-
Notifications
You must be signed in to change notification settings - Fork 0
Functional rural areas
Functional rural areas are the rural counterpart of functional urban areas. A functional urban area is defined by commuting into a city; outside the reach of any city that definition gives nothing, and large parts of the world are left as residual territory. The model builds a regionalisation for that territory instead, by merging catchments of rural centres into larger areas until they reach a workable size.
This work was carried out with the OECD in mind, and the country lists in the batch scripts reflect the countries requested for it.
The starting point is a catchment set, by default the m10_ex_fua variant: catchments of settlements that are a centre at 10 minutes, with the OECD functional urban areas excluded. See Catchments.
The analysis then runs per country, over the catchments of that country, using the same continental network for routing between them.
19 variants are configured, each a combination of three settings:
| Setting | Meaning |
|---|---|
| ideal population | the target population of a functional rural area, from 25 000 to 4 million |
| maximum time | the travel time within which merging is allowed at all |
| minimum time | the travel time below which an area is considered too small to stand on its own |
Variant names are built from these values, for example idealpop005_maxt60_mint20. A travel time aggregation method is defined as well, either the mean travel time or the travel time between the largest centres.
The wide range of the ideal population setting is deliberate. It is the parameter that decides whether the result is a fine grained regionalisation of a densely populated country or a coarse one for a sparsely populated territory, and the variants exist to make that trade off visible rather than to fix it in advance.
A travel time matrix is computed between all catchments of the country, cut off at the largest maximum time. A polygon connectivity graph gives which catchments physically touch.
The procedure then merges iteratively. In each round, an area that is below the minimum time or below the target population looks for a partner among its neighbours, scored on travel time, population and contiguity, and the best pairs are merged. The merged set becomes the input for the next round, and the process stops when no candidate scores above the acceptance rule.
The result is a set of areas that each reach a reasonable population and internal travel time, and that together tile the rural part of the country.
Per country and variant, to %LocalDataProjDir%/Functional_Rural_Areas/<settlement set>/Base_<threshold>/<variant>/<continent>/<country>.shp, with the area geometry, its name, its population and a flag telling whether it is a functional urban area rather than a rural one.
Two aggregations are produced on top of this:
- a mosaic per variant across countries, in
mozaiks/, together with statistics per country: the number of zones, the number of functional rural areas, how many reach the ideal population, and the mean and standard deviation of area size and population; -
Analyses/count_fra_results, a small container that counts the defined areas per variant for a fixed list of countries, used to compare the variants against each other.
GeoDmsRun.exe cfg\main.dms /Analyses/Functional_Rural_Areas/Per_Continent/<continent>/iterate_by_Memberstates/<country>/ResultsByVariant/button/runAllbatch/gen_functional_areas.cmd loops this over the country lists. Running all 19 variants for one country in one call is what runAll does.