GIS and the GLOBAL Distribution of CO2

Gary D. Shaw



"The mechanics of GIS were made a lot easier in the last decade, but the relationships and assumptions built into GIS remain mere sketches of uncharted intellectual terrain." (Berry 1995). After reviewing many articles and books it appeared that Geographical Information Systems, at least in the geophysical sciences, seemed to be an application waiting for a problem to solve but others see it as "... recognition that the analysis of patterns and relationships in geographical data should be a central function of geographical information systems, the sophistication of certain areas of analytical functionally in many existing GIS continues to leave much to be desired." (Bailey 1994) Often there are applications that are based in GIS but the methodologies are not credited to GIS the science. Philosophically GIS's evolutionary path where commercial applications have been the primary driving force may have hindered its growth in the geophysical sciences and particullarly those that exist almost entirely on governmental funding of one form or another such as climatology.

The lack of a profit motive should not relegate GIS to an inventory control device and as more applications are forthcoming the growth of GIS will hopefully accelerate. This is not to suggest that only the private sector uses GIS there are several very good applications on the internet and one of particular interest is the use of GIS to predict future changes in the Great Lakes region. The Great Lakes Environmental Research Laboratory (GLERL) division of the National Oceanic and Atmospheric Administration (NOAA) has used GIS to model possible climatic changes resulting from a doubling of atmospheric carbon dioxide. The Tasmanian government has used GIS to show flora and fauna distributions using climatic factors by comparing vegetative and feeding requirements to rainfall, temperature, elevation, etc. An animation showing an El Nino event using map layers versus time to show the beginning and strengthening of the incident. The examples are a just a small sample of what has been done with GIS but it does not appear to be very significant compared to what could be done.



Atmospheric carbon dioxide (CO2) has been steadily increasing since the beginning of the industrial revolution in the late 19th Century. Using previous studies and ice core samples the pre-industrial revolution level of CO2 has been determined to be between 260 to 280 parts per million by volume (ppmv). (Wigley 1983). Today's CO2 level averages near 365 ppmv or about a 1.5 ppmv annual increase. The concern about elevated CO2 has normally been focused on its greenhouse effect or global warming and the subsequent effects of what a significant increase in temperature would have on the environment and humankind. The dramatic rise in CO2 is attributable to human activities (anthropogenic) primarily from burning fossil fuels such as coal and petroleum products but also the decrease in forests which act as carbon sinks by using CO2 for production of vegetative tissue. More recently studies have shown that there will in all likelihood be significant changes in plant distribution and forage quality.

Several studies have shown that certain species of plants are favored in an enriched CO2 world. Marginal areas where water is limited appear to favor woody plants over herbaceous plants and certain grass species (Knapp 1996). Herbivores both ruminants and insects do not receive the same level of nutrition from forage species grown in elevated CO2 plots. (Lindroth 1993 & Owensby 1988) Digestive residence in sheep increases due to an increase in fiber and a reduction in sucrose (nutritive value) from plants grown in enriched CO2 which suggests a reduction in protein and dairy production per acre. The reduction of nutrition derived per unit area consumed by insects should result in more area consumed which will compound the affect of swarms still experienced in Africa. Still other studies show a disruption in the activity of oceanic phytoplankton which are capable of exchanging 15 times the CO2 man produces annually. (Noever 1994) The increase in CO2 has understandably received a great deal of attention in the last two decades but very little has been done in understanding how the gas is distributed globally and why the distribution seems relatively constant.

NOAA has since 1968 measured CO2 levels globally. In 1978 there were only six sampling sites and methods of sampling had not been standardized. At this time the United States Department of Energy (DOE) in response to growing concerns about increased CO2 levels created under NOAA the Geophysical Monitoring for Climatic Change (GMCC) program. By 1982 there were 23 stations monitoring CO2 levels throughout the free world and of the 23 only one could be considered continental the rest were on islands or along coasts. Ironically the station at Niwot Ridge, Colorado (University of Colorado site) does not show any evidence of variation due to its location. Today there are over 30 sites set up to monitor CO2 including one in the People's Republic of China.

Data from these sites is available on the world wide web and contains the exact location of the site, pertinent information about site peculiarities and monthly readings for the period monitored i.e. the site on Mauna Loa has continuous data since 1969. The data is available in graphical and tabular form but other than a contour map and a three dimensional display of distribution along the dateline little has been done with the information by GMCC. (Komhyr 1985)


GIS and the Global Distribution of CO2

To display the global distribution of CO2 with only 15 to 30 sites is probably more an exercise in GIS than a scientifically significant or meaningful production but it should have enough dimension to be useful in showing the relative differences in year to year or the annual site changes of CO2. Depicting the distribution would ideally done as a spherical interpolation, however, time and resource constraints dictated a more judicious approach and a equirectangular projection from GEOCART1 was chosen as the base map to represent CO2 distribution. The equirectangular map was rotated to show the 160( W. longitude and the equator as the center of the map. The area approximately bounded by the west coasts of North and South America to 150 E and from 80 S to 85 N has 14 of an average of 20 monitoring sites and is considered to more accurately portray CO2 distribution. The base map was exported to ADOBE ILLUSTRATOR were the map's grid was changed to one-half inch intervals. The equirectangular map was used because of its linear east-west and north-south features which would readily translate into Cartesian coordinates to locate sites. The sites were then located on the modified base map and the X-Y coordinates measured for use in preparing the spread sheet for interpolation. The base map was again cleaned and exported to MAP FACTORY where it would be used to blank out land masses on the distribution matrices. Trial and error proved that establishing map size prior to overlaying or performing algebraic functions on maps saves a great deal of effort in trying to scale maps after they are rasterized. Scaling of different size maps can be done using the RESHAPE function but the maps must be proportional.


The site data, including X-Y coordinates, for 1982, 1985, 1988 and 1992 was entered into a spreadsheet. The spreadsheet was read into SIMPLE TEXT and modified for export into SURFACE III. The data in sxx is shown as X,Y,Z,Z...Z values. The Z values represent the levels of CO2 and each spreadsheet had 12, one for each month. Sxx can develop a grid by one of three ways, KRIGING is considered the most accurate of the three but the application is for more complex applications than the static data on CO2 distribution. TREND is also avaliable but GRID using an inverse square interpolation was used to create the matrices. The grid was set at 365 columns by 211 rows which would fit the base map when exported to MAP FACTORY map extremes were also reset to the map size created in MAP FACTORY. The product was a 77,015 cell matrix of interpolated Z values which took nearly four minutes run time for each map. The data distribution or lack of data was such that the interpolation function blanked several cells in the 0,0 (S.E.) corner of the matrix this is the result of the inverse square function not calculating values due to the lack of a near enough neighbor. The matrices were then exported to MAP FACTORY. A note before proceeding: before exporting you will be cued to either turn off or leave on the header, turn it off if it is left in your data file maybe read as a single row of several thousand values, also a value for the blanked cells must be selected, use the value cued.

The matrices were imported into MAP FACTORY as text delimited, floating point files. They were then renamed and saved. The base map consisted of "void" VALUE land masses and "zero" VALUE oceans. Voids cannot have any algebraic functions performed within their area but any other numeric such as zero will add, subtract etc. CO2 elevation maps (Fig. 1) were added to the base map(Fig. 2) to create a DEM see Fig. 3a & 3b, which portrays the global distribution of CO2 for that month. The northern latitudes have the greatest amount of variability annually up to 16 ppmv. The low point comes around the summer solstice and the high around the winter solstice. By subtracting the January map from the August map an annual change map was created, see Fig 4. Using the "magnify" tool several elevations (CO2) were measured along the 160 W longitude to determine the North to South distribution the plotted data can be seen in Fig. 5, for January and August 1992 data.


CO2 Elevation Map of SURFACE III Matrix January 1992

Figure 1

Base Map

Figure 2

CO2 Distribution January 1992

Figure 3a

CO2 Distribution August 1992

Figure 3b

The trends are fairly well in keeping with what was recorded in 1985. (Komhyr 1985 ) The rise and fall in the northern hemisphere corresponds with the growing seasons. As photosynthesis increases the amount of CO2 fixed by plants increases. There is an equatorial bulge that persists throughout the year but vacillates about + or - 5 degrees this may be attributable to meteorological or oceanic phenomena. The slight rise near the ice shelf in Antarctica is the very cold CO2 rich water rising. There looks to be a sink in the central mid-Pacific and Atlantic where the warm water could absorb gases. The fluctuations north of the equator are as earlier stated in response to photosynthesis. There appears to be little if any migration of CO2 during the winter months between the latitudes one suggestion is the polar vortex holds the CO2 at the northern latitudes. (Kohmyr 1985)



Difference of CO2 Distribution January - August 1992

Figure 4



Unfortunately CO2 data is limited and virtually non-existent for continents but what is available does suggest a persistence in the patterns and distribution for the ten years from 1982 to 1992. Using the GIS resourses available at the University of Nebraska at Omaha's Geography-Geology Department one can only conclude that GIS does have application in the geophysical sciences. Certainly the distributions of the components of El Nino or droughts in central Africa could readily be described with GIS. The underlying reasons GIS seems to be lagging in many applications are probably no longer important other than to realize something caused it but what needs to be done now is to make certain the market is aware of GIS potential. There is an adage that goes something like "The people with the problems are the people with the solutions." and if armed with the proper tools those solutions will include GIS.



Bailey T. C., Spatial Analysis and GIS, Taylor and Francis, (1994)

Edited by: Fotheringham S. and Rogerson P.


Berry J. K., Spatial Reasoning for Effective GIS, pp 4, GIS World Books, (1995)


Knapp P. A. and SoulÈ P. T., Vegetation Change and the Role of Atmospheric CO2 Enrichment on a Relict Site in Central Oregon: 1960 - 1994, Annals of the Association of American Geographers, 86(3), pp 387 - 411, (1996)


Komhyr W. D., Gammon R. H., Harris T. B., Waterman L. S., Conway T. J.,

Taylor W. R. and Thoning K. W..

Global Atmospheric CO2 Distributions and Variations From 1968 - 1982

NOAA/GMCC CO2 Flask Sample Data, Journal of Geophysical Research, Vol. 90,

5567 - 5596, (1985)


Lindroth R. L., Kinney K. K. and Platz C. L., Responses of Deciduous Trees to Elevated Atmospheric CO2: Productivity, Phytochemistry and Insect Performance., Ecology,

Vol. 74, pp 763 - 777, (1993)


Owensby C. E., Ham J. M., Knapp A., Rice C. W., Coyne P. I. and Auen L. M.,

Ecosystem-Level Responses of Tallgrass Prairie to Elevated CO2, Carbon Dioxide and Terrestrial Ecosystems: Edited by G.W. Koch and H.A. Mooney, Academic Press,

pp 147 - 162, (1988).


Wigley T. M. L. The Pre-Industrial Carbon Dioxide Level, Climatic Change, Vol. 5,

pp 315-320, (1983)






1 All software will be shown in uppercase-italicized font and is considered copyrighted or Trade Mark ( protected).