Mon_Earl & Dargie_Interpreting Sand Dune Habitat change at

Report
Graham Earl & Tom Dargie: Part 1
Interpreting sand dune habitat change at Sandwich Bay using sequential NVC survey
Dr Tom Dargie – Boreas Ecology
1.
2.
3.
4.
5.
Nature Reserve
Context: primary slacks, adjacent effects
NVC mapping: 1989, 2001, 2008
Changes in habitat extent
Directions of habitat change
Quadrat analysis of change
Princes
Golf Course
Nature
Reserve
Sandwich
Royal St George’s
Golf Course
Bird
Observatory
Sandwich
Bay
Estate
Royal Cinque Ports
Golf Course
© Google Earth
Deal
Himantoglossum
hircinicum
Lizard Orchid
Ellenberg F = 3
Ellenberg N = 2
Orobanche
caryophyllacea
Clove-scented Broomrape
Ellenberg F = 3
Ellenberg N = 2
NVC Mapping: Initial Survey 1989 (Sand Dune Vegetation Survey of Great Britain)
(Doarks, Hedley, Radley & Woolven (1990) NCC CSD Report No. 1126)
Digitised version of 1989 map used
in Dargie (2002)
NVC Mapping and Mapping Change: Follow-up Survey (2001)
(T. Dargie (2002) Sandwich Bay Dunes, Kent: NVC Survey 2001. Unpublished Report to English Nature (Kent Team))
NVC Mapping and Mapping Change: Third Survey (2008)
Royal St George’s Golf Club
(T. Dargie (2009) National Vegetation Classification (NVC) survey of sand dunes at Sandwich Bay, Kent: 2008 Repeat NVC
Survey of Royal St. George’s Golf Club. Unpublished Report to Natural England. NE Contract 6/NE/0126)
Sandwich Bay Habitat Extents 1989 - 2001
Golf course (bunker, fairway, green, tee, other mown ground, tracks)
Slack: Wet neutral grassland (MG9, MG10, MG11, MG12, MG13)
Dry neutral grassland (MG1)
Improved grassland (MG6, MG7)
2001
Transitional dry dune grassland (SD8 with MG1, MG6, MG7)
1989
Fixed dry dune grassland (SD8, SD10, SD11, SD12)
Semi-fixed dune (SD7, SD9)
0
40
80
120
Extent (ha)
160
200
Royal St George’s GC Habitat Extents 1989 – 2001 - 2008
Golf course (bunker, fairway, green, tee, other mown ground, tracks)
Slack: Wet neutral grassland (MG9, MG10, MG11, MG12, MG13)
Dry neutral grassland (MG1)
2008
Improved grassland (MG6, MG7)
2001
Transitional dry dune grassland (SD8 with MG1, MG6, MG7)
1989
Fixed dry dune grassland (SD8, SD10, SD11, SD12)
Semi-fixed dune (SD7, SD9)
0
20
40
Extent (ha)
60
80
Main Pathways of Change
(hectares)
mainly
SB
1989
to
2001
mainly
mainly
SD8
MG1
Festuca rubra - Galium verum
Arrhenatherum elatius
123
49
25
MG12 slack
Schedonorus arundinaceus
(Festuca arundinacea)
8
20
9
11
7
RSG
2001
to
2008
51
5
1
in cattle grazing enclosure
15
Quadrat Analysis
Step 1 Pseudo-random quadrat (‘pseudoquadrat’) approach
Rhind et al. (2006) Proc. Royal Irish Academy 106B first applied it to dune habitat
Simulates species data for 25 quadrats per NVC sub-community based on
(i) Published range in Domin scores
(ii) Published constancy (frequency) class (ranging from I to V)
(iii) Published range and average number of species per quadrat
Dr Roy Sanderson (Newcastle) supplied pseudoquadrat data for dune NVC types
Step 2 Ellenberg Scores Gradient Analysis
Calculate F Moisture and N Nitrogen for pseudoquadrats and RSG quadrats
Calculate mean and standard deviation for F and N per NVC (sub)community
Graph probability ellipses (3 SD widths) for NVC communities on F & N axes
Graph individual quadrat positions, colour coded by year
Ellenberg Gradient Analysis: Pseudoquadrat Probability Ellipses
Ellenberg Gradient Analysis: Royal St George’s Golf Course Quadrats
Conclusions
•
GIS evidence suggests that SD8 fixed dune grassland is changing quite rapidly to MG1 and MG12
•
Quadrat change analysis suggests the same
•
(Repeat analysis of target notes shows the same)
•
The future for key Red Data Book species at this site seems threatened by hydroecological change

Hydro-ecological Investigation of Habitat
Change: Based on Sandwich Bay Sand
Dunes, Kent
Interpreting sand dune habitat change at Sandwich Bay using
sequential NVC surveys and hydrological analysis
Graham C J Earl
Coastal Eco-hydrological Researcher
Canterbury Christ Church University
Overview of Research
My research is investigating research recommended in the Dargie
(2009) report:
• The possible interactions of physicochemical values within the
groundwater;
• The interactions between vegetation communities and nutrient
levels;
• The possible influences from golf course management;
• A possible long-term increase in the water table, possibly driven
by sea-level rise.
The focus of this presentation:
– The relationship between hydrological dynamics,
physicochemical factors and overlying sand dune vegetation.
Dipwell Positions Across
Sandwich Bay
• Dipwells chosen using
Judgement Quota
sampling;
• Two tier approach, using
historic data and ground
truthing;
• Locations limited by access
and position of water table.
Sampling Schematic
Dipwell Elevation obtained
from LiDAR data, adjusted to
OS Newlyn datum
Chemical Analysis
• Water samples collected monthly June 2012 – November 2012,
with a further collection quarterly from March 2013 – June 2014,
and analysed for:
–
–
–
–
–
–
–
Sodium (NaCl)
Potassium (KCl)
Total oxidised nitrogen (TON; NO2 + NO3)
Phosphate (P)
Ammonia (NH3)
pH
Electrical conductivity (EC)
Observed Total Oxidised Nitrogen
Concentration
Summer
2012
Autumn
2012
Winter
2012
Spring
2013
Observed Ammonia Concentration
Summer
2012
Autumn
2012
Winter
2012
Spring
2013
Observed Phosphate Concentration
Summer
2012
Autumn
2012
Winter
2012
Spring
2013
Null Hypotheses
• There is no significant difference in nutrient concentrations
between vegetation communities;
– Preliminary results using a nonparametric Kruskal-Wallis test,
on chemical values comparing between different NVC classes
near dipwells;
Summer 2012 Summer 2013
NaCl
P = 0.415
P = 0.623
KCl
P = 0.032
P = 0.213
PO4
P = 0.797
P = 0.197
TON
P = 0.055
P = 0.884
NH3
P = 0.456
P = 0.141
Note: data collection is on going until June 2014, thus analysis is
currently on an incomplete data set.
Historic Weather Dynamics and NVC
Communities
Surveyed Years
Historic Weather Dynamics and NVC
Communities
Observed Water Table Levels
Summer
2012
Autumn
2012
Winter
2012
Spring
2013
Null Hypotheses
• There is no significant difference between ground
elevation and vegetation community cover;
– Nonparametric test Kruskal-Wallis, on LiDAR elevation data
comparing between different NVC classes near dipwells;
1989 NVC Cover
P = 0.038
2001 NVC Cover
P = 0.010
2012 NVC Cover
P = 0.012
2013 NVC Cover
P = 0.018
Adjusted data to ensure samples with ≥3 replicates were
analysed.
Summary
• Data collection on-going (20 months out of a 24 month
sample period has been collected);
• Only simple non-parametric methods have been applied
so far in data analysis, multivariate analysis will be run
on a complete data set;
• Repeat survey of tidal fluctuation effects on the water
table height.
Acknowledgements
• Funding support: RCPGC and Natural England;
• Access permission and other support: RCPGC, RSGGC,
Princes GC, Alan Husk, and Sandwich Bay Residents;
• Help in the field in sometimes poor conditions, plus
research advice (the sun does not always shine over
Sandwich!): Phil Buckley, Phil Williams, David
Ponsonby, Alex Kent, John Hills (CCCU), Phil Williams
(Natural England) , Tom Dargie (Boreas Ecology),
Students at CCCU, Friends and Family.
Thank you for listening
Questions Welcome
Contact:
[email protected]

similar documents