Aquatic applications of a HEROlike moderate resolution hyperspectral mission

1
Aquatic applications of a HERO-like moderate
resolution hyperspectral mission
Gary Borstad and Bob Bukata G. A. Borstad
Associates Ltd., Sidney BC gary_at_borstad.com Natio
nal Waters Research Institute, Environment
Canada, Burlington, Ontario robert.bukata_at_ec.gc.ca
BCCARMS Workshop on Resource and Environmental
Products January 25 and 26, 2005 Laurel Point
Inn, Victoria, BC
2
PRIORITIES COMPRISING THE ENVIRONMENTAL SCAN FOR
THE NATURE RESEARCH AGENDA OF ENVIRONMENT CANADA

• Wildlife/Biodiversity
• Aquatic Ecosystems (freshwater and marine
  ecosystems)
• Impacts of Atmospheric Change
• Impacts of Land Use
• Impacts of Highly Populated Areas
• Impacts of Pesticides and Toxic Substances
• Impacts of Resource Exploitation
• Impacts of Exotic Species
• Impacts of Biotechnology
• Cumulative Impacts

• HERO can contribute in some way to every priority
• Water is 2?? but
• Water integrates most impacts
• Water controls or impacts most priority issues

3
WHAT WATER-RELATED INFORMATION DO ENVIRONMENTAL
DECISION-MAKERS NEED?

• Water quantity (surface area, water level,
  water flow, diversions, etc.) every several
  months to every day in crisis conditions
• Water quality (presence of algae, turbidity,
  dissolved organic matter) every 10 days
• Aquatic bioproductivity every 10 days
• Delineation of regional groundwater
  discharge, recharge, and transition areas and
  temporal changes therein every few months
• Hydrologic modeling (prediction of water
  levels) every day as water levels rise during
  flooding

?
?
?
?
?
4
WHAT WATER-RELATED INFORMATION DO ENVIRONMENTAL
DECISION-MAKERS NEED?

• Natural and anthropogenic changes in
  shoreline features (annually)
• Mapping (and monitoring) for oil and other
  industrial spills
• Detection and spatial extent of noxious
  blue-green and other algal blooms as they occur
  and develop
• Identification of phytoplankton species
  (phytoplankton as an indicator of ecological
  dynamics invasive species)
• Bathymetry of some high use shallow waters
  (semi-annually)
• Location, identification, and spatial
  distribution of bottom vegetation
• Location, identification and distribution of
  riparian vegetation

?
?
?
?
?
?
?
5
The most important aquatic issues in Canada are
related to the quantity and quality of our water
supply
6
The biggest problems are people-related.

• Urbanisation, development and poor land use
  practicescan adversely affect watersheds,
  shorelines and riparian zones, and lead to
  degradation of water quality.
• There is an important need for detailed habitat
  and water quality monitoring, including
  interaction between activities in the watershed
  and water quality.

http//www.surreychamber.org
HERO can address both water and land issues
7
Water Problems
MERIS
MERISHERO
Landsat or HERO
Borstad et al, 2002. Trends and Opportunities in
Marine and Freshwater Remote Sensing, CSA report
number CSA-ST-CR-2001-0026

8
MERIS HERO
MERIS in large lakes HERO in smaller lakes and
water bodies

9
MERIS in large lakes HERO in smaller lakes
HERO in all water bodies discharges are
usually small

10
HERO
HERO
HERO

11
MERIS is the most appropriate sensor for wide
area coverage of coastal waters
HEROs coverage Is not optimal for mapping large
coastal waters
12
Heros spatial resolution is the same as Thematic
Mapper
13
MERIS FR data at 300 m resolution is adequate
for wide swath, high temporal resolution
monitoring of water events in the coastal zone
and in large lakes. It has the wide dynamic
range to image clouds and dark water at the same
time.
Lake Okanagan, BC - 5 km wide and 95 km long
14
MERIS also has sufficient SNR to image signals
like solar-stimulated fluorescence from from
phytoplankton. HERO will have this capability
also.
Note FLH sees through thin cloud
Fluorescence imaged in Lake Okanagan, BC October
2002 with MERIS FR Other smaller lakes are just
visible
15
HERO will have spectral, spatial and radiometric
resolution better suited to image inland waters
Turbidity in Lake Okanagan, BC September 2004
from Landsat 5 (water as avg of bands 1234
pseudocolour, land as RGB)
Kelowna
Kelowna
fire scar
fire scar
Kelowna
16
HEROs 30 m spatial resolution will be much
better than MERIS 300m resolution for a
cyanophyte bloom in Hamilton harbour, like this
one on Sept 8, 2001
MERIS barely adequate resolution (simulation)
Green/Blue ratio from LANDSAT Same resolution as
HERO
17
HERO will be capable of differentiating
freshwater phytoplankton on the basis of their
pigmentation Some blue green algae are toxic to
livestock and are the cause of taste and odour
problems in Lake Ontario
Diagnostic of Blue-green algae
http//envisat.esa.int/workshops/meris03/participa
nts/203/paper_30_pena.pdf
18
Canada has a lot of water HERO will never image
it all

• The surface area of lakes in North America
  exceeds that of any other continent
• Canada has 3 million lakes (as many as 30
  lakes/100km2)
• 80 of Canadian lakes are less than 10ha
  surface area (100 HERO pixels, 1 MERIS pixel)

www.vermilionestates.com
While the spatial resolutions match, we will
never use HERO to monitor all of those lakes. It
will be used to monitor representative areas, and
high value targets
19
HERO will be capable of monitoring water use by
Canadian industry
Black liquor effluents from Prince George BC pulp
mills can be resolved in Landsat imagery with
better spectral and radiometric resolution HERO
will do a better job
20
HERO place in the Time and Space Domains
HERO will be intermediate between frequent,
low-res imagers and targeting high-res sensors,
and well matched to the scales of lakes
Coasts Great Lakes
Lakes reservoirs
21
Observations on the role of HERO

• HERO therefore is neither a high res targeting
  sensor (like IKONOS), nor a low-res wide area
  surveillance sensor (like AVHRR, MODIS and
  MERIS), but a moderate resolution device that can
  support both tasks.
• An important advantage will be its
  hyperspectral resolution, allowing it to provide
  a closer, diagnostic look at the surface
  chemistry or pigmentation of targets located by
  other means.

22
Observations on the role of HERO

• Some but not all applications will use the
  hyperspectral capability of HERO.
• Some applications will only require a few
  spectral channels for a fast, simple
  calculation. (A hyperspectral sensor is a
  multispectral sensor for everyone - with
  different bands for different applications)
• A historical hyperspectral dataset provides
  the opportunity to develop new algorithms in
  the future

23
HERO will provide the capability to repetitively
monitor smaller high interest areas this
example uses SeaWiFS to look at L. Ontario
A change over time?

24
Important aquatic contributions by HERO

• Repetitively monitor high value, critical or
  representative sites in conjunction with lower
  resolution, wider area coverage and in situ
  monitoring
• Rapid response targeting to events blooms,
  plumes, spills, floods and other disasters
• Diagnostic tool for more complete (high
  spectral spatial resolution) description of
  features seen in lower resolution, wide area
  monitoring

25
HEROs role in the survey/study hierarchy

• Intensive site-specific studies to understand
  processes and to interpret broader scale Level 1
  and 2 efforts (lt10,000 km2)
• Regional, generally watershed-scale studies,
  issue driven (10,000 to 1,000,000 km2)
• Broad-based, generally national, inventory and
  monitoring surveys (gt1,000,000 km2)

HERO
Level 2 Resource surveys
LANDSAT
MERIS for largest areas
Level 1 Inventory and monitoring
26
One last thought
We need to be aware of the overlapping
responsibilities in Canada and elsewhere There
will be many users other than the federal
government.
NGOs
27
(No Transcript)