Section: New Results

Material flows, production and consumption at sub-national geographic levels

As explained earlier, estimating the actual environmental impact of an urban area on the one hand, and the efficiency of (local or national) policy options in reducing these impacts on the other, requires an understanding of the material flows and material uses generated by the considered urban area. It is important to realize that impacts (both local and distant) can vary greatly from one region or departement to the next, depending on its agricultural and industrial characteristics. The whole point of this work is to evaluate as best as possible these variations, in order to best adapt public policies in terms of environmental impacts, for given socio-economic conditions and objectives.

The first step in this analysis is to establish a database of production, consumption and exchanges (import and export) at the various geographic levels of interest, and for the various types of material of interest. In practice, the finest scale of available data is a French département, and the publicly available data refer to the national, regional or “departemental” level. Only major primary materials are accounted for, through the content of end products and waste in these primary materials (toxic waste are accounted for separately). For example, for cereals such as wheat, production at the departement level is available through the national Agreste database, variations of stock are small once averaged over a few years period, import and export are obtained from the Sitram database (a database initially elaborated by the ministry of transportation and now maintained by the Ministry of Environment), which follows all national and international transport by transportation mode and by type, through annual stratified polls of transportation companies. Productions of non-agricultural products in France is very low except for construction materials (most notably cement), for which the industry maintains its own publicly available database. Following transformations requires information from various industrial sectors, e.g., the flour trade and food industry for wheat use, taking into account animal farming which consumes a non-negligible fraction of primary agricultural products.

Once this database is constructed, one also needs to estimate production, consumption and imports and exports at finer scale than the departement. In practice, this is performed by correlating the desired information at the national, regional and departemental scale with another auxiliary quantity serving as proxy, that is also known at the desired smaller scale. For example, wheat production can easily be correlated with available surfaces in wheat growing areas, that are known from the Corine Land Cover database at scales of the order of a few hundred meters. More generally, auxiliary quantities are constructed from relevant demographic and economic and geographic data, that are mostly available through the various INSEE databases. This requires some educated guess-work to find the most likely auxiliary quantities, and evaluate their correlation with the quantities of interest at scales where data on both are available. This aspect of the problem has been completed only for food staples at this stage.

An important aspect of the problem is to estimate the errors in the data. Errors can be detected when quantities of a given material are not conserved through transportation and transformation processes. It appears that the largest source of error comes from the transportation database, because the stratified polling methodology is optimized with respect to total transport from a pair of origin and destination, independently of the nature of the transported goods. It is in principle possible to compute confidence intervals per type of material and not only on total volumes of exchanges, but this requires access to some non public information. Discussions have been initiated with the Ministry to have access to this information, in order to estimate the reliability of this method of transport quantification. If the Sitram database turns out to be too imprecise, the method described above to estimate lacking data can be applied to transport as well with appropriate auxiliary quantities, but the results also suffer from various sources of error.

This first stage of the Material Flow analysis is nevertheless largely underway. The two next steps consist in environmental impact evaluation on the one hand, at the present date, and in developing a method of analysis of changes of such impacts under various policy scenarios and options. Bith will rely on the use of Life Cycle Analysis databases, as mentioned above.