wood waste connected to the timbering on site. Most of the wood splinters belong to the final trimming of the upper parts of the piles after ramming them into the ground.

4.5.4 Backfill sediment storage

The sediment-storage procedure results in a map of backfill thickness covering the entire canal (Fig. 4.6, a detailed view of backfill distribution is shown in Supplemental Fig. S2; detailed information about the backfill thicknesses of the drillings are summarised in Supplemental Tab. S1). The absolute volume (c. 41,600 m³) of the backfill sediments is not equally distributed along the canal. The largest proportion is localised in the Central and West-East sections. Surprisingly, the highest thickness does not correspond to the Central European watershed and, therefore, the deepest depth of excavation. It is situated c. 150 m to the north.

4 Sediment budgeting of short-term backfilling processes

Figure 4.6: Spatial sediment budgeting model of the backfill thickness of the Carolingian canal Fossa Carolina. The colour ramp of the backfill volume is displayed with a gamma stretch of 2.

4.5.5 Backfill sediment budget

The catchment of the backfill sediments is defined by the surrounding dams created by the Carolingian constructors. Hence, the sediment budget is the comparison of the backfill volumes with the maximum excavation volume and the residual dam volume. Fifteen percent of the maximum excavation volume was eroded from the dams and accumulated in the canal trench (Tab. 4.3). In contrast, only c. 40% of the excavated material is still stored in the residual dams. Interestingly, the proportion of backfills to the maximum excavation depth over the different canal sections is similar. The ratio ranges from 12 to 19%. Surprisingly, we found the highest ratio not in the Central Section, but rather in the North-Eastern Section, where the lowest excavation volume occurs. In contrast, we found the lowest ratio in the West-East Section. The ratio between the dams and the maximum excavation volumes is lowermost in this section. Nevertheless, the total amounts of the backfill volume follow the maximum excavation volume.

4 Sediment budgeting of short-term backfilling processes

Maximu Minimum (Organic Initial Ratio backfill/ Dams Ratio Dams/ Length ma [m³] fills)b [m³] backfills [m³] maximum [%] [m³] Maximum [%] [m] Total 285455 243820 41635 15 119681 42 2829

Central 160815 136379 24436 15 85478 53 803 Section WE Section 84364 74573 9791 12 19268 23 494 Northern 26188 21409 4779 18 13838 53 738 Section North-Eastern 14088 11458 2630 19 5988 42 794 Section

4.6 Discussion

4.6.1 Sediment budget approach

Our sediment budgeting approach includes the volume information of the maximum Carolingian excavation, present remnants of the dams, and the initial backfill sediments. The quality and accuracy of the presented sediment storages on the local scale are excellent, due to plenty of ground truth data and a reproducible modelling approach. In contrast to large-scale and catchment-based sediment budget studies (Hoffmann et al. 2007), we could establish a dense network of subsurface data, but also other methodologies for the sediment storage quantification exist. Kesel (1989) used historical data to decipher the flooded areas and calculated them with mean sediment densities to estimate the sediment storage of the Mississippi River floodplain. Some studies use geostatistical interpolation of drilling results (Bussmann et al. 2014; Rommens et al. 2005), geometric forms of sediment bodies (derived from drillings; Suchodoletz et al. 2009), geophysical subsurface data (e.g., Guillocheau et al. 2012), or isopach maps (for submarine fan or deltas; Carvajal et al. 2009) to decipher the sediment storage.

Furthermore, the procedure of using synthetic transects with interpolated depth information, resulted in a precise canal geometry along the entire canal course. In contrast to studies in large-scale dimension, we derived a spatially differentiated sediment storage. In addition, we work on a small temporal scale with a short-term erosion event. Historical sources describe heavy precipitation in autumn 793 AD, which caused the erosion of excavated material (Hack 2014; Werther et al. 2020). Some studies on sediment budgets struggle with the dating and methodological errors (e.g., reworked material, broad 14C range due to radiocarbon plateau, insufficient bleaching for luminescence dating) of their archives (Bussmann et al. 2014; Brown et al. 2009). The dendrochronological results of our study give a clear interval of the canal construction and pre-dates the accumulation of the backfills. Furthermore, the concise radiocarbon data of the backfills reveal a specific geomorphological event that happened directly after the construction site abandonment during the construction.

4 Sediment budgeting of short-term backfilling processes

Furthermore, the biggest challenge of sediment budget studies is the estimation of sediment output. In colluvial systems (Bussmann et al. 2014), fluvial systems (Förster and Wunderlich 2009), and in deltas (Erkens et al. 2006), there is an unknown proportion of sediment loss. Even though there are models available that estimate the loss, an exact estimation is crucial (Bhattacharya et al. 2016). Only studiesthat work on continuous archives without sediment loss in endorheic geomorphological positions (sediment traps) can produce reliable information about the sediment storage (e.g., Suchodoletz et al. 2009; Zolitschka 1998; Breuer et al. 2013). The geomorphological system of the Fossa Carolina, with its surrounding dams and excavated trench, is endorheic. Therefore, it acts solely as a sediment trap and we can assume no sediment loss of the initial backfills. Moreover, the surrounding dams form small and explicit catchments for sediment flux.

4.6.2 Spatial Distribution of backfill sediments

Longitudinal sections of the Fossa Carolina have been published (Leitholdt et al. 2012; Leitholdt et al. 2014; Zielhofer et al. 2014; Kirchner et al. 2018), but now we are able to add two additional datasets, which attend to the erosional processes during construction site abandonment (Fig. 4.7). First, the black line, that shows the maximum Carolingian excavation depth, which is the result of Schmidt et al. (2019). Additionally, we added the orange line, which indicates the upper limit of the backfill sediments. The space between both lines reflects the thickness of these backfill sediments. The backfill sediment thickness along the entire canal varies considerably.

The largest amounts of backfills do not correspond with the maximum excavation depth in the middle part of the Central Section, as previously expected. The largest amount can be found c. 150 m north of the Central European Watershed. There is a missing percentage of material in the calculation. For example, 43% of material are neither stored in the dams (c. 42%) nor backfilled (c. 15%) in the canal trench. The outer slopes of the dams eroded into the surrounding areas, and only the inner slopes of the dams led to the transport of material in the canal trench. One could assume the same amount of erosion to the surrounding landscape as to the canal trench. Therefore, we have a distinct “loss” of material to the sediment cascade system of the surrounding landscape. Furthermore, during the Late Middle Ages and Early Modern time, lots of sediment accumulated in the upper layers of the canal trench fillings (Zielhofer et al. 2014). However, there is also a strong modern anthropogenic contribution to the negative sediment balance. During railway construction, huge amounts of material have been removed from the dams in the Central section (see Schmidt et al. 2018).

4 Sediment budgeting of short-term backfilling processes

4.6.3 Canal construction progress, abandonment and collapse

Beck (1911) discussed various possibilities for the abandonment of the construction; military conflicts, subsequent food shortages, drainage problems during construction, and general technical problems. Due to the lack of subsurface data, Beck (1911) could not tackle geomorphological causality. Birzer (1958) and Koch (1993) assume the possibility of eroded dams and the redeposition of dam material within the canal trench. Nevertheless, Koch (1996) conducted drillings, but mentions the difficulties of the detection of redeposited parent material, especially without geochemical analysis or numerical dates. Our study deals with the main abandonment reason discussed in the literature. The erosional collapse of the surrounding dams. The unambiguous radiocarbon results of the backfill sediments and the stratigraphic context, directly on top of the trench bottom (precisely dated by dendrochronology), reveal the initial erosional deposition of dam material. For this study, we subdivide the construction process into two major parts: (i) the Carolingian excavation and installation of timber (the partial completion of the canal construction); and (ii) the erosional collapse and site abandonment.

(i) After careful consideration of the surrounding landscape, the Carolingian constructors decided to reduce the excavation volume by a topography-based s-shape canal course (Schmidt et al. 2018). Within this

4 Sediment budgeting of short-term backfilling processes

objective, they build the canal as a summit canal (Zielhofer et al. 2014). The final step of the earthwork was the installation of timber along the edges of the canal, to enhance the embankment stability (Werther et al. 2015; Werther 2016). Werther et al. (2020) show that all timber fellings date to 792 and 793 AD. In general, the recovered timbers show almost no storage signs, such as a fungal attack. Hence, we interpret the felling date as also the installation date. The spatial comparison of dendrochronological results from three archaeological excavations shows a distinct differentiation of the felling dates (Fig. 4.8). The northernmost excavation (trench 3) reveals the oldest timber and the excavations more to the south (trench 2 and 1) reveal explicit younger ages. These results indicate construction progress from the North to South of the canal, in the direction from the tributary to the watershed. Concerning the artificial drainage of the Carolingian construction site, dry conditions in the construction pit were obligatory and, with the North-to South approach, feasible. Kirchner et al. (2018) disproved the southern part of the canal in the Altmühl floodplain, and subsequently, the construction was never finished in that section. Therefore, our construction progress direction is only reliable for the Northern sections of the canal.

(ii) The stratigraphy of nearly all drillings shows that the backfills cover the parent material. If the canal construction had been finished and no sudden collapse occurred, we would have found organic-rich sediments (e.g. peat, sapropel, finely layered sediments; Leitholdt et al. 2014). The lack of evidence for Carolingian stable (open water) conditions supports the research hypothesis of an initial erosional collapse of the canal construction. Hence, the precise dendrochronological results pre-date the erosional collapse. Furthermore, the sedimentary conditions (Fig. 4.4) show that the backfills originate from the parent material from the surrounding dams. The radiocarbon results of macro-remains within the backfills show Carolingian ages (Fig. 4.5). Only sample “S2”, which has a slightly older age than the time of the Fossa Carolina construction, was specified as Quercus sp. (Zielhofer et al. 2014). This age can be explained by reworking of older natural wood, as part of wood waste from Carolingian wood-working or even as part of a shoring system dating the inner part of an oak tree with a so-called old-wood-effect. The Carolingian constructors usually used oak for timber (Werther 2016). The scientific evidence of the sudden, colluvial collapse corresponds with the testimony of the revised version of the Frankish Annals, which describes the strong rainfalls in autumn 793 AD, which, in the afternoon, washed back the material, that the workers have excavated during the day (Werther et al. 2020; Nelson 2015). The results of the sediment budgeting show that 15% of the total excavated material was eroded (Tab. 4.3). The relative amounts of backfills are equally distributed along the canal, indicating that the collapse occurred on the whole canal structure. The highest absolute amounts of backfills can be found in the Central and West-East Section due to the maximum excavation depth. It is likely that the large amounts of backfills in these sections (c. 24,000 m³ and c. 10,000 m³) have led to an enormous additional workload for the workers that should not be underestimated. One can argue that, in total, c. 40,000 m³ of backfills has acted as the tipping point and resulted in the abandonment of the construction, as contemporary written sources report (Hack 2014; Nelson 2015; Werther et al. 2020).

4 Sediment budgeting of short-term backfilling processes

Figure 4.8: Spatial distribution of dendrochronological derived felling dates of timber from the three archaeological excavation along the Fossa Carolina.

4.6.4 Large-scale control or local feature?

The large amounts of sedimentary backfills accumulated in such a short time suggest a heavy precipitation event or another process that might have affected morphodynamics on a larger scale. Dotterweich (2008) and Dreibrodt et al. (2010) gave comprehensive overviews of historical soil erosion captured in slope, 97

4 Sediment budgeting of short-term backfilling processes

alluvial, or lake deposits in Central Europe. Both compilations did not identify distinct single erosion events for the Carolingian period. Even though the end of the 1st millenium AD is characterised by an increase of soil erosion (Bork et al. 1998; Dreibrodt et al. 2010), and a further increase in the High Middle Ages, the erosion rates and documented events are sparse. In contrast, the Merovingian land seems not to have caused enhanced soil erosion. Schreg (2014) argues that this might be due to a limited level of landscape cultivation and forest clearance. However, as our case study area was not affected by larger-scale to “land-use changes” during the Carolingian period (the dams have a bare earth surface regardless of the land-use in the catchment), which is discussed in the geomorphological community as a major trigger on the landscape activation (Dotterweich 2008; Dreibrodt et al. 2010; Kalis et al. 2003), one might suggest a major precipitation event causing the erosional collapse of the Fossa Carolina.

Büntgen et al. (2011) propose a European-scale hydroclimate reconstruction based on tree-rings. These authors reveal a slight precipitation increase after the end of the migration period and the beginning of the Early Middle Ages. However, the large spatial-scale approach is not sufficient to cope with the Fossa Carolina collapse. A reconstruction of precipitation using tree-ring based modelling approaches, among others, on the recovered timbers from Fossa Carolina was done by Muigg et al. (2020).Therefore, the reconstruction covers the study area perfectly, but oak tree ring growth is sensitive to spring to summer soil moisture (Pechtl and Land 2019), and the data does not cover autumn precipitation (Fossa Carolina collapse is assumed to have happened in the autumn of 793 AD). However, Muigg et al. (2020) have not been able to identify a wet phase in the years after 793 AD when the construction was abandoned (at least not during April-August), which is in accordance with the Old World Drought Atlas (Cook et al. 2015). Also, Land et al. (2019) reveal no distinct wet phases at the End of the 8th century and the beginning of the 9th century AD. In contrast, longer and pronounced wet phases (Muigg et al. 2020; Land et al. 2019; Büntgen et al. 2011) in the High Medieval period significantly affected the hydrographic situation of the Fossa Carolina and led to the development of peats in the Fossa Carolina canal trench (Leitholdt et al. 2014).

The excavated sediments stored in the dams during Carolingian canal construction are highly sensitive to erosion, due to the reworking and subsequent disturbance of sediment stabilising properties (e.g. aggregates, cohesivity; see Jewell 1963). Therefore, short-term precipitation events might have caused fast and significant erosion of the dams, without affecting the soils at the larger landscape level. Also, Bork et al. (2003), Leopold et al. (2011) and Lisá et al. (2015) report fast redistribution events of excavated sediments, stored in dams and ditches. Jewell (1963) discusses studies, where archaeological dams along ditches eroded due to steep slopes (angle of repose), bare earth surfaces, and subsequent erosion susceptibility to rainfall events. Therefore, backfill sediments must be a common feature in many geoarchaeological sites with ditches and ditch-like structures.

4 Sediment budgeting of short-term backfilling processes

4.7 Conclusion

Our study integrates various subsurface data (drillings, archaeological excavations, direct push sensing) and geospatial data (LiDAR DTM, pre-modern DTM, 3D-model of the maximum excavation depth of the canal) to create the spatial distribution of backfill sediments at the Fossa Carolina in South Germany. The multi-method data integration for calculating the spatial distribution model of the backfills was performed at high-resolution. We identified these backfills stratigraphically and via geochemical analysis (mainly, TOC and grain size distribution). Furthermore, radiocarbon ages of macro-remains within the backfill reveal a clear Carolingian age. The chronological framework was supported by precise dendrochronological results from excavated timbers of the canal construction. The oak timber should stabilise the canal trench embankments and mark the completion of the specific section. The dendrochronological analysis proved the canal construction was done during 792 - 793 AD.

Our modelling approach resulted in the spatial distribution of the backfill amounts. We have been able to show that the erosional redeposition occurred along the whole canal trench. Furthermore, the modelling reveals the quantity of the initially eroded material, that is now stored as backfills in the canal trench. With the information of the total excavated material and the sediment volume that remains in the present dams, we established a sediment budgeting approach. In total, 15% of the excavated material was redeposited as backfill in the canal trench. This large amount (c. 40,000 m³) could have acted as a tipping point to abandon the construction site, as contemporary written sources suggest. The major amounts of backfills in the Central Section may have exceeded the Carolingian manpower. Additionally, this tipping point could have led to the abandonment and non-beginning of the southernmost part of the canal (Altmühl floodplain Section). However, there is no evidence for large-scale climatic control of the erosional collapse, but rather a high erosion sensitivity of the dams. Therefore, local precipitation events might have caused the backfilling processes. Additionally, we compiled the dendrochronological data with the excavation position. For the first time, we could reveal the Carolingian construction progress of the Fossa Carolina; the construction started in the northernmost Section towards the Central European Watershed.

4.8 References

Bavarian Land Surveying Office (2018): Geländemodell. https://www.ldbv.bayern.de/produkte/ 3dprodukte/gelaende.html, accessed 11.04.2018.

Beck, F. (1911): Der Karlsgraben. Eine historische, topographische and kritische Abhandlung. Nürnberg: Verlag der Friedrich Kornschen Buchhandlung.

Bhattacharya, J. P.; Copeland, P.; Lawton, T. F.; Holbrook, J. (2016): Estimation of source area, river paleo-discharge, paleoslope, and sediment budgets of linked deep-time depositional systems and

4 Sediment budgeting of short-term backfilling processes

implications for hydrocarbon potential. In: Earth-Science Reviews 153, S. 77–110. DOI: 10.1016/j.earscirev.2015.10.013.

Birzer, F. (1958): Der Kanalbauversuch Karls des Großen. In: Geologische Blätter für Nordost-Bayern and angrenzende Gebiete 8, S. 171–178.

Bork, H.-R.; Bork, H.; Dalchow, C.; Faust, B.; Piorr, H.-P.; Schatz, T. (Hg.) (1998): Landschaftsentwicklung in Mitteleuropa - Wirkungen des Menschen auf die Landschaften. Gotha: Klett-Perthes.

Bork, H.-R.; Becker, A.; Bork, H.; Dotterweich, M.; Rasbach, G.; Schmidtchen, G. (2003): Die Umwehrung der Römersiedlung Waldgirmes bei Wetzlar. In: Bork, H.-R. Schmidtchen, G. and Dotterweich, M. (Hg.): Bodenbildungen, Bodenerosion and Reliefentwicklung im Mittel- and Jungholozän Deutschlands (Forschungen zur Deutschen Landesgeschichte), S. 187–194.

Breuer, S.; Kilian, R.; Baeza, O.; Lamy, F.; Arz, H. (2013): Holocene denudation rates from the superhumid southernmost Chilean Patagonian Andes (53°S) deduced from lake sediment budgets. In: Geomorphology 187, S. 135–152. DOI: 10.1016/j.geomorph.2013.01.009.

Brown, A. G.; Carey, C.; Erkens, G.; Fuchs, M.; Hoffmann, T.; Macaire, J.-J.; Moldenhauer, K.-M.; Des Walling, E. (2009): From sedimentary records to sediment budgets: Multiple approaches to catchment sediment flux. In: Geomorphology 108 (1-2), S. 35–47. DOI: 10.1016/j.geomorph.2008.01.021.

Büntgen, U.; Tegel, W.; Nicolussi, K.; McCormick, M.; Frank, D.; Trouet, V.; Kaplan, J. O.; Herzig, F.; Heussner, K.-U.; Wanner, H.; Luterbacher, J.; Esper, J. (2011): 2500 years of European climate variability and human susceptibility. In: Science (New York, N.Y.) 331 (6017), S. 578–582. DOI: 10.1126/science.1197175.

Carvajal, C.; Steel, R.; Petter, A. (2009): Sediment supply: The main driver of shelf-margin growth. In: Earth-Science Reviews 96 (4), S. 221–248. DOI: 10.1016/j.earscirev.2009.06.008.

Chen, C.-W.; Oguchi, T.; Hayakawa, Y. S.; Saito, H.; Chen, H.; Lin, G.-W.; Wei, L.-W.; Chao, Y.-C. (2018): Sediment yield during typhoon events in relation to landslides, rainfall, and catchment areas in Taiwan. In: Geomorphology 303, S. 540–548. DOI: 10.1016/j.geomorph.2017.11.007.

100

4 Sediment budgeting of short-term backfilling processes

Cook, E. R.; Seager, R.; Kushnir, Y.; Briffa, K. R.; Büntgen, U.; Frank, D.; Krusic, P.l J.; Tegel, W.; van der Schrier, G.; Andreu-H., L.; Baillie, M.; Baittinger, C.; Bleicher, N.; Bonde, N.; Brown, D.; Carrer, M.; Cooper, R.; Čufar, K.; Dittmar, C.; Esper, J.; Griggs, C.; Gunnarson, B.; Günther, B.; Gutierrez, E.; Haneca, K.; Helama, Sa.; Herzig, F.; Heussner, K.-U.; Hofmann, J.; Janda, P.; Kontic, R.; Köse, N.; Kyncl, T.; Levanič, T.; Linderholm, H.; Manning, S.; Melvin, T. M.; Miles, D.; Neuwirth, B.; Nicolussi, K.; Nola, P.; Panayotov, M.; Popa, I.; Rothe, A.; Seftigen, K.; Seim, A.; Svarva, H.; Svoboda, M.; Thun, T.; Timonen, M.; Touchan, R.; Trotsiuk, V.; Trouet, V.; Walder, F.; Ważny, T.; Wilson, R.; Zang, C. (2015): Old World megadroughts and pluvials during the Common Era. In: Science advances 1 (10), e1500561. DOI: 10.1126/sciadv.1500561.

Dietrich, P.; Leven, C. (2009): Direct Push-Technologies. In: Kirsch, R. (Hg.): Groundwater Geophysics. 2. Aufl. Berlin: Springer, S. 347–366.

Dotterweich, M. (2008): The history of soil erosion and fluvial deposits in small catchments of central Europe. Deciphering the long-term interaction between humans and the environment — A review. In: Geomorphology 101 (1-2), S. 192–208. DOI: 10.1016/j.geomorph.2008.05.023.

Dotterweich, M.; Schmitt, A.; Schmidtchen, G.; Bork, H.-R. (2003): Quantifying historical gully erosion in northern Bavaria. In: CATENA 50 (2-4), S. 135–150. DOI: 10.1016/S0341-8162(02)00142-X.

Dreibrodt, S.; Lubos, C.; Terhorst, B.; Damm, B.; Bork, H.-R. (2010): Historical soil erosion by water in Germany. Scales and archives, chronology, research perspectives. In: Quaternary International 222 (1-2), S. 80–95. DOI: 10.1016/j.quaint.2009.06.014.

Erkens, G.; Cohen, K. M.; Gouw, M. J. P.; Middelkoop, H.; Hoek, Wim Z. (2006): Holocene sediment budgets of the Rhine Delta (The Netherlands): a record of changing sediment delivery. In: Sediment Dynamics and the Hydromorphology of Fluvial Systems (Proceedings of a symposion held in Dundee, UK, July 2006) (306).

Förster, H.; Wunderlich, J. (2009): Holocene sediment budgets for upland catchments: The problem of soilscape model and data availability. In: CATENA 77 (2), S. 143–149. DOI: 10.1016/j.catena. 2008.09.004.

Gellis, A. C.; Myers, M. K.; Noe, G. B.; Hupp, C. R.; Schenk, E. R.; Myers, L. (2017): Storms, channel changes, and a sediment budget for an urban-suburban stream, Difficult Run, Virginia, USA. In: Geomorphology 278, S. 128–148. DOI: 10.1016/j.geomorph.2016.10.031.

Griffiths, R. E.; Topping, D. J. (2017): Importance of measuring discharge and sediment transport in lesser tributaries when closing sediment budgets. In: Geomorphology 296, S. 59–73. DOI: 10.1016/j.geomorph.2017.08.037.

101

4 Sediment budgeting of short-term backfilling processes

Guillocheau, F.; Rouby, D.; Robin, C.; Helm, C.; Rolland, N.; Le Carlier de Veslud, C.; Braun, J. (2012): Quantification and causes of the terrigeneous sediment budget at the scale of a continental margin: a new method applied to the Namibia-South Africa margin. In: Basin Research 24 (1), S. 3–30. DOI: 10.1111/j.1365-2117.2011.00511.x.

Hack, A. (2014): Der Bau des Karlsgrabens nach den Schriftquellen. In: Ettel, P.; Daim, F.; Berg- Hobohm, S.; Werther, L. and Zielhofer, C. (Hg.): Großbaustelle 793. Das Kanalprojekt Karls des Großen zwischen Rhein and Donau. Mainz: Verl. des Römisch-Germanischen Zentralmuseums (Mosaiksteine, 11), S. 53–62.

Haneca, K.; Čufar, K.; Beeckman, H. (2009): Oaks, tree-rings and wooden cultural heritage: a review of the main characteristics and applications of oak dendrochronology in Europe. In: Journal of Archaeological Science 36 (1), S. 1–11. DOI: 10.1016/j.jas.2008.07.005.

Herzig, F. (2018): Karlsgraben-Nordanschluss. Dendroarchäologische Untersuchungen. Bayerisches Landesamt für Denkmalpflege, Praktische Denkmalpflege, Archäologische Denkmäler, Referat BV - Restaurierung, Archäologie and Dendrolabor. Thierhaupten.

Hinderer, M. (2012): From gullies to mountain belts: A review of sediment budgets at various scales. In: Sedimentary Geology 280, S. 21–59. DOI: 10.1016/j.sedgeo.2012.03.009.

Hinderer, M.; Einsele, G. (2001): The world's large lake basins as denudation-accumulation systems and implications for their lifetimes. In: Journal of Paleolimnology (26), S. 355–372.

Hoffmann, T.; Erkens, G.; Cohen, K. M.; Houben, P.; Seidel, J.; Dikau, R. (2007): Holocene floodplain sediment storage and hillslope erosion within the Rhine catchment. In: The Holocene 17 (1), S. 105– 118. DOI: 10.1177/0959683607073287.

Jewell, P. A. (1963): The experimental earthwork on Overton Down Whiltshire 1960. London: British Association for the Advancement of Science.

Kalis, A. J.; Merkt, J.; Wunderlich, J. (2003): Environmental changes during the Holocene climatic optimum in central Europe - human impact and natural causes. In: Quaternary Science Reviews 22 (1), S. 33–79. DOI: 10.1016/S0277-3791(02)00181-6.

Kesel, R. H. (1989): The Role of the Mississippi River in Wetland loss in Southeastern Louisiana, U.S.A. In: Environmental Geology and Water Sciences 13 (3), S. 183–193.

102

4 Sediment budgeting of short-term backfilling processes

Kesel, R. H.; Yodis, E. G.; McCraw, D. J. (1992): An approximation of the sediment budget of the lower mississippi river prior to major human modification. In: Earth Surf. Process. Landforms (17), S. 711– 722.

Koch, R. (1993): Fossa Carolina - 1200 Jahre Karlsgraben. München: Denkmalpflege Informationen.

Koch, R. (1996): Neue Beobachtungen and Forschungen zum Karlsgraben. In: Jahrbuch des Historischen Vereins für Mittelfranken 97, S. 1–16.

Köhn, D.; Wilken, D.; Nil, D. de; Wunderlich, T.; Rabbel, W.; Werther, L.; Schmidt, J.; Zielhofer, C.; Linzen, S. (2019): Comparison of time-domain SH waveform inversion strategies based on sequential low and bandpass filtered data for improved resolution in near-surface prospecting. In: Journal of Applied Geophysics 160, S. 69–83. DOI: 10.1016/j.jappgeo.2018.11.001.

Kondolf, G. M.; Schmitt, R. J. P.; Carling, P.; Darby, S.; Arias, M.; Bizzi, S.; Castelletti, A.; Cochrane, T. A.; Gibson, S.; Kummu, M.; Oeurng, C.; Rubin, Z.; Wild, T. (2018): Changing sediment budget of the Mekong: Cumulative threats and management strategies for a large river basin. In: The Science of the total environment 625, S. 114–134. DOI: 10.1016/j.scitotenv.2017.11.361.

Lacquement, C. H. (2010): Recalculating mound volume at moundville. In: Southeastern Archaeology 29 (2), S. 341–354.

Land, A.; Remmele, S.; Hofmann, J.; Reichle, D.; Eppli, M.; Zang, C.; Buras, A.; Hein, S.; Zimmermann, R. (2019): Two millennia of Main region (southern Germany) hydroclimate variability. In: Clim. Past 15 (5), S. 1677–1690. DOI: 10.5194/cp-15-1677-2019.

103

4 Sediment budgeting of short-term backfilling processes

Leopold, M.; Hürkamp, K.; Völkel, J.; Schmotz, K. (2011): Black soils, sediments and brown calcic luvisols: A pedological description of a newly discovered neolithic ring ditch system at Stephansposching, Eastern Bavaria, Germany. In: Quaternary International 243 (2), S. 293–304. DOI: 10.1016/j.quaint.2010.11.021.

Linzen, S.; Schneider, M. (2014): Der Karlsgraben im Fokus der Geophysik. In: Ettel, P.; Daim, F.; Berg- Hobohm, S.; Werther, L. and Zielhofer, C. (Hg.): Großbaustelle 793. Das Kanalprojekt Karls des Großen zwischen Rhein and Donau. Mainz: Verl. des Römisch-Germanischen Zentralmuseums (Mosaiksteine, 11), S. 29–32.

Linzen, S.; Schneider, M.; Berg-Hobohm, S.; Werther, L.; Ettel, P.; Zielhofer, C.; Schmidt, J.; Fassbinder, J. W.E.; Wilken, D.; Fediuk, A.; Dunkel, S.; Stolz, R.; Meyer, H.-G.; Sommer, C. S. (2017): From magnetic SQUID prospection to excavation - investigations at Fossa Carolina, Germany. In: Jennings, B.; Gaffney, C.; Sparrow, T. and Gaffney, S. (Hg.): 12th International Conference of Archaeoloigcal Prospection. Bradford. The University of Bradford: Archaeopress, S. 144–145.

Lisá, L.; Komoróczy, B.; Vlach, M.; Válek, D.; Bajer, A.; Kovárník, J.; Rajtár, J.; Hüssen, C. M.; Šumberová, R. (2015): How were the ditches filled? Sedimentological and micromorphological classification of formation processes within graben-like archaeological objects. In: Quaternary International 370, S. 66–76. DOI: 10.1016/j.quaint.2014.11.049.

Muigg, B.; Seim, A.; Tegel, W.; Werther, L.; Herzig, F.; Schmidt, J.; Zielhofer, C.; Land, A.; Büntgen, U. (2020): Tree rings reveal dry conditions during Charlemagne’s Fossa Carolina construction in 793 CE. In: Quaternary Science Reviews 227, S. 106040. DOI: 10.1016/j.quascirev.2019.106040.

Nelson, J. (2015): Evidence in question: dendrochronology and early medieval historians. In: Kano, O. and Lemâitre, J.-L. (Hg.): Entre texte et histoire: études d'histoire médiévale offertes au professeur Soichi Sato. Paris: Éditions de Boccard, S. 227–249.

Pechtl, J.; Land, A. (2019): Tree rings as a proxy for seasonal precipitation variability and Early Neolithic settlement dynamics in Bavaria, Germany. In: PloS one 14 (1), e0210438. DOI: 10.1371/journal.pone.0210438.

104

4 Sediment budgeting of short-term backfilling processes

Qiao, S.; Shi, X.; Wang, G.; Zhou, L.; Hu, B.; Hu, L.; Yang, G.; Liu, Y.; Yao, Zh.; Liu, S. (2017): Sediment accumulation and budget in the Bohai Sea, Yellow Sea and East China Sea. In: Marine Geology 390, S. 270–281. DOI: 10.1016/j.margeo.2017.06.004.

Rabiger-Völlmer, J.; Schmidt, J.; Linzen, S.; Schneider, M.; Werban, U.; Dietrich, P.; Wilken, D.; Wunderlich, T.; Fediuk, A.; Berg, S.; Werther, L.; Zielhofer, C. (2020): Non-invasive prospection techniques and direct push sensing as high-resolution validation tools in wetland geoarchaeology – Artificial water supply at a Carolingian canal in South Germany? In: Journal of Applied Geophysics 173, S. 103928. DOI: 10.1016/j.jappgeo.2019.103928.

Rascher, E.; Rindler, R.; Habersack, H.; Sass, O. (2018): Impacts of gravel mining and renaturation measures on the sediment flux and budget in an alpine catchment (Johnsbach Valley, Austria). In: Geomorphology 318, S. 404–420. DOI: 10.1016/j.geomorph.2018.07.009.

Reimer, P. J.; Bard, E.; Bayliss, A.; Beck, J. W.; Blackwell, P. G.; Ramsey, C. B.; Buck, C. E.; Cheng, H.; Edwards, R. L.; Friedrich, M.; Grootes, P. M.; Guilderson, T. P.; Haflidason, H.; Hajdas, I.; Hatté, C.; Heaton, T. J.; Hoffmann, D. L.; Hogg, A. G.; Hughen, K. A.; Kaiser, K. F.; Kromer, B.; Manning, S. W.; Niu, M.; Reimer, R. W.; Richards, D. A.; Scott, E. M.; Southon, J. R.; Staff, R. A.; Turney, C. S. M.; van der Plicht, J. (2013): IntCal13 and Marine13 Radiocarbon Age Calibration Curves 0– 50,000 Years cal BP. In: Radiocarbon 55 (04), S. 1869–1887. DOI: 10.2458/azu_js_rc.55.16947.

Rommens, T.; Verstraeten, G.; Poesen, J.; Govers, G.; van Rompaey, A.; Peeters, I.; Lang, A. (2005): Soil erosion and sediment deposition in the Belgian loess belt during the Holocene: establishing a sediment budget for a small agricultural catchment. In: The Holocene 15 (7), S. 1032–1043.

Schmidt-Kaler, H. (1993): Geologie and Landschaftsentwicklung im Rezat-Altmühl Bereich. In: Bau intern Special Issue, S. 8–10.

Schreg, R. (2014): Uncultivated landscapes or wilderness? Early medieval land use in low mountain ranges and flood plains of Southern Germany. In: European Journal of Post - Classical Archaeologies (4), S. 69–98.

105

4 Sediment budgeting of short-term backfilling processes

Schweingruber, F. H. (1988): Tree Rings. Basics and Applications of Dendrochronology. Dordrecht: Springer Netherlands.

Smetanová, A.; Verstraeten, G.; Notebaert, B.; Dotterweich, M.; Létal, A. (2017): Landform transformation and long-term sediment budget for a Chernozem-dominated lowland agricultural catchment. In: CATENA 157, S. 24–34. DOI: 10.1016/j.catena.2017.05.007.

Stele, A.; Fassbinder, J. W. E.; Härtling, J. W.; Bussmann, J.; Schmidt, J.; Zielhofer, C. (2019): Genesis of magnetic anomalies and magnetic properties of archaeological sediments in floodplain wetlands of the Fossa Carolina. In: Archaeological Prospection 4 (4), S. 1–11. DOI: 10.1002/arp.1761.

Stuiver, M.; Reimer, P. J.; Reimer, R. W. (2019): CALIB. (www program). Version 7.1. http://calib.org, accessed 08.03.2019.

Suchodoletz, H. von; Faust, D.; Zöller, L. (2009): Geomorphological investigations of sediment traps on Lanzarote (Canary Islands) as a key for the interpretation of a palaeoclimate archive off NW Africa. In: Quaternary International 196 (1-2), S. 44–56. DOI: 10.1016/j.quaint.2008.03.014.

Voiculescu, M.; Ianăş, A.-N.; Germain, D. (2019): Exploring the impact of snow vole (Chionomys nivalis) burrowing activity in the Făgăraș Mountains, Southern Carpathians (Romania): Geomorphic characteristics and sediment budget. In: CATENA 181, S. 104070. DOI: 10.1016/j.catena.2019.05.016.

Völlmer, J.; Zielhofer, C; Hausmann, J.; Dietrich, P.; Werban, U.; Schmidt, J.; Werther, L.; Berg, S. (2018): Minimalinvasive Direct-push Erkundung in der Feuchtboden(geo)archäologie am Beispiel des Karlsgrabens (Fossa Carolina). In: Archäologisches Korrespondenzblatt 48 (4), S. 577–608.

Weber, M. D.; Pasternack, G. B. (2017): Valley-scale morphology drives differences in fluvial sediment budgets and incision rates during contrasting flow regimes. In: Geomorphology 288, S. 39–51. DOI: 10.1016/j.geomorph.2017.03.018.

Werther, L. (2016): Großbaustelle Karlsgraben. Eine Chaine opératoire für den Umgang mit der Ressource Holz in der Karolingerzeit. In: Mitteilungen der Deutschen Gesellschaft für Archäologie des Mittelalters and der Neuzeit 29, S. 103–112.

Werther, L. (2017): Karlsgraben doch schiffbar? Aktuelles aus der Landesarchäologie. In: Archäologie in Deutschland (5), S. 41–42.

106

4 Sediment budgeting of short-term backfilling processes

Werther, L.; Nelson, J.; Herzig, F.; Schmidt, J.; Berg, S.; Ettel, P.; Linzen, S.; Zielhofer, C. (2020): 792 or 793? Charlemagne's canal project: craft, nature and memory. In: Early Medieval Europe 28 (3), S. 444–465.

Zielhofer, C.; Leitholdt, E.; Werther, L.; Stele, A.; Bussmann, J.; Linzen, S.; Schneider, M.; Meyer, C.; Berg-Hobohm, S.; Ettel, P. (2014): Charlemagne's summit canal. An early medieval hydro- engineering project for passing the Central European Watershed. In: PloS one 9 (9), e108194. DOI: 10.1371/journal.pone.0108194.

Zolitschka, B. (1998): A 14,000 year sediment yield record from western Germany based on annually laminated lake sediments. In: Geomorphology 22 (1), S.1–17. DOI: 10.1016/S0169-555X(97)00051-2.

107

5 Discussion

Chapter 5 – Discussion

5.1 Methodological approaches

5.1.1 Palaeo-surface modelling

Palaeo-surface models are important for the interpretation of (geo)archaeological sites (Carey et al. 2017) and their landscape context. Palaeo-surfaces can be reconstructed by inductive (Smedt et al. 2013b; Baubinienė et al. 2015; Verhegge et al. 2017), deductive (Werbrouck et al. 2011; van Loon et al. 2009) or combined approaches (Vermeer et al. 2014; Zwertvaegher et al. 2010). Deductive approaches usually have the advantage to model the study area without time-consuming fieldwork and large areas can be computed (e.g. Werbrouck et al. 2011). Further, there is no need of ground-truth data in beforehand and therefore, this approach is highly adaptable for other study areas. On the other hand, inductive approaches as well as combined approaches have the advantage that they are able to date their specific paleosurface by numerical or archaeological approaches. (Kirchner et al. 2018; Schneider et al. 2017; Smedt et al. 2013a).

I decided to use a deductive modelling approach, that works by “deconstruction” of a present DTM. The specific removal of modern anthropogenic features allows the pre-dating of the modern period (pre-modern DTM). However, published studies with deductive approaches mostly do not give a temporal frame of their palaeosurfaces. For the first time, our approach allows a controlled methodology and therefore concise conclusions.

To check the accuracy of the modelling result, we compared our modelling result with ground-truth information derived from several drillings. By validation of the modelled palaeosurface levels with the true palaeosurface levels (expressed by fossil topsoils/palaeosols) we estimated the model inaccuracies. In some case (below the Fossa Carolina dams) the buried A-horizon was missing, and therefore we have to assume a pre-Carolingian erosion. Taken this into account, the overall offset of the model accuracy is low to moderate (root mean square error RMSE = 0.69 [m]). Further, we checked the model error in two separate landscape parts, where specific data is available; in the Altmühl floodplain (RMSE = 0.71 [m]) and below the Fossa Carolina dams along the canal course (RMSE = 0.62 [m]). The main offset are wider scale erosion and accumulation processes, because this spatial impact is taken into account in the modelling, due to a lack of data. However, our validation approach does not show significant offsets and the general geomorphological characteristic remains undisturbed. Nevertheless, we have to take landscape dynamic into account. Prominent studies (Dotterweich 2008; Dreibrodt et al. 2010) show omnipresent erosion- accumulation dynamics throughout Germany, especially in hilly landscapes. Therefore, we have to assume post-Carolingian sediment dynamics, that might disturb the modelling result and create a spatial offset. The calculated total RMSE of 0.62 may reflect this process. However, the study site is relatively less affected

108

5 Discussion

from erosion, compared with loess covered landscapes. The quantitative results of the error estimation show that the pre-modern DTM is reliable for further steps.

5.1.2 3D-modelling

3D-modelling and GIS applications in archaeological issues have become common in the last years (Carey et al. 2018; Earley-Spadoni 2017). Studies dealing with 3D-reconstructions of archaeological features (earth work and construction scale) are sparse. Especially, studies dealing with filled features (e.g. Bork et al. 2003). Smedt et al. (2011) and Diamanti et al. (2005) used high-resolution geophysical approaches to reconstruct archaeological and geoarchaeological features in detail. However, the Fossa Carolina canal has a length of c. 2.9 km and a maximum depth of more than 10 m below the present surface. Such a huge archaeological feature has not been reconstructed or modelled before. We developed a new approach combining archaeological excavation, direct push sensing and drilling data. Both, archaeological excavations and direct push sensing produce depth accurate information (Hausmann et al. 2016; Werther 2016; Völlmer et al. 2018). Therefore, we could use them as reference cross-sections and used their canal trench geometric information. Drillings have the disadvantage, that they have a coarser vertical resolution and depth inaccuracies, due to the compaction of sediments (Hausmann et al. 2018), but the application is fast, flexible and cost-effective. Due to the possible depth inaccuracies, slight offsets have to be taken into account by interpreting the modelling result. Unfortunately, the error cannot be calculated quantitatively due to a lack of the true depth information along the canal. But roughly estimated the error varies on a centimetre to decimetre scale. However, the overall picture of the canal trench bottom is consistent (Fig. 5.1). The presented approach is applicable for other study sites, due to the new conceptual approach using reference cross-sections and a spatial interpolation guided by single drilling information. Also, geophysical techniques with suitable depth accuracies, such as electromagnetic induction (Smedt et al. 2011), seismics (Köhn et al. 2019), or SQUID magnetic inversions (Linzen et al. 2017), or ground penetrating radar (GPR) (Chapman et al. 2009) have the opportunity to be used for a guided interpolation. Corradini et al. (2020) used extensive GPR measurements to map different sediment facies. They could show, that GPR measurements are suitable for extensive mapping and reconstruction. An integration of such precise geophysical data in a geoarchaeological 3D-modelling approach is conceivable.

109

5 Discussion

Figure 5.1: Canal longitudinal section with the present surface (green line; derived from the present LiDAR DTM), pre-modern surface (grey line; derived from the pre-modern DTM from Schmidt et al. 2018), the maximum Carolingian excavation depth (black line, derived from the 3D-model of the Fossa Carolina from Schmidt et al. 2019) and the top level of the backfills (orange line; Schmidt et al. 2020). The grey area shows the thickness of the backfills for the entire Fossa Carolina. The y-axis is 80-fold superelevated to show prominent information (modified after Schmidt et al. 2020).

5.1.3 Geoarchaeological sediment budgeting

The geoarchaeological sediment budgeting of the initial backfill sediments comprises information about the reference palaeosurface (Schmidt et al. 2018), the maximum excavation depth/ maximum canal trench volume (Schmidt et al. 2019), the volume of backfill sediments (Schmidt et al. 2020) and the sediment storage in the present dams (Schmidt et al. 2018). As developed and discussed in the specific studies, the input data of the budgeting approach are of high accuracy. Prominent studies of sediment budgeting approaches focus on large catchments, with less dense data networks (Hoffmann et al. 2007; Erkens et al. 2006). But also, other methodologies are discussed in the scientific community. Sediment storages are quantified by spatial interpolation of drilling results (Bussmann et al. 2014; Rommens et al. 2005), geometric assumptions of sediment bodies (Suchodoletz et al. 2009) or geophysical surveys (Guillocheau et al. 2012). We were able to establish an equidistant spacing of cross-section with high resolution depth information. In contrast to the studies mentioned above, we work on a local scale and plenty of ground- truth data. Comparable studies (geoarchaeological, local scale, dense network of ground-truth data, etc.) are sparse. For example, Bork et al. (2003) describe the sedimentation of a roman ditch, that is pronounced smaller than Charlemagne’s summit canal.

110

5 Discussion

The case of Fossa Carolina has the main advantage that it does not have to take sediment output into account. For example, colluvial systems (Bussmann et al. 2014; Bussmann 2014), fluvial systems (Förster and Wunderlich 2009), or deltas (Erkens et al. 2006) have an unknown amount of sediment loss through time. Lakes or endorheic positions can act solely as sediment traps, but mostly they are part of larger catchments (Zolitschka 1998; Breuer et al. 2013). The Fossa Carolina also acts solely as a sediment trap without sediment loss is not an issue. Furthermore, the backfills sediments of the Fossa Carolina were dated numerically by AMS radiocarbon dating of macro-remains and charcoals. The ages plot well within their 2-sigma range in the time of the construction site abandonment (Schmidt et al. 2020). Dating of sediments can be difficult due to reworked material (age overestimation), radiocarbon plateaus or insufficient bleaching (for luminescence dating approaches; Brown et al. 2009). Moreover, the precise dendrochronological results of the recovered construction timber give a chronological framework of the construction itself. This terminus post quem (all refilled sediments have to be younger than 793 AD) and the fact that the backfill sediment are stratigraphically on top of the construction timber highlight the chronological precision of the canal construction, its collapse and subsequent for the sediment budgeting approach.

5.2 Key findings

5.2.1 Canal course

The advantage of the pre-modern DTM is to provide a precise approximation of the historical topography for palaeohydrographical analysis (hydrographic correct landscape connectivity). Thus, a least cost path analysis was conducted, to predict the best canal course between the Altmühl and Swabian Rezat rivers concerning the absolute height levels. The GIS-analysis compared to the real Carolingian canal revealed quite similar courses (Fig. 5.2). Koch (1996) assumed geological conditions as determining factor for the general canal course. For the first time, quantitative analysis revealed solely the topography as the main reason. However, two considerable deviations are prominent. Therefore, this general S-shape can be linked to the topography. In the West-East Section the real canal course has a more southern course and in the Northeastern Section it has a more western course. The sharp bend between to West-East Section and the Northern Section were discussed before as results of a connection of the Swabian Rezat river and the canal for water supply (Zielhofer et al. 2014). Hitherto, no remains of a connection structure were found and proven. Therefore, the water supply reconstruction remains unclear (Rabiger-Völlmer et al. 2020).

The topographic wetness index shows potential wet areas. It is striking that the canal course deviations are linked to the wet areas (Fig. 5.2). The real canal course avoids to run through the depth contour, but rather at the edges of the wetlands. These areas are dominated by organic-rich sediments (Zielhofer and Kirchner 2014), which are unsuitable for excavation work, due to unstable embankments. Also, Werther (2019) reports the location of the Carolingian canal at the transition from the wetland to the lower slope. 111

5 Discussion

Furthermore, the Carolingian constructors had to keep the construction pit free of water. They had to avoid the lowest points in the landscape to create an artificial drainage system. The planers and constructors of the canal had to inspect the alignment on-site beforehand. This is a typical procedure for river and floodplain prospection in the 8th century AD, as it was recorded in written sources (Werther 2019). Nevertheless, no specific information on the planning process of the Fossa Carolina are known. It is likely, that such hydraulic engineers came from the water mills. At least, several water mill constructions are known in the late 8th century AD in Bavaria (Werther et al. 2015).

5.2.2 Canal trench volume

Multiple authors assumed or estimated the absolute excavation volume of the Carolingian canal trench (Birzer 1958; Koch 1993; Hofmann 1976). But no published study could use quantitative data. Thus, the estimations range from 80,000 m³ to 450,000 m³ (or several 100,000s m³). This variability is mainly due to the specific level of knowledge. Birzer (1958) and Hofmann (1976) did not take the Northern and

112

5 Discussion

Northeastern Section into account; just because it was unknown and unproved in that time. All authors had no indications for the hydro-technical concept of the canal. Therefore, Koch (1993) assumed a constant trench bottom level. First, Zielhofer et al. (2014) had evidence for a summit canal concept, due to several drillings along the canal.

For the first time, the 3D-modelling approach with various ground truth data and reproducible methodology revealed the geometry and topography of the canal trench. Based on this precise excavation morphology and the pre-modern DTM, an exact canal trench volume was calculated (c. 285,000 m³; Tab. 5.1). Furthermore, the sediment volume of the dam remnants was calculated using the pre-modern DTM and the present LiDAR DTM (c. 120,000 m³; Tab. 5.1). The largest amounts of the excavation volume are related to the Central Section, where the deepest canal trench is located (Fig. 5.3). The quantitative results are basic part of the budgeting approach of the backfill sediments (see chapter 5.2.5). Furthermore, the absolute quantity of the maximum excavation volume can be the basic part for a workload/manpower calculation, that will give a range of the number of workers needed to dig the canal trench.

113

5 Discussion

5.2.3 Summit level of the Fossa Carolina

Numerous authors and studies investigated the Fossa Carolina. Earlier studies that did not have subsurface information assumed a constant canal trench bottom level (Koch 1993; Birzer 1958; Hofmann 1976). First, Koch (1996) found in his drillings, that the canal trench bottom level is not continuous between both tributaries. The concept of a summit canal was described and proven by Zielhofer et al. (2014). According to the available drillings, Zielhofer et al. (2014) assumed the summit section, and therefore, the hydrotechnical control systems, in the eastern part of the West-East Section (knickpoint to the Northern Section). If an artificial water supply does exist, it has to drain into the summit section of the canal. The 3D-modelling of the maximum excavation level reveals that the summit section is c. 500 m more to the north (Fig. 5.1). This is accompanied with a study from Rabiger-Völlmer et al. (2020), where no artificial water inlet to the Fossa Carolina in the knickpoint area has been found. With the new quantitative result with a precise prediction of the summit area, a new purposive prospection of a potential artificial water supply can be developed and might answer the essential question of the Carolingian water controlling system.

5.2.4 Canal construction, collapse and backfill sediment budget

The reconstruction of the canal construction, its collapse and a quantification approach of the tipping point is the overall connection of the individual papers, presented in this doctoral thesis. General, the reconstruction of a canal construction is a matter of scale. The general reconstruction of the construction pit and involved ex-situ labour (logging, timber production, etc.) was described by Werther (2016). A precise idea of the canal trench construction design was identified by several archaeological excavations, where the timber installation at the trench edges was recognised (Werther et al. 2018; Werther et al. 2015). For the first time, a sub-annual and high-resolution information of the spatial construction progress was developed using the recovered timbers and subsequent dendrological analysis (Schmidt et al. 2020; Herzig 2018). In general, all dendrochronological results show, that the logging and timber installation was done in 792 and 793 AD (Werther et al. 2020). The spatial distribution of the felling dates shows explicit differences (Fig. 5.4). The northern archaeological excavation trenches have slightly older felling date distributions than the southern one. Therefore, a north-to-south construction progress is likely.

114

5 Discussion

Furthermore, the dendrochronological dates serve a chronological framework of the post-construction time processes. Sedimentological analyses revealed backfills (eroded sediments from surrounding dams) directly on top the excavated timbers (Zielhofer et al. 2014). Therefore, these backfill sediments were accumulated during or directly after the construction site abandonment. The sediments were dated by radiocarbon-dating of incorporated macro-remains (charcoal, wood, plant remains). The results show that the backfill date precisely (within the 2-sigma range) to the construction period or rather the abandonment (Schmidt et al. 115

5 Discussion

2020). Within this doctoral thesis, the backfill sediments were quantified (c. 40,000 m³) and spatially reconstructed along the whole proven canal trench (Fig. 5.5).

Figure 5.5: Spatial distribution of backfill sediment thickness along the Fossa Carolina canal course (modified after Schmidt et al. 2020).

The map shows an uneven spatial distribution of the backfills, with the major amounts in the Central and West-East Sections. Concerning the contemporary written sources (Werther et al. 2020; Nelson 2015), the backfilling processes (the erosion of the dams) and subsequent the recurrent replenishment of the construction it in the Central and West-East Sections could have acted as a tipping point. The large amounts (c. 15 %) of the total excavation volume was transported back into the construction pit as initial backfills (Tab. 5.1). For the first time, this phenomenon was described and mapped systematically. Furthermore, the quantification approach revealed the large amounts that can be discussed with a sediment budgeting approach.

Table 5.1 shows the most important numbers of the quantified Carolingian canal trench and its collapse derived from the three papers of this doctoral thesis. This result allows a quantitative sediment budgeting of the backfills. In total, 42 % of the maximum excavation volume is stored in the present dams and 15 % 116

5 Discussion

was washed back into the canal trench as initial backfills. Surprisingly, the ratios of the backfills and maximum volume (Tab. 5.1, column 5) show similar values in all canal sections. Nevertheless, the absolute amounts of backfills are present in the Central Section. For the first time, this information allows quantitative statements of the construction collapse. Furthermore, it is possible to identify a collapse-zone. The large amounts of backfilled sediments in the Central Section might have caused the abandonment.

Ratio Ratio Maximuma Minimum (Organic Initial Dams Length backfill/maximum Dams/Maximum [m³] fills)b [m³] backfills [m³] [m³] [m] [%] [%] Total 285455 243820 41635 15 119681 42 2829 Central Section 160815 136379 24436 15 85478 53 803 WE Section 84364 74573 9791 12 19268 23 494 Northern 26188 21409 4779 18 13838 53 738 Section North-Eastern 14088 11458 2630 19 5988 42 794 Section

117

6 Conclusion

Chapter 6 – Conclusion

For the first time, quantitative approaches were used on geoarchaeological processes at the early medieval canal construction Fossa Carolina. This phd-thesis were conducted in the frame of the wider “Fossa Carolina” project dealing with historical, archaeological, geophysical and geoscientific research about the canal and its supra-regional significance (critical conjunction in the early medieval waterway network). The present phd-thesis resolved the concepts of construction planning (canal alignment) and implementation (excavation) of the canal at the local scale. Furthermore, the hypothesis of a collapsed canal by erosion of the dams was localised and quantified.

The modelling of the pre-modern topography of study area allowed the subsequent calculation of the shortest path between the Altmühl and Swabian Rezat rivers. The shortest path is similar to the real, proven canal course (S-shape), but has also significant deviations. These deviations are a consequence of avoiding the wet areas and near surface organic-rich sediments with unsuitable properties for canal trench embankments. The Carolingian constructors almost made the perfect alignment of the canal trench concerning the minimum earth movement and maximum geotechnical stability of the construction. For the first time, this impressive hydro-engineering knowledge, surveying ability and large-scale construction site organisation was proven by quantification approaches of spatial data in a geoarchaeological context. For this purpose, a modelling procedure was developed, tested and validated. Furthermore, this approach applicable to other study areas, which is documented by recent citations of the manuscript.

For the first time, a 3D-modelling approach, integrating different geoarchaeological subsurface data with a pre-modern DTM, was developed. Direct-push sensing transects and archaeological excavations served as references for the section-wise canal trench geometry. Drilling transects were used to interpolate the geometric information along the total canal course. Compared to other 3D-modelling approaches, it is highly transparent and benefits from the strengths of the input methods. Therefore, the conceptual approach is applicable to other study areas, especially in geoarchaeological wetland sites. The final 3D-model shows the maximum excavation situation of the Fossa Carolina within the pre-modern landscape. For the first time, the volume of the excavated earth of the canal was calculated. Almost 300,000 m³ of material were moved during construction. Furthermore, these amounts are not equally distributed along the canal trench. Despite, only c. 120,000 m³ of material is still stored in the dam remnants. The lack of both values reflects the post-construction reworking, erosion and usage processes. Furthermore, the exact summit section of the canal was defined. It is situated more to the north, than anticipated in the published literature. Hence, new research on the hydro-technical control system has to be done in this predicted area. Therefore, the chance to identify such a structure is enhanced and could give new insights in the hydro-technical concept and technical state.

118

6 Conclusion

The high-resolution reconstruction of the backfill sediments was performed by the same modelling approach than the 3D-model. The identification of the backfills was done by sediment characteristics such as grain size distributions and organic carbon content. The results show that the backfills were eroded from the dams surrounding the canal trench. Furthermore, the backfills cover the timbering installation of the canal directly. The stratigraphical information, accompanied with dendrochronological information from the timber and radiocarbon ages from the backfills itself, indicate a concurrent erosion of the dams with abandonment of the construction. The spatial distribution of the backfills show that the major amounts are situated in the Central and West-East Sections. The erosional collapse of the canal construction is visible along the whole canal trench (even in the same local relative amounts), but the large absolute quantities may indicate a collapse zone in the Central and West-East Sections. It could have acted as tipping point to abandon the construction site, as contemporary written sources suggest.

Due to the wealth of quantitative data about the pre-modern landscape, the canal construction (including the dams), the canal trench volume and refilled sediments it was possible to set up a sediment budget to quantify the relative amounts of the eroded dams and backfilled material compared to the absolute canal trench section volumes. In total 42 % of the excavated material is still stored in the present dams, where as 15 % was accumulated as backfills back in the canal trench. Therefore, 43 % cannot be explained by the models. Qualitatively, a lot of material was accumulated in the canal trench in late medieval and early modern times. Further, it is known, that a lot of material was intentional removed from the dams as part of the railway construction works. Finally, the phd-thesis demonstrated, that quantitative approaches have a clear merit in geoarchaeological issues.

119

References

Agapiou, A.; Lysandrou, V. (2015): Remote sensing archaeology. Tracking and mapping evolution in European scientific literature from 1999 to 2015. In: Journal of Archaeological Science: Reports 4, S. 192–200. DOI: 10.1016/j.jasrep.2015.09.010.

Bavarian Land Surveying Office (2018a): Flurkarte. https://www.ldbv.bayern.de/produkte/kataster/ flurkarte.html, accessed 11.04.2018.

Bavarian Land Surveying Office (2018b): Geländemodell. https://www.ldbv.bayern.de/produkte/ 3dprodukte/gelaende.html, accessed 11.04.2018.

Bavarian Land Surveying Office (2018c): Historische Flurkarten. https://www.ldbv.bayern.de/ produkte/historisch/hist-flurkarten.html, accessed 11.04.2018.

Bavarian Land Surveying Office (2018d): Orthofotos. https://www.ldbv.bayern.de/produkte/ luftbild/orthophotos.html, accessed 11.04.2018.

Bavarian Land Surveying Office (2018e): Tatsächliche Nutzung. https://www.ldbv.bayern.de/produkte/ kataster/tat_nutzung.html, accessed 11.04.2018.

Bavarian State Department of Cultural Heritage BLfD (1985): Luftbildarchiv, Archivnummer: 7130_027, Filmnummer: 3840B, Bild 12, 19.02.1985. O. Braasch (Regie).

Beck, F. (1911): Der Karlsgraben. Eine historische, topographische and kritische Abhandlung. Nürnberg: Verlag der Friedrich Kornschen Buchhandlung.

Beck, H. E.; Zimmermann, N. E.; McVicar, T. R.; Vergopolan, N.; Berg, A.; Wood, E. F. (2018): Present and future Köppen-Geiger climate classification maps at 1-km resolution. In: Scientific data 5, S. 180214. DOI: 10.1038/sdata.2018.214.

Berg-Hobohm, S. (2014a): Archäologische Forschungsgeschichte der Fossa Carolina. In: Ettel, P.; Daim, F.; Berg-Hobohm, S.; Werther, L. and Zielhofer, C. (Hg.): Großbaustelle 793. Das Kanalprojekt Karls des Großen zwischen Rhein and Donau. Mainz: Verl. des Römisch-Germanischen Zentralmuseums (Mosaiksteine, 11), S. 1–4.

120

References

Berg-Hobohm, S. (2014b): Der Karlsgraben als Defensionslinie während des spanischen Erbfolgekriegs. In: Ettel, P.; Daim, F.; Berg-Hobohm, S.; Werther, L. and Zielhofer, C. (Hg.): Großbaustelle 793. Das Kanalprojekt Karls des Großen zwischen Rhein and Donau. Mainz: Verl. des Römisch-Germanischen Zentralmuseums (Mosaiksteine, 11), S. 13–14.

Berg-Hobohm, S.; Werther, L. (2014): Das rezente Erscheinungsbild des Karlsgrabens. In Ettel, P.; Daim, F.; Berg-Hobohm, S.; Werther, L. and Zielhofer, C. (Hg.): Großbaustelle 793. Das Kanalprojekt Karls des Großen zwischen Rhein and Donau. Mainz: Verl. des Römisch-Germanischen Zentralmuseums (Mosaiksteine, 11), S. 9–12.

Beuzen-Waller, T.; Stock, F.; Kondo, Y. (2018): Geoarchaeology: A toolbox for revealing latent data in sedimentological and archaeological records. In: Quaternary International 483, S. 1–4. DOI: 10.1016/j.quaint.2018.05.029.

Birzer, F. (1958): Der Kanalbauversuch Karls des Großen. In: Geologische Blätter für Nordost-Bayern and angrenzende Gebiete 8, S. 171–178.

Bork, H.-R.; Becker, A.; Bork, H.; Dotterweich, M.; Rasbach, G.; Schmidtchen, G. (2003): Die Umwehrung der Römersiedlung Waldgirmes bei Wetzlar. In: Bork, H.-R.; Schmidtchen, G. and Dotterweich, M. (Hg.): Bodenbildungen, Bodenerosion and Reliefentwicklung im Mittel- and Jungholozän Deutschlands (Forschungen zur Deutschen Landesgeschichte), S. 187–194.

Bussmann, J. (2014): Holozäne Sedimentdynamik im Umfeld der Varusschlacht. Ph.D. thesis. Osnabrück.

121

References

Campana, S. (2017): Remote sensing in archaeology. In: Gilbert, A. S.; Goldberg, P.; Holliday, V. T.; Mandel, R. D. and Sternberg, R. S. (Hg.): Encyclopedia of geoarchaeology. Dordrecht, Heidelberg, New York, London: Springer Reference (Encyclopedia of earth sciences series), S. 703–725.

Canti, M.; Huisman, D. J. (2015): Scientific advances in geoarchaeology during the last twenty years. In: Journal of Archaeological Science 56, S. 96–108. DOI: 10.1016/j.jas.2015.02.024.

Carey, C.; Howard, A. J.; Knight, D.; Corcoran, J.; Heathcote, J. (2018): Deposit Modelling and Archaeology. Brighton: University of Brighton.

Carson, M. T. (2014): Paleo-Terrain Research. Finding the First Settlement Sites of Remote Oceania. In: Geoarchaeology 29 (3), S. 268–275. DOI: 10.1002/gea.21457.

Climate data (2020): Station Weissenburg. https://de.climate- data.org/europa/deutschland/bayern/weissenburg-22320/, accessed 28.05.2020.

Corradini, E.; Wilken, D.; Zanon, M.; Groß, D.; Lübke, H.; Panning, D.; Dörfler, W.; Rusch, K.; Mecking, R.; Erkul, E.; Pickartz, N.; Feeser, I.; Rabbel, W. (2020): Reconstructing the palaeoenvironment at the early Mesolithic site of Lake Duvensee: Ground-penetrating radar and geoarchaeology for 3D facies mapping. In: The Holocene 30 (6), S. 820–833. DOI: 10.1177/09596836209022.

122

References

Dietrich, P.; Leven, C. (2009): Direct Push-Technologies. In: Kirsch, R. (Hg.): Groundwater Geophysics. 2. Aufl. Berlin: Springer, S. 347–366.

Ehlers, C. (2014): Der reisende Herrscher: Organisation and Kommunikation. In: Pohle, K. (Hg.): Karl der Große. Charlemagne. Orte der Macht. Essays. Dresden: Sandstein Kommunikation, S. 40–47.

Elmshäuser, K. (1992): Kanalbau and technische Wasserführung im Mittelalter. In: Technikgeschichte 59 (1), S. 1–26.

Erkens, G.; Cohen, K. M.; Gouw, M. J. P.; Middelkoop, H.; Hoek, W. Z. (2006): Holocene sediment budgets of the Rhine Delta (The Netherlands): a record of changing sediment delivery. In: Sediment Dynamics and the Hydromorphology of Fluvial Systems (Proceedings of a symposion held in Dundee, UK, July 2006) (306).

Fischer, P.; Wunderlich, T.; Rabbel, W.; Vött, A.; Willershäuser, T.; Baika, K.; Rigakou, D.; Metallinou, G. (2016): Combined Electrical Resistivity Tomography (ERT), Direct-Push Electrical Conductivity (DP-EC) Logging and Coring - A New Methodological Approach in Geoarchaeological Research. In: Archaeol. Prospect. 23 (3), S. 213–228. DOI: 10.1002/arp.1542.

123

References

Geofabrik (2016): Open Street Map data: Mittelfranken. https://download.geofabrik.de/europe/germany/bayern/mittelfranken.html, accessed 11.04.2018.

Grimm, U.; Heinrich, J. (2019): Leipzig 1015 ce – a multiproxy study to reconstruct the palaeorelief of Leipzig’s city centre. In: Archaeological Prospection 26 (3), S. 225–237. DOI: 10.1002/arp.1736.

Hausmann, J. (2013): Parametrisation of the near surface by combined geophysical and direct push techniques in the frame of geotechnical site investigation. Dissertation. Tübingen.

124

References

He, M.; Zheng, H.; Clift, P. D.; Tada, R.; Wu, W.; Luo, C. (2015): Geochemistry of fine-grained sediments in the Yangtze River and the implications for provenance and chemical weathering in East Asia. In: Prog. in Earth and Planet. Sci. 2 (1), S. 353. DOI: 10.1186/s40645-015-0061-6.

Herzig, F.; Werther, L. (2014): Der Karlsgraben im Fokus der Dendroarchäologie. In: Ettel, P.; Daim, F.; Berg-Hobohm, S.; Werther, L. and Zielhofer, C. (Hg.): Großbaustelle 793. Das Kanalprojekt Karls des Großen zwischen Rhein and Donau. Mainz: Verl. des Römisch-Germanischen Zentralmuseums (Mosaiksteine, 11), S. 41–44.

Hesse, R. (2010): LiDAR-derived Local Relief Models - a new tool for archaeological prospection. In: Archaeol. Prospect. 79 (3), n/a-n/a. DOI: 10.1002/arp.374.

Hilgart, M. (1999): Geomorphologisch-bodenkundliche Untersuchung im Umfeld des Karlsgrabens (Fossa Carolina), Lkr. Weißenburg-Gunzenhausen. Solum, Sachverständigenbüro für angeandte Geowissenschaften, Boden- and Grundwasserschutz GmbH. Sinzing.

Hinderer, M. (2012): From gullies to mountain belts: A review of sediment budgets at various scales. In: Sedimentary Geology 280, S. 21–59. DOI: 10.1016/j.sedgeo.2012.03.009.

Hinderer, M.; Einsele, G. (2001): The world’s large lake basins as denudation-accumulation systems and implications for their lifetimes. In: Journal of Paleolimnology (26), S. 355–372.

Hoffmann, T. (2006): Modelling the Holocene Sediment Budget of the Rhine System. Dissertation. Bonn.

125

References

Kirchner, A.; Herrmann, N.; Stadtmann, R.; Lahmer, T.; Hille, L.-M.; Steinbrecher, T.; Sauerwein, M. (2020): Spatial analysis of hollow ways in the Hildesheimer Wald Mountains (Lower Saxony, Germany) as a model for mountainous regions of Central Europe. In: Erdkunde 74 (1), S. 1–14. DOI: 10.3112/erdkunde.2020.01.01.

Koch, R. (1993): Fossa Carolina - 1200 Jahre Karlsgraben. München: Denkmalpflege Informationen.

Koch, R. (1996): Neue Beobachtungen and Forschungen zum Karlsgraben. In: Jahrbuch des Historischen Vereins für Mittelfranken 97, S. 1–16.

Koch, R.; Leininger, G. (1993): Der Karlsgraben - Ergebnisse neuer Erkundungen. In: Heimatkundliche Beiträge (1), S. 6–9.

Kokalj, Ž.; Hesse, R. (2017): Airborne laser scanning raster data visualization. A guide to good practice. 1st e-ed. Ljubljana: Založba ZRC (Prostor, kraj, čas, 14).

Lacquement, C. H. (2010): Recalculating mound volume at moundville. In: Southeastern Archaeology 29 (2), S. 341–354.

Larsen, B. P.; Holdaway, S. J.; Fanning, P. C.; Mackrell, T.; Shiner, J. I. (2017): Shape as an outcome of formation history: Terrestrial Laser Scanning of shell mounds from far north Queensland, Australia. In: Quaternary International 427, S. 5–12. DOI: 10.1016/j.quaint.2015.06.066.

Lee, S.; Wolberg, G.; Shin, S. Y. (1997): Scattered data interpolation with multilevel B-splines. In: IEEE Trans. Visual. Comput. Graphics 3 (3), S. 228–244. DOI: 10.1109/2945.620490.

Leitholdt, E. (2014): Der Karlsgraben. Überwindung der Europäischen Hauptwasserscheide auf dem Wasserweg. Rekonstruktion der Kanalanlage and deren Sedimentationsgeschichte vom Frühmittelalter bis heute. Dissertation: Leipzig.

126

References

Linzen, S.; Schneider, M.; Berg-Hobohm, S.; Werther, L.; Ettel, P.; Zielhofer, C.; Schmidt, J.; Fassbinder, J. W. E.; Wilken, D.; Fediuk, A.; Dunkel, S.; Stolz, R.; Meyer, H.-G.; Sommer, C. S. (2017): From magnetic SQUID prospection to excavation - investigations at Fossa Carolina, Germany. In: Jennings, B.; Gaffney, C.; Sparrow, T. and Gaffney, S. (Hg.): 12th International Conference of Archaeoloigcal Prospection. Bradford. The University of Bradford: Archaeopress, S. 144–145.

Luo, L.; Wang, X.; Guo, H.; Lasaponara, R.; Zong, X.; Masini, N.; Wang, G.; Shi, P.; Khatteli, H.; Chen, F.; Tariq, S.; Shao, J.; Bachagha, N.; Yang, R.; Yao, Y. (2019): Airborne and spaceborne remote sensing for archaeological and cultural heritage applications: A review of the century (1907–2017). In: Remote Sensing of Environment 232, S. 111280. DOI: 10.1016/j.rse.2019.111280.

Magnani, M.; Schroder, W. (2015): New approaches to modeling the volume of earthen archaeological features: A case-study from the Hopewell culture mounds. In: Journal of Archaeological Science 64, S. 12–21. DOI: 10.1016/j.jas.2015.09.001.

McCormick, M. (2010): The origin of the European economy. Communications and commerce A.D. 300 - 900. Cambridge: Cambridge Univ. Press.

McKitterick, R. (2008): Charlemagne. The formation of a European identity. Cambridge: Cambridge Univ. Press.

Missiaen, T.; Verhegge, J.; Heirman, K.; Crombé, P. (2015): Potential of cone penetrating testing for mapping deeply buried palaeolandscapes in the context of archaeological surveys in polder areas. In: Journal of Archaeological Science 55, S. 174–187. DOI: 10.1016/j.jas.2015.01.003.

127

References

Nelson, J. (2015): Evidence in question: dendrochronology and early medieval historians. In: Kano, O. and Lemâitre, J.-L. (Hg.): Entre texte et histoire: études d’histoire médiévale offertes au professeur Soichi Sato. Paris: Éditions de Boccard, S. 227–249.

Pröschel, B.; Lehmkuhl, F. (2019): Paleotopography and anthropogenic deposition thickness of the city of Aachen, Germany. In: Journal of Maps 15 (2), S. 269–277. DOI: 10.1080/17445647.2019.1590248.

Saey, T.; Simpson, D.; Vitharana, U. W. A.; Vermeersch, H.; Vermang, J.; van Meirvenne, M. (2008): Reconstructing the paleotopography beneath the loess cover with the aid of an electromagnetic induction sensor. In: CATENA 74 (1), S. 58–64. DOI: 10.1016/j.catena.2008.03.007.

128

References

Schmidt, J.; Werther, L.; Rabiger-Völlmer, J.; Herzig, F.; Schneider, B.; Werban, U.; Dietrich, P.; Berg, S.; Linzen, S.; Ettel, P.; Zielhofer, C. (2020): Sediment budgeting of short-term backfilling processes - the colluvial collapse of a Carolingian canal construction. accepted. In: Earth Surf. Process. Landforms.

Schmidt-Kaler, H. (1993): Geologie and Landschaftsentwicklung im Rezat-Altmühl Bereich. In: Bau intern Special Issue, S. 8–10.

Schulmeister, M. K.; Butler, J. J.; Healey, J. M.; Zheng, L.; Wysocki, D. A.; McCall, G. W. (2003): Direct-Push Electrical Conductivity Logging for High-Resolution Hydrostratigraphic Characterization. In: Groundwater Monitoring & Remediation 23 (3), S. 52–62. DOI: 10.1111/j.1745- 6592.2003.tb00683.x.

Seeliger, M.; Pint, A.; Frenzel, P.; Weisenseel, P.; Erkul, E.; Wilken, D.; Wunderlich, T.; Başaran, S.; Bücherl, H.; Herbrecht, M.; Rabbel, W.; Schmidts, T.; Szemkus, N.; Brückner, H. (2018): Using a Multi-Proxy Approach to Detect and Date a Buried part of the Hellenistic City Wall of Ainos (NW Turkey). In: Geosciences 8 (10), S. 357. DOI: 10.3390/geosciences8100357.

Smedt, P. de; van Meirvenne, M.; Davies, N. S.; Bats, M.; Saey, T; Reu, J. de; Meerschman, E.; Gelorini, V.; Zwertvaegher, A.; Antrop, M.; Bourgeois, J.; Maeyer, P. de; Finke, P. A.; Verniers, J.; Crombé, P. (2013a): A multidisciplinary approach to reconstructing Late Glacial and Early Holocene landscapes. In: Journal of Archaeological Science 40 (2), S. 1260–1267. DOI: 10.1016/j.jas.2012.09.004.

129

References

Smedt, P. de; van Meirvenne, M.; Herremans, D.; Reu, J. de; Saey, T; Meerschman, E.; Crombé, P.; Clercq, W. de (2013b): The 3-D reconstruction of medieval wetland reclamation through electromagnetic induction survey. In: Scientific reports 3, S. 1517. DOI: 10.1038/srep01517.

Squatriti, P. (2002): Digging Ditches in Early Medieval Europe. In: Past & Present 176 (1), S. 11–65. DOI: 10.1093/past/176.1.11.

Stele, A. (2017): Magnetometerprospektion and magnetische Eigenschaften von braunen Plaggeneschen (Oberesch; Lechtinger Esch) and von semiterrestrischen Sedimenten (Fossa Carolina). Dissertation. Osnabrück.

Trögl, H. (2004): Weißenburg and der Main-Donau-Kanal. In: Weißenburger Blätter für Geschichte, Heimatkunde and Kultur von Stadt and Weißenburger Land (3), S. 14–25. van der Meulen, B.; Cohen, K. M.; Pierik, H. J.; Zinsmeister, J. J.; Middelkoop, H. (2020): LiDAR- derived high-resolution palaeo-DEM construction workflow and application to the early medieval Lower Rhine valley and upper delta. In: Geomorphology, S. 107370. DOI: 10.1016/j.geomorph. 2020.107370. van Loon, A. H.; Schot, P. P.; Griffioen, J.; Bierkens, M.F.P.; Wassen, M. J. (2009): Palaeo-hydrological reconstruction of a managed fen area in The Netherlands. In: Journal of Hydrology 378 (3-4), S. 205– 217. DOI: 10.1016/j.jhydrol.2009.09.014.

Verbrugghe, G.; Clercq, W. de; van Eetvelde, V. (2017): Routes across the Civitas Menapiorum. Using least cost paths and GIS to locate the Roman roads of Sandy Flanders. In: Journal of Historical Geography 57, S. 76–88. DOI: 10.1016/j.jhg.2017.06.006.

130

References

Werbrouck, I.; Antrop, M.; van Eetvelde, V.; Stal, C.; Maeyer, P. de; Bats, M.; Bourgeois, J.; Court- Picon, M.; Crombé, P.; Reu, J. de; Smedt, P. de; Finke, P. A.; van Meirvenne, M.; Verniers, J.; Zwertvaegher, A. (2011): Digital Elevation Model generation for historical landscape analysis based on LiDAR data, a case study in Flanders (Belgium). In: Expert Systems with Applications 38 (7), S. 8178–8185. DOI: 10.1016/j.eswa.2010.12.162.

Werther, L. (2014): Siedlungsentwicklung and Kulturlandschaft im Umfeld des Karlsgrabens. In: Ettel, P.; Daim, F.; Berg-Hobohm, S.; Werther, L. and Zielhofer, C. (Hg.): Großbaustelle 793. Das Kanalprojekt Karls des Großen zwischen Rhein and Donau. Mainz: Verl. des Römisch-Germanischen Zentralmuseums (Mosaiksteine, 11), S. 45–52.

Werther, Lukas (2016): Großbaustelle Karlsgraben. Eine Chaine opératoire für den Umgang mit der Ressource Holz in der Karolingerzeit. In: Mitteilungen der Deutschen Gesellschaft für Archäologie des Mittelalters and der Neuzeit 29, S. 103–112.

Werther, L. (2017): Karlsgraben doch schiffbar? Aktuelles aus der Landesarchäologie. In: Archäologie in Deutschland (5), S. 41–42.

Werther, L. (2019): Der Karlsgraben and andere Kanalbauten als Schnittstellen frühgeschichtlicher Verkehrsnetzwerke. Habilitation: Jena.

131

References

Werther, L.; Berg, S.; Ettel, P.; Linzen, S.; Schmidt, J.; Zielhofer, C. (2019): Fossa Carolina / Karlsgraben. Hg. v. Historisches Lexikon Bayern. http://www.historisches-lexikon- bayerns.de/Lexikon/Fossa_Carolina_/_Karlsgraben, last update am 26.06.2019.

Werther, L.; Kröger, L.; Kirchner, A.; Zielhofer, C.; Leitholdt, E.; Schneider, M.; Linzen, S.; Berg- Hobohm, S.; Ettel, P. (2018): Fossata Magna - a Canal Contribution to Harbour Construction in the 1st Millenium AD. In: von Carnap-Bornheim, C.; Daim, F.; Ettel, P. and Warnke, U. (Hg.): Harbours as object of interdisciplinary research. Archaeology + history + geosciences. Mainz: Verl. des RGZM (RGZM Tagungen, 34), S. 355–372.

Werther, L.; Nelson, J.; Herzig, F.; Schmidt, J.; Berg, S.; Ettel, P.; Linzen, S.; Zielhofer, C. (2020): 792 or 793? Charlemagne’s canal project: craft, nature and memory. In: Early Medieval Europe 28 (3), S. 444–465.

Zielhofer, C.; Leitholdt, E.; Werther, L.; Stele, A.; Bussmann, J.; Linzen, S.; Schneider, M.; Meyer, C.; Berg-Hobohm, S.; Ettel, P. (2014): Charlemagne’s summit canal. An early medieval hydro- engineering project for passing the Central European Watershed. In: PloS one 9 (9), e108194. DOI: 10.1371/journal.pone.0108194.

132

References

Zielhofer, C.; Wellbrock, K.; al-Souliman, A. S.; Grafenstein, M. von; Schneider, B.; Fitzsimmons, K.; Stele, A.; Lauer, T.; Suchodoletz, H. von; Grottker, M.; Gebel, H. G. K. (2018): Climate forcing and shifts in water management on the Northwest Arabian Peninsula (mid-Holocene Rasif wetlands, Saudi Arabia). In: Quaternary International 473, S. 120–140. DOI: 10.1016/j.quaint.2018.03.001.

Zielhofer, C.; Werther, L.; Dietrich, P.; Berg-Hobohm, S.; Ettel, P. (2017): Der Kanal Karls des Großen. Ein frühmittelalterliches Projekt europäischer Dimension. In: Geographische Rundschau (9), S. 12– 18.

Zolitschka, B. (1998): A 14,000 year sediment yield record from western Germany based on annually laminated lake sediments. In: Geomorphology 22 (1), S. 1–17. DOI: 10.1016/S0169-555X(97)00051- 2.

133

Appendix

A1 – Minimalinvasive Direct-oush Erkundung in der Feuchtboden(geo)archäologie am Beispiel des Karlsgrabens (Fossa Carolina)

Appendix

Minimalinvasive Direct-push-Erkundung in der Feuchtboden(geo)archäologie am Beispiel des Karlsgrabens (Fossa Carolina)

Johannes Völlmer1*, Christoph Zielhofer1, Jörg Hausmann2+, Peter Dietrich2 5, Ulrike Werban2, Johannes Schmidt1, Lukas Werther3, Stefanie Berg4

1Lehrstuhl Physische Geographie, Institut für Geographie, Universität Leipzig 2Department Monitoring- and Erkundungstechnologien, Helmholtz Zentrum für Umweltforschung, UFZ, Leipzig 3Institut für Ur- and Frühgeschichte, Friedrich-Schiller-Universität, Jena 4Bayerisches Landesamt für Denkmalpflege, München 5Zentrum für Angewandte Geowissenschaften, Eberhard-Karls-Universität Tübingen +aktuelle Anschrift: Basalt-Actien-Gesellschaft, Wiedemar *Korrespondenz: [email protected]

Kurzfassung Archäologische Grabungen in Flussauen, Feuchtböden and feuchtbodenarchäologischen Fundstätten sind aufgrund von Grundwasserzustrom and instabilem Untergrund oft teuer and schwer durchführbar. Dennoch bieten diese Standorte sehr wichtige Fundstellen and Bodenarchive. Alternativ kommen Rammkernsondierungen zum Einsatz, bei denen jedoch Kompressionseffekte in den organischen Schichten zu ungenauen Tiefenangaben führen. In dieser Studie wird der Einsatz von tiefengenauen, minimalinvasiven in situ Direct-push-Methoden zur hochaufgelösten Erkundung von Bodendenkmälern im grundwassergesättigten Bereich vorgestellt. Exemplarisch wurde dazu die Grabenfüllung der frühmittelalterlichen Fossa Carolina (Karlsgraben) untersucht. Die Fossa Carolina repräsentiert den ersten Versuch eine durchgehende Schiffsverbindung vom Rhein zur Donau zu schaffen and gilt als eines der bedeutendsten frühmittelalterlichen Bodendenkmäler in Europa. Mithilfe der hochauflösenden and tiefengenauen in-situ-Direct-push-Sondierungen (Elektrische Leitfähigkeit and Sedimentfarbe) wurden die räumlichen Dimensionen and anhand der dokumentierten Stratigraphie die Verlandungsgeschichte der ehemaligen Kanalrinne höhengenau rekonstruiert.

Abstract Archaeological excavations in floodplains and wetlands are often expensive and difficult to carry out due to groundwater inflow and an unstable subsoil. However, these locations offer very important archaeological sides and archives. Alternatively, driving core techniques lack in imprecise depth accuracy caused by high compaction rates of the organic sediments. This study presents the application of depth- accurate, minimally invasive in situ direct push sensing techniques for the high-resolution exploration of archaeological structures below the groundwater level. In this study, we reconstruct the trench fillings of C

Appendix

the early medieval Fossa Carolina (Charlemagne’s Summit Canal). The Fossa Carolina represents the first attempt to create a continuous shipping lane from the Rhine to the Danube and is considered as one of the most important early medieval ground monuments in Europe. The spatial extend and the aggradation history of the Carolingian trench were reconstructed by the application of high-resolution and depth- accurate direct push sensing technics (electrical conductivity and sediment colour).

Résumé Les fouilles archéologiques dans les plaines alluviales et les zones palustres sont souvent coûteuses et difficiles à réaliser en raison de l'afflux d'eau souterraine et des dépôts instables. Néanmoins, ces endroits offrent des sites et archives archéologiques très importants. Alternativement, carottages ne donnent pas les résultats escomptés à cause des effets de compression dans les couches organiques ces qui entraînent des profondeurs inexactes. Cette étude présente l'utilisation de méthodes d’in situ direct push, minimalement invasives et à profondeur précise pour l'exploration à haute résolution de monuments archéologiques dans la zone saturée d'eau souterraine. A titre d'exemple, le remplissage des tranchées de la Fossa Carolina (Karlsgraben) ont été examinés. La Fossa Carolina est une construction du début du Moyen Âge, qui a été initié par Charlemagne. La construction représente la première tentative de créer une connexion continue entre le Rhin et le Danube et est considérée comme l'un des monuments médiévales les plus importants d'Europe. À l'aide des sondages d’in situ direct push (conductivité électrique et couleur des sédiments), les dimensions et l’historique de l'alluvionnement de l'ancien canal ont été reconstitués à la hauteur exacte.

Einführung Besonders in Gebieten mit geringen Grundwasserflurabständen ist die Durchführung von archäologischen Grabungen an Bodendenkmälern mit vielen Schwierigkeiten verbunden. Neben hohen Kosten and großem Aufwand für die Stabilisierung der Grabungsränder (Werther and Feiner 2014) sind umfangreiche Maßnahmen zur Wasserhaltung (Bates and Bates 2000; Doran 2013) nötig. Entgegen des Leitbildes der Bodendenkmalpflege wird das Bodendenkmal durch den Eingriff häufig zerstört and Funde müssen nach der Bergung aufwendig konserviert werden. Rammkernsondierungen sind mit einem geringeren Eingriff in das Bodendenkmal verbunden and bieten eine alternative Methode zur archäologischen Grabung. Diese erfassen das Bauwerk jedoch nur punktuell and liefern aufgrund von Stauchungseffekten insbesondere bei organischen Schichten keine einwandfreien Tiefenangaben (Leitholdt et al. 2012; Leitholdt et al. 2014a; Hausmann et al. 2018). Die Anwendung von Direct-push-Methoden (wörtl. Übersetzung: „direktes Drücken“) bietet hier neue and umfangreiche Möglichkeiten zur in-situ-Erkundung durch das Eindrücken verschiedener Messsonden mit geringem Durchmesser in den Untergrund (Dietrich and Leven 2009). Die vorliegende Studie liefert einen Einblick in die flexible, minimalinvasive and kosteneffiziente Technologie für Anwendungen in der D

Appendix

Feuchtbodenarchäologie. Am Beispiel des Karlsgrabens (Fossa Carolina) soll die tiefengenaue Beschreibung der Aushubstrukturen erfolgen. Weiterhin soll geklärt werden, welche Arten von Grabenfüllungen vorhanden sind and wie die Verfüllung erfolgte. Für diese Analyse werden sowohl die verschiedenen Direct-push-Methoden als auch konventionelle Ansprachen aus Rammkernsondierungen gegenübergestellt and ausgewertet.

Historische and geographische Grundlagen Der Karlsgraben ist ein frühmittelalterlicher Kanal auf Höhe der Europäischen Hauptwasserscheide in Mittelfranken and liegt etwa 6 km südwestlich von Weißenburg (Abb. A.1a). Das Bauwerk wurde 793 n. Chr. durch Karl den Großen initiiert, wie durch die historischen Quellen belegt ist (Hack 2014; Nelson 2015). Durch die dendrochronologische Datierung (Werther et al. 2015) wurde dieser Zeitraum naturwissenschaftlich bestätigt. Der Kanal sollte als schiffbarer Verkehrsweg die Europäische Hauptwasserscheide queren and eine Verbindung zwischen der Schwäbischen Rezat (Rhein-Main- Einzugsgebiet) and der Altmühl (Donau-Einzugsgebiet) herstellen. Somit wäre eine durchgehende Schifffahrtsverbindung zwischen der Nordsee and dem Schwarzen Meer möglich gewesen. Die Wasserscheide verläuft nördlich des Treuchtlinger Talknotens als Talwasserscheide innerhalb des Schichtstufenlandes der Südlichen Frankenalb. An dieser Stelle beträgt der Abstand der beiden zu verbindenden Fließgewässer lediglich ca. 2 km (Abb. A.1b) and bietet somit eine ausgesprochen günstige Geländeposition für einen Kanalbau (Leitholdt et al. 2012; Leitholdt et al. 2014a; Zielhofer and Kirchner 2014).

Abbildung A.1a: Der Karlsgraben südlich von Nürnberg quert die Europäische Hauptwasserscheide zwischen Schwäbischer Rezat (Rhein-Main-Einzugsgebiet) and Altmühl (Donau-Einzugsgebiet). – b Die Wasserscheide verläuft dabei in einem Tallauf zwischen den Ortschaften Grönhart and Graben. Das Untersuchungsgebiet liegt im West-Ost-Bereich des Karlsgrabens. – (Grafik J. Völlmer, Leipzig; a verändert Zielhofer et al. 2014, 2 Abb. 1b; b Datengrundlage LIDAR/Bayerische Vermessungsverwaltung, OpenStreetMap). E

Appendix

Der oberflächennahe Untergrund im Bereich der Talwasserscheide ist durch spätpleistozäne sandig-tonige Talfüllungen gekennzeichnet (Schmidt-Kaler 1976). Ein geringer Grundwasserflurabstand charakterisiert den Talraum and führte zur Bildung eines Niedermoors in der Rezatniederung (Zielhofer and Kirchner 2014). Der Karlsgraben gliedert sich im Untersuchungsgebiet in verschiedene Abschnitte (Abb. A.1b). Der Zentralbereich liegt nördlich des Ortsteils Graben in der Gemeinde Treuchtlingen. An dieser Stelle wird die Wasserscheide durchbrochen, was durch die heute noch erhaltenen bis zu 10 m hohen Aushubwälle deutlich wird. Die Überreste des Bauwerks werden heute als Teich genutzt. Der West-Ost-Bereich ist vollständig verlandet, wobei ein schmaler Entwässerungsgraben in der Mitte der ca. 30 m breiten Grabenverfüllung verläuft and somit bereits auf den geringen Grundwasserflurabstand hinweist. Im West- Ost-Bereich wurde unsere Studie durchgeführt. Den dritten Abschnitt bildet die nördliche Verlängerung des Bauwerks bis zur heutigen Bahnstrecke. Hier ist der Graben beinahe vollständig eingeebnet and durch Grünland oder Ackerflächen überprägt. Die weitere Verlängerung des Karlsgrabens Richtung Weißenburg verläuft nordöstlich der Bahnstrecke and ist heute an der Oberfläche kaum noch sichtbar. Der nördliche and der nordöstliche Bereich werden östlich von der Schwäbischen Rezat and dem damit einhergehenden Rezat-Ried flankiert. Südlich des Ortes Graben beginnt die Aue der heutigen Altmühl (Koch 1996; Berg- Hobohm and Werther 2014; Zielhofer et al. 2014).

Forschungsstand Am Karlsgraben fanden seit 2010 umfangreiche geophysikalische, geoarchäologische and archäologische Untersuchungen zum karolingerzeitlichen Ausbauzustand aber auch zur nachfolgenden Verlandungsgeschichte statt. Diese kann für den West-Ost-Bereich in fünf Phasen (vgl. Zielhofer et al. 2014) gefasst werden, die auch durch Ausgrabungen im nördlichen Bereich bestätigt wurden (Werther et al. 2015): Der präkarolingische Untergrund (I) besteht aus fluvial verlagerten, organikarmen, sandig-tonigen Sedimenten des Quartärs (Schmidt-Kaler 1993; Zielhofer and Kirchner 2014), in welchen das Bauwerk angelegt wurde. Aufgrund unterschiedlicher Sohlniveaus wird eine Weihertreppenkonstruktion für den West-Ost-Bereich angenommen (Zielhofer et al. 2014). Die ersten Grabenfüllungen nach der frühmittelalterlichen Aushubphase (II) sind geprägt durch erste umgelagerte organische Reste in den initialen Verfüllungen. Bei der Datierung von Holzresten ergaben sich kalibrierte 14C-Alter im Zeitraum von 660 bis 730 n. Chr., bei denen mit Altholzeffekten zu rechnen ist (Zielhofer et al. 2014; Leitholdt et al. 2014b). Weiter im Norden wurde diese Schicht dendrochronologisch anhand von Eichenbohlen (Fälldatum) auf das Jahr 793 n. Chr. datiert (Werther et al. 2015). Darüber folgen weitere sandige Verfüllungen aus Rutschungen kurz nach dem Bau sowie feinere Sedimente frühmittelalterlicher Weiher (III) (Leitholdt et al. 2014a). Faulschlamm- and Torfbildungen weisen auf ein semiterrestrisches Milieu and Stillwasserbedingungen hin. Während des Hochmittelalters bis zum 13. Jahrhundert (IV) kam es zu F

Appendix

weiteren Torf- and Faulschlammbildungen. In der letzten Phase (V) verlandete der Bereich endgültig (Leitholdt et al. 2014a). Aufgrund von Sohltiefen der karolingerzeitlichen Baugrube deutlich über 5 m unter der rezenten Oberfläche war es bislang nicht möglich, diesen Bereich des Bauwerkes mit einer höheren Detailgenauigkeit zu untersuchen oder archäologische Grabungen durchzuführen. Besonders die genauen Dimensionen der Rinne, konstruktive Merkmale and eine höhengenaue Auflösung von Kanalsohle and Verfüllschichten konnten bislang nicht erfasst werden (Werther et al. 2015). Um dieses Defizit zu beheben, wurde seit 2014 begonnen, Direct-push-Methoden am Karlsgraben einzusetzen (Hausmann et al. 2018).

In situ Direct-push-Methoden Die Direct-push-Technologie bezeichnet generell eine Reihe an Methoden, bei der ein Hohlgestänge mit geringem Durchmesser in den oberflächennahen and nicht verfestigten Untergrund gedrückt oder gehämmert werden (Dietrich and Leven 2009; Leven et al. 2011). Es steht eine Auswahl von unterschiedlichen minimalinvasiven Verfahren and Sonden zur Verfügung. Mittels mobiler Sondiergeräte, z. B. der in dieser Studie eingesetzten mobilen Sondierraupe (Abb. A.2), werden diese räumlich flexibel and zeiteffektiv angewendet. Die hochauflösenden Messungen finden in situ statt, wobei die Daten vor Ort erstausgewertet werden, was eine schnelle Anpassung der Beprobungsstrategien ermöglicht. Es ist eine hohe Anzahl an Sondierungen in relativ kurzer Zeit durchführbar (ca. 20 - 80 Sondiermeter/Tag, je nach Methode). Eine Kombination der in situ Direct-push-Daten mit Rammkernsondierungen bzw. den daraus gewonnenen laboranalytischen Daten kann durchgeführt werden, um die Direct-push-Daten zu verifizieren (EPA 1997; Leven et al. 2011; Hausmann et al. 2016; Hausmann et al. 2018). Die ursprünglich für die Altlasten- and Baugrunderkundung entwickelten Systeme finden aufgrund flexibler Anwendungen and guter Kombinierbarkeit verstärkt auch Anwendung in der Geoarchäologie (Dalan et al. 2011; Fischer et al. 2016; Koster 2016; Hausmann et al. 2018). In der vorliegenden Studie wurden eine Sonde zur Farbmessung (Color-Logging-Tool, CLT) and eine Leitfähigkeitssonde (Electrical Conductivity, EC) eingesetzt.

Appendix

Abbildung A.2: Sondierarbeiten mit der Geoprobe 6610DT am Karlsgraben.. – (Foto F. Möckel).

Für die Studie wurde der West-Ost-Bereich des Karlsgrabens gewählt, bei dem der Grundwasserflurabstand ca. 20 cm beträgt, was die Durchführung einer Grabung in hohem Maße erschwert hätte. Weiterhin konnte mit Rammkernsondierung (Zielhofer et al. 2014) aufgrund von Kompressionseffekten der organikreichen Sedimente keine ausreichende Tiefengenauigkeit erzielt werden. Das zu untersuchende Direct-push- Transekt (Abb. A.3) wurde initial 2014 (Hausmann et al. 2018) angelegt and in den Jahren 2016 and 2017 sukzessive durch Arbeiten unserer Arbeitsgruppe erweitert. Die Durchführung der Direct-push- Sondierungen erfolgte mittels einer mobilen Sondierraupe (Geoprobe 6610DT) in Abständen von 12,5 bis 150 cm. Ausgegeben werden die Leitfähigkeitdaten mit einer Auflösung von 3 - 4 cm and die Messwerte der Farbsonde mit 1 - 2 cm, bei einer Tiefe von bis zu 10 m.

Appendix

Abbildung A.3: Überblick über die Sondierungen mittels Direct-push im West-Ost-Bereich des Karlsgrabens. – (Grafik J. Völlmer; Datengrundlage LIDAR/Bayerische Vermessungsverwaltung, Raumbezug: WGS84, GK4).

Die Farbsonde Die Farbsonde (Soil Color Optical Screening Tool, SCOSTTM, Dakota Technologies, Fargo, USA) ähnelt in ihrer Funktionsweise einer Digitalkamera. Für die Messung wird weißes Licht vom Steuergerät über ein Glasfaserkabel in die Sonde übertragen and durch ein Saphirglasfenster an der Seite der Sonde auf das Sediment gelenkt. Die Reflektion des Lichtes wird anschließend über eine weitere Glasfaser zurück zum Steuergerät übertragen, wo es über einen Detektorchip gemessen wird (Wellenlängenbereich: 350 - 1000 nm). Während des Vortriebs (ca. 2 cm/s) wird im Zeitintervall von ca. 300 ms (variabel) ein Farbwert interpoliert and die Tiefe zentimetergenau über ein Potentiometer aufgezeichnet. Somit kann über die Sondiergeschwindigkeit die Auflösung (Werte/cm) gesteuert werden. Ein Anhaften and Verschleppen von Material an der Sonde wird durch die hohe Reibung der Sonde im Sediment verhindert. Um die Vergleichbarkeit der Farbmessungen zu gewährleisten, wird vor den Messungen ein standardisierter Weißabgleich durchgeführt. Die Ausgabe der Daten erfolgt für die Vorortauswertung als RGB- Vorschaubild and für die spätere Datenverarbeitung numerisch. Die Daten werden durch die Software I

Appendix

(OST, Dakota Technologies, Fargo, USA) in den Farbsystemen Munsell, RGB and CIE XYZ (Tristimuli) bereitgestellt. In der numerischen Verarbeitung ist eine Umrechnung der Farbdaten in andere Farbsysteme für weitere Analysen möglich (Barrett 2002; Dalan et al. 2011; Hausmann et al. 2016). Die aufgenommenen Daten können dazu verwendet werden, Schichten tiefengenau and hochaufgelöst anhand ihrer Farbunterschiede zu identifizieren and ihre Eigenschaften, z. B. organische, reduzierte and oxidierte Horizonte and Verfüllungen etc., zu beschreiben (Cornell and Schwertmann 2003; Blume et al. 2010; Hartemink and Minasny 2014; Hausmann et al. 2016). Besonders die Anwendung mehrerer Farbsondierungen im Bereich eines Transekts erlauben die Rekonstruktion des horizontalen Schichtaufbaus (Hartemink and Minasny 2014; Hausmann et al. 2016; Hausmann et al. 2018).

Die Leitfähigkeitssonde Die Leitfähigkeitssonde (SC-500, Keijr Engineering Inc., Geoprobe Systems, USA) misst den spezifischen elektrischen Widerstand des Sediments vertikal entlang des Sondierkanals über ein Integral von 10 cm and gibt den Wert invers als elektrische Leitfähigkeit in mS/m wieder. Die Widerstandsmessung erfolgt mit vier in gleichmäßigem Abstand in der Sonde verbauten Elektroden in einer Wenner-α-Anordnung (Christy et al. 1994; Schulmeister et al. 2003). Auf diese Weise können Korngrößenzusammensetzungen relativ voneinander unterschieden werden, wobei tonige Sedimente in der Regel hohe and sandige Sedimente vergleichsweise niedrige Werte aufweisen (Butler et al. 1999; Schulmeister et al. 2003). Die Ergebnisse hängen jedoch nicht ausschließlich von der Korngröße ab, sondern werden generell durch Porenraum, Sättigung sowie Chemismus and Salinität des Porenfluids beeinflusst. Aufgrund der Elektrodenanordnung an der Sonde können Schichten ab einer Mächtigkeit von ca. 5 cm detektiert werden. Die Datenreihen bilden einen relativen Parameter, welcher im vertikalen Profil Schichten bzw. Schichtgrenzen aufzeigt (Butler et al. 1999; Schulmeister et al. 2002; Weidelt 2005; Lange 2005). Zur Ergänzung, Präzisierung and Validierung der Direct-push-Dateninterpretation wurden Rammkernsondierungen interpretiert and laboranalytische Ergebnisse in den Ansatz mit aufgenommen (vgl. Matney et al. 2014; Hausmann et al. 2018).

Ergebnisse and Interpretation

Aus 37 Direct-push-Farbsondierungen, 12 Direct-push-Leitfähigkeitaufnahmen and acht Rammkernsondierungen wurde ein zweidimensionaler Schnitt durch den Graben erstellt. Abbildung A.4 a-c zeigen die Ergebnisse, welche die Grundlage für die stratigraphische Rekonstruktion der Grabenfüllungen darstellen.

Appendix

Abbildung A.4: a Transekt durch den West-Ost-Bereich des Karlsgrabens: Durch die hochauflösenden Messungen der Farbsonde können die verschiedenen Sedimente and die Geometrie des Grabens tiefengenau bestimmt werden. Im unteren Teil sind die rötlichen Farben des oxidierten sandig tonigen Untergrundes zu erkennen. Darüber and an den Rändern befindet sich feineres reduziertes (graues) Sediment. Im Zentrum befindet sich eine mächtige organische Grabenfüllung. An den Rändern treten kolluviale Verlagerungen aus Richtung der Aushubwälle auf (Hausmann u. a. 2018 and eigene Daten).. –b Transekt durch den West-Ost-Bereich des Karlsgrabens: Dargestellt sind die Daten der Leitfähigkeitssonde, welche verschiedene Schichten and Fazies durch ähnliche Kurvenverläufe aufzeigen. Eingezeichnet sind sichtbare bauwerksbedingte and sedimentationsbedingte bzw. quasinatürliche Strukturen (Hausmann u. a. 2018 and eigene Daten). – c Transekt durch den West-Ost-Bereich des Karlsgrabens: Sedimentansprachen aus Rammkernsondierungen. Im Gelände wurden Bodenart, Redoxverhältnisse, Faziestypen and organische Bestandteile angesprochen bzw. abgeschätzt. Lücken in den Sondierungen sind durch n/a gekennzeichnet. – (Grafik J. Völlmer; a nach Hausmann u. a. 2018, 10 Abb. 8c; and eigene Daten, b nach Hausmann u. a. 2018, 10 Abb. 8b; and eigene Daten, c nach Zielhofer u. a. 2014; Hausmann u. a. 2018, 10 Abb. 8a; and eigene Daten).

Direct-push-Farbsondierungen Die vertikal hochauflösenden Farbsondierungen (Abb. A.4a) bringen verschiedene anthropogene and natürliche Strukturen im Untergrund zum Vorschein. Der tiefere, horizontal gelagerte Untergrund ist durch eine graue Schicht an der Basis des Schnitts charakterisiert. Im Bereich der Sondierungen KGCLT08 bis 12 befindet sich eine rötliche Fazies, welche die graue Schicht unterbricht. Darüber folgen im gesamten Transekt rötliche Schichten. Im zentralen Bereich befindet sich darüber eine zweigliedrige, dunkelbraune bis schwarze Verfüllung mit einer Mächtigkeit von ca. 2 m. Besonders auffällig ist hier im unteren Bereich eine rinnenartige Struktur (Basis 412,5 m ü. NN) mit einer Breite von ca. 2,5 m and einer Mächtigkeit von ca. 1 m. Auf dem Niveau von 413,5 m ü. NN erweitert sich die dunkle Verfüllung um ca. 3 m lateral in nördlich Richtung. Im südlichen Bereich des Transekts ist die seitliche Ausweitung weniger stark ausgeprägt. Auf die dunkelbraune bis schwarze Verfüllung folgt stratigraphisch eine grau-grüne Sedimentfolge zwischen ca. 414,5 and 416 m ü. NN. Weitere Besonderheiten bilden Verzahnungen innerhalb der Grabenfüllung. Im Norden greift ab der Höhe von 415 m ü. NN ein hellgrauer Keil in die dunkle Struktur in der Mitte. Eine weitere Einschaltung stellt ein keilförmiger hellgrauer Bereich nahe der Oberfläche dar. Ebenso findet sich im Süden nahe der Oberfläche eine hellgraue Verzahnung. Die sondierten Sedimentfarben können in verschiedene Einheiten eingeteilt werden and weisen auf natürliche and anthropogene Prozesse hin. Der geogene Untergrund (I) ist durch wechselhafte Redoxverhältnisse gekennzeichnet. Sowohl die unterste graue als auch die darüber folgende rötliche Schicht können diesem zugeordnet werden. Letztere unterliegt jedoch im oberen Bereich indirekt dem Einfluss des Baus des Karlsgrabens. Dies betrifft die Redoxverhältnisse, welche durch die Anlage des Kanals and dem damit einhergehenden Einfluss von Sauerstoff and Wasser verändert wurden. Die trogförmige Unterbrechung der grauen Schicht im zentralen unteren Bereich des Schnitts weist darauf hin.

Appendix

In der darüberliegenden grauen Schicht erfolgte der Aushub des Kanals im Frühmittelalter. Die Färbung weist auf reduzierende Verhältnisse and einen wassergesättigten Zustand hin. Es kann davon ausgegangen werden, dass der Aushub bis mindestens an die Unterkante der darüber liegenden dunkelbraunen bis schwarzen Schicht erfolgte (rote Pfeile in Abb. A.4a). Die Aushubtiefe beläuft sich innerhalb der Rinne auf ein Niveau von mindestens 412,5 m ü. NN. Deutliche Folgen des Baus stellen keilförmige kolluviale Verfüllungen (II? + III) an den Aushubrändern (orange Pfeile Abb. A.4a) dar. Die dunkelbraunen bis schwarzen Sedimente (IV) repräsentieren die karolingischen bis hochmittelalterlichen Grabenfüllungen. Bei diesen handelt es sich um organikreiche Ablagerungen, welche durch offene Stehgewässer oder unter semiterrestrischen Bedingungen entstanden sind. Die scharfe seitliche Abgrenzung der zentralen Rinne weist wahrscheinlich auf eine Stabilisierung z. B. mit Bauhölzern hin. Innerhalb der Farbsondierungen können Hölzer ebenso wie Bruchtorfe durch helle Farben angezeigt werden (rote Kreise in Abb. A.4a). Die dunkelbraunen bis schwarzen Sedimente (IV) können in zwei Schichten untergliedert werden. Die etwas hellere Färbung der oberen Schicht lässt möglicherweise auf einen leicht höheren Anteil an klastischen Sedimenten schließen. Oberhalb eines deutlich erkennbaren organikreichen Bodenhorizonts (schwarz Pfeile in Abb. A.4a), welcher die organischen Grabenfüllungen abschließt, beginnt die jüngste Sedimentationsphase (V). Das graue Sediment beschreibt wiederum reduzierende and somit wassergesättigte Verhältnisse. Aufgrund des Rückgangs organischen Materials erfolgte die Sedimentation vermutlich relativ schnell.

Direct-push-Leitfähigkeitssondierungen Ergänzend zu den Farbsondierungen lieferten die Leitfähigkeitssondierungen (Abb. A.4b) Interpretationsansätze zur Korngrößendifferenzierung innerhalb des sondierten Transekts. In der untersten Sedimentsequenz (I) finden sich heterogene Werte zwischen 10 and 40 mS/m. Die Kurvenverläufe der Leitfähigkeitswerte ähneln sich im nördlichen Teil and weisen auf eine einheitliche Schichtfolge hin. Richtung Süden wird diese Schichtfolge deutlich mächtiger. Die Werte nehmen im südlichen Bereich teils höhere and heterogenere Verteilungen an. Der Übergang zur darüberliegenden Schicht ist durch eine deutliche and abrupte Abnahme der Leitfähigkeitswerte gekennzeichnet. Im Süden sind die Werte wiederum deutlich heterogener. In der folgenden Sequenz (II-IV) wurden im Zentrum relativ konstante Werte im Bereich von ca. 50 mS/m gemessen. Der Bereich wird durch niedrige Werte eingegrenzt. Abschließend folgt nahe der Oberfläche (V) eine heterogene Fazies. Hier werden Minima dicht unter der Oberfläche sichtbar. Darunter erreichen die Messungen maximale Werte von ca. 70 bis 100 mS/m. Aus den Leitfähigkeitsdaten (Lf) kann der sedimentologische Aufbau des Grabens and der umliegenden Ablagerungen rekonstruiert werden. Der geogene Untergrund (I) ist horizontal gelagert and besteht aus einer Wechsellagerung von sandigen (Lf-Minima) and tonigeren Sedimenten (Lf-Maxima). Sandige Lagen finden sich beispielsweise an der Basis des Profils sowie am Übergang zum nächsten Abschnitt.

Appendix

Der Übergang (oberer Bereich der Lage I) zur Grabenverfüllung zeichnet sich durch niedrige Lf-Werte aus. Bei diesem Sediment handelt es sich wahrscheinlich um geogenes, sandiges Material. Ob dieses Material umgelagert wurde oder beim Bau verrutscht ist, kann nicht eindeutig geklärt werden. Die organikreiche Verfüllung des Karlsgrabens (III-IV) zeichnet sich durch gleichmäßige Leitfähigkeitswerte aus, welche auf einen homogenen Sedimenttypus schließen lassen. Der Anstieg der Leitfähigkeitswerte in der abschließenden Sedimentlage (V) weist auf eine deutlich feinere Matrix hin. Auch in dieser Sequenz können somit intern verschiedene Schichten ausgewiesen werden. Eine sandige Einschaltung ist in der mittleren Sequenz der Lage V durch geringere Leitfähigkeitswerte zu erkennen.

Rammkernsondierungen Ergänzend zu den in situ Direct-push-Daten wurden mehrere Rammkernsondierungen im Bereich des Direct-push-Schnitts durchgeführt (Abb. A.4c). Verschiedene fluviale and limnische Fazies wurden durch die Ansprache der Korngrößen, der Organikgehalte, der Redoxmerkmale and der daraus resultierenden Schicht- and Horizontabfolge identifiziert. In der ältesten Sequenz (I) unterhalb der organischen Füllung werden teils oxidierte, teils reduzierte sandig-tonige Sedimente vorgefunden. Hierbei handelt es sich um das geogene Ausgangsmaterial fluvialen Ursprungs. Die Aushubphase (II) ist gekennzeichnet durch umgelagertes Material mit ersten organischen Resten. Innerhalb der organikreichen Grabenfüllung (Sedimentlage III, IV) wurden mächtige Torflagen and vereinzelt Faulschlämme erkundet. Oberhalb (Lage V) and innerhalb der organikreichen Grabenfüllungen befinden sich feinkörnige and reduzierte Sedimente einer Stillwasserfazies. Die Ränder des Transekts weisen kolluviale Durchmischungen von sandig-tonigem sowie oxidiertem and reduziertem Material auf.

Diskussion Ausbauzustand and Verfüllung des Karlsgrabens Innerhalb des Grabens konnten verschiedenen Schichten (Abb. A.4a-c; Tab. A.1) ausdifferenziert werden, die auf unterschiedliche Ablagerungebedingungen and Prozesse zurückzuführen sind, welche schematisch in Abbildung A.5 zu sehen sind. Einen großen Einfluss bei der Verfüllung hatten Sedimentverlagerungen während and nach der Bauzeit (III, IV, V), was an den Rändern durch Verzahnungen mit dem erodierten Aushub- bzw. Wallmaterial deutlich wird. Der erodierte Aushub kann mit der in den Schriftquellen (Hack 2014) and aus archäologischen Grabungen (Werther u. a. 2015) beschriebenen Rutschungsanfälligkeit des Aushubmaterials während der Bauphase in Zusammenhang stehen. Die Farbmessungen lieferten hierzu Hinweise in Form von keilförmigen Einschüben vom Rand (Abb. A.4a, gelbe Pfeile).

Appendix

Tabelle A.1: Sedimentationsphasen aus den Abb. 4a-b anhand der in situ Direct-push- and Rammkernsondierungen and ergänzenden archäologischen Daten. Phasen Eigenschaften Interpretation Zeitspanne Zeitmarker (aus Zielhofer Farbe Leitfähigkeit überwiegendeFazies u. a. 2014; Werther u. a. 2015) V grau 70 - 100 mS/m limnische and Verlandung seit Spätmittelalter 19.-20. Jh. (Funde) umgelagerte Sedimente 1631+/- 8 n. Chr. (Dendrochonologie) 1523-1619 n. Chr. (14C, cal) 16.-17. Jh. (Hufeisen) 15.-16. Jh. (Keramik) IV braun- ca. 50 mS/m limnische Sedimente Weiher karolingerzeitlich bis 1266-1286 (14C, cal) schwarz Hochmittelalter 1022-1153 (14C, cal) 1043-1206 (14C, cal)

III grau ca. 50 mS/m limnische Sedimente Weiher karolingerzeitlich 990-1020 (14C, cal) 913-933 (14C, cal) 903-971 (14C, cal) 792-870 (14C, cal)

II grau ca. 50 mS/m limnische and maximale karolingerzeitlich 793 n. Chr. umgelagerte Sedimente Aushubtiefe and (Dendrochronologie, erste Verfüllungen Schriftquellen) 716-836 (14C, cal) 689-761 (14C, cal) 666-708 (14C, cal) I grau- 10 - 40 mS/m fluvial geogener präkarolingisch rötlich Untergrund

Die maximale Aushubtiefe wurde im Bereich zwischen 412 and 412,5 m ü. NN (Abb. A.4a, rote Pfeile) verortet. Die Tiefe der organischen Lage kann durch die Farbmessungen eindeutig identifiziert werden. Die organischen Lagen belegen eine Mindestaushubtiefe von 412,5 m ü. NN. Da sich jedoch vor der Bildung der organischen Grabenverfüllung basale kolluviale Ablagerungen durch Rückführung des Aushubmaterials gebildet haben (Zielhofer et al. 2014; Werther et al. 2015; Hausmann et al. 2018), ist eine Festlegung auf die Tiefe von 412,5 m ü. NN zwangsläufig nicht exakt. Die tatsächliche Aushubtiefe ist nach den Befunden aus den Rammkernsondierungen vermutlich 50 - 70 cm tiefer anzusetzen, was durch das Auffinden von ersten organischen Resten deutlich wird. Die Leitfähigkeitswerte zeigen in diesen basalen kolluvialen Ablagerungen ein Minimum, was auf eine sandige Fazies schließen lässt. Wären die basalen Ablagerungen durch Rutschungen bei den Bauarbeiten entstanden, so kann aufgrund der Leitfähigkeitswerte eine maximale Aushubtiefe von ca. 412 m ü. NN angenommen werden. Die sandige Schicht kann in Richtung Süden horizontal verfolgt werden, was ebenfalls auf eine geogene Sedimentation schließen lässt. Frühere Interpretationen der Direct-push-Daten in diesem Bereich (nördliches Teiltransekt: Hausmann et al. 2018), welche die Aushubtiefe am Übergang von rötlichen zu grauen Farben (411,5 m ü. NN) sahen, müssen somit einer alternativen These gegenübergestellt werden. Der Übergang zwischen dem rötlichen and grauen Sedimentbereich weist demnach eher auf eine Veränderung der Redoxverhältnisse (I zu II) hin als auf einen tieferen Grabenaushub.

Appendix

Abbildung A.5: Schematischer Ablauf des Baus and der sukzessive Verfüllung des West-Ost-Bereichs: I ungestörter, präkarolingischer Untergrund. – II Aushubphase mit veränderten Redoxverhältnissen im Bereich der Rinne and organischen Einträgen durch die Bauaktivität. – III Weiher mit fortschreitender Veränderung der Redoxverhältnisse and einsetzendem Torfwachstum. – IV Weiher and weiteres Torfwachstum. – V jüngste Verlandung. – (Grafik J. Völlmer mit Informationen aus Zielhofer et al.. 2014; Werther et al. 2015).

Die dunkelbraune bis schwarze organische Grabenfüllung (III, IV) ist auf semiterrestrische and teils limnische Sedimentation zurückzuführen. In den Rammkernsondierungen wurden entsprechend Bruchtorfe, Niedermoortorfe and Faulschlämme (Ad-hoc-AG Boden 2005; Blume et al. 2010) angesprochen. Durch die Anwendung der Direct-push-Sondierungen wurde die Ausdehnung der organischen Grabenfüllung genau bestimmt. Dabei hatte die zentrale organisch verfüllte Rinne eine Breite von ca. 2,5 m and eine Tiefe von 1 m. Falls es sich um die Fahrrinne in ihrem finalen Bauzustand handelt, hätte dies ein Passieren von größeren karolingischen Booten in eine Richtung erlaubt (Hoffmann and Ellmers 1990/1991; Bockius 2014; Zielhofer et al. 2014). Dem gegenüber stehen Ergebnisse, welche an Standorten mit einer geringeren notwendigen Aushubtiefe eine Kanalbreite von 5 - 6 m ergaben (Zielhofer et al. 2014; Werther et al. 2015). Somit müssen verschiedene Modelle vorgestellt and Indizien (Abb. A.4a) für eine schmalere oder vergleichbar breite Rinne diskutiert werden. Von besonderer Bedeutung ist dabei die Trennung von in situ befindlichen spätpleistozänen Talfüllungen and erodiertem Aushubmaterial aus eben diesen Talfüllungen: P

Appendix

Im Norden des Transekts befindet sich im Bereich der Sondierung KGCLT06 ein deutlicher Höhenversatz der organischen Verfüllung von ca. 80 cm, welche somit wesentlich tiefer reicht als in den nördlich and südlich angrenzenden Sondierungen. Wenn dieser, tiefer liegende, Bereich den nördlichen Rand der Baugrube markiert, ergibt sich eine ursprüngliche Breite der Baugrube von insgesamt von 5 - 6 m. Für die unmittelbar angrenzende graue Fazies gibt es in diesem Fall zwei Möglichkeiten: Entweder, es handelt sich um kolluvial verlagertes Aushubmaterial oder die Rinne wurde in diesem Bereich nicht fertiggestellt and Sondierung KGCLT06 hat eine gräbchenartige Vertiefung am Rand der Baugrube erfasst. Auf der südlichen Seite der zentralen Fahrrinne befinden sich in der Sondierung WOG_CLT_1 organische Verfüllungen (Abb. A.4a, bei 412,50 m ü. NN), welche von grauen Sedimenten überlagert sind. Gegebenenfalls zeigt sich auch hier ein Indiz, dass die zentrale Fahrrinne ursprünglich breiter angelegt war. Gegen eine Sohlbreite der zentralen Rinne von 5 - 6 m spricht allerdings die scharfe laterale Abgrenzung der organischen Rinnenfüllung (Abb. A.4a, Sondierung KG_CLT_21) zum angrenzenden nördlichen grauen Sediment (Abb. A.4a, Sondierung KG_CLT_08), welche eine hölzerne Böschungssicherung zwingend voraussetzt, wie sie in anderen Bereichen gefunden wurde (Werther et al. 2015) Wenn es sich nicht um ein Provisorium aus der Bauzeit handelt, sondern um einen finalen Verbau, würde dies bedeuten, dass der tiefste Teil der Baugrube and damit auch die Fahrrinne in diesem Bereich nicht breiter als 2,5 bis 3 m war. Hinweise für stratigraphisch jüngere Rückführungen von Böschungs- and Aushubmaterial zeigen sich durch dünne Lagen mit erhöhtem Organikanteil in den Rammkernbohrungen O4 (Abb. A.4c, bei ca. 414,50 m ü. NN) and O5 (Abb. A.4c, bei ca. 415,40 m ü. NN), etwa auf Höhe der Sondierung WOG_CLT_01 (Abb. A.4a) im südlichen Böschungsbereich der Baugrube. Diese Lagen zeigen an, dass es sich bei den darüberliegenden Schichten um verlagertes Aushubmaterial handelt. Gelbe Pfeile in Abbildung A.4a zeigen auffällige Positionen der Verzahnung von organischen Grabenfüllungen and erodiertem Aushubmaterial. Damit muss festgehalten werden, dass die organische Verfüllung im Zentralbereich zwar die maximale laterale Ausdehnung einer potentiell intakten Fahrrinne markiert, im Böschungsbereich allerdings nur eine Minimalausdehnung des ehemaligen Bauwerks nachzeichnet. Erst die Berücksichtigung möglicher kolluvialer Massenverlagerungen im Böschungsverlauf and ergänzende Befunde aus Rammkernbohrungen ermöglichen dort eine Erfassung der maximalen Dimensionen der Baugrube. Die Direct-push-Daten in Verbindung mit Rammkernsondierungen liefern damit für den Standort präzise Modelle für die Geometrie der Baugrube, die deutlich über den bisherigen Kenntnisstand hinausgehen.

In situ Direct-push-Sondierungen versus Rammkernsondierungen Vorteilhaft bei Rammkernsondierungen sind die umfassende and genaue Ansprache der Sedimente sowie die Möglichkeit der Probennahme bzw. daraus folgende Analysen im Labor. Hierbei kann eine gute vertikale Auflösung erreicht werden. Allerdings sind engmaschige Rammkernsondierungen mit einem hohen Arbeits- and Zeitaufwand verbunden (ca. 15 - 25 Sondiermeter/Tag) and die Tiefenangaben sind Q

Appendix

aufgrund von Sedimentverschleppungen in der Bohrkammer sowie Kompressionseffekten nicht genau (Leitholdt et al. 2012; Leitholdt et al. 2014a; Hausmann et al. 2018). Für die Zusammenstellung von Transekten mit hoher lateraler Auflösung bieten In situ Direct-push-Sondierungen deutliche Vorteile durch höhere Tiefengenauigkeit, methodische Breite, Geschwindigkeit and Kosteneffizienz (20 - 80 Sondiermeter/Tag, je nach Methode). In den vorgestellten Untersuchungen wurden zwei Sondentypen (Farbe and Leitfähigkeit) für die Erfassung eines Querschnitts verwendet. Es besteht jedoch die Möglichkeit des Einsatzes weiterer Sonden (Leven et al. 2011; Zschornakc and Leven-Pfister 2012). Beispielsweise bietet sich der Einsatz von Drucksonden zur Erfassung der Festigkeit des Untergrundes and Abschätzung der Korngrößenverteilung der Sedimente an (Robertson 1990; Lunne et al. 1997; Prinz and Strauß 2011; Vienken et al. 2012).

Das Potential der Direct-push-Methoden in der (Geo-)Archäologie Minimalinvasive geophysikalische Erkundungsmethoden spielen in der (Geo-)Archäologie bei der Untersuchung von Bodendenkmälern zunehmend eine Rolle. Sie dienen insbesondere zur Lokalisierung der Bodendenkmäler and zur näheren Erfassung ihrer Strukturen (z. B. Bates and Bates 2000; Grasmueck et al. 2004; Papadopoulos et al. 2006; Dalan et al. 2011; Fischer et al. 2016). Ziel ist es wichtige Informationen zur Tiefe, Erhaltung and Ausdehnung z. B. bei der Planung von Baumaßnahmen oder auch bei der Flächenausweisung von Bodendenkmälern (z. B. Welterbe Pfahlbauten) zu gewinnen. Die Direct-push-Sondierung steht allgemein für ein methodisch umfangreiches, minimalinvasives in situ- Messverfahren, welches nichtinvasive Erkundungstechniken (Döberl et al. 2012) sowie geringinvasive Rammkernsondierungen ergänzt. Im Gegensatz zur elektrischen Widerstandstomographie, zur Magnetik, zu seismischen Erkundungstechniken oder auch zum Bodenradar liefert die Direct-push-Sondierung allerdings hochauflösende Daten in einer exakten Tiefengenauigkeit. Direct-push-Messungen können zur Validierung von nicht- and minimalinvasiven Erkundungstechniken herangezogen werden. Hierdurch entstehen Synergieeffekte and potentiell auch mehrdimensionale Modellierungsansätze (Bates and Bates 2000; Hausmann et al. 2013; Fischer et al. 2016; Hausmann et al. 2018). Insbesondere in der Feuchtbodenarchäologie, bei der Grabungen nur erschwert oder auch unmöglich sind, bietet die Direct-push-Sondierung einen alternativen, zeit- and kosteneffizienten Ansatz. Weiterhin erlauben Direct-push-Sondierungen die Analyse eines Bodendenkmals auf verschiedenen räumlichen Skalenebenen. Bauwerke and Strukturen mit mehreren hundert Metern Ausdehnung, wie z. B. der vorgestellte Karlsgraben, können in relativ kurzer Zeit an verschiedenen Stellen untersucht werden um großskalige Erkenntnisse zu gewinnen. Die Reduzierung der Sondierabstände erlaubt die Detailuntersuchung in ausgewählten Transekten wichtiger Bereiche. Generell ist der Informationsgehalt, z. B. der Direct-push-Farbsonde, nur bedingt vergleichbar mit der umfassenden Untersuchung des Bodendenkmals durch eine detailreiche archäologische Grabung. Dennoch können durch eine zielgerichtete and flexible Kombination von Methoden (in situ Direct-push, Labor etc.)

Appendix

verschiedene Fragestellungen zum Aufbau and der Verfüllung geklärt werden, wie diese Studie belegt. Somit ergeben sich neue Möglichkeiten für bis dato nicht erkundete Standorte.

Zusammenfassung and Ausblick Über In situ Direct-push-Sondierungen wurde am Karlsgraben ein hochaufgelöstes Transekt durch die Grabensohle in einem Bereich erstellt, in dem Ausgrabungen nicht mit vertretbarem Aufwand durchführbar sind. Das minimalinvasive Verfahren bietet insbesondere in der Feuchtbodenarchäologie eine mögliche Alternative zur logistisch aufwendigen archäologischen Grabung. Die kontinuierliche Sondierung der Sedimentfarben im Bereich des Bodendenkmals ergab eine hochpräzise Rekonstruktion des Bauwerkquerschnitts inklusive der Verfüllungen. Über den Einsatz der Leitfähigkeitssonde wurden zusätzliche Informationen zur Textur gewonnen. Ergänzende Rammkernsondierungen unterstützten and validierten die Interpretation. Es lassen sich verschiedene Verfüllungsstadien rekonstruieren: An der Basis des Transekts befinden sich rötliche, horizontal gelagerte, fluviale (teils tonige) Sande (I). In dieser Sequenz wurde die maximale Aushubtiefe erreicht. Darüber folgen initiale Verfüllungen (II) aus der Bauzeit mit sandigem Material. Torf- and Faulschlammablagerungen weisen auf Weiher (III, IV) während des Früh- bis Hochmittelalters hin. Deutlich werden besonders die verfüllte Rinne, die kolluvialen Einträge von den Rändern sowie Strukturen durch Redoxveränderungen. In der Verlandungsphase (V) haben sich reduzierte, feinkörnige, limnische Sedimente während des Spätmittelalters bis zur rezenten Grünlandnutzung abgelagert. Einträge von den Aushubwällen weisen auf die Instabilität des Bauwerks bis heute hin. Die Untersuchungen zeigen einen breiten Aushub bis 413,5 m ü. NN (4 m Tiefe) and eine mindestens 2,5 m breite zentrale Rinne mit einer Sohle bei ca. 412,5 m ü. NN. Anzeichen für einen Holzverbau an den Rändern sind in den Farbsondierungen nicht direkt erkennbar, wohl aber machen abrupte laterale Farbwechsel and die vertikale Schichtgrenze einen Holzverbau wahrscheinlich. Verschiedenen Hinweise deuten eine breitere Rinne (3 - 6 m) im zentralen Bereich an and müssen diskutiert werden. Die Anwendung der Direct-push-Leitfähigkeits- and Farbsonden liefern somit einen Beitrag zur tiefengenauen Beschreibung des Bauwerks sowie zur Rekonstruktion der Verfüllungen. Die laterale Auflösung von bis zu 25 cm and die vertikale Auflösung im Zentimeterbereich zeigen wichtige Details and bieten umfangreiche Möglichkeiten zur Interpretation der Verfüllungsprozesse.

Literaturverzeichnis

Ad-hoc-AG Boden (2005): Bodenkundliche Kartieranleitung. Mit 103 Tabellen and 31 Listen. Unter Mitarbeit von Herbert Sponagel. 5., verbesserte and erweiterte Auflage. Stuttgart: E. Schweizerbart‘sche Verlagsbuchhandlung (Nägele and Obermiller).

Appendix

Barrett, L. R. (2002): Spectrophotometric color measurement in situ in well drained sandy soils. In: Geoderma 108 (1-2), S. 49–77. DOI: 10.1016/S0016-7061(02)00121-0.

Bates, M. R.; Bates, C. R. (2000): Multidisciplinary Approaches to the Geoarchaeological Evaluation of Deeply Stratified Sedimentary Sequences: Examples from Pleistocene and Holocene Deposits in Southern England, United Kingdom. In: Journal of Archaeological Science 27 (9), S. 845–858. DOI: 10.1006/jasc.2000.0584.

Blume, H.-P.; Brümmer, G.; Horn, R.; Kandeler, E.; Kögel-Knabner, I.; Kretzschmar, R.; Stahr, K.; Wilke, B.-M. (2010): Scheffer/Schachtschabel. Lehrbuch der Bodenkunde. Unter Mitarbeit von Sören Thiele- Bruhn, Gerhard Welp and Rolf Tippkötter. 16. Auflage. Heidelberg: Springer Spektrum Akademischer Verlag. http://site.ebrary.com/lib/alltitles/docDetail.action?docID=10359930.

Bockius, R. (2014): Binnenfahrzeuge im Karolingerreich. In: Ettel, P.; Daim, F.; Berg-Hobohm, S.; Werther, L. and Zielhofer, C. (Hg.): Großbaustelle 793. Das Kanalprojekt Karls des Großen zwischen Rhein and Donau. Mainz: Verl. des Römisch-Germanischen Zentralmuseums (Mosaiksteine, 11), S. 81– 86.

Christy, C. D.; Christy, T. M.; Wittig, V. (1994): A percussion probing tool for the direct sensing of soil conductivity. In: Proceedings of the 8th Outdoor Action Conference, S. 381–394.

Cornell, R. M.; Schwertmann, U. (2003): The iron oxides. Structure, properties, reactions, occurrences, and uses. 2nd, completely rev. and extended ed. Weinheim: Wiley-VCH.

Dalan, R. A.; Bevan, B. W.; Goodman, D.; Lynch, D.; Vore, S. de; Adamek, S.; Martin, T.; Holley, G.; Michlovic, M. (2011): The Measurement and Analysis of Depth in Archaeological Geophysics: Tests at the Biesterfeldt Site, USA. In: Archaeol. Prospect. 18 (4), S. 245–265. DOI: 10.1002/arp.419.

Dietrich, P.; Leven, C. (2009): Direct Push-Technologies. In: Kirsch, R. (Hg.): Groundwater Geophysics. 2. Aufl. Berlin: Springer, S. 347–366.

Döberl, G.; Müller, D.; Dörrie, T. (2012): Guideline. Section 1. In: Kästner, M.; Braeckevelt, M.; Döberl, G.; Cassiani, G.; Papini, M. P.; C. Leven-Pfister, C.; Van Ree, C. C. D. F. (Hg.): Model driven soil probing, site assessment and evaluation. Guidance on technologies. Rome.

Appendix

Doran, G. H. (2013): Excavating wet sites. In: Menotti, F. and O’Sullivan, A. (Hg.): The Oxford handbook of wetland archaeology. Oxford (Oxford handbooks in Archaeology), S. 483–494.

EPA (1997): Expedited Site Assessment Tools for Underground Storage Tank Sites: A Guide for Regulators. Direct Push Technologies - Chapter 5: United States Environmental Protection Agency, Office of Underground Storage Tanks (USTs) (EPA 510-B-97-001/1997).

Grasmueck, M.; Weger, R.; Horstmeyer, H. (2004): Three-dimensional ground-penetrating radar imaging of sedimentary structures, fractures, and archaeological features at submeter resolution. In: Geol 32 (11), S. 933. DOI: 10.1130/G20776.1.

Hack, A. (2014): Der Bau des Karlsgrabens nach den Schriftquellen. In: Ettel, P.; Daim, F.; Berg-Hobohm, S.; Werther, L. and Zielhofer, C. (Hg.): Großbaustelle 793. Das Kanalprojekt Karls des Großen zwischen Rhein and Donau. Mainz: Verl. des Römisch-Germanischen Zentralmuseums (Mosaiksteine, 11), S. 53– 62.

Hartemink, A. E.; Minasny, B. (2014): Towards digital soil morphometrics. In: Geoderma 230-231, S. 305– 317. DOI: 10.1016/j.geoderma.2014.03.008.

Hausmann, J.; Steinel, H.; Kreck, M.; Werban, U.; Vienken, T.; Dietrich, P. (2013): Two-dimensional geomorphological characterization of a filled abandoned meander using geophysical methods and soil sampling. In: Geomorphology 201, S. 335–343. DOI: 10.1016/j.geomorph.2013.07.009.

Hoffmann, P.; Ellmers, D. (1990/1991): Ein Frachter aus der Zeit Karls des Großen. In: Bremer Archäologische Blätter Neue Folge (N.F.) (1), S. 33–37.

Koch, R. (1996): Neue Beobachtungen and Forschungen zum Karlsgraben. In: Jahrbuch des Historischen Vereins für Mittelfranken 97, S. 1–16.

Koster, K. (2016): Cone Penetration Testing: A Sound Method for Urban Archaeological Prospection. In: Archaeol. Prospect. 23 (1), S. 55–69. DOI: 10.1002/arp.1531. U

Appendix

Lange, G. (2005): 5.3 Gleichstromgeoelektrik. In: Knödel, K.; Krummel, H. and Lange, G. (Hg.): Handbuch zur Erkundung des Untergrundes von Deponien and Altlasten. Band 3: Geophysik. Berlin: Springer Berlin Heidelberg, S. 128–174.

Leitholdt, E.; Krüger, A.; Zielhofer, C. (2014a): The medieval peat layer of the Fossa Carolina – Evidence for bridging the Central European Watershed or climate control? In: Zeit fur Geo Supp 58 (1), S. 189– 209. DOI: 10.1127/0372-8854/2012/S-00113.

Leitholdt, E.; Zielhofer, C.; Bussmann, J.; Stele, A. (2014b): Durchbruch der Europäischen Hauotwasserscheide. Der zentrale Bereich der Fossa Carolina - offene Wasserflächen and Verlandungsgeschichte. In: Ettel, P.; Daim, F.; Berg-Hobohm, S.; Werther, L. and Zielhofer, C. (Hg.): Großbaustelle 793. Das Kanalprojekt Karls des Großen zwischen Rhein and Donau. Mainz: Verl. des Römisch-Germanischen Zentralmuseums (Mosaiksteine, 11), S. 15–20.

Lunne, T.; Robertson, P. K.; Powell, J. J. M.; Powell, J. (1997): Cone penetration testing in geotechnical practice. London: Spon Press. https://ebookcentral.proquest.com/lib/gbv/detail.action?docID=5379100.

Matney, T.; Barrett, L. R.; Dawadi, M. B.; Maki, D.; Maxton, C.; Perry, D. S.; Roper, D. C.; Somers, L.; Whitman, L. G. (2014): In situ shallow subsurface reflectance spectroscopy of archaeological soils and features: a case-study of two Native American settlement sites in Kansas. In: Journal of Archaeological Science 43, S. 315–324. DOI: 10.1016/j.jas.2013.11.027.

Nelson, J. (2015): Evidence in question: dendrochronology and early medieval historians. In: Kano, O. and Lemâitre, J.-L. (Hg.): Entre texte et histoire: études d‘histoire médiévale offertes au professeur Soichi Sato. Paris: Éditions de Boccard, S. 227–249.

Papadopoulos, N. G.; Tsourlos, P.; Tsokas, G. N.; Sarris, A. (2006): Two-dimensional and three- dimensional resistivity imaging in archaeological site investigation. In: Archaeol. Prospect. 13 (3), S. 163–181. DOI: 10.1002/arp.276.

Prinz, H.; Strauß, R. (2011): Ingenieurgeologie. 5., bearb. and erw. Aufl. Heidelberg: Spektrum Akad. Verl. http://site.ebrary.com/lib/alltitles/docDetail.action?docID=10528437.

Robertson, P. K. (1990): Soil classification using the cone penetration test. In: Can. Geotech. J. 27 (1), S. 151–158. DOI: 10.1139/t90-014.

Appendix

Schmidt-Kaler, H. (1976): Geologische Karte von Bayern. 7031 Blatt Treuchtlingen. München.

Schmidt-Kaler, H. (1993): Geologie and Landschaftsentwicklung im Rezat-Altmühl Bereich. In: Bau intern Special Issue, S. 8–10.

Schulmeister, M. K.; Healey, J. M.; Butler Jr., James J.; McCall, G. W.; Birk, S. (2002): High-resolution characterization of chemical heterogeneity in an alluvial aquifer. In: Calibration and Reliability in Groundwater Modelling: A Few Steps Closer to Reality (Proceedings of ModelCARE2002, Prague, Czech Republic, June 2002) (277), S. 419–424.

Vienken, T.; Leven, C.; Dietrich, P. (2012): Use of CPT and other direct push methods for (hydro-) stratigraphic aquifer characterization — a field study. In: Can. Geotech. J. 49 (2), S. 197–206. DOI: 10.1139/t11-094.

Weidelt, P. (2005): 5.1 Grundlagen. In: Knödel, K.; Krummel, H. and Lange, G. (Hg.): Handbuch zur Erkundung des Untergrundes von Deponien and Altlasten. Band 3: Geophysik. Berlin: Springer Berlin Heidelberg, S. 71–100.

Zielhofer, C.; Leitholdt, E.; Werther, L.; Stele, A.; Bussmann, J.; Linzen, S.; Schneider, M.; Meyer, C.; Berg-Hobohm, S.; Ettel, P. (2014): Charlemagne‘s summit canal. An early medieval hydro-engineering project for passing the Central European Watershed. In: PloS one 9 (9), e108194. DOI: 10.1371/journal.pone.0108194.

Appendix

Zschornakc, L.; Leven-Pfister, C. (2012): Direct Push Tools fo Geophysical Measurements. In: Kästner, M.; Braeckevelt, M.; Döberl, G.; Cassiani, G.; Papini, M. P.; Leven-Pfister, C. and Van Ree, C. C. D. F. (Hg.): Model driven soil probing, site assessment and evaluation. Guidance on technologies. Rome, S. 165–174.

Appendix

A2 - Short Curriculum vitae

Personal data

Name: Johannes Schmidt Date of birth: 08.12.1988 Address 1: Erdmannstrasse 18, 04229 Leipzig Address 2: Institute of Geography, Leipzig University Johannisallee 19a, 04103 Leipzig Phone: +49 341 97 32966 Phone (mobile): +49 170 24 68 762 eMail: [email protected] ORCID: https://orcid.org/0000-0002-4584-7382

Professional career

2015 – today Scientific assistant at Leipzig University, Institute of Geography, Chair of Physical Geography (Prof. Dr. Christoph Zielhofer)

University education

2017 – today Doctoral studies (Leipzig University) Work title: “Archaeohydrological modelling of the Fossa Carolina”

2012 – 2015 Master of Science, Physical Geography/Geoecology (Leipzig University) Thesis: “Reconstruction and comparison of the development of two thermokarst depressions in Central Yakutia, Russia. Sedimentological, geochemical and statistical analysis”

2008 – 2012 Bachelor of Science, Geography (Leipzig University) Thesis: “Statistical modelling of carbon and nitrogen stocks of the organic layer in a tropical mountain rain forest in South Ecuador”

Selected memberships

Permafrost Young Researchers Network (PYRN) Young Geomorphologists German working group “Permafrost” German working group “Geoarchaeology” German working group “Geomorphology” German working group “High Mountains”

Appendix

A3 - Author publication list

Peer-reviewed

2016

Ließ, M.; Schmidt, J.; Glaser, B. (2016): Improving the Spatial Prediction of Soil Organic Carbon Stocks in a Complex Tropical Mountain Landscape by Methodological Specifications in Machine Learning Approaches. In: PloS one 11 (4), e0153673. DOI: 10.1371/journal.pone.0153673.

2017

Ulrich, M.; Wetterich, S.; Rudaya, N.; Frolova, L.; Schmidt, J.; Siegert, C.; Fedorov, A. N.; Zielhofer, C. (2017): Rapid thermokarst evolution during the mid-Holocene in Central Yakutia, Russia. In: The Holocene 27 (12), S. 1899–1913. DOI: 10.1177/0959683617708454.

2018

Appendix

2019

Ulrich, M.; Matthes, H.; Schmidt, J.; Fedorov, A. N.; Schirrmeister, L.; Siegert, C.; Schneider, B.; Strauss, J.; Zielhofer, C. (2019): Holocene thermokarst dynamics in Central Yakutia – A multi-core and robust grain-size endmember modeling approach. In: Quaternary Science Reviews 218, S. 10–33. DOI: 10.1016/j.quascirev.2019.06.010.

2020

Appendix

Other

2017

Linzen, S.; Schneider, M.; Berg-Hobohm, S.; Werther, L.; Ettel, P.; Zielhofer, C.; Schmidt, J.; Fassbinder, J. W. E.; Wilken, D.; Fediuk, A.; Dunkel, S.; Stolz, R.; Meyer, H.-G.; Sommer, C. S. (2017): From magnetic SQUID prospection to excavation - investigations at Fossa Carolina, Germany. In: Benjamin Jennings, Christopher Gaffney, Thomas Sparrow and Sue Gaffney (Hg.): 12th International Conference of Archaeoloigcal Prospection. Bradford. The University of Bradford: Archaeopress, S. 144–145.

2019

Berg, S.; Ettel, P.; Herzig, F.; Linzen, S.; Schmidt, J.; Werther, L.; Zielhofer, C. (2019): Römische Daten fehlen völlig. Stellungnahme zum Karlsgraben-Beitrag. In: Bayerische Archäologie (4), S. 44.

Werther, L.; Berg, S.; Ettel, P.; Linzen, S.; Schmidt, J.; Zielhofer, C. (2019): Fossa Carolina / Karlsgraben. Hg. v. Historisches Lexikon Bayern. http://www.historisches-lexikon-bayerns.de/Lexikon/ Fossa_Carolina_/_Karlsgraben, last update 26.06.2019.

Appendix

A4 - Paper Contributions

Total contribution: 95 %

Literarture review and conceptualization: 95 % Data acquisition and analysis: 85 % Interpretation and discussion: 90 % Writing original draft and coordination of the final manuscript: 95 %

Total contribution: 95 %

Literarture review and conceptualization: 95 % Data acquisition and analysis: 85 % Interpretation and discussion: 90 % Writing original draft and coordination of the final manuscript: 95 %

Total contribution: 90 %

Literarture review and conceptualization: 90 % Data acquisition and analysis: 75 % Interpretation and discussion: 85 % Writing original draft and coordination of the final manuscript: 95 %

Appendix

Total contribution: 20 %

Literature review and conceptualization: 20 % Data acquisition and analysis: 15 % Interpretation and discussion: 20 % Writing original draft and coordination of the final manuscript: 10 %

Appendix

A5 – Declaration of Originality / Eigenständigkeitserklärung

I hereby confirm that I have independently prepared the dissertation I have submitted, that I have given full details of the sources and aids used and that I have marked the places where the work - including tables, maps and illustrations - was borrowed from other works, either in its wording or in its sense, in each individual case; that this dissertation has not yet been submitted to any other faculty or university for examination; that it has not yet been published, apart from the partial publications listed below; and that I will not publish it before the end of the doctoral procedure. I am also aware of the provisions of the doctoral regulations. The dissertation submitted by me has been supervised by Univ. Prof. Dr. Christoph Zielhofer.

The following sub-publications do exist:

Völlmer, J.; Zielhofer, C.; Hausmann, J.; Dietrich, P.; Werban, U.; Schmidt, J.; Werther, L. Berg, S. (2018): Minimalinvasive Direct-push Erkundung in der Feuchtboden(geo)archäologie am Beispiel des Karlsgrabens (Fossa Carolina). In: Archäologisches Korrespondenzblatt 48 (4), S. 577–608.

Leipzig, 28.09.2020 Johannes Schmidt

Appendix

A6 - Bibliographic information

Schmidt, Johannes Quantification and modelling approaches of geoarchaeological processes – The course, construction and collapse of the Carolingian canal Fossa Carolina Universität Leipzig, Dissertation 188 S., 192 Lit., 41 Abb., 12 Tab., 5 Anhänge (inkl. 1 CD)

Referat

Die Fossa Carolina ist ein frühmittelalterlicher Kanalbau, der 792/793 n. Chr. im Auftrag Karls des Großen gebaut wurde. Er sollte die Europäische Hauptwasserscheide überbrücken, die das Untersuchungsgebiet zwischen den Städten Weißenburg und Treuchtlingen in Bayern, in ein Rhein-Main-System und in ein Donau-System teilt. Da die Schifffahrtswege im Frühmittelalter sowohl für den Güter- und Personenverkehr als auch für militärische Zwecke genutzt wurden, war ein Kanal von höchstem geostrategischen Interesse. Bisher veröffentlichte Studien befassten sich vor allem mit der qualitativen Rekonstruktion und Beschreibung der Kanalreste. Die vorgestellte Dissertation konzentriert sich hingegen auf Modellierungs- und Quantifizierungsansätze zur Generierung numerischer Werte verschiedener geoarchäologischer Prozesse, wie z.B. die Landschaftsabhängigkeit des Kanalverlaufs, die dreidimensionale Kanalrekonstruktion und anschließende Berechnung der beim Bau bewegten Erdvolumina sowie die Sedimentbudgetierung von Kollapssedimenten des Kanals.

Abstract

The Fossa Carolina is an early medieval canal construction built in 792/793 AD on order of Charlemagne. It should bridge the Central European Watershed, which divides the study area between the cities of Weißenburg and Treuchtlingen in Bavaria, in the Rhine-Main System and the Danube System. Because early medieval shipping routes were used for the movement of goods and people as well as for military purposes, a canal was of highest geostrategic interest. Published studies mainly dealt with the qualitative reconstruction and description of the canal remnants. The presented Phd-thesis focuses on modelling and quantification approaches to generate numerical values of different related geoarchaeological processes, such as the landscape dependence of the canal course, the three-dimensional canal reconstruction and subsequent calculation of the earth volumes moved during the construction, as well as the sediment budgeting of collapse sediments of the canal.