Thoughts on the Role of Remote Sensing in Monitoring Ecological Health

1
Thoughts on the Role of Remote Sensing in
Monitoring Ecological Health
CENS Weekly Seminar, UCLA 6 August 2004
Gary Geller NASA Jet Propulsion
Laboratory California Institute of
Technology Gary.N.Geller_at_jpl.nasa.gov
2
Overview

• Protected areas
• Monitoring and indicators
• Role of remote sensing
• Operationalizing the monitoring process
• Use of sensor networks in
• Remote sensing
• Ecological monitoring
• Overcoming obstacles to operationalizing new
  technology

Purpose Stimulate thought / plant seeds on
3
Protected Areas
4
Protected Areas
"An area dedicated to the protection and
maintenance of biological diversity and natural
resources"

• Globally gt100,000 protected areas
• Effectiveness of management varies greatly
• Unprotected areas also ecologically important
• Will focus on PAs
• Applicable to all natural areas

5
Sundarbans, India and Bangladesh
6
Rondonia, Brazil
7
Protected Areas
8
Monitoring and Indicators
9
Importance and Status

• Why is monitoring important?
• Identifies PA management problems
• Leads to greater PA health (hopefully)
• No global monitoring program for PAs exists
• Monitoring of individual PAs is rare
• Remote sensing has appeal
• PAs cover large areas
• Periodic assessment is required

• What needs to be measured?
• Can it be measured using RS techniques?
• Can RS-based methods be operationalized?

10
Indicators
An informational device that provides a measure
of the health of a PA or ecosystem

• Basis for monitoring
• Indicator development is a new field
• No single set of standard indicators exists
• MA, Heinz Center ecosystem indicators
• CBD starting to develop one for PAs
• Reflect success of the PA systemand of the
  convention

11
Many Indicators Are Possible

• 22 Area and percentage of forest area affected by
  natural disasters (insect attack, disease, fire
  and flooding)
• 22 Number and extent of invasive species
• 21 Percentage forest protected areas (by forest
  type, age, class, and successional stage)
• 21 Wood harvesting intensity
• 21 Managed forest ratio
• 20 Changes in the proportions of stands managed
  for conservation and utilization of genetic
  resources (gene reserves, seed collection stands,
  etc.)
• 20 Per capita wood consumption
• 19 Extent of mixed stands
• 18 Estimate of carbon stored
• 18 Percentage forest land managed for recreation
  and tourism to total forest area
• 17 Number of forest dependent species whose
  populations are declining
• 17 Fragmentation of forests
• 16 Threatened tree species as a percentage of the
  20 most used for commercial purposes
• 15 Area and extent of degraded lands reclaimed
  through forest operations
• 14 Area and percentage of forests managed for
  catchment protection
• 14 Self-regenerating area as a percentage of
  total area
• 13 Population levels of representative species
  from diverse habitats monitored across their
  range
• 12 Self-regenerating area per habitat type
• 10 Ratio between exotic species and native
  species in plantation area

• 22 Area and percentage of forest area affected by
  natural disasters (insect attack, disease, fire
  and flooding)
• 22 Number and extent of invasive species
• 21 Percentage forest protected areas (by forest
  type, age, class, and successional stage)
• 21 Wood harvesting intensity
• 21 Managed forest ratio
• 20 Changes in the proportions of stands managed
  for conservation and utilization of genetic
  resources (gene reserves, seed collection stands,
  etc.)
• 20 Per capita wood consumption
• 19 Extent of mixed stands
• 18 Estimate of carbon stored
• 18 Percentage forest land managed for recreation
  and tourism to total forest area
• 17 Number of forest dependent species whose
  populations are declining
• 17 Fragmentation of forests
• 16 Threatened tree species as a percentage of the
  20 most used for commercial purposes
• 15 Area and extent of degraded lands reclaimed
  through forest operations
• 14 Area and percentage of forests managed for
  catchment protection
• 14 Self-regenerating area as a percentage of
  total area
• 13 Population levels of representative species
  from diverse habitats monitored across their
  range
• 12 Self-regenerating area per habitat type
• 10 Ratio between exotic species and native
  species in plantation area

Selecting the right ones is a difficult task
12
Reduced List

• Land cover
• Land use (inside PA and adjacent lands)
• Fragmentation
• Vegetation health
• Vegetation parameters (NDVI, NPP, LAI)
• Frequency of invasive species
• Climate change impacts
• Water quality
• Streamflow
• Air quality
• Inputs (nitrogen, mercury)

13
Role of Remote Sensing
14
What Is It?
Measuring from a substantial distance, often
from space
Blue
Green
Red
15
Example
16
For Conservation Users

• How can satellite images helpconservation
  practitioners?
• Observe land cover, boundaries, threats
• Classify into vegetation and land use
  categories
• Monitor change in forest cover, range
  condition, land use
• Measure areas and distances
• Detect fires

17
Classification

• Importance Central to many indicators
• Goal Create a map of different habitats
• Utility
• Defines where species can live and travel
• Shows land use patterns
• Approach Visual or automated
• Caveats
• Labor intensive
• Typically requires expertise
• Ground truth / knowledge needed for sufficient
  accuracy

18
Paris Area Classifcation
Green Vegetated/Rural Blue Water Red
Built Material/Urban
Built Material cover 31
19
Synthesis
20
The Triad
Satellite Remote Sensing
Sensor Networks
Indicators and Monitoring
21
Major Problems to Solve

• Classified data is expensive, of limited
  availability, and ages poorly
• Ground truthing is expensive
• Access to images requires expertise
• Selecting and developing indicators is difficult
• Monitoring program must be operationalized

22
PA Health Indicators

• Land cover
• Land use (inside PA and adjacent lands)
• Fragmentation
• Vegetation health
• Vegetation parameters (NDVI, NPP, LAI)
• Frequency of invasive species
• Climate change impacts
• Air quality
• Water quality
• Streamflow
• Inputs (nitrogen, mercury)

Remote Sensing? Sensor networks?

Sensor networks?
23
Breather
24
SideBar Validation and Ground Truthing

• Validation used to develop accurate data product
  algorithms
• Ground truthing to confirm/correct/interpret
  remote sensing data using ground measurements
• Both are
• A lot of work
• Essential
• Require site visits / revisits
• Could robotic sensors / sensor networks help?

25
SideBar Validating Surface Radiance
Barreal Blanco, Argentina, 22 Jan 2001
26
SideBar Lake Tahoe Study
27
Operationalization
28
Monitoring

• Two possibilities
• Use RS experts
• Use PA site managers
• Only Possibility 2 is sustainable and scaleable
• But
• Many PAs are in developing countries
• Requires appropriate methods and tools
• Must gain acceptance by PA community
• How do we do Possibility 2?

29
Enabling Site Managers to Do Monitoring

• Build tools that overcome the constraints
• Low/no cost
• Require no remote sensing expertise
• Require no monitoring expertise
• Easy to learn
• Require no Internet
• Infuse them into the culture
• Provide useful service
• Must feel familiar
• Outreach to PA manager community
• Training

30
Example RS Accessability

• Loads of zero- and low-cost data available
• BUTCurrent data access and analysis systems are
  for experts
• Result UNDERUTILIZATION

31
Solutions

• Two solutions are possible
• Training
• Make satellite data easier to
• Obtain
• Use

Protected Area Archive
32
Protected Area Archive Tool

• Approach
• Create collections of images of PAs
• Package with simple tools to find and use them
• Intended audience
• PA managers with no experience using satellite
  data

33
Demonstration
34
Desired Additional Capabilities

• Push-button indicators
• Forest cover/cover change
• Land use classification
• Logged area and intensity
• Range condition (NDVI)
• Visualization toolsso users can use an image to
  tell a story
• Custom CDs user selects scenes/PAs
• Interactive website

35
Summary and Conclusions

• Indicator development is in its infancy
• Remote sensing has a role in PA health monitoring
• Sensor networks will have a role if they can
  reduce the cost of doing it
• Remote sensing validation and ground truthing
• PA health monitoring
• Infusing a new technology into the user community
  requires patience, thought, and careful listening

36
Thank You