Summary of Working Group Proceedings
Prepared by Donat Agosti, American Museum of Natural History and
Chris Rodstrom, Conservation International
Problem Statement
The conservation NGO/museum community currently lacks the capacity to perform global analyses of changes in biological diversity at a variety of levels of biological organization, and to couple these analyses with remote-sensing data.
Assessing global conditions is needed to provide the broader context for ecosystem and site-level priority-setting or monitoring and to understand the overall global conditions of biodiversity. For instance, it is difficult to answer the commonly-asked question, "how much forest remains on the planet, and how fast is it disappearing?" Similar questions are also raised for grasslands, xeric, and coastal/marine ecosystems, or at a continental or national level. However, these results would be a very powerful tool to pinpoint specific areas and their changes in the form of a highly-educational map.
The currently available remote-sensing land-use classifications do reflect classes such as "forest", and within forests even some subclasses, but they do not adequately reflect their biological composition. It is however within this level, which includes the historically-known communities of organisms of a specific area, where the mechanisms of the maintenance of biodiversity lie and where biodiversity has to be assessed. This is the major focus of conservation.
We also lack the ability to easily link this global information to local and regional data sets that contain specific information on habitat, communities, and species. This link would benefit both the regional/site-level studies (e.g., what are the relationships between a temporal change in a species' distribution and deforestation?) and the global assessment (e.g., taken collectively, what is the distribution of certain species or genera, and how can they indicate the health or condition of the remaining habitat globally?).
The recent example of the global change in amphibian populations illustrates this point. Scientists have identified a decline in many species of frogs around the world, perhaps indicating a global trend linked to change in habitat, climatic conditions, or pollution (the debate as to the causes is still far from over). There is no mechanism in place to monitor changes in the population of frogs globally, or study the relationship between population fluxes and changes in habitat, global climate, or other related biological or physical conditions. Additionally, we cannot easily determine if similar trends are occurring in other taxonomic groups.
Developing the ability to measure and monitor global biological diversity at a variety of levels of biological organization will likely occur in phases. In the short-term we must be able to answer broader questions of global changes in habitat and related biophysical conditions. In the longer-term we should strive for a more comprehensive system of measuring and monitoring key taxonomic groups, to provide greater depth to and meaning from the global assessments of habitat and biophysical conditions.
Background
The background to the discussion of measuring and monitoring of global biodiversity can be divided into two categories: data gathering and analysis, and dissemination of data.
Data Gathering and Analysis
Remote sensing of the environment is based on the combination and analysis of spectral bands that result in a thematic map, including land-cover classes such as forests, disturbed areas, water, etc. The biological diversity described by these maps and the different communities they might include (e.g. xeric site vs. heavily degraded grass lands, natural forest vs. tree plantations) cannot easily be determined from remote-sensing studies, but must also include some field surveys.
Global scale remotely-sensed data have historically been accessible to a small community of researchers. Data such as Advanced Very High Resolution Radiometer (AVHRR), Thematic Mapper (TM), and Multi-spectral Scanner (MSS) have generally been unavailable to the broader conservation community due to many reasons such as the high cost of data, the expense of storing and processing the data, and the high level of expertise and facilities required to manage the data.
The sophistication of global biodiversity assessments from ground surveys and monitoring programs differs across the world. Whereas areas such as the United States and Mexico have an extensive knowledge of the distribution and state of their biological diversity, this is not the case for most developing countries where most of the biological diversity is.
Dissemination
There have been few successful projects for distributing remotely-sensed data in a form that is usable by broader, non-technical audiences. By their very nature, remote-sensing data sets are very large, and difficult to process or manage. Recent efforts in remote sensing analysis (NASA Pathfinder, the EPA North America Landscape Characterization Program) and distribution via the Internet (EOSDIS, GLIS) have increased access to these data in a usable format. However, accessing and managing this information still remains limited to a narrow community.
Another technology available for disseminating data, especially in the developing world, is CDROM. Some global data sets are distributed using this technology, including the Digital Chart of the World (ESRI), the Tropical Moist Forests and Protected Areas (WCMC), Digital Soil Map of the World (FAO), and the EPA/NOAA Global Ecosystems Database. However, use of these data still requires advanced technical expertise, and a significant amount of processing.
Discussion of the Issue
While advances in analysis and distribution of remotely-sensed data have made global assessments possible by a small community of experts in government, academia, and the private sector, the majority of the NGO/Museum organizations are still unable to access and use the data, or to combine them with their own data holdings. This larger group requires information and tools specifically tailored to a less technically-proficient audience, to both conduct global assessments and link to their regional and site-level data.
Any approach must define varying scales, from broad global monitoring to a more integrated approach that includes more detailed data at higher resolution (species information), and integrating spatial data sets of varying scales (Landsat TM Pathfinder and AVHRR forest cover). At each scale, definition of correct indicators, from classification schemes of land use/land cover, spatial extent, integration of aquatic and marine, will be necessary.
In addition, the approach to providing access to data and tools to use the data should take into account different types of access (e.g. CDROM vs. Internet), with an eye towards evolving technologies that can eventually provide advanced analytic tools (e.g., MapObjects, map internet servers, JAVA). Finally, this problem and the solutions discussed are primarily for an information-providing system. In later versions, decision making tools could be included.
How to Address the Issue
The development of a global measuring and monitoring approach might be best thought of in components, where each addresses a separate aspect of the problem: global spatial data, integrating ground data, and providing access and analytic tools. For each component the challenge is to define both short- and long-term objectives that will provide usable results quickly, while allowing for more sophisticated monitoring in the future.
Global Spatial Data Sets
- Define the questions to answer with global spatial data sets, such as how much forest remains on the planet, and how fast is it disappearing?
- Determine what information is needed to answer the questions: what to measure (e.g. land cover, forest cover, climatic conditions), resolution (1degree, 1km, 30m), time periods (annual, 5 year, 10 year), and geographic coverage. One recent example is the Global 1km Land Cover Project.
- Identify appropriate sources for a global spatial data set (AVHRR, TM, MSS, etc.) based on needed resolution, coverage, time period, etc. Explore nesting regional, higher resolution data with global data to provide greater accuracy (e.g., Pathfinder data).
- Define classification schemes for analysis with appropriate indicators (forest vs. non-forest, land cover, etc.).
- Identify other important existing global spatial data sets, such as soils, hydrography, national/sub-national boundaries, population centers, etc.
Ground Data
- Identify important ground data to complement global data sets. Link regional/site specific information to global data sets.
- Include geo-referenced ground data and develop algorithms to produce classifications that reflect biodiversity distribution patterns as accurately as possible. Incorporate into new classifications of global data sets, as is possible. Identify predictive population and distributional models.
Provide Access and Analytic Tools
- Make the global data available georeferenced to conduct global and continental/regional/national assessments, produce standard thematic maps, tables, and reports. Explore appropriate media, interface, languages, etc.
- Produce standard algorithms to measure magnitude and type of changes over time.
- Include additional analyses, query and decision-making tools for predictive modeling of expected changing conditions.
Pilot Projects
Meeting the goals of monitoring global change can be done in steps through one or several pilot projects, done concurrently or sequentially. The simplest system might be based on a set of already-existing global spatial data such as the global 1km land cover, combined with ground data such as biodiversity hot spots, ecoregions, etc., or georeferenced data from organisms with well know distribution such as tiger beetles. In additional phases, those areas will be revised, based on additional data. This allows one to compare summary information with actual conditions on the ground.
A more advanced system would facilitate global analyses and local analyses in the context of the global data, survey trends in biodiversity hot spots and global change, support the analyses of hot spots and centers of endemism, and develop tools for efficient and systematic studies of large, unexplored areas, such as the Amazon or Congo basins.
The system will be based on a tool combining data collecting, analyses and distribution. The tool will be developed and customized, based on remote-sensing input from NASA and standardized ground data from NGOs. Designed as interactive, the system will also include data input facilities and descriptions of the standardized sampling protocols.
Potential Pilot Projects
- Land-cover distribution maps that are easily queried, interactive, and presented
- Calculations of changes in land cover over time by continent, country, globe, etc.
- Calculations of other global data sets by continent, country, globe, etc.
- Assess actual vs. predicted distribution ranges and patterns for specific organisms using gap analysis, habitat modeling, or other predictive modeling techniques (e.g. ERIN)
- Construct global distribution maps for selected organisms, both actual and potential maps, based on the developed mapping tools and the geo-referenced data.
back to main
