GLOBOLAKES
Global Observatory of Lake Responses to Environmental Change
Andrew Tyler1 & Peter Hunter2
Biological and Environmental Sciences, University of Stirling
Group on Earth Observation (GEO) Webinar
26 April 2013
1 a.n.tyler@stir.ac.uk
2 p.d.hunter@stir.ac.uk
•
•
•
•
•
•
Research Rationale
Aims and Objectives
Data Acquisition
Lake Selection
Coherence in Lake Response
Summary
The Consortium
Andrew Tyler, Peter Hunter, Evangelos
Spyrakos: University of Stirling, UK
Steve Groom, Victor Vicente-Martinez,
Gavin Tilstone, Giorgio Dall’Olmo:
Plymouth Marine Laboratory, UK
Christopher Merchant, Stuart MacCallum:
University of Edinburgh/University of
Reading, UK
Mark Cutler, John Rowan, Terry Dawson,
Eirini Politi: University of Dundee, UK
Stephen Maberly, Laurence Carvalho,
Stephen Thackery, Alex Elliott: Centre for
Ecology & Hydrology, UK
Claire Miller, Marion Scott:
University of Glasgow, UK
Rationale
• >300 million lakes globally
• Providing essential ecosystem goods &
services
• Fundamental to global food security
• Global concerns over future water
security (Unsustainable use; MEA
2005)
• Important in global biogeochemical
cycling (Bastviken et al. 2011, Science)
Geographic distribution of the world’s 200 largest lakes. Data courtesy of ESA-funded ARCLakes project (PI Christopher
Merchant, University of Edinburgh)
Rationale
• Lakes are ‘sentinels’ of environmental change
• These can trigger internal interactions & direct
responses leading to:
–
–
–
–
loss of habitat
eutrophication
fish kills
loss of species (highest proportion of species threatened
with extinction; MEA 2005)
– altered communities & shifts to less desirable species
Rationale
SeaWifs: 1997-2010
Timeliness:
• Increasing robustness of algorithms and ensemble
approaches
• Capability for processing huge data volumes in near real
time
• MERIS: spectral and temporal resolution (until April 2012)
• GMES: ESA planned launches – superior capabilities (2014)
Opportunity:
Access to nearly 20 years of data on 1000 lakes of different
types across the globe will give a paradigm shift in our ability to
ask fundamental ecological questions in relation to the status
and change in the condition of the world’s lakes
MERIS: 2002-2012
Sentinel: 2 2014
Sentinel: 3 2014
Questions
What controls the differential sensitivity of
lakes to environmental perturbation?
Some pressing questions:
•
What is the present state & evidence for long-term change for the 1000
lakes?
•
To what extent are patterns temporally coherent & what are the causes?
•
Is there evidence for phenological change & what are the causes?
•
What factors control cyanobacterial blooms?
•
What factors control the concentration of coloured DOC?
•
How sensitive are different lake types to varying environmental
perturbation?
•
Can we forecast the future response of phytoplankton composition &
abundance, & risk of cyanobacterial blooms, for lakes in different
landscapes?
Aims and Objectives
Derwent Water
Investigate the state of lakes & their response to
environmental change drivers:
•
Near real time processing satellite based
observatory
•
Processing archived data for up to 20-year time
series
•
Including:
(i) LSWT; (ii) TSM; (iii) CDOM; (iv) Chl a; (v) PC
•
Detect spatial & temporal trends & attribute causes of
change for 1000 lakes worldwide (1/3 of inland water,
2/3 of all inland water > 1km2)
•
Forecast lake sensitivity to environmental change
•
Apply findings into lake management
Lake Balaton
Ice Cover
Project structure
WP1 and WP2
• Selection and operationalization
of algorithms for retrieval of lake
biogeochemical properties
• Initially MERIS archive, but later
Sentinel-3 OLCI
• Extension of global lake surface
water algorithms (from ArcLakes)
to smaller lakes (ATSR & SLSTR)
• In situ data from research cruises,
long-term monitoring programmes
and instrumented buoys
WP2: Operational
processing
WP1: Algorithm validation
Study lakes
Level 1 Case study lakes
(~15)
Intensive field campaigns to
test algorithms
Testing
In situ optics and biogeochemical validation data
UK lakes plus selected international lakes with good data
availability (IOPs, AOPs, in situ data for MERIS match-ups)
study lakes
Refining
Level 2 International
lakes (~50)
Validation of most
promising algorithms using
partner data sets from
contrasting
lakes
Validation
Level 3 Global lake
population (~1000)
Operationalization of
selected algorithms
In situ biogeochemical validation data:
UK; Estonia; Hungary, Italy; Spain
USA; Canada; Australia; South Africa, Kenya,
High temporal resolution buoy data
UKLEON – NERC funded buoy sensor network
GLEON – NSF funded international buoy sensor network
Level-2/-3 products delivered to end-users
Time-series biogeochemical data provided to project
partners and end-users
GloboLakes data
Partner data
Atmospheric correction
Additional credits: Stephanie Palmer, Jose Antonio Dominguez, Brockmann Consult
Chlorophyll algorithms
Chla mg m-3
Fluorescence
= 0.75
RMSE = 4.3 mg m3
Bias= 0.5 mg m3
20
10
0
10
0
40
10
20
30
Lab-analyzed Chl-a (mg m-3)
40
R2
= 0.41
RMSE = 8.1 mg m3
Bias= -4.5 mg m3
20
10
0
10
10
20
30
Lab-analyzed Chl-a (mg m-3)
40
FUB/WeW Chl-a
120
R2 = 0.64
RMSE = 15.9 mg m3
Bias= 6.7 mg m3
105
90
75
60
45
30
15
0
0
Case
C2R
Chl-a2 Regional
30
MERIS-derived Chl-a (mg m-3)
MERIS-derived Chl-a (mg m-3)
20
R2 = 0.65
RMSE = 4.9 mg m3
Bias= -0.001 mg m3
20
0
R2 = 0.42
RMSE = 7.3 mg m3
Bias= -0.8 mg m3
30
0
30
40
Boreal
Lake
Boreal
Lake
Chl-a
40
MERIS-derived Chl-a (mg m-3)
10
20
30
Lab-analyzed Chl-a (mg m-3)
MERIS-derived Chl-a (mg m-3)
R2
30
0
Maximum
Chlorophyll
Index
40
L1b
FLHHeight
Chl-a
Line
MERIS-derived Chl-a (mg m-3)
MERIS-derived Chl-a (mg m-3)
40
40
15
30 45 60 75 90 105 120
Lab-analyzed Chl-a (mg m-3)
Eutrophic
LakeLake
Chl-a
Eutrophic
30
R2 = 0.38
RMSE = 8.2 mg m3
Bias= -4.4 mg m3
20
10
0
0
0
10
20
30
Lab-analyzed Chl-a (mg m-3)
40
0
10
20
30
Lab-analyzed Chl-a (mg m-3)
40
Additional credits: Stephanie Palmer, Astrium
LIMNADES
• Lake bIo-optica​l MeasuremeNts And matchup Data
for rEmote Sensing (LIMNADES)
• LIMNADES provides a repository for:
1) inherent and apparent optical properties for
algorithm development
2) in situ water constituents measurements for
algorithms validation
• Vision is to provide long-term archive running
beyond lifetime of GloboLakes
• Find out more: www.globolakes.ac.uk/limnades
LIMNADES
Partners
Quality control
Related projects
Other contributors/users
Uncertainties
Privately shared
Publicly shared
DATA ACCESS & POLICY:
http://www.globolakes.ac.uk/limnades/data_access_policy.html
LIMNADES
GloboLakes partner lakes
Other data contributed to LIMNADES (Apr 2013)
WP3: Lake selection
Lake Landscape Context
(after e.g., Sorrano et al. 2009)
•
Incorporate a wide range of catchmentto-lake-surface-area ratios
•
Span a wide range of water quality
parameters and ecological
characteristics, e.g. pH, alkalinity,
eutrophication status and mixing
regime
•
Include lakes of special scientific
interest
•
Ideally a pool from which lakes
selected using a randomised,
probability-based approach (Stevens
1994).
Global lake population
Currently 1,811 under consideration
WP3: Drivers of change
• Spatial database of drivers of catchment change, including:
–
–
–
–
–
–
climate
land cover
catchment morphology
productivity
development
lake hydromorphology etc.
• Trends in catchment change derived from global datasets (30
years of change)
• Run-off modelled for each catchment using a lumped GISbased model
WP3: Catchment data
Road network
River network & Dams
Climatic ecoregions
Soil type
NDVI
Land cover
Annual mean temperature
Annual mean precipitation
Elevation
Case Study: ARC-Lake data
LSWT Time Series (257 standardized LSWT time series)
Finazzi, F., Miller, Cl., Scott, M.
Case Study: ARC-Lake data
Clustering Result
Finazzi, F., Miller, Cl., Scott, M.
Case Study: ARC-Lake data
Clustering Result – Global map
Finazzi, F., Miller, Cl., Scott, M.
Engagement
Success of GloboLakes will rely on contributions from
across the EO and end-user communities
•
More than 20 scientific partners from over 15 nations
•
CSIRO, Australia; CSIR, South Africa; VITO, Belgium
•
Environment Canada; Estonian Marine Institute;
•
EC Joint Research Centre; CNR-IREA, Italy;
•
INTA, Spain; CUNY, USA; Creighton, USA
•
South Florida, USA; Institute of Limnology, Nanjing…
•
Engagement with end-users including UK environmental regulators
(EA, SEPA, NIEA)
•
Engagement with UK National Centre for Earth Observation (NCEO),
European Environment Agency, ESA and GEO
Linkages and Developments
•
Brockmann Associates
•
Lake Algal Bloom Pilot Project
•
KTAMOP: Ecological Status and Function of Lakes (Hungarian Academy of
Sciences)
– HICO (Hypespectral Imager of the Coastal Ocean)
•
Validation team for ESA Sentinel 3
•
•
INFORM: Improved monitoring and forecasting of ecological status of European
Inland waters by combining Future earth ObseRvation data and Models (FP7
project pending): VITO, Belgium
GLaSS: Global Lakes Sentinel Services: Water Insight, Wageningen
•
DANCERS: DANube macroregion: Capacity building and Excellence in River
Systems (basin, delta and sea): GeoEcoMar, Romania
•
CYANOCOST: Training and networking for EO
•
NETLAKE: COST Action for in-situ sensors
www.globolakes.ac.uk
Thank you!