Durham E-Theses
Reconstruction of Holocene sea-levels using diatom- and pollen-based microfossil transfer functions, west coast of Scotland, UK
Gregory, Caroline
How to cite: Gregory, Caroline (2007) Reconstruction of Holocene sea-levels using diatom- and pollen-based microfossil transfer functions, west coast of Scotland, UK, Durham theses, Durham University. Available at Durham E-Theses Online: http://etheses.dur.ac.uk/2916/
Use policy
The full-text may be used and/or reproduced, and given to third parties in any format or medium, without prior permission or charge, for personal research or study, educational, or not-for-pro t purposes provided that:
• a full bibliographic reference is made to the original source
• a link is made to the metadata record in Durham E-Theses
• the full-text is not changed in any way
The full-text must not be sold in any format or medium without the formal permission of the copyright holders.
Please consult the full Durham E-Theses policy for further details. Academic Support O ce, Durham University, University O ce, Old Elvet, Durham DH1 3HP e-mail: [email protected] Tel: +44 0191 334 6107 http://etheses.dur.ac.uk
2 Reconstruction of Holocene sea-levels using diatom- and pollen-based microfossil transfer functions, west coast of Scotland, UK
Volume 2: Figures, Tables, Plates and Appendices
The copyright of this thesis rests with the author or the university to which it was submitted. No quotation from it, or information derived from it may be published without the prior written consent of the author or university, and any information derived from it should be acknowledged.
Caroline Gregory
Thesis submitted for the degree of Doctor of Philosophy
Department of Geography Durham University
August 2007
= 4 JUN ZUOB LIST OF CONTENTS
Volume2 Page No LIST OF FIGURES LIST OF TABLES vi LIST OF PLATES viii APPENDICES ix
LIST OF FIGURES
CHAPTER 1 -Introduction 1 Figure 1.1 Quadratic trend surface showing isobases (in metres) 1 on the Main Postglacial Shoreline (c. 6.5 ka BP) in Scotland (Firth eta/., 1993). Figure 1.2 Observations and model predictions of relative sea-level 2 change 16000 cal yr BP to present. Figure 1.3 Cross section of a sediment profile illustrating the data 3 used to reconstruct RSL using SLI points in comparison to a transfer function which can be applied to the whole fossil sequence.
CHAPTER 2 - Field Sites 4 Figure 2. 1 Location of the local sample areas on the northwest of 4 Scotland. The approximate location of the study areas at Kintail, Morar, Appin and Knapdale are shown as rectangles and labelled accordingly. Figure 2.2 The Kintail area and the location of the field sites 5 Nonach, Eilean nan Gall and Balmacara. Figure 2.3 The Nonach field site. The location of the sample 6 transect is shown on the tidal marsh as a red line. Figure 2.4 Eilean nan Gall field site. The location of the sample 7 transect is shown on the tidal marsh as a red line. Figure 2.5 Balmacara field site. The location of the sample transect 8 is shown on the tidal marsh as a red line.
Figure 2.6 The Morar area and the location of the Sa ideal nan 9 Ceapaich field site. Figure 2.7 The Saideal nan Ceapaich field site. The location of the 10 sample transect is shown on the tidal marsh as a red line. Figure 2.8 The Appin area and the location of Loch Creran field site. 11 Figure 2.9 The Loch Creran field site. The location of the sample 12 transect is shown on the tidal marsh as a red line. Figure 2.10 The Knapdale area and the location of the Barr-na-Criche 13 field site. Figure 2.11 The Barr-na-Criche field site. The location of the sample 14 transect is shown on the tidal marsh as a red line. Figure 2. 12 The Nonach field site. The location of the fossil core NN04 15 is shown as a red diamond and annotated accordingly. Figure 2.13 Location of monoliths MMN01A & B at Mointeach Mhor 16 North, Morar. Figure 2. 14 The Barr-na-Criche field site. The location of the fossil 17 core BC006a is shown as a red diamond and annotated accordingly.
Chapter 4 - Results: development of a transfer function Figure 4.1 Regional diatom training set 22 Figure 4.2 CCA Ordination diagram for samples and environmental 23 Variables for the regional diatom training set. Figure 4.3 Regression results for the regional diatom training set 24 using WAPLS components 1 - 3 are shown on the left and the associated regression residuals are shown on the right. Figure 4.4 Kintail diatom training set 25 Figure 4.5 Morar diatom training set 26 Figure 4.6 Argyll diatom training set 27 Figure 4.7 CCA Ordination diagram for samples and environmental 28
11 variables for the diatom training sets. Figure 4.8 Regression results for the Kintail diatom training set 29 using WAPLS component 3 are shown on the left and the associated regression residuals are shown on the right. Figure 4.9 Regression results for the Morar diatom training set 29 using WAPLS component 3 are shown on the left and the associated regression residuals are shown on the right. Figure 4.10 Regression results for the Argyll diatom training set 30 using WAPLS component 3 are shown on the left and the associated regression residuals are shown on the right. Figure 4.11 Regional pollen training set 31 Figure 4.12 Kintail pollen training set 32 Figure 4.13 Morar pollen training set 33 Figure 4.14 Argyll pollen training set 34 Figure 4.15 CCA Ordination diagram for samples and environmental 35 variables for the pollen training sets. Figure 4.16 Regression results for the regional pollen training set 36 using PLS component 2 are shown on the left and the associated regression residuals are shown on the right. Figure 4.17 Regression results for the Kintail pollen training set using 36 PLS component 3 are shown on the left and the associated regression residuals are shown on the right. Figure 4.18 Regression results for the Morar pollen training set using 37 PLS component 3 are shown on the left and the associated regression residuals are shown on the right. Figure 4.19 Regression results for the Argyll pollen training set 37 using WAPLS component 3 are shown on the left and the associated regression residuals are shown on the right. Figure 4.20 Regional multi-proxy training set 38 Figure 4.21 Kintail multi-proxy training set 39 Figure 4.22 Morar multi-proxy training set 40 Figure 4.23 Argyll multi-proxy training set 41 Figure 4.24 CCA Ordination diagram for samples and environmental 42
lll variables for the multi-proxy training sets. Figure 4.25 Regression results for the regional multi-proxy training 43 set using WAPLS component 3 are shown on the left and the associated regression residuals are shown on the right. Figure 4.26 Regression results for the Kintail multi-proxy training 43 set using WAPLS component 3 are shown on the left and the associated regression residuals are shown on the right. Figure 4.27 Regression results for the Morar multi-proxy training set 44 using WAPLS component 3 are shown on the left and the associated regression residuals are shown on the right. Figure 4.28 Regression results for the Argyll multi-proxy training set 44 using WAPLS component 2 are shown on the left and the associated regression residuals are shown on the right.
Chapter 5 - Results: reconstruction of former sea-levels Figure 5. 1 Summary of the results of the diatom RWL reconstructions 70 for core NN04. Figure 5.2 Summary of the results of the pollen RWL reconstructions 71 for core NN04. Figure 5.3 Summary of the results of the multi-proxy RWL reconstructions 72 for core NN04. Figure 5.4 Scatter plots comparing the regional diatom reference water 73 level reconstruction with the local reconstructions for core NN04. Figure 5.5 Scatter plots comparing the regional pollen reference water 74 level reconstruction with the local reconstructions for core NN04. Figure 5.6 Scatter plots comparing the regional diatom and pollen reference 75 water level reconstruction with the local reconstructions for core NN04. Figure 5. 7 Summary of the results of the diatom RWL reconstructions 76 for monolith MMN01A & B. Figure 5.8 Summary of the results of the pollen RWL reconstructions 77 for monolith MMN01A & B. Figure 5.9 Summary of the results of the multi-proxy RWL reconstructions 78 for monolith MMN01A & B.
lV Figure 5.10 Scatter plots comparing the regional diatom reference water 79 level reconstruction with the local reconstructions for monoliths MMN01A & B. Figure 5.11 Scatter plots comparing the regional pollen reference water 80 level reconstruction with the local reconstructions for monoliths MMN01 A & B. Figure 5.12 Scatter plots comparing the regional diatom and pollen reference 81 water level reconstruction with the local reconstructions for monolith MMN01 A & B. Figure 5.13 Bi-plots comparing the RWL reconstruction values produced for 82 the regional diatom, pollen and multi-proxy training sets using only the samples within the multi-proxy training set to allow direct comparison. Figure 5.14 Summary of the results of the pollen RWL reconstructions 83 for core BC006a. Figure 5.15 Scatter plots comparing the regional pollen reference water 84 level reconstruction with the local reconstructions for core BC006a.
Chapter 6 - Discussion Figure 6.1 Summary of the results of the diatom-inferred reference water 90 level reconstruction for fossil monolith sample MMN01A&B. Figure 6.2 Summary of the results of the pollen-inferred reference water 91 level reconstruction for fossil monolith sample MMN01A&B. Figure 6.3 Summary of the results of the multi-proxy-inferred reference water 92 level reconstruction for fossil monolith sample MMN01A&B.
v LIST OF TABLES
CHAPTER 2 - Field Sites Table 2.1 Summary of the tide levels from the Admiralty Tide 18 Tables (1986) transformed into m OD by Shennan (pers comm.) for the ports nearest to the field sites.
CHAPTER 4 - Results: development of a transfer function 45 Table 4.1 Detrended Canonical Correspondence Analysis results 45 Table 4.2 Correlation matrices for environmental variables from 45 the regional diatom training set Table 4.3 CCA results showing summary of eigenvalues for the 45 Regional diatom training set. Table 4.4 Summary of the results of CCA analysis listing the 46 percentage of explained and unexplained variance. Table 4.5 Summary of the transfer function performance statistics for 46 the regional diatom training set. WA-PLS component 3 is the chosen model. Table 4.6 List of? values from other studies. 47 Table 4. 7 Correlation matrices for environmental variables for diatom 48 training sets. Table 4.8 CCA results showing summary of eigenvalues for the diatom 49 training sets. Table 4.9 Summary of the transfer function performance statistics for 50 the diatom training sets for each model chosen. The results are shown in full in appendix 4.1. Table 4.10 Correlation matrices for environmental variables for the 51 pollen training sets. Table 4.11 CCA results showing summary of eigenvalues for the 52 pollen training sets. Table 4.12 Summary of the performance statistics for the chosen 53 transfer function model for each of the pollen training sets. Table 4.13 Correlation matrices for environmental variables from the 54 multi-proxy training sets. Table 4.14 CCA results showing summary of eigenvalues for the 55 multi-proxy training sets. Vl Table 4.15 Summary of the performance statistics for the chosen 56 transfer function model for each of the multi-proxy training sets. Table 4.16 Summary of the performance statistics for the chosen 57 transfer function models for each of the training sets. Table 4.17 Summary of the performance statistics for the most 58 appropriate models for the multi-proxy training set and diatom and pollen training sets comprising just the diatom and pollen samples present in the multi-proxy training set.
Chapter 5- Results: reconstruction of former sea-levels Table 5.1 Lists the number of microfossils recorded in the fossil cores and 85 the percentages of microfossils also found in the modern training set(% of modern analogues). Table 5.2 Lists the number of microfossils recorded in the fossil cores and 86 the percentages of microfossils also found in the modern training sets(% of modern analogues) for the diatoms and pollen samples within the multi-proxy training sets.
Chapter 6 - Discussion Table 6.1 For each transfer function model RMSEP in SWLI units have 93 been converted into mOD Table 6.2 For each transfer function model RMSEP in SWLI units has 94 been converted into m OD for the training sets that contain only samples within the multi-proxy training sets. Table 6.3 Arisaig sea-level index point. 95 Table 6.4 Shows the results of the diatom, pollen and multi-proxy RWL 96 Reconstructions and associated error terms and the RWL and error terms for the sea-level index points from Mointeach Mhor North (after Shennan eta/., 2005). Table 6.5 Shows the results of the diatom and pollen RWL reconstructions 97 For the samples within the multi-proxy training set and associated error terms and the RWL and error terms for the sea-level index points from Mointeach Mhor North (after Shennan eta/., 2005).
Vll LIST OF PLATES Page No. CHAPTER 2 - Field Sites Plate 2.1 Eilean nan Gall field site. The location of the sample 19 Transect is shown as a red line on the photographs. Plate 2.2 Loch Creran field site. The location of the sample 20 Transect is shown as a red line on the photographs. Plate 2.3 Barr-na-Criche field site. The location of the sample 21 Transect is shown as a red line on the photograph.
Vlll APPENDICES Page No. 4. 1 Transfer function results. 59 5.1 Stratigraphy. 87
IX Figures, Tables & Plates - Chapter 1
Figure 1.1 Quadratic trend surface showing isobases (in metres) on the Main Postglacial Shoreline (c. 6.5 ka BP) in Scotland (Firth eta/., 1993). Figures, Tables & Plates - Chapter 1
60
50
40 -11-
....J (/) IY 30 E • ••- 1•\odltl A - Modt-18 20
10
0 2000 4000 6000 8000 10000 12000 14000 16000 18000 cat yr BP
Figure 1.2 Observations and model predictions of relative sea-level change 16000 cal yr BP to present. Age error bars for sea-level index points indicate the maximum and minimum calibrated ages and the vertical bar depicts the median calibrated age. The vertical error calculation includes the sum of all the quantified or estimated height errors (see table 1 in Shennan eta/., 2003). Relative sea-level must lie at or below limiting dates, shown as solid symbols • · Model predictions come from: (A) the modified version of Lambeck (1995) described in Shennan et at. (2000); and (B) the modified version of Peltier et at. (2002) described as model 3 in Shennan et at. (2002).
2 Figures, Tables & Plates - Chapter 1
Sea-level index points
Intertidal sand, the sediment that records the peak of marine influence.
Figure 1.3 Cross section of a sediment profile illustrating the data used to reconstruct RSL using SLI points in comparison to a transfer function which can be applied to the whole fossil sequence (vertical black lines represent a sediment core from which microfossil data can be used in a transfer function).
3 Figures, Tables & Plates - Chapter 2
g1n • • Inverness Kintail D
Morar LJ·1allaig Kingussie
Fort William Dundee AppinD • Oban Perth• Stirting •
() KnapdaleO Glasgow EDINBURGH
Jedburgh
/
Figure 2.1. Location of the local sample areas on the northwest coast of Scotland. The approximate location of the study areas at Kintail, Morar, Appin and Knapdale are shown as black rectangles and labelled accordingly.
4 Figures, Tables & Plates - Chapter 2
Scotland
Isle of Skye Eilean nan
Figure 2.2 The Kintail area and the location of the field sites Nonach, Eilean nan Gall and Balmacara.
5 Figures, Tables & Plates - Chapter 2
CreagDh~ j ~ ~
Q ~p $ Creag a' Chaisil Q Q $ Q $ ,, $
Q Q 0 __?; Q 0 0
0 1km
Figure 2.3. The Nonach field site. The location of the sample transect is shown on the tidal marsh as a red line.
6 Figures, Tables & Plates - Chapter 2
.. ·: .. ·:
.. ·.·. .', .. , . f •
Loch Dulch
~ 11111,
.till I I
"""·
Jetty .Uilt ~
0 1km
Figure 2.4. Eilean nan Gall field site. The location of the sample transect is shown on the tidal marsh as a red line.
7 Figures, Tables & Plates - Chapter 2
$ ~ ollllf •
...... --.- . ·.-..· -',·...·.., ....
ML
LochAish
0------1km
Figure 2.5. Balmacara field site. The location of the sample transect is shown on the tidal marsh as a red line.
8 Figures, Tables & Plates - Chapter 2
N l
Loch ,.n Uamh
Figure 2.6. The Morar area and the location of the Sa ideal nan Ceapaich field site.
9 Figures, Tables & Plates - Chapter 2
0 0 Keppoch 0 0 0
Sa/deal • • CNJM/Ch • • • / ,-(--::--~-- ) Q,,,/------' ' / / -- ' ~~\ . / ------"\ ,r-.y U' ) v.)(..- Morroch0 \\,, ~ . 0 Loch nan Ceall I • ~ 0 I I • 0 0 0 0 ~-~· 0 0 ~ -o.J . 0 0 Strath of Arisaig Morroc:h PoJt • ,. 0 0 1- ,. 0 / \ " \ D 0 0 /' 0 1 km
Figure 2.7. The Saideal nan Ceapaich field site. The location of the sample transect is shown on the tidal marsh as a red line.
10 Figures, Tables & Plates - Chapter 2
APPIN
Figure 2.8. The Appin area and the location of Loch Creran field site.
11 Figures, Tables & Plates - Chapter 2
0
0&. 0 00 D
MLWS
,, ,,,, Loch Creran 'I '
0 1km
Figure 2.9. The Loch Creran field site. The location of the sample transect is shown on the tidal marsh as a red line.
12 Figures, Tables & Plates - Chapter 2
N • Cretshengan l KNAPDALE
Sound of Buts
4/' Loch Ciaran
8642 Sound of Glgha .., Ballochroy KINTYRE Crossal{(l I
1 ' Isle of Arran I Rhunahaorine
Figure 2.10. The Knapdale area and the location of the Barr-na-Criche field site.
13 Figures, Tables & Plates- Chapter 2
.,u .. .
c..;...... oiiiii, Q \. olllllo
L::::::;:.../ o I III J
• llltt •
.till It '"'"'(;./ Muir Achaidli''"
o1 lit I, ~d o ...... ((~ . 0 0 ·"'"' ~Barr ·"'". ( Reamhar tlllllt l. _,p '-""'' '"'"'
West Loch Tarbert
0 0.5km
Figure 2.11. The Barr-na-Criche field site. The location of the sample transect is shown on the tidal marsh as a red line.
14 Figures, Tables & Plates - Chapter 2
N
I c 4• ' 0 $0 j ...... , ~ Creag a' Chaisil ·. '~ Loch Long ...... • •• 0 $ Q !fl
0 0
Figure 2.12. The Nonach field site. The location of the fossil core NN04 is shown as a red diamond and annotated accordingly.
15 Figures, Tables & Plates - Chapter 2
Mointeach Mhor
\ ... -., \ \ \ \ '
t;? BldtJal na Ceapaich
0 2km
Figure 2.13 Location of monoliths MMN01A & Bat Mointeach Mhor North, Morar.
16 Figures, Tables & Plates - Chapter 2
olllllo
olllllt \. .ollll. ~ ·•1111 !!IIIIo 11 ... u, Fossil core ·"' • dllll. BC006a .. 1111. /. dllll. Q (I dllll.?) • f {;.)-' t' dllllo 11 Muir Achaidft' '' S":rr a' IIIII II c .tllllt Achadh-~ .tllllo Chaorann .11111, dlllt. ~d o Q ~ .11111,((~ . 0 Q olllllt .11111. I; Barr
., 1111 • ( Reamhar tlllllt L Jf> 1...11111. dllll.
olllll1
Rubh' a' Bharr Ruaidh West Loch Tarbert
0 0.5km
Figure 2.14. The Barr-na-Criche field site. The location of the fossil core BC006a is shown as a red diamond and annotated accordingly.
17 Figures, Tables & Plates - Chapter 2
Nearest tidal Fie'ld Site LAT LON MHWS MHWN MLWN MLWS MTL ' prediction I
Balmacara Kyle of Lochalsh 5717 543 2.57 1.17 -0.53 -1.93 0.32 Eilean nan Gall Dornie Bridge 5717 531 2.62 1.12 -0.58 -1.98 0.3 Nonach Dornie Bridge 5717 531 2.62 1.12 -0.58 -1.98 0.3 Saideal nan Mallaig 5700 550 2.38 1.18 -0.52 -1.92 0.28 Ceapaich Loch Creran Port Appin 5633 525 2.25 1.15 -0.05 -1.15 0.55 East Loch Barr-na-Criche 5552 524 1.78 1.28 -0.52 -1.32 0.31 Tarbert
Table 2.1. Summary of the tidal predictions for the ports nearest to the field sites. The tidal predictions have been transformed into mOD from the Admiralty Tide Tables (1986) by Shennan (pers comm.).
18 Figures, Tables & Plates - Chapter 2
Plate 2.1 Eilean nan Gall field site. The location of the sample transect Is shown as a red line on the photographs.
19 Figures, Tables & Plates - Chapter 2
Plate 2.2 Loch Creran field site. The location of the sample transect Is shown as a red line on the photographs.
20 Figures, Tables & Plates - Chapter 2
Plate 2.3 Barr-na-Criche field site. The location of the sample transect is shown as a red line on the photograph.
21 Figures, Tables & Plates - Chapter 4
• I F' ......
I 1!11" I • = : . :" . I : • 0 :. t I ... I I I I L " ------=- . • -- I I I .. - : ii_: : 1!1 - • I . !!!!!!:"'" !!!!: - - - . -· -r;~~r-;rr-;"",....,,,...... ,."'....,"'""~~~~~~""'";~'"~~~r-o;;"f""''"l"~~~--~--~~~'""'"!-"!"'"!:""'~'-;,"'-;r-:-'~~r-;, ll 40 ro ~~~~«l ....-;~~--;,~ ...... ~,.....:J)r-;;"~~r-;, 'l.•bundanc• of dla1oms
Figure 4.1 Regional diatom training set, dominant diatom species exceeding 10% of the total are shown. Species are ordered according to their salinity tolerance. The diatom samples are ordered on the y-axis according to the elevation (SWU) of the assemblage and the histograms show the percentage abundance of each of the diatom species.
22 Figures, Tables & Plates - Chapter 4
F aw Sl 0 ~ I --&- i _ .,..A .. ~· ····-·· '":-:~~ ....• -···· ~ -."'f i -- . ~ a ' -~ ! ... - -pH i s. .,
0 T
,...~ I ~------~------1 -1 .0 1.0
Key: SW - Standardised Water Level Index LOI - Loss on Ignition Ph -pH sa - Salinity Sa -Sand Si -Silt Cl -Clay
Figure 4.2 CCA Ordination diagram for samples and environmental variables for the regional diatom training set.
23 Figures, Tables & Plates - Chapter 4
300 100 ... 80 ~.. 250 ...;:- 5i ••• 5i 60 c •••• c &. 200 &. 40 E E 0 20 ' 2. .~·· 150 e. .!!! 0 i Ia (/) #f.;#: :II 50 2 300 :2 100 -20 a::~ I (/) -40 ..J ~· 11. . ~ 50 -60 • • ~ ~ -80 . 0 0 100 200 300 -100 Observed SWLI WAPLS Residuals Observed SWLI
300 100 80 N'... 250 N"... 5i c 60 c Gl ~ :J}/;• 40 • &. 200 &. E E 0 0 20 0 2. 150 • .!!! i Ia 0 :I 50 300 4*~ :2 -20 ·' XI 1100 ~·~~ a:: (/) -40 ·; ..J w 50 11. -60 ~ -80 0 0 100 200 300 -100 Observed SWLI WAPLS Residuals Observed SWLI
300 100 80 ;:;... 250 r;; i · ~ ..5i 60 c . c 8. 200 &. 40 • E E 8 ~~· 0 20 . 2. 150 .!!! i Ia 0 :I 50 300 :2 -20 f( XI 1100 a:: ~·~· (/) -40 .~ ..J 11. ~ 50 -60 •• ~ • -80 0 0 100 200 300 -100 Observed SWU WAPLS Residuals Observed SWLI
Figure 4.3 Regression results for the regional diatom training set using WAPLS components 1 - 3 are shown on the left and the associated regression residuals are shown on the right. 24 Figures, Tables & Plates- Chapter 4
I l I I I I . I ~~ .. . , .. . . I : ' - : I I ; I I . . I . - ... .I I . I . . I I .. I . I . ··- -· .. ·- ,. I I .I I I . . __ . . I . I .I I - I - I I I . I I . .. . ' r: . . - - ·I • L.lo :.I I I ~ • • -...... I .I • • . ' I I I I I -· I I . : .I .0 I . :. ! I .I . I .I . . .. - . I : .. . . I .. ' . . . . ' . . . .I . - • .. • ' • . • i ·· ·· l:'"r ~t I " : -· I ; ; I . L r .. -~ ~ r •r I • I I ~ I I I : • • - . . . • f . I • r • : r::- ,.. : · r :. 1: ; ' . I t I I - ' .I I I 1 I 1 ' 1 I .. • I • • lao .. ~ I. it I - I •• --· · -r- .r• - · . r I. ·-· ,... . I I I. • I ...... I • • lo ' l.ooo '
,..,..,...,.. ,...... ,...,,..,..,..,.,,.,,..,...... ,.. ,..,..,,..,,..,l"""i"'T'I""T'...... ,..,...""1f""lii",....,...,..,...,...,...,...... ,,...,...,..,...,..,.,,..,.,,..,..'""1i"i",..,..,..,...,.,,....,...... I"""Ji'...... ,.., IJ) lj iblil> ,..,..,..,..,.. ,..,...~....,...... ,,..,...,.. ,..,.. '"'"i"'l.,...... ,.. .. ~., ......
Figure 4.4 Kintail diatom training set, dominant diatom species exceeding 2% of the total are shown. Species are ordered according to their salinity tolerance. The diatom samples are ordered on the y-axis according to their elevation (SWLI) and the histograms show the percentage abundance of each of the dominant diatom species.
25 Figures, Tables & Plates- Chapter 4
• - - • . - • D I I I I I • 0 I • • 0 0 0 • . . .. . I 0 . . . -I I I I I -I I I - - - • • • • I. I- I I I I • II I I - I I - I I- I I I • I I I ...... • . . :J1 -• - -• I I I I I I I -... I - I I 5: . .. 0 . !!!.• - • I •• - 0 .... 0 - .. 1: - - .2 .. -I I I I I I I I- I I I- - -I • I I I • I • I I 11 I I I I • • • -• I I I I I I I
I I p I I' 0 I I p ...... _.- I I I I I • • - ' • .-·--
rrrr;,rn rn f"TTT;J 1 • • 'in' •'.io' •'i.J' •rn rn rn rn f"TTT;J rn rn rn rn rn rn rn rn rrrr;,rn rn rrrr;,rn rn,..., ~rn f"TTT;J f"'"T"20,...... rn rn rn rn rn rn rn rn rn rn ~rn rrrr;,rn rn rn rn rn rn rn rn rn rn,...... ,.... rn ~,...... ~ ~rn % abundance of diatoms
Figure 4.5 Morar diatom training set, dominant diatom species exceeding 2% of the total are shown. Species are ordered according to their salinity tolerance. The diatom samples are ordered on the y-axis according to their elevation (SWLI) and the histograms show the percentage abundance of each of the dominant diatom species.
26 Figures, Tables & Plates- Chapter 4
I I I I I I ' . I o ,..,:=_ • ; : • : ' . . : ~ . ' ...... "= ." I ...... :=- . . ' - ... ~ . . . . ' ' .. .' -' . . . .-· ' --' . I I . ;. I • I I I I • I : I p .. I ' ... --=- - .' . . • ..... : - , r ,. I I .._I _...... ILl l I:.. I. I ' P 1.", •. r I •, E.., . • .• L • loll : ' .: ~rr- . . . . ' . I : .._ p==o • L. . P II' IPL.P .. I ~ .__..,_I L •• -r .... . ! ! ! ' ~ ! ! : • r . • , ;=-r • · . • r !'"'1. I i r. · iiLr: & 1 f f ~.f. ~ i I r I , ~. _. I ••r ; I • I F . .. , I .. ' .. •P I t I ~ r F f Dill 'I ~--·- .. ·- ~ - - : - I = I I ==aa r-r-....-;,r"""i"" ...... ,...,...... -;,r-;-- r-o~.,...... J""" ,....,...,.... f""" r-;-",... .- ,...... ~r ,.....rr--r-o,...,... ,....,...,.... r-;"' ..... ~,...... ~r-r-r-;,r-r,...,... .-.-r-r-;-,... ,...,...... -;~ ,_~ r--r .--:J""2o,...... ~,....l"""r- ~~"""' ~""'~"""'"~~,...r-;;-"""rr-;""" ,...I"'" ... 'lloablnlutce d dldGIIII
Figure 4.6 Argyll diatom training set, dominant diatom species exceeding 2% of the total are shown. Species are ordered according to their salinity tolerance. The diatom samples are ordered on the y-axis according to their elevation (SWLI) and the histograms show the percentage abundance of each of the dominant diatom species.
27 Figures, Tables & Plates- Chapter 4 a) Morar
0 Sa
0
~ ,.... Cl · ~----~----~--~------~ -1.5 1.0 b) Argyll
~ -·· , "
«=! .....' ,______..____ -----1 -1.0 1.0
Key: SW - Standardised Water Level Index LOI - Loss on Ignition Ph - pH sa - Salinity Sa - Sand Si -Silt Cl - Clay
Figure 4.7 CCA Ordination diagram for samples and environmental variables for the diatom training sets.
28 Figures, Tables & Plates- Chapter 4
300 80 >< c:; .. >< 60 Iii 250 c:;.. l i 40 • •• c • • i 200 ··' 8. •• • •• 20 If) • ..I •• Q. 150 ·.:~ i • . ~ 0 ...... ~ . ... :I 50 ••20'+ 300 "' 100~ ~~ i~ i 100 ··r . ~ -20 If) •••• Gl • • IX • • •• ~ -40 • 50 Q. • C( ·~· l:Q wIll 3: -60 0 0 50 100 150 200 250 300 -80 Observed SWU WAPLS Residuals Observed SWLI
Figure 4.8 Regression results for the Kintail diatom training set using WAPLS component 3 are shown on the left and the associated regression residuals are shown on the right.
300 80
60 250 >< c:; ..c 40 cGl i 200 0 Q. 20 If) E 0 s. # ...... 'i 150 ~ 0 ~ . ~ j :I 100 .~oo 300 'g"' • 0 100 · ~ -20 • • 41 IX If) ..I -40 • 50 Q. ~ -60 0 1---.----,---,----.-----,----, 0 50 100 150 200 250 300 -80 Estimated SWLI WAPLS (component 3) X WAPLS Residuals Observed SWL.I
Figure 4.9 Regression results for the Morar diatom training set using WAPLS component 3 are shown on the left and the associated regression residuals are shown on the right.
29 Figures, Tables & Plates- Chapter 4
300 60 >< ;:; • • 40 ~ 250 >< ;:; •• • l • ' .• -i 20 • !200 ...~· . i • ~ 0 Q. 150 ,.s l ~ 50 250 300 ~ • ii :i -20 100 ifJ) .,,~· l1 . IX ~• • • ~ -40 • Q. I 50 <( • ~ ~ -00 • 0 0 50 100 150 200 250 300 -80 Observed SWLI WAPLS Residuals Observed SWLI
Figure 4.10 Regression results for the Argyll diatom training set using WAPLS component 3 are shown on the left and the associated regression residuals are shown on the right.
30 Figures, Tables & Plates - Chapter 4
.. _ I _. ' I II= ' ...... • I I I I • ..• a - "" I - I r • • ~r ~ I- • I •, E. I I I 1: t'~ & r ~ I I • .. I .. I •I I Iii' I I I L I I • r; ..I t . L II I • b• ~ ~ .... [ I I -I' Ip r I: .. ~ • • ~ II~ ~: [ ' Ji~ • r I ~ I I ... J .. I r .. ~ I - r ~~ ~ r I II!= r tr '- r I t .... I .a ...... I ~ • I I .. • I ... I I .• • I II .I I -• I I I - • - - I I I :::i - - I ,__ - c p ... ~ , - I ~ 270 I" Ia r I I B I = I !!!. • ~-- c 290 0 ... I • I a b' I l!o iii 310 • ... > • iii 330 II I 0 350 I • • • - • - -• • I 370 - • 41 - -
470 I
490
20 40 (lfJr-> I"""' I"""' I"""' I""" r-> r"' I"""' 20 40 ~I"""' I"""' 20 40 6fJ BlJr-> I"""' r"' I""" r-> r-;;J I"""' I"""' I"""' I""" r-> ~I""" r-> ~I'""' r-> f"""T2'' ~~I""" r-> % abundance pollen
Figure 4.11 Regional pollen training set, dominant pollen species exceeding 5% of the total are shown and species are ordered according to their life form. The pollen samples are ordered on the y-axis according to their elevation (SWLI) and the histograms show the percentage abundance of each of the pollen species.
31 Figures, Tables & Plates - Chapter 4
Trees Shrubs Herbs Spores Other / ;;;.. -~~ 9><' ~ ~'I> ~llj 9> 0~ "'0 0 ~~# vo/~/ ~f ~~ "#'.?. .. "' ~.,. !>.'1>":; ~ ~ift #..~Vv ~tf .9>'<$<$' ff' ~t§J #t/' #.~ ~'1>,:§> .Iff 0<' It~~ ,;:,~ ,;:,+ ~ ~#~# ~~#'I>,(}~.f ~rl? §;~<:-" ~'I>~~"' :::."' fl 9><' ~<' 9><' ~llj !;. ~ • Iff s,'IJ ~ ~'IJ ~<' q;~q_~ ~~ 't"~ ~lljc!' eft§(}. &'60~ ~" ~ ~" ~0 ....'1> ~ «.~ e:;llJ e:;<:l Figure 4.12 Kintaill pollen training set, dominant pollen species exceeding 2% of the total are shown and species are ordered according to their life form. The pollen samples are ordered on the y-axis according to their elevation (SWLI) and the histograms show the percentage abundance of each of the pollen species. 32 Figures, Tables & Plates - Chapter 4 100 105 - I • • • • 110 -- I o=:oll • • 115 - - - - - 120 • • • • I .. II • II - - - I • • Ill I 125 - - - II I - 130 • • • • • • II • -• • • -• • • 135 • • - • • .. a • = .- 140 - - - ::i • I a .. a • rn~ 145 - - --c 150 .2 155 l • • - • I • wGl 160 - - - 165 • II I • • I 170 • I I - I I - • • • • • I • 175 I I • -• I - - B I -I I 180 - 165 • • Ill - - I - - 11!1 B I 190 195 I I • • • I • I • • • 200 - r~r~rrrr~~rr- rrr'"'2orrrrrrrrrr~r 20 20 40 % abundance of pollen Figure 4.13 Morar pollen training set, dominant pollen species exceeding 2% of the total are shown and species are ordered according to their life form. The pollen samples are ordered on the y-axis according to their elevation (SWLI) and the histograms show the percentage abundance of each of the pollen species. 33 Figures, Tables & Plates - Chapter 4 Trees Shrubs 140 160 180 I I~ • a=~' • • I I I • • ... . • I I r r-~ - - - I I t I ,.... ' I I r 220 tl: t r E: t • I I • I I I II I I I I ... I I I • I I 240 .. ' ' .• • - ' I• •I I I • I - • I • I • 260 • •I • - I - -= - • • - - I I - I • - ~280 • - I I I -c 300 • I • • ;:3200 I I I - Ill I - = • > J!340 - - w • • I I 360 - • - 360 • 400 - - 440 II I 480 - - I 'io' I '.io' I 'tlo' I l'"""j"T'"I ,...,.,io..-of'"TI <40"T'"I ,...,., l.,.,'tlor-' r-' r-' j""""1j o I 'io' I '.io' I r-' r-' -j o I 'io' I '.io' I tlo I tlofT' r-' r-' r-' ~ r-' r-' fT' r-' ~r-' r-' r-' ~ ~~fT' %abundance of pollen Figure 4.14 Argyll pollen training set, dominant pollen species exceeding 2% of the total are shown and species are ordered according to their life form. The pollen samples are ordered on they-axis according to their elevation (SWLI) and the histograms show the percentage abundance of each of the pollen species. 34 Figures, Tables & Plates - Chapter 4 a) Regional training set <0 9 ~------~----~------~ -1.0 1.0 b) Morar training set ;i '"• Key: sw - Standardised Water Level Index LOI - Loss on Ignition ·~ Ph -pH v sa -Salinity ...... ~········· ...... Ycr:.... . Sa -Sand .I Si - Silt Cl -Clay ! ! .. ! &W •• ~ ~ · ~------~------~ -1.0 1.0 c) Argyll training set l.OI Figure 4.15 CCA Ordination diagram for samples and environmental variables for the pollen training sets. 35 Figures, Tables & Plates - Chapter 4 500 250 )( 450 200 N' 1i 400 )( 150 N' §.. 350 .. • ~ 100 i 300 • §.. • 50 • ~ 250 ~ i 0 # • .!Q ~ 200 ·~·~ IV 50 100 300 3~ 400 450 500 ·. :ll -50 rn • :2 ~ 150 - XI • • Ill:: -100 E 100 • • • :p ~ ~ -150 81 50 -200 • 0 +-~~--~~~~--~~~-- • 0 50 100 150 200 250 300 350 400 450 500 -250 Observed SWLI PLS Residuals Observed SWLI Figure 4.16 Regression results for the regional pollen training set using PLS component 2 are shown on the left and the associated regression residuals are shown on the right. 500 250 >< 450 200 M' .. 400 >< 150 i M' t: ~ 350 41 100 E c ! 300 &. ••• .. 50 •• ~ 250 IL l 0 , . Ill .... ii 50 300 350 400 450 500 i 200 • :ll -50 rn • • :2 ' ~ 150 • ~• ! -100 ... rn ~ 100 • ., it -150 • Jl 50 -200 0 ---~~~~~~~~--~~-- 0 50 100 150 200 250 300 350 400 450 500 -250 PLS Residuals Observed SWLI Observed SWLJ Figure 4.17 Regression results for the Kintail pollen training set using PLS component 3 are shown on the left and the associated regression residuals are shown on the right. 36 Figures, Tables & Plates - Chapter 4 500 250 )( 450 200 ;:; 400 )( 150 i ;:; 1350 ..c 100 cGl 1300 &. 50 E ~ 250 0 ~. G. ~ 0 .Ja so 100 150~00~~ 300 350 400 450 500 i 200 ICI -50 VI -a:I ) 150 .iii -100 ~ ·i 100 ~ -150 w 50 G. -200 0 50 100 150 200 250 300 350 400 450 500 -250 Observed SWU PLS Residuals Observed SWLI Figure 4.18 Regression results for the Morar pollen training set using PLS component 3 are shown on the left and the associated regression residuals are shown on the right. 500 250 ~ 450 200 ...I') )( i 400 ;:; 150 • c 350 -Gl 100 1E c • 0a. ! 300 E 50 • • • 0 • ~ ~ ·:· ~ 250 0 . .Ja ~ 200 _..!_~~·· ~ 50 100 1=~t.3~0 3~ 4~0 4~0 5~0 ·-a;;; -50 Gl • •• ~ 150 if: a: -100 • VI ..J 1100 G. -150 • Figure 4.19 Regression results for the Argyll pollen training set using WAPLS component 3 are shown on the left and the associated regression residuals are shown on the right. 37 Figures, Tables & Plates - Chapter 4 ,__ I I • . ... . • • ' po .. • . ' . • • . . I •I I-• ; ; ; t • -- • - . ·- . -~ : ~· ~ ~ ~ ~ !~ ·I !!!!!!!!!!!!!..- ·~ . - ' . -~ · - - ...... ---' - · - -·. ' ...... ,.....,~,....,...,.~,...,,...,.~~ ,.. ,.....m~,...... ,~~,...,'"'"l~~,...,f'"1T""o\r"'.\~f'T"A""'l"'""r'~~I""'T'~~,...,~,....,,...,,.....,....,.,..,lr'T'...t Figure 4.20 Regional multi-proxy training set, dominant species exceeding 10% of the total are shown. Species are ordered according to their salinity tolerance or life form. The samples are ordered on they-axis according to their elevation (SWLI) and the histograms show the percentage abundance of the diatom and pollen species. Diatoms are shown in black and pollen in green. 38 Figures, Tables & Plates- Chapter 4 . . ' .- - - - .- - - . ------. . . . I ; I I ; ; ' ; I ; 0 . ;, - . .. . I- .I I . • . ~ ::. :. ~ : . .. : :. . I (II I . II II lt-~r-~••l•tlll'• o , I , ,... , I I 0 r ' . .. ••-- • r ' I •- ·• i I I • I . . I I I j I ~ I ~ , • I - I • ... '-! ...... , • : I I • I. • I ' I - - =- ,. = . : .. .. &.. r -~:r- . Jlf ' ~ 1: I • t • t - I I I . . J a t. ' . . ... • .. . .. • 0 - - - ' - ...... • I . . .._ ...._ . r-- . ,. r- . . r - .. -· ...... - - .. -·• - -·-· - I • ' t If o - -_. I -· I o .I - - - ··. I - .I ··-·a ···-a · - -I...I at .. ~,....,_"-"li"",...,...~~r-;.r-"'t"',....,...,...~~-~,-.r-r~r-or-o,-.,....,....~,.,.~,...f""""')f'"'",...t"'T,....~,...,_,....,...~~,...~.,-~-ra:l···trr-r,...,....r-r-r,....~,...~,.....,....- .... ~~,...~-~ .. ~., ...... ,....,. Figure 4o21 Kintail multi-proxy training set, dominant species exceeding 2% of the total are shown. Species are ordered according to their salinity tolerance or life form. The samples are ordered on the y-axis according to their elevation (SWLI) and the histograms show the percentage abundance of the diatom and pollen species. Diatoms are shown in black and pollen in green. 39 Figures, Tables & Plates- Chapter 4 ...... I t • t - ·· • ftc a- I . ··- ...... • I I I • • ~ . . -.... I •• I .. .-· I I I t • I I I - . . - -. " ' 0 0- I •' :. : : .:~:·:r - 0 I - 00 . . : : 0 - ::-.:· :- . -. I a.. I J 0 0 --- - ...... ' .==- - -. ' . . . . ·· - ... -- . - -... - · ...... ,.....,.,,....~,..,,..,..,...... ,...... ,...,,...,., .... ,..,...... ,.,,....,...I'"TT',...... ,,...... ,....,..,,..,,....,.,...,..., ...... ,.,...,...,...,,..,,...,..,,..,,.,,.,,....,...,.....f""'''T'~,...,..... ,.....,...... ,,..,,..,....,,..,r-r-r',..,,..,....,~,....,....,..,,..,,..., .... ,., ..... ,., ,..,,.,,.., ,....,...~ ~ •und~ of dl•o-m. ..a polltn Figure 4o22 Morar multi-proxy training set. dominant species exceeding 2% of the total are shown. Species are ordered according to their salinity tolerance or life form. The samples are ordered on the y-axis according to their elevation (SWLI) and the histograms show the percentage abundance of the diatom and pollen species. Diatoms are shown in black and pollen in green. 40 Figures, Tables & Plates- Chapter 4 f I I I 1 I I I a ... a • I t a I I .. I 1 • I I lf l • L,. t I ' ·- 1 : a. . 1 1 r ~ ~ • • r • . .. . :. l ~ ; ~ r-. ' ~ ' . ~ :: - =- !..:-1"' ' ~ ~ ~ f .. f" ~; !Ill-" r ~ .. , r . , . . ' -.•• to I If :-· ; ·~ . .. -• • • I : : r - - . • ' · - · :::=:- • I t , • •• I t I I I .. .' . - I • II I ... • • .. . . ' ...___ .. , - - . . ' . ··- ···· ·· -·- ..... ··--- '' . . ' ...... • I I I - - • I I - • I I - - .. - .. ,... r-:"",..,,...~,.. ,_...r:",... ~.,r,.. r-r-~r-r- r-~r;.,...r-r-r~,... r--;r r-;rr ,.....-;,-:,-rr- ,_,... ~,.... ,... r-~r-r""i':" l"'"jrr-.,r-:r-7r-r,.. ~,.. ,..r-:, r-;:-.r-,rr-;:-:,rr-·r-,...... rr-::"r-:-~.r ,... r-1""r-;-::-,....r-~.r,... ,...,... ~r-r-r-r- ,..._.r ..... r-r-r--.r-,r ~ ------·-.... Figure 4.23 Argyll multi-proxy training set, dominant species exceeding 2% of the total are shown. Species are ordered according to their salinity tolerance or life form. The samples are ordered on the y-axis according to their elevation (SWLI) and the histograms show the percentage abundance of the diatom and pollen species. Diatoms are shown in black and pollen in green. 41 Figures, Tables & Plates - Chapter 4 a) Regional training set q ~ ~ ~------~ -1.0 1.0 b) Morar training set ...... , ...... 0 0 0 8 «=! ~ I -1.5 1.0 c) Argyll training set .....«=! 0 0 0 0 aw I 0 Key: SW - Standardised Water Level Index LO I - Loss on Ignition fo Ph -pH 0 i sa - Salinity i Sa - Sand Si -Silt Cl - Clay 0 -1.0 1.0 Figure 4.24 CCA Ordination diagram for samples and environmental variables for the multi-proxy training sets. 42 Figures, Tables & Plates - Chapter 4 300 150 ..,)( )( 100 1250 •• c:;-.. i 1200 j~.•·~··· §. 50 • ~ 150 1 • .!!1 0 ~ •• :II 50 100 300 :2" i 100 ~ fl) a: -50 • •• ~ • Q. I 50 ct ~ ~ -100 0 +---~------~--~--~--~ 0 50 100 150 200 250 300 -150 Observed SWU WAPLS Residuals Observed SWU Figure 4.25 Regression results for the regional multi-proxy training set using WAPLS component 3 are shown on the left and the associated regression residuals are shown on the right. 300 150 )( c:;- i 250 .... )( 100 l ".. 1200 ••• l 50 fl) • ...I ~· • Q. 150 1 .. 0 • ... .!!1 ...... ~ ·~~ :II 50 100 25t. 300 :2" ~·,~~ ~ 100 ~ -50 •• ~ Q. 50 01( JIll ~ -100 w • 0 0 50 100 150 200 250 300 -150 Observed SWU WAPLS Residuals Observed SWU Figure 4.26 Regression results for the Kintail multi-proxy training set using WAPLS component 3 are shown on the left and the associated regression residuals are shown on the right. 43 Figures, Tables & Plates - Chapter 4 300 150 )( c:; c 250 )( 100 ..c:; l .;. i 1200 J i 50 ~ G. 150 - i • ~ 0 .... ~ ~ ·~· :sz 50 100 150 ~00 ~~0 300 ~ ~ 100 « -50 • f/)_, G. !I( I " ?: -100 0 ~----~--~--~--~--~---. 0 50 100 150 200 250 300 -150 Observed SWLI WAPLS Residuals Observed SWLI Figure 4.27 Regression results for the Morar multi-proxy training set using WAPLS component 3 are shown on the left and the associated regression residuals are shown on the right. 300 150 )( N' i 250 )( 100 c: t •• N' 8. .. : . .... 1200 l 50 f/)_, ....~ G. 150 l • ~ 0 Ill ~ :;, :sz 50 100 300 ~ 100 XI « -50 ~ • G. 50 < ?: -100 !w 0 0 50 100 150 200 250 300 -150 Observed SWU WAPLS Residuals Observed SWLI Figure 4.28 Regression results for the Argyll multi-proxy training set using WAPLS component 2 are shown on the left and the associated regression residuals are shown on the right. 44 Figures, Tables & Plates - Chapter 4 OCCA axis 1 OCCA axis 1 OCCAaxis 1 Training Environmental length (SO length (SO length (SO Set Name Variable units) diatoms units) diatoms units) pollen and pollen Regional SWLI 4.12 1.57 2.64 Kintail SWLI 6.78 1.52 2.66 Morar SWLI 4.48 1.52 2.36 Argyll SWLI 4.95 2.05 2.38 Table 4.1 Detrended Canonical Correspondence Analysis results. pH 0.18 LOI -0.19 -0.29 ~Sand -0.44 0.19 -0.41 %Silt 0.33 -0.19 0.48 -0.98 %Clay 0.66 -0.12 0.14 -0.87 0.76 SWLI -0.48 -0.47 0.59 -0.07 0.17 -0.22 Salinity pH LOI %Sand %Silt %Clay Table 4.2 Correlation matrices for environmental variables from the regional diatom training set. Total ~es 1 2 3 4 inertia Eigenvalues: 0.57 0.38 0.31 0.28 7.63 Species-environment porrelations: 0.95 0.89 0.86 0.88 Cumulative percentage variance of species data: 7.4 12.5 16.5 20.3 of species-environment relation: 28.8 48.4 64.3 78.8 Sum of all eigenvalues: 7.63 Sum of all canonical eigenvalues: 1.96 Table 4.3 CCA results showing summary of eigenvalues for the regional diatom training set. 45 Figures, Tables & Plates - Chapter 4 Training Set Name Explained Unexplained Variance (%) Variance (%) Regional diatom training set 26 74 Kintail diatom training set 11 89 Morar diatom training set 51 49 Argyll diatom training set 34 66 Regional pollen Training Set 28 73 Kintail pollen Training Set 16 84 Morar pollen Training Set 53 47 Argyll pollen Training Set 37 63 Regional multi-proxy training set 28 72 Kintail multi-proxy training set 18 82 Morar multi-proxy training set 43 57 Argyll multi-proxy_ training set 38 62 Table 4.4 Summary of the results of CCA analysis listing the percentage of explained and unexplained variance. Model Jack Jack Boot Boot Name RMSE R2 Number R2 RMSEP R2 RMSEP WAPLS Component 1 for SWLI 27.07 0.75 0.69 30.02 0.70 30.59 WAPLS Component 2 for SWLI 18.86 0.88 0.79 24.85 0.79 26.43 ----- 1 13.97 0:93 . 21.84 0.83 24.83 WAPLS Component 3 for SWLI -----·- - (),84 ------WAPLS Component 4 for SWLI 11.35 0.96 0.85 20.77 0.85 23.88 WAPLS Component 5 for SWLI 9.13 0.97 0.87 19.65 0.87 23.32 Weighted averaging model (inverse 27.07 0.75 0.69 29.94 0.70 30.55 deshrinking) for SWLI Weighted averaging model (classical 31.25 0.75 0.70 33.56 0.70 33.68 deshrinking) for SWLI 2 Weighted averaging model (tolerance downweighted, inverse deshrinking) 21.51 0.84 0.74 27.87 0.74 29.48 for SWLI Weighted averaging model (tolerance downweighted, classical deshrinking) 23.44 0.84 0.74 29.21 0.74 30.24 for SWLI Table 4.5 Summary of the transfer function performance stat1st1cs for the reg1onal diatom training set. WA-PLS component 3 is the chosen model. 46 Figures, Tables & Plates - Chapter 4 Paper r Gasse eta/. (1995) 0.77 to 0.92 Campeau et a/. ( 1999) 0.87 to 0.89 Zong and Horton (1999) 0.65 to 0.72 Gehrels (2000) 0.77 to 0.85 Sawai et a/. (2004) 0.88 to 0.94 Table 4.6 List of f2 values from other studies. 47 Figures, Tables & Plates - Chapter 4 a) Morar pH 0.61 LOI -0.04 0.41 Yo Sand -0.22 -0.48 -0.91 Yo Silt 0.16 0.48 0.93 -0.99 Yo Clay 0.43 0.41 0.75 -0.90 0.85 SWLI -0.49 -0.16 0.49 -0.40 0.45 0.16 Salinity pH LOI %Sand %Silt %Clay b) Argyll pH 0.42 LOI -0.34 -0.56 Yo Sand -0.21 0.40 -0.53 Yo Silt 0.24 -0.38 0.52 -0.99 Yo Clay 0.28 -0.23 0.37 -0.91 0.88 SWLI -0.44 -0.80 0.59 -0.39 0.38 0.23 Salinity _pH LOI %Sand %Silt %Clay Table 4.7 Correlation matrices for environmental variables for the local diatom training sets. 48 Figures, Tables & Plates - Chapter 4 a) Kintail diatom training set Total ~es 1 2 3 4 inertia Eigenvalues: 0.70 0.88 0.61 0.52 6.52 Species-environment correlations: 0.92 0.00 0.00 0.00 Cumulative percentage variance of species data: 10.70 24.30 33.60 41.50 of species-environment relation: 100.00 0.00 0.00 0.00 Sum of all eigenvalues 6.52 Sum of all canonical eigenvalues 0.70 b) Morar diatom training set Total Axes 1 2 3 4 inertia Eigenvalues: 0.69 0.41 0.32 0.23 3.52 Species-environment correlations: 0.99 0.95 0.86 0.931 Cumulative percentage variance of species data: 19.6 31.3 40.4 46.9 of species-environment relation: 38.5 61.4 79.4 92.1 Sum of all eigenvalues: 3.52 Sum of all canonical eigenvalues: 1.79 c) Argyll diatom training set Total ~es 1 2 3 4 inertia Eigenvalues: 0.70 0.43 0.38 0.29 6.21 Species-environment correlations: 0.97 0.90 0.88 0.86 Cumulative percentage variance of species data: 11.20 18.10 24.10 28.80 of species-environment relation: 33.6 53 69.8 84.5 Sum of all eigenvalues 6.21 Sum of all canonical eigenvalues 2.11 Table 4.8 CCA results showing summary of eigenvalues for the diatom training sets. 49 Figures, Tables & Plates - Chapter 4 . . · ~o of I . Jack Boot Trammg ~et Name samples Component Name RMSE R2 Jack R2 RMSE.P Boot R2 RMSEP 1 I . - Regional diatoms 116 3 WAPLS Component 3 for SWLI 13.97 0.93 0.84 21.84 0.83 24.83 Kintail diatoms 43 3 WAPLS Component 3 for SWLI 11.23 0.96 0.82 24.19 0.82 27.79 Morar diatoms 22 3 WAPLS Component 3 for SWLI 2.76 0.99 0.93 10.28 0.91 12.46 Ar II diatoms 51 3 WAPLS Com onent 3 for SWLI 9.97 0.97 0.86 21.13 0.87 24.06 Table 4.9 Summary of the transfer function performance statistics for the diatom training sets for each model chosen. The results are shown in full in appendix 4.1. 50 Figures, Tables & Plates - Chapter 4 a) Regional training set pH 0.33 LOI -0.32 -0.46 ~Sand -0.58 -0.18 -0.01 ~Silt 0.43 0.20 0.08 -0.96 Yo Clay 0.79 0.20 -0.21 -0.86 0.71 SWLI -0.34 -0.81 0.49 0.25 -0.26 -0.2g Salinity pH LOI %Sand %Silt %Clay b) Morar training set pH 0.55 LOI -0.22 0.30 %Sand -0.38 -0.44 -0.69 %Silt 0.21 0.47 0.79 -0.96 %Clay 0.65 0.21 0.22 -0.74 0.52 SWLI -0.87 -0.55 0.06 0.57 -0.39 -0.78 Salinity pH LOI %Sand %Silt %Clay c) Argyll training set pH 0.58 LOI -0.52 -0.57 %Sand -0.32 -0.13 -0.19 %Silt 0.36 0.15 0.17 -0.99 %Clay 0.43 0.37 -0.06 -0.88 0.85 SWLI -0.60 -0.87 0.51 0.23 -0.23 -0.44 Salinity pH LOI %Sand %Silt %Clay Table 4.10 Correlation matrices for environmental variables for the pollen training sets. 51 Figures, Tables & Plates - Chapter 4 a) Regional training set Total !Axes 1 2 3 4 inertia Eigenvalues: 0.17 0.13 0.07 0.07 1.90 Species-environment correlations: 0.88 0.79 0.61 0.77 Cumulative percentage variance of species data: 9.00 15.70 19.50 22.90 of species-environment relation: 32.50 56.80 70.60 83.00 Sum of all eigenvalues 1.90 Sum of all canonical eigenvalues 0.53 b) Kintail training set Total !Axes 1 2 3 4 inertia Eigenvalues: 0.16 0.25 0.09 0.08 0.96 Species-environment correlations: 0.76 0.00 0.00 0.00 Cumulative percentage variance of species data: 16.30 42.90 52.20 61.00 of species-environment relation: 100.00 0.00 0.00 0.00 Sum of all eigenvalues: 0.96 Sum of all canonical eigenvalues: 0.16 c) Morar training set Total Axes 1 2 3 4 inertia Eigenvalues: 0.24 0.14 0.11 0.06 1.21 Species-environment correlations: 0.87 0.91 0.94 0.76 Cumulative percentage variance of species data: 19.8 31.2 40.6 45.5 of species-environment relation: 37.5 59.1 76.7 86.1 Sum of all eigenvalues: 1.21 Sum of all.canonical eigenvalues: 0.64 d) Argyll training set Total Axes 1 2 3 4 inertia Eigenvalues: 0.19 0.14 0.06 0.04 1.39 Species-environment correlations : 0.92 0.89 0.84 0.82 Cumulative percentage variance of species data: 13.80 24.20 28.60 31.60 of species-environment relation: 37.30 65.30 77.10 85.40 Sum of all eigenvalues: 1.39 Sum of all canonical eigenvalues: 0.52 Table 4.11 CCA results showing summary of eigenvalues for the pollen training sets. 52 Figures, Tables & Plates - Chapter 4 I, r, - - I . . No of Jack Boot Trammg' S,et Name _: _ samples Compon:nt Name , RMS~ R2 Jack R2 RMSEP Boot R2 RMSE, Regional pollen 82 2 PLS Component 2 for SWLI 39.44 0.56 0.34 48.19 0.36 50.89 Kintail pollen 25 3 PLS Component 3 for SWLI 34.63 0.66 0.30 52.92 0.31 57.61 Morar pollen 15 3 PLS Component 3 for SWLI 6.49 0.94 0.59 17.97 0.65 18.44 Arg II allen 42 3 WAPLS Component 3 for SWLI 18.74 0.92 0.59 42.20 0.67 43.29 Table 4.12 Summary of the performance statistics for the chosen transfer function model for each of the pollen training sets. A table of the results for each model is shown in Appendix 4.1. 53 Figures, Tables & Plates - Chapter 4 a) Regional training set pH 0.30 LOt -0.26 -0.16 %Sand -0.55 -0.06 -0.10 %Silt 0.42 0.10 0.16 -0.96 %Clay 0.75 0.03 -0.11 -0.87 0.71 SWLI -0.37 -0.65 0.18 0.13 -0.17 -0.11 Salinity pH LOt %Sand %Silt %Clay b) Morar training set IPH 0.58 LOt -0.14 0.33 ~Sand -0.37 -0.48 -0.76 ~Silt 0.22 0.50 0.82 -0.97 ~Clay 0.63 0.28 0.35 -0.77 0.58 jsWLt -0.85 -0.59 -0.03 0.58 -0.44 -0.76 Salinity pH LOt %Sand %Silt %Clay c) Argyll training set pH 0.55 LOt -0.42 -0.33 ~Sand -0.30 -0.05 -0.27 Yo Silt 0.35 0.07 0.25 -0.99 ~Clay 0.35 0.17 0.11 -0.89 0.85 ISWLI -0.79 -0.68 0.28 0.24 -0.24 -0.3( Salinity pH LOt %Sand %Silt %Clay Table 4.13 Correlation matrices for environmental variables from the multi-proxy training sets. 54 Figures, Tables & Plates - Chapter 4 a) Regional training set Total ~es 1 2 3 4 inertia Eigenvalues: 0.31 0.21 0.18 0.10 3.47 Species-environment correlations: 0.95 0.85 0.90 0.78 Cumulative percentage variance ot species data: 9.00 15.10 20.20 23.00 of species-environment relation: 31.40 52.90 70.50 80.30 Sum of all eigenvalues: 3.47 Sum of all canonical eigenvalues: 0.99 b) Kintail training set Total Axes 1 2 3 4 inertia Eigenvalues: 0.45 0.51 0.33 0.21 2.47 Species-environment correlations: 0.93 0.00 0.00 0.00 Cumulative percentage variance of species data: 18.00 38.60 51.80 60.10 of species-environment relation: 100.00 0.00 0.00 0.00 Sum of all eigenvalues: 2.47 Sum of all canonical eigenvalues: 0.45 c) Morar training set Total ~es 1 2 3 4 inertia Eigenvalues: 0.37 0.21 0.17 0.12 1.76 Species-environment correlations: 0.97 0.94 0.93 0.85 Cumulative percentage variance of species data: 21.10 32.80 42.40 49.20 of species-environment relation: 37.10 57.80 74.70 86.50 Sum of all eigenvalues 1.76 Sum of all canonical eigenvalues 1.00 d) Argyll training set Total ~es 1 2 3 4 inertia Eigenvalues: 0.35 0.23 0.17 0.11 2.62 Species-environment correlations: 0.96 0.92 0.80 0.83 Cumulative percentage variance of species data: 13.30 22.00 28.50 32.80 of species-environment relation: 34.80 57.60 74.70 85.70 Sum of all eigenvalues: 2.62 Sum of all canonical eigenvalues: 1.00 Table 4.14 CCA results showing summary of eigenvalues for the multi-proxy training sets. 55 Figures, Tables & Plates - Chapter 4 I I - - I . . , No of Jack Boot 1 Tramm; Set Name :, samples Compon~nt Name . __ . RMSE R2 Jack R2 RMSEP Boot R2 RMSE~ Regional multi-proxy 75 3 WAPLS Component 3 for SWLI 6.10 0.97 0.82 15.66 0.81 17.90 Kintail multi-proxy 23 3 WAPLS Component 3 for SWLI 5.06 0.99 0.73 25.00 0.74 26.40 Morar multi-proxy 15 3 WAPLS Component 3 for SWLI 1.21 1.00 0.87 10.51 0.88 11.65 Ar II multi- roxy 37 2 WAPLS Component 2 for SWLI 7.34 0.95 0.80 14.42 0.80 15.93 Table 4.15 Summary of the performance statistics for the chosen transfer function model for each of the multi-proxy training sets. A table of the results for each model is shown in Appendix 4. 1. 56 Figures, Tables & Plates - Chapter 4 - I I I ' No of Jack Boot [ Training Set Name Component Name RMS~ R2 Jack R2 Boot R2 samples I I RMS~P RMSEP, I I J I- Regional diatoms 116 3 WAPLS Component 3 for SWLI 13.97 0.93 0.84 21.84 0.83 24.83 Kintail diatoms 43 3 WAPLS Component 3 for SWLI 11.23 0.96 0.82 24.19 0.82 27.79 Morar diatoms 22 3 WAPLS Component 3 for SWLI 2.76 0.99 0.93 10.28 0.91 12.46 Argyll diatoms 51 3 WAPLS Component 3 for SWLI 9.97 0.97 0.86 21.13 0.87 24.06 Regional pollen 82 2 PLS Component 2 for SWLI 39.44 0.56 0.34 48.19 0.36 50.89 Kintail pollen 25 3 PLS Component 3 for SWLI 34.63 0.66 0.30 52.92 0.31 57.61 Morar pollen 15 3 PLS Component 3 for SWLI 6.49 0.94 0.59 17.97 0.65 18.44 Argyll pollen 42 3 WAPLS Component 3 for SWLI 18.74 0.92 0.59 42.20 0.67 43.29 Regional multi-proxy 75 3 WAPLS Component 3 for SWLI 6.10 0.97 0.82 15.66 0.81 17.90 Kintail multi-proxy 23 3 WAPLS Component 3 for SWLI 5.06 0.99 0.73 25.00 0.74 26.40 Morar multi-proxy 15 3 WAPLS Component 3 for SWLI 1.21 1.00 0.87 10.51 0.88 11.65 Argyll multi-proxy 37 2 WAPLS Component 2 for SWLI 7.34 0.95 0.80 14.42 0.80 15.93 Table 4.16 Summary of the performance statistics for the chosen transfer function models for each of the training sets. 57 Figures, Tables & Plates - Chapter 4 I \ \. Multi-proxy Training ?et No of C ' t N RMSE R J k R Jack t R Boot ' Name samples ompon~n ame 2 ac 2 RMSEP 8 00 2 RMSE~ I I I ' " - I ' Regional Diatom and WAPLS Component 3 for SWLI 15.66 0.81 17.90 pollen 75 3 6.10 0.97 0.82 Regional Diatoms 75 3 WAPLS Component 3 for SWLI 8.11 0.95 0.78 17.51 0.78 20.35 Re ional Pollen 75 2 WAPLS Com onent 2 for SWLI 18.39 0.76 0.58 24.40 0.61 24.73 Table 4.17 Summary of the performance statistics for the most appropriate models for the multi-proxy training set and diatom and pollen training sets comprising just the diatom and pollen samples present in the multi-proxy training set. 58 Figures, Tables & Plates - Chapter 4 APPENDIX 4.1 Transfer function results. Regional Diatom Training Set Jack Boot Model Code Name RMSE R2 Jack R2 BootR2 RMSEP RMSEP Component 1 WAPLS Component 1 for SWLI 27.07 0.75 0.69 30.02 0.70 30.59 Component 2 WAPLS Component 2 for SWLI 18.86 0.88 0.79 24.85 0.79 26.43 1 Component 3 WAPLS Component 3 for SWLI 13.97 0.93 0.84 21..84 0.83 24.83 Component 4 WAPLS Component 4 for SWLI 11.35 0.96 0.85 20.77 0.85 23.88 Component 5 WAPLS Component 5 for SWLI 9.13 0.97 0.87 19.65 0.87 23.32 Jack Boot Model Code Name RMSE R2 Jack R2 BootR2 RMSEP RMSEP WA_Inv Weighted averaging model (inverse deshrinking) for SWLI 27.07 0.75 0.69 29.94 0.70 30.55 WA_Cia Weighted averaging model (classical deshrinking) for SWLI 2 31.25 0.75 0.70 33.56 0.70 33.68 Weighted averaging model (tolerance downweighted, inverse WATOL_Inv deshrinking) for SWLI 21.51 0.84 0.74 27.87 0.74 29.48 Weighted averaging model (tolerance downweighted, classical WATOL_Cia deshrinking) for SWLI 23.44 0.84 0.74 29.21 0.74 30.24 59 Figures, Tables & Plates- Chapter 4 Kintail Diatom Training Set Jack Boot Model Code Name RMSE R2 Jack R2 BootR2 RMSEP RMSEP Component 1 WAPLS Component 1 for SWLI 27.04 0.77 0.68 32.07 0.69 34.27 Component 2 WAPLS Component 2 for SWLI 16.18 0.92 0.79 25.76- 0.80 28.93 1 Component 3 WAPLS Component 3 for SWLI i 11.23 0.96 0.82 24.1.9 0.82 27.79 Component 4 WAPLS Component 4 for SWLI 8.29 0.98 0.86 21.54 0.85 26.32 Component 5 WAPLS Component 5 for SWLI 6.05 0.99 0.87 20.79 0.86 26.39 Jack Boot Model Code Name RMSE R2 Jack R2 BootR2 RMSEP RMSEP WA_Inv Weighted averaging model (inverse deshrinking) for SWLI 27.04 0.77 0.68 31.94 0.68 33.43 WA_Cia Weighted averaging model (classical deshrinking) for SWLI 30.76 0.77 0.69 35.71 0.69 36.12 Weighted averaging model (tolerance downweighted, inverse 2 WATOL_Inv deshrinking) for SWLI 24.10 0.82 0.70 31.30 0.69 34.67 WATOL Cl Weighted averaging model (tolerance downweighted, classical - a deshrinking) for SWLI 26.63 0.82 0.70 33.90 0.70 35.97 Morar Diatom Training Set Jack Boot Model Code Name RMSE R2 Jack R2 BootR2 RMSEP RMSEP Component 1 WAPLS Component 1 for SWLI 8.67 0.94 0.84 15.00 0.84 15.64 Component 2 WAPLS Component 2 for SWLI 4.38 0.99 0.89 12.75 0.89 13.68 1 Component 3 WAPLS Component 3 for SWLI 2.76 0.99 0.93. 10.28 0.91 12.46 Component 4 WAPLS Component 4 for SWLI 1.56 1.00 0.93 10.17 0.92 12.29 Component 5 WAPLS Component 5 for SWLI 0.99 1.00 0.93 10.05 0.92 12.18 Jack Boot Model Code Name RMSE R2 Jack R2 BootR2 RMSEP RMSEP WA_Inv Weighted averaging model (inverse deshrinking) for SWLI 8.67 0.94 0.86 14.01 0.88 13.81 WA_Cia Weighted averaging model (classical deshrinking) for SWLI 8.92 0.94 0.86 13.75 0.88 13.48 Weighted averaging model (tolerance downweighted, inverse 2 WATOL_Inv deshrinking) for SWLI 6.38 0.97 0.89 12.21 0.91 14.48 WATOL Cl Weighted averaging model (tolerance downweighted, classical - a deshrinking) for SWLI 6.47 0.97 0.89 11.98 0.91 14.30 60 Figures, Tables & Plates - Chapter 4 Argyll Diatom Training Set Jack Boot Model Code Name RMSE R2 Jack R2 BootR2 RMSEP RMSEP Component 1 WAPLS Component 1 for SWLI 24.46 0.81 0.74 28.69 0.75 29.51 Component 2 WAPLS Component 2 for SWLI 15.73 0.92 0.82 23.66 0.82 26.20 1 Component 3 WAPLS Component 3 for SWLI I 9.97 0.97 0.86 21.13 0.87 24.06 Component 4 WAPLS Component 4 for SWLI 7.34 0.98 0.87 20.13 0.87 23.87 Component 5 WAPLS Component 5 for SWLI 5.18 0.99 0.88 19.90 0.88 24.33 Jack Boot Model Code Name RMSE R2 Jack R2 BootR2 RMSEP RMSEP WA_Inv Weighted averaging model (inverse deshrinking) for SWLI 24.46 0.81 0.74 28.73 0.73 30.22 WA_Cia Weighted averaging model (classical deshrinking) for SWLI 27.18 0.81 0.74 31.48 0.74 32.19 Weighted averaging model (tolerance downweighted, inverse 2 WATOL_Inv deshrinking) for SWLI 14.58 0.93 0.80 25.44 0.80 30.02 WATOL Cl Weighted averaging model (tolerance downweighted, classical - a deshrinking) for SWLI 15.10 0.93 0.81 24.84 0.80 29.82 61 Figures, Tables & Plates - Chapter 4 Regiona . I Po II en T raining Set Model Code Name Jack Boot RMSE R2 Jack R2 RMSEP BootR2 RMSEP Component 1 PLS Component 1 for SWLI 47.30 0.36 0.17 54.82 0.19 55.71 Component2 PLS Component 2 for SWLI I 39.44 0.56 0.34 48.19 0.36 50.89 1 Component 3 PLS Component 3 for SWLI 37.91 0.59 0.33 49.26 0.32 54.01 Component4 PLS Component 4 for SWLI 36.32 0.62 0.23 57.29 0.27 58.28 Component 5 PLS Component 5 for SWLI 35.65 0.64 0.19 58.96 0.24 61.21 Jack Boot Model Code Name RMSE R2 Jack R2 RMSEP BootR2 RMSEP Imbrie & Kipp Factor Analysis Regression model for SWLI using PCAR_01 axes 1 -1 59.12 0.00 0.61 62.69 0.55 64.85 Imbrie & Kipp Factor Analysis Regression model for SWLI using PCAR_02 axes 1 -2 52.52 0.21 0.09 57.46 0.09 60.09 2 Imbrie & Kipp Factor Analysis Regression model for SWLI using PCAR_03 axes 1 -3 46.94 0.37 0.22 52.68 0.22 56.13 Imbrie & Kipp Factor Analysis Regression model for SWLI using PCAR_04 axes 1 -4 46.31 0.39 0.23 52.42 0.25 54.43 Imbrie & Kipp Factor Analysis Regression model for SWLI using PCAR_05 axes 1 -5 40.73 0.53 0.35 47.96 0.32 52.47 Model Code Name RMSE R2 PCAR_01 Principal Components Regression model for SWLI using axes 1 - 1 57.59 0.05 PCAR_02 Principal Components Regression model for SWLI using axes 1 - 2 48.94 0.32 3 PCAR_03 Principal Components Regression model for SWLI using axes 1 - 3 42.05 0.49 PCAR_04 Principal Components Regression model for SWLI using axes 1 - 4 39.35 0.56 PCAR 05 Principal Components Regression model for SWLI using axes 1 - 5 39.35 0.56 62 Figures, Tables & Plates - Chapter 4 Kintail Pollen Training Set Jack Boot Model Code Name RMSE R2 Jack R2 BootR2 RMSEP RMSEP Component 1 PLS Component 1 for SWLI 46.71 0.38 0.27 51.12 0.29 52.14 Component2 PLS Component 2 for SWLI 37.36 0.60 0.35 48.84 0.37 50.86 1 Component 3 PLS Component 3 for SWLI 34.63 0.66 0.30 52.92 0.31 57.61 Component4 PLS Component 4 for SWLI 32.76 0.70 0.11 68.96 0.20 69.13 Component 5 PLS Component 5 for SWLI 28.02 0.78 0.05 81.47 0.16 77.25 Jack Boot Model Code Name RMSE R2 Jack R2 BootR2 RMSEP RMSEP Imbrie & Kipp Factor Analysis Regression model for SWLI using PCAR_01 axes 1 - 1 54.46 0.16 0.07 58.44 0.16 59.96 Imbrie & Kipp Factor Analysis Regression model for SWLI using PCAR_02 axes 1 -2 44.43 0.44 0.28 51.11 0.33 53.01 Imbrie & Kipp Factor Analysis Regression model for SWLI using 2 PCAR_03 axes 1 -3 40.62 0.53 0.36 48.39 0.40 50.74 Imbrie & Kipp Factor Analysis Regression model for SWLI using PCAR_04 axes 1 -4 36.76 0.62 0.40 46.49 0.41 50.33 Imbrie & Kipp Factor Analysis Regression model for SWLI using PCAR_05 axes 1 -5 35.92 0.63 0.40 46.79 0.38 55.50 Model Code Name RMSE R2 PCAR_01 Principal Components Regression model for SWLI using axes 1 - 1 49.44 0.31 PCAR_02 Principal Components Regression model for SWLI using axes 1 - 2 44.29 0.44 3 PCAR_03 Principal Components Regression model for SWLI using axes 1 - 3 38.93 0.57 PCAR_04 Principal Components Regression model for SWLI using axes 1 - 4 35.80 0.64 PCAR 05 Principal Components Regression model for SWLI using axes 1 - 5 35.80 0.64 63 Figures, Tables & Plates - Chapter 4 Morar Pollen Training Set Jack Boot Model Code Name RMSE R2 Jack R2 BootR2 RMSEP RMSEP Component 1 PLS Component 1 for SWLI 17.58 0.59 0.39 21.53 0.47 22.02 Component 2 PLS Component 2 for SWLI 9.67 0.88 0.60 17.50 0.61 19.18 1 Component 3 PLS Component 3 for SWLI !6.49 0.94 0.59 17.97 0.65 18.44 Component4 PLS Component 4 for SWLI 3.67 0.98 0.72 14.93 0.72 16.91 Component 5 PLS Component 5 for SWLI 2.70 0.99 0.80 12.35 0.74 16.42 Jack Boot Model Code Name RMSE R2 Jack R2 BootR2 RMSEP RMSEP Imbrie & Kipp Factor Analysis Regression model for SWLI using PCAR_01 axes 1 - 1 25.17 0.17 0.08 48.41 0.11 51.19 Imbrie & Kipp Factor Analysis Regression model for SWLI using PCAR_02 axes 1 -2 20.16 0.46 0.18 27.54 0.42 43.05 2 Imbrie & Kipp Factor Analysis Regression model for SWLI using PCAR_03 axes 1 -3 16.55 0.64 0.39 21.64 0.54 40.40 Imbrie & Kipp Factor Analysis Regression model for SWLI using PCAR_04 axes 1 -4 14.25 0.73 0.49 21.26 0.52 43.39 Imbrie & Kipp Factor Analysis Regression model for SWLI using PCAR_05 axes 1 -5 13.78 0.75 0.51 19.37 0.38 47.79 Model Code Name RMSE R2 PCAR_01 Principal Components Regression model for SWLI using axes 1 - 1 23.38 0.28 PCAR_02 Principal Components Regression model for SWLI using axes 1 - 2 16.50 0.64 3 PCAR_03 Principal Components Regression model for SWLI using axes 1 - 3 13.94 0.74 PCAR_04 Principal Components Regression model for SWLI using axes 1 - 4 11.79 0.82 PCAR 05 Principal Components Regression model for SWLI using axes 1 - 5 9.66 0.88 64 Figures, Tables & Plates - Chapter 4 Argyll Pollen Training. Set Model Code Jack Boot Name RMSE R2 Jack R2 RMSEP BootR2 RMSEP WA_Inv Weighted averaging model (inverse deshrinking) for SWLI 26.50 0.84 0.69 37.29 0.70 40.84 WA_Cia Weighted averaging model (classical deshrinking) for SWLI 29.00 0.84 0.70 36.42 0.71 39.82 1 Weighted averaging model (tolerance downweighted, inverse WATOL_Inv deshrinking) for SWLI 23.04 0.88 0.67 38.84 0.72 41.99 Weighted averaging model (tolerance downweighted, classical WATOL_Cia deshrinking) for SWLI 24.62 0.88 0.68 37.91 0.73 40.72 Model Code Name RMSE R2 Jack R2 Jack Boot RMSEP BootR2 RMSEP Component 1 WAPLS Component 1 for SWLI 26.50 0.84 0.69 37.30 0.72 40.99 Component 2 WAPLS Component 2 for SWLI 21.78 0.89 0.66 38.35 0.70 40.83 2 Component 3 WAPLS Component 3 for SWLI 18.74 0.92 0.59 42.20 0.67 43.29 Component 4 WAPLS Component 4 for SWLI 16.72 0.93 0.56 44.75 0.64 45.66 Component 5 WAPLS Component 5 for SWLI 14.92 0.95 0.47 50.39 0.62 48.35 Jack Boot Model Code Name RMSE R2 Jack R2 RMSEP BootR2 RMSEP Component 1 PLS Component 1 for SWLI 26.22 0.84 0.58 43.61 0.55 53.52 Component 2 PLS Component 2 for SWLI 25.44 0.85 0.67 37.88 0.69 41.91 3 Component 3 PLS Component 3 for SWLI 23.48 0.87 0.65 39.44 0.69 41.73 Component 4 PLS Component 4 for SWLI 22.39 0.88 0.66 38.45 0.68 42.40 Component 5 PLS Component 5 for SWLI 21.88 0.89 0.55 44.80 0.67 44.84 65 Figures, Tables & Plates - Chapter 4 Regional Multi-proxy Training Set Jack Boot Model Code Name RMSE R2 Jack R2 BootR2 RMSEP RMSEP WA_Inv Weighted averaging model (inverse deshrinking) for SWLI 14.16 0.86 0.76 18.16 0.75 19.45 WA_Cia Weighted averaging model (classical deshrinking) for SWLI 15.31 0.86 0.77 18.73 0.76 19.72 1 Weighted averaging model (tolerance downweighted, inverse WATOL_Inv deshrinking) for SWLI 13.98 0.86 0.65 22.39 0.73 22.49 Weighted averaging model (tolerance downweighted, classical WATOL_Cia deshrinking) for SWLI 15.08 0.86 0.65 24.00 0.73 22.91 Jack Boot Model Code Name RMSE R2 Jack R2 BootR2 RMSEP RMSEP Component 1 WAPLS Component 1 for SWLI 14.16 0.86 0.76 18.26 0.75 19.47 Component2 WAPLS Component 2 for SWLI 8.40 0.95 0.80 16.65 0.79 18.53 2 Component 3 WAPLS Component 3 for SWLI [, 6.10 0.97 0.82 15.66 0.81 17.90 Component4 WAPLS Component 4 for SWLI 4.79 0.98 0.83 15.33 0.81 17.97 Component 5 WAPLS Component 5 for SWLI 3.92 0.99 0.83 15.55 0.81 18.45 Jack Boot Model Code Name RMSE R2 Jack R2 BootR2 RMSEP RMSEP Component 1 PLS Component 1 for SWLI 19.23 0.73 0.69 20.68 0.69 21.28 Component 2 PLS Component 2 for SWLI 14.98 0.84 0.71 20.20 0.72 20.76 3 Component 3 PLS Component 3 for SWLI 12.80 0.88 0.72 19.73 0.74 20.30 Component4 PLS Component 4 for SWLI 10.68 0.92 0.75 18.65 0.76 19.98 Component 5 PLS Component 5 for SWLI 9.29 0.94 0.77 18.13 0.77 19.83 66 Figures, Tables & Plates - Chapter 4 Kintail Multi-proxy Training Set Jack Boot Model Code Name RMSE R2 Jack R2 BootR2 RMSEP RMSEP WA_Inv Weighted averaging model (inverse deshrinking) for SWLI 16.16 0.88 0.63 29.13 0.65 29.81 WA_Cia Weighted averaging model (classical deshrinking) for SWLI 17.21 0.88 0.63 30.36 0.65 30.43 1 Weighted averaging model (tolerance downweighted, inverse WATOL_Inv deshrinking) for SWLI 17.73 0.86 0.67 27.75 0.69 29.31 Weighted averaging model (tolerance downweighted, classical WATOL_Cia deshrinking) for SWLI 19.13 0.86 0.67 30.00 0.70 29.68 Jack Boot Model Code Name RMSE R2 Jack R2 BootR2 RMSEP RMSEP Component 1 WAPLS Component 1 for SWLI 16.16 0.88 0.63 29.07 0.64 30.16 Component 2 WAPLS Component 2 for SWLI 8.18 0.97 0.73 25.32 0.74 26.87 2 Component 3 WAPLS Component 3 for SWLI 5.06 0.99 0~73 25.00 0.74 26.40 Component4 WAPLS Component 4 for SWLI 3.41 0.99 0.68 27.15 0.71 27.55 Component 5 WAPLS Component 5 for SWLI 2.27 1.00 0.58 31.53 0.68 28.84 Jack Boot Model Code Name RMSE R2 Jack R2 BootR2 RMSEP RMSEP Component 1 PLS Component 1 for SWLI 23.72 0.75 0.63 28.64 0.65 29.10 Component2 PLS Component 2 for SWLI 11.13 0.94 0.78 22.35 0.77 23.97 3 Component 3 PLS Component 3 for SWLI 8.61 0.97 0.77 22.98 0.78 23.78 Component 4 PLS Component 4 for SWLI 6.16 0.98 0.76 23.46 0.77 23.99 Component 5 PLS Component 5 for SWLI 4.92 0.99 0.76 23.37 0.77 24.15 67 Figures, Tables & Plates - Chapter 4 Morar Multi-proxy Training Set Jack Boot Model Code Name RMSE R2 Jack R2 BootR2 RMSEP RMSEP WA_Inv Weighted averaging model (inverse deshrinking) for SWLI 6.20 0.95 0.83 11.47 0.84 12.20 WA_Cia Weighted averaging model (classical deshrinking) for SWLI 6.37 0.95 0.84 11.24 0.84 11.81 1 Weighted averaging model (tolerance downweighted, inverse WATOL_Inv deshrinking) for SWLI 5.39 0.96 0.88 9.71 0.86 13.45 Weighted averaging model (tolerance downweighted, classical WATOL_Cia deshrinking) for SWLI 5.50 0.96 0.89 9.48 0.86 13.21 Jack Boot Model Code Name RMSE R2 Jack R2 BootR2 RMSEP RMSEP Component 1 WAPLS Component 1 for SWLI 6.20 0.95 0.81 12.34 0.83 13.02 Component2 WAPLS Component 2 for SWLI 2.79 0.99 0.88 10.72 0.88 11.88 2 Component 3 WAPLS Component 3 for SWLI I . 1.21 1.00 0.87 10.51 0.88 11.65 Component4 WAPLS Component 4 for SWLI 0.57 1.00 0.88 10.36 0.88 11.63 Component 5 WAPLS Component 5 for SWLI 0.28 1.00 0.88 10.32 0.88 11.62 Jack Boot Model Code Name RMSE R2 Jack R2 BootR2 RMSEP RMSEP Component 1 PLS Component 1 for SWLI 12.50 0.79 0.68 15.64 0.68 16.64 Component 2 PLS Component 2 for SWLI 4.09 0.98 0.81 12.23 0.78 14.14 3 Component 3 PLS Component 3 for SWLI 2.10 0.99 0.84 11.47 0.81 13.59 Component 4 PLS Component 4 for SWLI 1.18 1.00 0.86 10.81 0.82 13.34 Component 5 PLS Component 5 for SWLI 0.79 1.00 0.87 10.63 0.82 13.22 68 Figures, Tables & Plates - Chapter 4 Argyll Multi-proxy Training Set Jack Jack Boot Boot Model Code Name RMSE R2 R2 RMSEP R2 RMSEP WA_Inv Weighted averaging model (inverse deshrinking) for SWLI 10.88 0.88 0.77 14.96 0.79 15.84 WA_Cia Weighted averaging model (classical deshrinking) for SWLI 11.59 0.88 0.78 15.39 0.80 15.72 1 WATOL_Inv Weighted averaging model (tolerance downweighted, inverse deshrinking) for SWLI 8.42 0.93 0.74 16.06 0.81 19.37 WATOL_Cia Weighted averaging model (tolerance downweighted, classical deshrinking) for SWLI 8.74 0.93 0.74 16.03 0.81 19.17 Jack Jack Boot Boot Model Code Name RMSE R2 R2 RMSEP R2 RMSEP Component WAPLS Component 1 for SWLI 1 10.88 0.88 0.77 14.97 0.78 16.09 Component WAPLS Component 2 for SWLI I 2 7.34 0.95 0.80 14.42 0.80 15.93 2 I Component WAPLS Component 3 for SWLI 3 5.10 0.97 0.78 15.52 0.80 16.76 Component WAPLS Component 4 for SWLI 4 3.40 0.99 0.76 16.50 0.79 17.34 Component WAPLS Component 5 for SWLI 5 1.97 1.00 0.75 17.12 0.79 17.53 Jack Jack Boot Boot Model Code Name RMSE R2 R2 RMSEP R2 RMSEP Component PLS Component 1 for SWLI 1 15.14 0.77 0.70 17.19 0.71 17.95 Component PLS Component 2 for SWLI 11.69 0.86 0.74 15.95 0.75 16.85 3 2 Component PLS Component 3 for SWLI 3 9.59 0.91 0.74 16.36 0.76 17.27 Component PLS Component 4 for SWLI 4 7.94 0.94 0.75 15.82 0.77 17.44 Component PLS Component 5 for SWLI 5 6.59 0.96 0.76 16.00 0.78 17.83 69 Figures, Tables & Plates- Chapter 5 Reference water level (m 00) 1.00 1.50 2.00 2.50 3.00 3.50 4.00 1.00 2.00 3.00 38 38 .. I I 40 o I I 40 I I • • • • .I .I I -- I I - - 42 42 • I • I - - 44 44 • I • I -• I - • • - 46 • I • - • --- 46 §...... 48 • • • • • 48 L; - Ci CD 50 0 • • • • 50 • • I • - - 52 • • - - • • - 52 • • - -• - 54 I • -• • - - - 54 56 • I • • - - 56 58 60 !" !" jl fT""' !"Ill"!"...... , , I I I I I 601 I ~20 I' )"""""""""201' !" 20 40 20 20 20 40 Percentage of diatoms -+- Regional Diatom Ref Water Level (m) 0 organic silt-sand peat ...._ Kintail Diatom Ref Water (m) Morar Diatom Ref Water Level (m) EJ organic sand ...._ Argyll Diatom Ref Water Level (m) Figure 5.1 Summary of the results of the diatom RWL reconstruction showing the transfer function inferred values of RWL for core NN04. The diatom training set reconstructions are shown to the immediate right of the diatom diagram and the training sets containing only the samples in the multi-proxy training set are shown on the far right. The grey vertical line shows MHWS at the location of the nearest tidal prediction. 70 Figures, Tables & Plates- Chapter 5 0 .00 1 .00 2 .00 3 .00 4 00 5 .00 6 .0( 38 -- 40 40 I I • - ' . I ' 42 42 • . • :[44 44 . . ' - • • • . I I 0 -g_4 ' ' ' Ill 46 - - - - 'C 48 I I 48 . - I - - -- 5 50 • -· - I ' -· - - - 52 -- f I 52 ' ' I 54 - - 54 2D 40 EDrrrrr--;-'r~rrrrrrr~rr~rr~rrrrrr r rr r r Percenla90 of pollen organic silt-sand peat D -+- Regional pollen ref water level (m -.a- Kintail pollen ref water level (m) 0 organic sand Morar pollen ref water level (m) ~ AI I ollen ref water level m Figure 5.2 Summary of the results of the pollen RWL reconstruction showing the transfer function inferred values of RWL for core NN04. The pollen training set reconstructions are shown to the immediate right of the pollen diagram and the training sets containing only the samples in the multi-proxy training set are shown to the far right. The grey vertical line shows MHWS at the location of the nearest tidal prediction. 71 Figures, Tables & Plates- Chapter 5 Refetence Water Level (m 00) 1.60 2.00 2.40 2.80 3.20 38 40 - • • • - • - I 40 42 • • - • I - I 42 • • I I • I I • • • 44 • • I - • - I • - -- I .c 46 • I I I I • • • - - • ------"6. 46 -E ~ ~48 48 £ Q. <1)50 I I 0 • • • • 50 I I • -• - - • I • --• • - -• 52 • • - • • • • - • I • - - 52 • • • • ------• • • • • • I - -• • I - - I - • - - - - 54 56 58 60 1 1 20r ,...,-. rr r ro 1 • i>' ~~ro r r ~ r r ~rr;(," ~r-r;(,",.,,., Percentage of diatoms & pollen organic silt-sand peat Regional D & P Ref water level (m) D -...... Kintail D & P Ref water level (m) Morar D & P Ref water level (m) D organic sand Argyll 0 & P Ref water level (m) Figure 5.3 Summary of the results of the multi-proxy RWL reconstruction showing the transfer function inferred values of RWL for core NN04. The multi-proxy training set reconstructions are shown to the immediate right of the microfossil diagram. The grey vertical line shows MHWS at the location of the nearest tidal prediction. 72 Figures, Tables & Plates - Chapter 5 0.00 +--<-----.-----.--~-~-~~ 0.00 0.50 ' .00 1.!10 2.00 2.60 3.00 3.60 Klntall Reforenoe Water Low! (m 00) 0 .00 ~--~--~- ODO 1.00 2.00 3DO 400 M orar Retafance Waltl( uwfll (mOO) 400 l 9'3oo j !... .§ ! ~ 2 00 ~ I a:~ 1100 a: 0.00 I 000 0.50 1.00 1 50 2 00 2.50 3 00 3.60 Argyll Rei4Kence Wllt4K Lewl (m 00) Figure 5.4 Scatter plots comparing the regional diatom reference water level reconstruction with the local reconstructions for core NN04. 73 Figures, Tables & Plates- Chapter 5 ~------800 • 500 • 1.00 0.00 +-'---~--~--~ o.oo 1.00 2.00 3.00 •.oo s.oo e.oo Klnlall FWior.- Watar L.... (m 00) 600 • 500 ~ • .• .... • • 100 0.00 ¥------..--~ 0.00 1.00 2.00 3.00 4 00 5 00 11.00 Molar Refefenee Water Lew! (m 00) 0 00 ¥------.--"-T'--,..---.- o.oo 1.00 2.00 3.00 • oo 5 oo _:je oo Argyll Retwenc. W.W Level (m 00) Figure 5.5 Scatter plots comparing the regional pollen reference water level reconstruction with the local reconstructions for core NN04. 74 Figures, Tables & Plates -Chapter 5 0.00 ~--....----~---~-- 0.00 1.00 2.00 300 4.00 Klma~ Relefence Water leYI!I (m 00) 8300s. ! I ~ 2.00 j ! ~ 1.00 ex: 0.00 .IL..----,----~----.- 0.00 1.00 2.00 300 Morar Reference Watt!!' l~ (m 00) 4.00 ~300 ~ J i 2.00 ~ I }•oo 000 L+-.---..---~- 0.00 0.50 1.00 1.50 2.00 2 50 300 350 Algyl1 R-enee Willet LIM!I (m 00) Figure 5.6 Scatter plots comparing the regional diatom and pollen reference water level reconstruction with the local reconstructions for core NN04. 75 Figures, Tables & Plates- Chapter 5 Reference water level (rn 00) 1.00 2.00 3.00 1.00 1.50 2.00 2.50 3.00 0 0 5 5 10 10 10 15 15 15 20 - • • - '. - 20 20 25 • • a- I - 25 E ~ - . - - .t:. ' . - ' . . . - ·-- - 30 ~5 - .. - . - .- 35 0 - .•-- • . -• . -. -• • 35 40 . -.- - 40 40 ,' Jr ~ --~~ L_, . L ., 45 45 - . - · · 50 50 55 60 65 rr;;rrrrr~rrrrr~rr ·~ 40 •rr~rr-;;; 20 •o 60 rrrr Percentage of dtatoms ~ peat D sand egional diatom Ref Water Level (m) D organic silt-day ntail diatom Ref Water Level (m) orar diatom Ref Water Level (m) gy!l diatom Ref Water Level ('!1_ Figure 5. 7 Summary of the results of the diatom RWL reconstructions showing the transfer function inferred values of RWL for monolith MMN01A&B. The diatom training sets reconstructions are shown to the immediate right of the diatom diagram and the diatom training sets containing only the samples in the multi-proxy training set are shown to the far right. The grey vertical line shows MHWS at the location of the nearest tidal prediction. 76 Figures, Tables & Plates - Chapter 5 ~ ~~ Reference water level (m 00) $~~ ,;§> ~~ S',p~fl) ~ (J ~~~ ~ -+- Regional Pollen Ref Water Level (m rn peat 0 sand - • - Kintail Pollen Ref Water Level (m) organic silt-day D Morar Pollen Ref Water Level (m) - e- Argyll Pollen Ref Water Level (m) Figure 5.8 Summary of the results of the pollen RWL reconstruction showing the transfer function inferred values of RWL for monolith MMN01A&B. The pollen training set reconstructions are shown to the immediate right of the pollen diagram and the training sets containing only the samples in the multi-proxy training set are shown to the far right. The grey vertical line shows MHWS at the location of the nearest tidal prediction. 77 Figures, Tables & Plates - Chapter 5 Reference Water Level {m 00 1.00 1.50 2.00 2.50 3.00 0 5 10 10 15 20 • • I • • • • • • ~20 • • • - • - - • - - • • I 25 • • t I - I I I I I I • • - •-· • • - - ~ • - • I • • - I I I • I - 30 I~ - - - I • • - • - I - • • • .c. - • - - • • • • • - 0.35 I - - - I I I CD - - - • • • • • • • - •I - I I I I I I I - - 0 - - - - • • • - 40 -40 - - -. I I I I • - I - • • • - • • - - • I I I • • • - • • I -• 45 ------r • - -I ,. - I I • I 50 ~ L I r • It r t r ~~ .. • • l .. r 5 55 65 1 f"""'"""'r"' rrrf"""'"""' r ~ rr~r rrr r f"""'"""'r"'r"' r I'' ' I • • r-o r r r rf"""'"""' 20 20 20 20 20 20 40 20 Percentage of diatoms and pollen Regional D & P Ref Water Level (m} peat [JJ sand Kintail D & P Ref Water Level (m) organic silt-clay Morar D & P Ref Water Level (m) Argyll D & P Ref Water Level (m) Figure 5.9 Summary of the results of the multi-proxy RWL reconstructions showing the transfer function inferred values of RWL for monolith MMN01A&B. The multi-proxy training sets are shown to the right of the microfossil diagram. The grey vertical line shows MHWS at the loC".ation of the ne::!rest tid::! I nredidion 78 Figures, Tables & Plates- Chapter 5 400 000 ~~-~-~-~- 0.00 0.50 1.00 1.50 2.00 2.50 3.00 350 KlntaJI Rafllrence Wo.. r L...,..l (m 00) 400 , S3oo .!'. ~ ,200 - i J,oo 000 -.------,- 0.00~' 100 200 300 ~00 • 000 +L---.....----- 000 1.00 200 300 400 Atgyll Reference Weier L-(m 00) Figure 5.10 Scatter plots comparing the regional diatom reference water level reconstruction with the local reconstructions for monoliths MMN01A & B. 79 Figures, Tables & Plates - Chapter 5 500 1 500 J •• 8 .. !. ••• t~oo l -" ,-. i ~.. ,300 ~ i • l200 1 a 100 000 0.00 1.00 2.00 300 ~ .00 5.00 5.00 K1111Bi Reference Water L- (m 00) 800 - •••• • ~-.,., 0 00 ,...----- ,.....---- 0.00 1.00 2.00 3.00 ~ 00 5.00 8.00 Morar R.,...noe Water LIM!4 (rn 00) 800 1 5.00 8 E .. J •oo •• ,300 1 • 1200J 100 000 0.00 1.00 2.00 3.00 4.00 500 800 Argyll Ro-W-low! (m 00) Figure 5.11 Scatter plots comparing the regional pollen reference water level reconstruction with the local reconstructions for monoliths MMN01 A & B. 80 Figures, Tables & Plates - Chapter 5 0.00 ~~-~,---~ 0.00 1.00 2.00 3.00 4.00 KJntalf Reference Wallf LM (m 00) 000 ~--~--- 000 1.00 2.00 300 400 Moror Rl!ferenoe Wotef LM (m 00) • .00 • 000~--~------000 1.00 200 300 •oo Figure 5.12 Scatter plots comparing the regional diatom and pollen reference water level reconstruction with the local reconstructions for monolith MMN01 A & B. 81 Figures, Tables & Plates - Chapter 5 5.00 400 3.5:) 83.00 .E. ~ 25:) It: j 2.00 0 IL 1.5:) 1.00 0.5:) 0.00 ~--~----~--~----~--~ 0.00 1.00 200 3.00 •t 00 5.00 l\\.dtl1>rc»ty R'M... (m cq 5.00 4SJ 4.00 s3SJ 0 + .§_ 300 ~ 2SJ a: j 2CD ~ 1.SJ 1.00 QSJ 000 ~--~----~--~----~--~ oro 100 200 300 4.00 5.00 Dabn IW.. (m CI:J 5.00 450 4.00 0'350 0 .§.300 ~ 200 1 200 o 1.50 1.00 050 000 *----.----r---~--~~--~ 0.00 1.00 200 lOO 4.00 5.00 MJI:I..paKy IW...(mCQ Figure 5.13 Bi-plots comparing the RWL reconstruction values produced for the regional diatom, pollen and multi-proxy training sets using only the samples within the multi-proxy training set to allow direct comparison. 82 Figures, Tables & Plates- Chapter 5 - • • - - • • 50 -+----- . .... • • • • 52 oi---1--- · -.. -· · ·.--- •. - I I a I • • • 54 .. - E' ~ • • ~56 1-+- . .. . - - .s:: Ci. I I • ~ ~58 : G) I I • • fso Ill -~ --+-- - - - (/) • • • -- • 62 r- • - . - - 64 t- ~ • • • • - • 66 -r • - - 68 r-rnrnrnrnrornrnrnrrr,-r,i'''''''''''''' •rrnr'"'"""rl'' , ,, •rrrn 20 40 ro ao 20 20 20 40 ro ao 20 20 40 Percentage of pollen Regional Pollen Ref Water Level (m) Kintail Pollen Ref Water Level (m) m organic silt-sand peat Morar Pollen Ref Water Level (m) Argyll Pollen Ref Water Level (m) Figure 5.14 Summary of the results of the pollen RWL reconstructions showing the transfer function inferred values of RWL for core BC006a. The pollen training set reconstructions are shown to the immediate right of the pollen diagram. The grey vertical line shows MHWS at the location of the nearest tidal prediction. 83 Figures, Tables & Plates - Chapter 5 500 ••••• 000 ¥---~----~--~- 0.00 1.00 200 3.00 400 500 0.00 +'I'----...---~--...--- 000 1.00 2.00 3.00 4.00 5.00 5.00 8 4 00 g, ! 300 ~ I ex1200 ~ }100 ~ 000 1-!-----1---,.- 0.00 1.00 200 300 -400 500 Figure 5.15 Scatter plots comparing the regional pollen reference water level reconstruction with the local reconstructions for core BC006a. 84 Figures, Tables & Plates - Chapter 5 No. of m1crofosslls Modem Core Training Set Microfossil In fOSSil analogues (%) dataset NN04 Regional Diatom 65 89.2 NN04 Kintail Diatom 65 83.1 NN04 Morar Diatom 65 58.5 NN04 Argyll Diatom 65 72.3 NN04 Regional Pollen 33 87.9 NN04 Kintail Pollen 33 75.8 NN04 Morar Pollen 33 78.8 NN04 Argyll Pollen 33 81 .8 NN04 Regional Diatom and pollen 91 92.3 NN04 Kintail Diatom and pollen 91 79.1 NN04 Morar Diatom and pollen 91 72.5 NN04 Argyll Diatom and pollen 91 81 .3 No. of microfossils Modern Core Training Set Microfossil in fossil analogues(%) dataset MMN01 Regional Diatom 26 73.1 MMN01 Kintail Diatom 26 53.8 MMN01 Morar Diatom 26 57.7 MMN01 Argyll Diatom 26 57.7 MMN01 Regional Pollen 45 68.9 MMN01 Kintail Pollen 45 55.6 MMN01 Morar Pollen 45 62.2 MMN01 Argyll Pollen 45 71 .1 MMN01 Regional Diatom and pollen 65 78.5 MMN01 Kintail Diatom and pollen 65 60.0 MMN01 Morar Diatom and pollen 65 64.6 MMN01 Argyll Diatom and pollen 65 67.7 I No. of . microfossils Modern Core Trammg Set Microfossil in fossil analogues (%) dataset I BC006a Regional Pollen 21 76.2 BC006a Kintail Pollen 21 71.4 BC006a Morar Pollen 21 71.4 BC006a Ar II Pollen 21 71 .4 Table 5.1 Lists the number of microfossils recorded in the fossil cores and the percentages of microfossils also found in the modern training set (% of modem analogues). 85 Figures, Tables & Plates- Chapter 5 No. of m1 crofosslls Modern Core Tramrng Set Microfossil m fossil analogues ('~o) dataset NN04 Regional Diatom 65 87.7 NN04 Kintail Diatom 65 76.9 NN04 Morar Diatom 65 56.9 NN04 Argyll Diatom 65 67.7 NN04 Regional Pollen 33 75.8 NN04 Kintail Pollen 33 69.7 NN04 Morar Pollen 33 66.7 NN04 Argyll Pollen 33 78.8 No. of microfossils Modern Core Trammg Set M1crofossil 0 111 fossil analogues ( /..) dataset MMN01 Regional Diatom 26 73.1 MMN01 Kintail Diatom 26 46.2 MMN01 Morar Diatom 26 50.0 MMN01 Argyll Diatom 26 50.0 MMN01 Regional Pollen 42 73.8 MMN01 Kintail Pollen 42 54.8 MMN01 Morar Pollen 42 59.5 MMN01 Argyll Pollen 42 66.7 Table 5.2 Lists the number of microfossils recorded in the fossil cores and the percentages of microfossils also found in the modern training sets (% of modern analog ues) for the diatoms and pollen samples within the multi-proxy training sets. 86 Figures, Tables & Plates - Chapter 5 APPENDIX 5.1 -Stratigraphy ~ Site: Nonach, Core NN04 Altitude m Corrected Observed Description OD depth, em depth, em nig, straf, sic, elas, lim 3.97- 0-1 In-situ sphagnum (substratum confusum) 3.96 3.96- 1-45 30200 3.52 Sh2 Th22 Ag+ Dark brown herbaceous peat with occasional silt. 3.52- 45-52 30200 3.45 Ag1 As1 Th2+ Sh2 Brown organic silt and sand. 3.45- 52-65 30200 3.32 As3 Ag+ Dl+ Gg+ Brown organic silt sand with occasional woody fragments and gravel. 87 Figures, Tables & Plates - Chapter 5 Site: Mointeach Mhor, Core MMN01/AB Altitude m Corrected Observed Description OD depth, em depth, em nig, straf, sic, elas, lim 9.37- 0-7 30200 9.30 Sh3 Th21 Very dark brown soft wet humified peat with rootlets and visible organic fragments. 9.30- 7-33 20200 9.04 Sh3 Th21 Ga+ Dark brown humified peat with frequent rootlets and occasional sand. 9.04- 33-45 30200 8.92 Sh3 Ga1 Ag+ Th2+ Organic sand with silt and occasional rootlets. 8.92- 45-50 30200 8.87 Sh 2 Ag 1 Th2 1 As+ Organic silt clay with organic fragments and rootlets. 8.87- 50-65 30200 8.72 Sh 3 Th2 1 Very dark brown humified peat with large organic fragments and rootlets. 88 Figures, Tables & Plates - Chapter 5 ~ Site: Barr na Criche, Core BC006a Altitude m Corrected Observed Description 00 depth, em depth, em nig, straf, sic, elas, lim 3.34- 37-60 3020- 3.11 Sh 3, Th 3 1 Humified herbaceous peat. 3.11- 60-79 20200 2.92 Sh 1, Ag 2, Ga 1, Th 3 + Slightly laminated organic silt and sand with herbaceous rootlets. 2.92- 79-87 20200 2.84 Ga 4, Ag +, Gg +, Sh +, Th3 + Slightly organic sand with rare rootlets, silty towards top, gravel at base. (Core ended on gravel). 89 Figures, Tables & Plates - Chapter 6 Reference Water Level (mOD) 1.00 200 3.00 1.00 1.50 2.00 2.50 3.0C ./ 0 --'---"-- 5 5 8.452-6716 10 10 15 15 - -• - ' - - - 20 20 -- I 25 25 - • • • - • - - 7309-7551 - 30 ' I • • • I • - 30 - . • • • - ' . .- . - . . - 35 35 - --• -• -• • - • - .- I - I - - - 40 40 l r ~ r r 45 45 • L r 7560-7786 J= ~ r ~ . . . !. - ll~ r 50 50 ~r r r r r ~ r r r r r ~ r r""'"":I0..,...... ,.~.0~ r r ~r 1 -·;; .-.-20...-~.o-.....so r r r r Percentage or diatoms ~ peat sand D -+-Regional diatom Ref Water Level (m) D organic silt-day Kintail diatom Ref Water Level (m) Morar diatom Ref Water Level (m) _._ II diatom Ref Water Level m Figure 6.1 Summary of the results of the diatom-inferred reference water level reconstruction for fossil monolith sample MMN01A & B. The depths of the sea-level index sample points are shown as black squares and the associated radiocarbon ages are shown adjacent to the lithology. The RWL reconstruction for the regional diatom training set is shown to the immediate right of the diatom diagram and the RWL reconstruction for diatom training set samples from the multi-proxy training set are shown on the far right. The dotted vertical line indicates the approximate position of the predicted MHWS at Mallaig (2.38 m 00) 90 Figures, Tables & Plates- Chapter 6 yrBP Reference Water Level (m OD) 3.00 5.00 7.00 1.00 3.00 5.00 7.00 9.( 0 . • . ' 5 5 . - ' . - 10 10 2 - 10 . . - I 6 • I I -- I I I - ' I 15 15 15 • • - .• ' . - ' - --- • 20 20 - - - • ' 20 • • - -I • -• - - - . ' E 25 25 1 25 - - - - .2. I I • . to ' =Q. 3o 30 1 I g. £: -- . - . - Ql . 35 1 ~ - - - . - - 0 35 I - 40 <10 . . - 40 45 ' 45 > • .• . --I .• - -r ' - 45 0 r I' : f - -I : . - =- . . . ~ f r. . : • . 50 -·.. ·. ~ -= ··· ·· ·... 50 .- . ~ ~ . - . = - - s 55 . - so l 55 -· ' 55 ' . - - 60 • I • - -- · -· 60 ------60 ·•·· 65 r'"'f"'"'r'"'r->r~~r->rr-r;or-> <10 r r;or->rf"'"'r f"'"' f"'"'rr;orr;o ~"'"' ~ ~"'"' 40 20 60 Percentage of polen !-+- Regional Pollen Ref Water Level (m [8 peat D sand Kintail Pollen Ref Water Level (m) D organic silt-day Morar Pollen Ref Water Level (m) - - Argyll Pollen Ref Water Level (m) Figure 6.2 Summary of the results of the pollen-based reference water level reconstruction for fossil monolith sample MMN01A & B. The depths of the sea-level index sample points are shown as black squares and the associated radiocarbon ages are shown adjacent to the lithology. The RWL reconstruction for the regional pollen training set is shown to the immediate right of the pollen diagram and the RWL reconstruction for pollen training set samples from the multi-proxy training set are shown on the far right. The dotted vertical line indicates the approximate position of the predicted MHWS at Mallaig (2.38 m 00). 91 Figures, Tables & Plates - Chapter 6 Cal yrBP Reference Water Level (m 00) 1.00 1.50 2.00 2.50 3.00 0 5 10 10 6452-6716 15 20 • • • • • • • • • • • • • - • - I - I - - I • I • 25 • I • I - I - I • • --I I I - 'E' • I I • - • • • • • • • I 30 ,£.30 - - - - I - - - .r; • - - • • • • • • • - 7309-7551 0.35 • I - I I • • • I • -I I I - - - I - I I - (I) • • • • • • • • - I - I - I - - I I - 0 • • • - • • • • • - - - 40 - 40 I • - I - - • • • • • I - I I - - I • I - • - • • • • • • - - • I 45 - .. r , - r • - I r I" r - -r 7515-nto - - .. I I .. I I ... I I r I' I , -. I I I 50 • I I I .. r 1 I r I ..r r .. .. ' • -- " 50 55 ---· 60 65 ,....,_..., r r r r ,....,_..., r r-'"'"'"'"' r r ~ r r r r r r-r""" r r r 1 • • • • • • • • r r r r r r-r""" 20 20 20 20 20 20 40 20 Percentage of diatoms and pollen ~ Regional D & P Ret Water Level (m) ~ peat D sand Kintail D & P Ref Water Level (m) D organic silt-clay Morar D & P Ref Water Level (m) Argyll D & P Ref Water Level (m) Figure 6.3 Summary of the results of the multi-proxy-based reference water level reconstruction for fossil monolith sample MMN01A & B. The depths of the sea-level index sample points are shown as black squares and the associated radiocarbon ages are shown adjacent to the lithology. The dotted vertical line indicates the approximate position of the predicted MHWS at MallaiQ (2.38 m 00). 92 Figures, Tables & Plates - Chapter 6 116 WAPLS Component 13.97 0.93 0.83 24.83 0.58 0.52 0.36 Regional diatoms 3 for SWLI 43 WAPLS Component 11 .23 0.96 0.82 27.79 0.64 0.58 0.41 l 22 WAPLS Component 2.76 0.99 0.91 12.46 0.29 0.26 0.18 ~orar diatoms 3 for SWLI 51 WAPLS Component 9.97 0.97 0.87 24.06 0.56 0.51 0.35 3 82 PLS Component 2 39.44 0.56 0.36 50.89 1.18 1.07 0.75 Regional pollen for SWLI 25 PLS Component 3 34.63 0.66 0.31 57.61 1.34 1.21 0.85 Kintail pollen for SWLI 15 PLS Component 3 6.49 0.94 0.65 18.44 0.43 0.39 0.27 forSWU 42 WAPLS Component 18.74 0.92: 0.67 43.29 I 1.00 0.91 0.64 . 3 SWLI 75 WAPLS Component 6.10 0.97 0_81 17.90 0.42 0.38 0.26 3 forSWLI 23 WAPLS Component 5.06 0.99 0.74 26.40 0.61 0.55 0.39 3 for SWLI 15 WAPLS Component 1.21 1.00 0.88 11 .65 0.27 0.24 0.17 3 for SWLI 37 WAPLS Component 7.34 0.95 0.80 15.93 0.37 0.33 0.23 2 for SWU Table 6.1 For each transfer fu nction model that was considered the best model, the RM SEP in SWLI units has been converted into m OD for each of the fossil sample sites taking into account the local tidal parameters. 93 Figures, Tables & Plates - Chapter 6 75 8.11 0.95 0.78 20.35 0.47 0.43 0.30 WAPLS Component 23 7.29 0.98 0.71 27.91 0.65 0.59 0.41 2 for SWLI WAPLS Component 15 1.56 1.00 0.87 11.22 0.26 0.24 0.16 3 for SWLI WAPLS Component 37 6.61 0.96 0.76 18.92 0.44 0.40 . 0.28 3 for SWLI WAPLS Component 75 18.39 0.76 0.61 24.73 0.57 0.52 0.36 2 for SWLI WAPLS Component 23 13.98 0.91 0.55 35.71 0.83 0.75 0.53 3 for SWLI WAPLS Component 15 3.53 0.98 0.76 16.99 0.39 0.36 0.25 3 for SWLI WAPLS Component 37 6.55 0.96 0.74 17.93 0.42 0.38 0.26 3 for SWLI Table 6.2 For each transfer function model that was considered the best model for the training sets containing only samples from the multi-proxy training set, the RMSEP in SWLI units has been converted into m OD for each of the fossil sample sites taking into account the local tidal parameters. 94 Figures, Tables & Plates - Chapter 6 Indicative 14C age BP and Elevation Median, maximum, minimum Relative sea level and Sea-level Site Lab code meaning error (moD) calibrated age (yr BP) error tendency (m) Mointeach Mhor North MMN01 AA54113 5771 47 9.27 2.73 6575 6716 6452 6.54 0.2 Negative Mointeach Mhor North MMN01 AA54112 6499 47 9.04 2.58 7394 7551 7309 6.46 0.2 Negative Mointeach Mhor North MMN01 AA54111 6786 49 8.91 2.18 7635 7710 7515 6.73 0.2 Positive Mointeach Mhor North MMN01 AA54110 6836 59 8.87 2.28 7671 7786 7580 6.59 0.2 Positive Table 6.3 Sea-level index points from Arisaig area (Shennan et a/., 2005). 95 Figures, Tables & Plates - Chapter 6 Regional Training Set (m Regional Training Set Regional Training Set Table 6.4 Shows the results of the diatom, pollen and multi-proxy RWL reconstructions and associated error terms and the RWL and error terms for the sea-level index (SLI) points from Mointeach Mhor North (after Shennan et a/., 2005). 96 Figures, Tables & Plates - Chapter 6 Regional I Training Training Set (m) Regional Training Set RWL Error (m) (m) Table 6.5 Shows the results of the diatom and pollen RWL reconstructions for the samples within the multi-proxy training set and associated error terms and the RWL and error terms for the sea-level index (SLI) points from Mointeach Mhor North (after Shennan eta/., 2005). 97