-
Notifications
You must be signed in to change notification settings - Fork 13
Calculating input data for development potential
The following outlines the preprocessing of different input data that can possibly be used to explain and drive urbanization in the FUTURES model including different environmental, infrastructural and socioeconomic factors. Based on a function of these different attributes the probability of urbanization is calculated. The influence of these different factors can be empirically derived, for example in a multilevel model, or based on prescribed heuristics. In both cases data layers are required for spatially deriving locations of probable urbanization. Raster or gridded data is used as the main input, which often requires pre-processing.
Development pressure is an important factor driving urbanization related to proximity to existing urban land use. The influence of the proximity of urban areas can be calculated in the grass module provided with the GRASS FUTURES extension to GRASS found here. FUTURE user are encouraged to calculate different level of urban influence based distance decay testing for the appropriate spatial outcomes. A factor of between 0.5 and 1.0 has been found to be the most reliable for application of FUTURE simulation in case studies thus far.
Transportation networks highly influence urbanization patterns and are important inputs in the FUTURES Model. In previous case studies we have used and developed map layers depicting road density, distance to interchanges and travel cost to urban centers as predictors of urbanization. To derive these layers the user will need to obtain shapefiles of the road networks for their study area. Shapefiles of road networks can be obtained from many different sources. Road files can be obtained for the US from the US census bureau here. Open Street map also offers transport networks for numerous countries. Road density can be derived by aggregating the number of roads within a neighborhood. This requires the road network shapefile to be converted to a raster file. The GRASS tool v.to.rast is advisable for converting the shapefile. The resulting raster layer can be processed using r.neighbors using the mean function to develop a surface of the percentage roads in a neighborhood. Different neighborhood extents (e.g. 1km) can be tested for their influence on urbanization. Distance to interchanges can be derived using interchange points. Euclidean distance from these locations can be used to quantify the magnitude of influence these road junction can have for urbanization. Points can be converted into raster like the above described methods and r.grow.distance used to calculate the distance to these features. Travel cost is calculated by the cumulative cost of moving between different geographic locations on an input raster map whose cell category values represent cost. The raster map whose cells represent cost can be based on the actual road network with roads being given the approximate time cost based on speed limits. For example a 1 km pixel would be assign a value of 36 seconds if the speed limit was 100km/hour. Traveling 100 pixels on that same road (100 km) would result in the pixel being assigned 3600 seconds or 1 hour travel cost. Application for the FUTURES model can use to calculate travel time to location hypothesized to influence urbanization (e.g. urban centers, lakes). The r.cost extension in GRASS is recommended for this calculation. The weighted cost layer can be made by converting the road network to a raster surface assigning coast based on logical speeds for travel on these roads (e.g. interstates have higher values than secondary roads). It is also advisable to assign no road areas a typical travel cost as a least distance path is assumed for all cumulative network calculations.
The conversion of land to urban areas is driven by proximate land use. Land use can influence transaction cost for instance with forested land being more expensive to develop. It may also act as an influence that attracts land conversion as an attraction for instance in forested area or by lakes and streams. There are several ways to obtain land cover/use data. The National Land Cover Database (NLCD) offers consistent land use data for the entire US. Similar databases are available for many different countries. Alternatively user can provide their own land cover data based on classification of aerial photo/imagery. Such data can be used as dummies indicting where urbanization can occur. Landscape scale analysis can leverage neighborhood tools in the GRASS module r.neighbors to assess percentage land cover for an appropriate spatial extent. For example in previous case studies percentage forest cover for a 1 km neighborhood has proven to be an important factor explaining urbanization.
Topography influences urbanization interacting as a feature related to agricultural suitability and as a potential attraction related to urban developments featuring 'good views'. To access its influence as a driver of urbanization different characteristics can be developed from digital elevation models (dem). DEMs can be obtained for the US from USGS and various other avenues globally. Different topographic characterisitics can be derived using the GRASS tool r.slope.aspect including slope, aspect, curvatures and partial derivatives from a elevation raster map.
#Attractions Amenity landscape are increasingly sought for land development, which is increasingly shaping exurbanization. Proximity to feature deem aesthetically pleasing including water bodies and protected areas can be calculated using network or euclidean distance similar to the above descriptions. Shapefiles of protected areas can be found at the GLCF: World Database on Protected Areas. Water bodies can be found for the US from the US census bureau here with European resources available from the European Envrionmental agency here
#Other Different factors can influence urbanization that are not summarized here. Soil fertility, socioeconomic factors including employment, GDP, can also be used as factor explaining/used for characterizing development potential in the FUTURES model. Different GRASS modules can be leveraged for preprossessing these data. Users of the r.futures are advised to consult the GRASS Manual for processing these data.