Municipality classification by land use patterns

This is an ongoing investigation of the land use patterns of my homeland, Finland.

Finland is a country with a huge variability in the land use patterns within administrative municipalities (cf. LocalFinland.fi for background information). The major towns are packed with infrastructural and residential areas filled with productive economical activities, whereas the administrative burden of the northern municipalities extend over large, practically uninhabitated areas of forests, and only a small portion of the administrative area is affected by economically active infrastructure. kuopio_corine

The local authorities have a broad responsibility for the provision of basic services to citizens over the whole spatial extent of the municipality. From this responsibility, the artificially generated administrative borders define the spatio-economical balance setting the prerequisite for both the level of services provided and for the ultimate growth potential and fault tolerance of the society.

There has been an ongoing political pressure for extensive municipal mergers for the sake of optimizing the economic structure of the country as a whole (see this article for more information and comparison with Canada), which further hampers the local authorities from providing accessible services for the whole population without major tax effects. In the end, the municipal mergers are an unavoidable consequence of the two factors: less-than-optimal spatial structure of the land use and the local economical responsibility to provide basic services.

finland_urbanproportion
Urbanization index from land use. (coloring by quantile)
finland_unhabited
Natural resources index from land use.
finland_pop_dens_2009_4quant
Population density.
property_tax
Property taxation

 

To assess some of this variety in land usage within the municipalities, I applied the CORINE 2006 land use and land cover classification data (SYKE) to analyse some patterns of the land use characteristics of the administrative regions. Using Quantum GIS with Python-interface and MATLAB for cluster analysis, I produced a characterization of the relative land use patterns of each municipality in Finland. Note that the population density of a given administrative region is strictly proportional to the residential land use characteristics of the municipality, hence the population density is neglected in the analysis.

Below is a detailed example of the relative land use patterns of five municipalities ordered from South to North (the towns are marked on the map at the right hand side). As the point of interest travels from South to North, the area of the administrative region and the sparsity of societal infrastructure increase hand in hand.

The land use characteristics of five municipalities

From South to North with respect to the finlandmap above: For each region, the leftmost map presents the spatial pattern of the land use of the given municipality (maps are not in scale). The central square presents the normalized relative areas that each land use class occupies, and the same data is presented by the bar graph at the right with land use class legend embedded.

Helsinki

TampereKuopioOuluRovaniemiTurku

Land Use Index and Cluster analysis

The given interpretation of the administrative regions by their relative land use characteristics introduces a way to measure and compare the municipalities regardless of differences in their spatial scales. It also allows to investigate the effect that the artificial administrative boundaries have on the economically coupled societies, as they are separated by the municipality boundaries.

The land use index is here defined as a vector composed of the land use ratios – in practice the land use index is the numerical representation of the bar graphs on the image sequence above. With this index, it is possible to perform cluster analysis to find stereotypical land use indices and classify municipalities by their land use patterns. I performed a simple k-means clustering and ended up with three optimal clusters, see the figure below.

finland_kmeans3
Classification of the municipalities in 3 classes. The clusters are presented on the right along with the cluster spread (standard deviation of the values within the cluster). Note that the scales are adjusted for normalization purposes.

 

finland_kmeans3_clusters

The clustering reveals the three expected differentiating factors between the municipalities: residential/industrial, agricultural and natural state land use. These classes may be characterized as

  1. Rural or uninhabitated areas (light brown). These regions have only minor infrastructure and natural resources are mostly unharvested.
  2. Urban areas, water basis (brown). These regions have begun urbanization and utilize mostly the major water bodies to provide the residential activity with resources (this is the Finnish Lakeland, which has developed through the wood processing industry with transport routes via the water channels).
  3. Urban areas, agricultural basis (turquoise). These regions have constructed an active urban infrastructure and begun to transform existing natural resources into agricultural activity.

One may argue that the rural areas (light brown) do not possess any real prospects to cope with the economical and infrastructural burden of serving their societies. Nor does these regions have resources to monetize the available local resources, and without governmental subvention the multinational industries have an edge over the local administration to harvest the land. This is a problem which has required governmental actions, most notably regionalization (VATT).

So what?

So, how to deal with this absent child of Finland? Cut the whole part off and wish for the best (Finland A below)? Or leave the disconnected part of the country on its own and wait for the organic growth (Finland B)? Or do nothing (Finland C)?

new_finland_1
Finland A
new_finland_2
Finland B
new_finland_3
Finland C

 

To be continued – this is an ongoing investigation of the homeland of strange retired people.

Published by

Vesa

Flâneuring through experiments in data, science and artsy digital things.