The spatial dimension of Indicators of Sustainable Development:
The role of Geographic Information Systems (GIS) and cartography

What's so special about spatial?

Sustainable development is by nature hundred per cent temporal as implied by the term 'development'. Whether sustainable or not, development by nature also is hundred per cent spatial or territorial, since all development takes place at certain spot on Earth, which can be represented by x, y (and z) co-ordinates. Thus, for a region, whether a municipality, country, transboundary region, continent or the whole world, there can be important spatial heterogeneity or differences, which in many cases deserve special attention. Surprisingly, the spatial dimension has often been neglected or been given low priority by groups preparing environmental or sustainable development indicators. The reasons for this are several. One main reason is missing or low capacities and skills to handle georeferenced data and information among those in charge of indicator work. This is increasingly becoming a reduced concern with the advent of more user-friendly GIS software, and also with the increasing tendency of merging traditional GIS software with commonly used 'office type' software. The recent bundling of a MapInfo subset into MS Excel 7.0 is a good example of this. Another principal reason is that the spatial heterogeneity for many ISDs have been considered rather irrelevant compared to the temporal dimension. In the set of proposed ISDs from the United Nations Commission on Sustainable Development (UN/CSD 1996), the spatial domain of Social, Economic and Institutional indicators have in most cases normally not been considered important. The concern empirically has been on getting time series data for the area considered. A notable exception is if sub-regional differences, whether internationally or nationally, are of interest.

The environmental indicators are the ones for which due attention has been paid with regard to the spatial domain. Furthermore, among those, the spatial focus has been on the state indicators. The driving forces (pressure) and response indicator related to human activities in the economic and institutional spheres of society have not been considered spatially to the same extent.

Thus, among the four main groups of Indicators of Sustainable Development, proposed by the UN/CSD, it is in particular the set Environmental Indicators, and State Indicators among those, that first and foremost benefit from using spatial indicators. Table 1 shows the environmental indicators from the proposed UN/CSD set that will most strongly benefit from the use of Geographic Information Systems (GIS).

Table 1. Indicators of Sustainable Development as proposed by UN/CSD (1996) particularly suited for elaboration in a spatial context using Geographic Information System.

However, most ISDs, whether Social, Economic, Environmental or Institutional, can visually be shown cartographically for sub-regions, as indicated earlier - NB. Therefore, the list above should be considered conservative.

Geographical Information Systems in Indicators of Sustainable Development work

The term Geographical Information Systems have traditionally been applied to software bundled with hardware particularly suited to deal with spatial data and information. Recent trends show that GIS are increasingly becoming integrated with of other types of information systems and tools, such as general Database Management Systems (DBMS), traditional office spreadsheet packages, and, very recently, Internet and WWW components thereof (Langaas, 1996a,b). However, this integration tends to happen at the expense of loss of many valuable functionalities found in a full-fledged GIS. Therefore, the advantages of using dedicated GIS packages most likely will remain for many years to come.

The merits of using GIS in indicator work are several. We will here briefly review some of the important functionalities. For those interested in an introduction to GIS and spatial analysis, we refer to text-books (see, e.g., Bernhardsen 1992, Star and Estes 1990 , Tomlin 1990)

• Sampling

In many cases, when dealing with environmental indicators, there is a need to obtain a spatially representative sample of an indicator in question, possibly for deduction of an average value for the overall area considered. Many GIS can assist in making spatially unbiased averages from geographically distributed sample measurements by offering geostatistical analytical tools.

• Analysis

One of the key features of dedicated GIS are the range of tools offered for spatial analysis, and frequently also tools (e.g., statistical, SQL) for non-spatial analysis of data associated with the geographic features, so-called attribute data. In the context of ISD work one should not exaggerate the importance of many of these possibilities, since they rarely will be used. The most frequently used 'analytical operation' is the use of Boolean overlay techniques for the derivation of statistics. For instance, for the issue Land affected by desertification in chapter 12 in UN/CSD (1996), it is proposed that the -

".. creation of an index which combines degrees of severity will require the following measures:

(i) Area subjected to severe land degradation x_Km2 (severe here includes both the severe and very severe categories of UNEP.

(ii) Area subjected to moderate land degradation y_Km2.

(iii) Area subjected to slight land degradation = z_Km2.

(iv) National area (excluding surface water bodies) n_Km2.

(v) National area of drylands (vulnerable to desertification, assuming that all drylands are potentially vulnerable to desertification. Hyper-arid lands are excluded), consisting of arid, semi-arid, and dry subhumid land = d_Km2.

From the above measurements, the following sets of numbers can be derived:

Indicator computations:

a. National area affected by desertification

= (x + y + z) Km2

b. Percent of national area affected by desertification

= [(x + y + z)/n] 100%

c. Percentages of national area affected by severe, moderate and slight desertification respectively can be calculated in the same way.

d. Percent of national drylands affected by desertification

= [(x + y + z)/d] 100 %

e. National area not affected by desertification

= [n - (x + y +z)] Km2

f. National dryland area not affected by desertification

= [d - (x + y + z)] Km2 "

The development of these indicators clearly requires analysis of GIS data using Boolean overlay techniques, and hardly can be obtained without it.

• Database management

Many of the more comprehensive GIS software packages are connected with powerful database management systems (DBMS). Thus, it may be considered to keep and maintain a ISD database, including the raw statistical data, using the DBMS found in GIS and thereby improve the opportunities for cartographic visualisation.

• Visualisation

The production of cartographic outputs is a key feature of any GIS. In a ISD framework, at least two types of cartographic out-puts may be useful to produce spatial indicator maps and 'Reference' maps. Figures 1 and 2 show good examples of spatial indicator maps, whereas figure 1 also incorporates a temporal aspect .

Figure 1. Wilderness territory in Norway 1900, 1940, 1980 and 1994. Wilderness territory is defined as an area lying more than 5 kilometres from roads, railways and regulated water-courses. In 1994, power lines and tractor tracks have been included. Wilderness is a state indicator for cultural and natural landscapes in the Norwegian state-of-the-environment report on the Internet. Source: UNEP/GRID-Arendal, 1996

Figure 2. Latin American and Caribbean coastal ecosystems threatened by development. For detailed descriptions of threat potential, see source. This indicator has recently been included in the Global Environment Outlook-1, UNEP's global state of the environment report 1997 (UNEP/DEIA 1997). Source: Bryant et al. 1995.

Another types of cartographic illustrations that may be useful in the context of ISD reporting and visualisation are reference or index maps, showing the locations of measurement stations, if a few measurement stations are supposed to represent a larger region.

• GIS - WWW Interlinkage

The potential use of GIS in ISD work should not be seen in technological isolation. As mentioned earlier, there is a trend that GIS software is being directly connected to Internet WWW server software (Langaas 1996b, Steinke et al. 1996). In this respect we are not referring to the publication of static maps, which in a environmental indicator context has been forcefully demonstrated by several national indicator based state of environment reports, such as for Australia, Canada, Georgia and Norway (Environment Canada 1996, ERIN 1997, UNEP/GRID-Arendal 1996, Georgian Geoinformation Centre 1996), as well as global ones (UNEP/DEIA 1997). We rather suggest the possibilities for interactive analysis by Internet users of on-line ISD databases managed by GIS software linked to Web servers allowing the Internet user to individually define and visualise spatial indicators better suiting the user's needs. The range of choices in such cases must be carefully determined by those responsible for the ISD database to prevent the generation of 'odd' outputs. Such on-line interactivity is not restricted to GIS - WWW connected databases, but also for DBMS - WWW ones, thus allowing for the individual definition and creation of graphs and charts. One research project examining these opportunities is the international State of the Environment Cities on Internet project lead by UNEP/GRID-Arendal, Norway.

Conclusions

In an ISD work context, whether on the municipal, national, international, or global level, the question is not whether to think spatially and use GIS. The question is - to what extent? Knowing that cartography is an extremely powerful means of conveying messages, the answer to this question needs careful consideration. Generally, though, we recommend increased spatial thinking and the use of GIS as a most appropriate technology.

References

Bernhardsen, T. 1992. Geographic Information Systems. VIAK IT, Norway, 318 pages.

Environment Canada. 1996. The State of Canada's Environment - 1996. <http://199.212.18.12/~soer/>

ERIN. 1997. Australian Environment On-line. <http://www.erin.gov.au/erin.html>

Georgian Geoinformation Centre. 1996. State of the Environment Georgia. <http://www.grida.no/prog/cee/enrin/htmls/georgia/soegeor/hp_soege.htm>

UNEP/GRID-Arendal. 1996. State of the Environment Norway 1995. Internet: <http://www.grida.no/soe95/>, last updated 27. September 1996.

Langaas, S. 1996a. Transboundary European GIS databases: A review of the Baltic Sea region experiences. Paper presented at GISDATA Specialist Meeting on "Geographic Information: the European Dimension", Buoux, France, 8-12 May 1996. < http://www.grida.no/prog/norbal/docs/gisdata/ >

Langaas, S. 1996b. Commercial, off-the-shelf solutions for GIS - WWW interfacing. <http://www.grida.no/prog/norbal/docs/giswww/giswww.htm >

Star, J. and Estes, J. 1990. Geographic Information Systems. An Introduction. Prentice Hall: Englewoods Cliff, 303 pages.

Steinke, A., Green, D.G., and Peters, D.J. 1996. On-line environmental and geographic information systems. In: Saarenmaa, H. and Kempf, A. (eds.), Internet applications and electronic information resources in forestry and environmental sciences. EFI Proceedings 10: 89-98, Joensuu,

Tomlin, C. Dana. 1990. Geographic Information Systems and Cartographic Modelling. Prentice Hall: Englewoods Cliff, 249 pages.

UNEP/DEIA. 1996. Rump, P.. State of the Environment reporting: Source Book of Methods and Approaches - UNEP/DEIA/TR.96-1, Nairobi.

UNEP/DEIA. 1997. Global Environment Outlook-1. New York: Oxford University Press, 265 pages. Also on Internet at <http://www.grida.no/geo1/>, <http://www.grid.unep.ch/geo1/>, <http://grid2.cr.usgs.gov/geo1/>, <http://www.rolac.unep.mx/geo1/>, <http://www-cger.nies.go.jp/geo1/>, <http://www.unep.org/unep/eia/geo1/>

UN Commission on Sustainable Development. 1996. Indicators of Sustainable Development. Framework and Methodologies. United Nations, New York, 428 pages.