Remote Sensing and Climate Change
Will We Bridge the Gap?
An Inconvenient Truth, a movie featuring former U.S. Vice President Al Gore has been seen by a fair number of folks around the planet. My colleagues in New Zealand were exposed to this cinematic clarion and are asking some pragmatic questions as to what they should do to preserve their island nation in the face of certainty for rising oceans and radical weather shifts.
Dr. Timothy W. Foresman is President of the International Centre for Remote Sensing Education. He has been director of United Nations Environment Programme’s Division of Early Warning and Assessment (Nairobi, Kenya) and national program manager for NASA’s Digital Earth (Washington, D.C.). He is editor of The History of Geographic Information Systems, 1998, Prentice Hall. Dr. Foresman is currently the Director-General for the 5th International Symposium on Digital Earth (www.isde5.org).
The IPCC (Intergovernmental Panel on Climate Change) model projections indicate a three-foot sea level rise by the end of this century, therein denuding vast acreages of New Zealand’s small island nation. They are taking this issue seriously and have hosted the first Digital Earth Summit on Sustainability in August of this year, bringing together clever folks for this dialog. The New Zealanders are keen to consider the realities of global climate change and to address the ameliorating and coping strategies necessary to sustain the quality of life for their four million residents.
For the remote sensing community, we might ask, what have we done to support the science of climate change and what are the challenges facing us in working with this high-profile community? This dialog came about subsequent to seeing the aforementioned documentary when a group of remote sensing types congregated in Colorado to discuss the use of our sky-borne technologies for measuring melting glaciers and other visible signs of temperature change phenomena. The question is compelling and caused me to ponder further where we have been on this issue and where we are heading.
In 1985, a special edition of The Science of the Total Environment captured papers from a conference titled “Man’s Role in Changing the Global Environment,” held in Venice, Italy. At the conference, the facts of climate change due to increasing atmospheric CO2 from anthropogenic emissions were presented matter-of-factly. Attendees were treated by our Italian hosts with views of the encroaching water into St. Mark’s Square (Piazza San Marco), enough to satisfy any dissenters (there were none).
But how would remote sensing be featured in this topic? Jet Propulsion Lab’s Fred Billingsley succinctly noted, “Global problems require global models, which for the most part are not available or are inadequate. Global models, in turn, require global data for their solution. Utilization of global data requires that they be planned to be sensed, then located and assembled.” (See note below.) Therefore the question is, how well have we been planning to use remote sensing data for climate change models?
One decade after this, in 1995, Initiative 15 of the National Center for Geographic Information and Analysis, titled “Multiple Roles for GIS in U.S. Global Change Research,” attacked the issue of bridging the spatial data community and the climate change community. Much to the chagrin of most participants, the bridge was not built and, indeed, a blueprint for such a bridge could not be found. “Besides the semantic differences between modelers and the geo-spatial community, what is needed is establishing time series with remote sensing that unfortunately has questionable life span,” noted Mike Goodchild of U.C. Santa Barbara, who co-led Initiative 15. “Look at Landsat and its continuity issues. MODIS (Moderate Resolution Imaging Spectroradiometer) is a short-term experiment. While we do have AVHRR (Advanced Very High Resolution Radiometer) and NDVI (Normalized Difference Vegetation Index) for small-scale greening phenomena, long-term series is the key. If we were to design a sensor for climate change, what would it look like? How long would it be up?” Considering the time frame that we are discussing for temperature rise, his responses were not encouraging.
Kate Beard, of the University of Maine, also co-leader of Initiative 15, remarked that there appeared to be more interest in applying in-situ sensors than in applying remote sensors of late. However, she did emphasize the necessity to figure out how to integrate in-situ with space-based sensors. “We really need more spatial and temporal details of our planet. For example, our work with the Gulf of Maine Observation System uses ten buoys for meteorological and oceanographic sensing. This is such an interdisciplinary issue and yet we still don’t know how to bring all these different disciplines together,” she remarked. “If we could only pull together all these inputs we could perhaps get the big picture.” Now, ten years later we still don’t have remarkable progress.
Another colleague, Matt Nolan, from University of Alaska Fairbanks, offered an interesting perspective: “The climate change folks — that is, those studying climate, are in their own little world in terms of data formats. They invented their own data that is inaccessible to others. Their compressed formats just don’t work with our field data very well. You have to know someone and be told how to use it.” Progress might be considered glacial in terms of the pace of accessing data in these environments.
Dr. Nolan explained how he painstakingly extracts individual text files from large area climate data sets to correlate with selected nodes that define his field sampling sites where he measures ice-mass wasting dynamics. He noted that the National Snow and Ice Data Center is assimilating a lot of remote sensing data looking at sea ice change. And they use NDVI for arctic vegetation changes as well, although a lot more could be done. “The real problem is that everyone has to deal with the serious investment in using these tools. It takes so long to become competent at using a sensor, or even a software package, that we don’t have enough time to focus on the problems,” he wryly noted.
Back at the University of Maryland, I asked professor John Townshend what he thought of the gap between the remote sensing and climate change communities. “Terminology remains a major hurdle. In addition, the climate change folks are so heavily invested in their models and their output, as are the oceanographers, that they really haven’t made much demand on our community. Such fundamental points as ‘What is an observation?’ separate us somewhat. We do need a systematic and comprehensive way to accomplish data assimilation and integration of multiple data forms. And we need different data that will reinforce each other. Look at MODIS, for example; the data are not even internally consistent. One sensor creates a continuous field and another creates classifications that are not compatible. If we are to succeed, we really should be designing proper systems with rigorous outcomes and proper data assimilation techniques. We must know a lot more about the nature of errors and what the errors are,” he offered.
When asked the same question, Bob Corell, Chair of the Arctic Climate Impact Assessment, immediately responded. “We need to define columnar CO2, CO, and CH4. We need to know exactly where these greenhouse gases are coming from and how much there is. We have just arrived at the technology; now if only we had a space program that would launch these new sensors, we would be in a much better position scientifically and policy-wise. Right now the best we can do is sit around watching the ice melt.”
Let’s hope that our community can do better than that, but after three decades with so little to show, we might start looking into mirrors for answers. Just look at what was stripped from the NPOESS (National Polar-orbiting Operational Environmental Satellite System) specifications: all the climate change instruments.
Note: Billingsley, F. 1986. Some Factors Affecting the Extraction of Global Resource Information from Remotely Sensed Data, Science of the Total Environment, Vol. 55, pp. 41-55.
Join us at the 5th International Symposium on Digital Earth in San Francisco, 5-9 June 2007. International experts in Earth observation technologies and applications will share the latest advances in 3-D visualization and web-enabled innovations. Check the web site for the call for papers and other information (www.isde5.org).