FUTURES v3.0.0

What is FUTURES

FUTURES is an open source urban growth model designed to address the regional-scale ecological and environmental impacts of urbanization; it is one of the few land change models that explicitly captures the spatial structure of development in response to user-specified scenarios. See also official website with list of publications, manual pages and tutorials.

What's Changed

New major release of FUTURES brings a climate forcing component that models flood impact on future development and integrates human response to flooding using a response function that relates adaptive capacity and level of flooding damage sustained. This component does not have to be used, in that case the simulation behaves the same way as in FUTURES v2. More information on how to use this component is coming. Eventually, this version will replace v2 in GRASS GIS addons.

Compatibility between v2 and v3:

Module r.futures.pga has been renamed to r.futures.simulation. This release includes a wrapper module that ensures tutorials are still working.
Also note that column names in potential file header are now expected to match the names of the predictor rasters, but the order can be different.

FUTURES v2.4.0

What's Changed

Module r.futures.pga:

• fix memory estimation (aeee528)
• give up searching for seed when it can't find it soon enough (92486c4)
• reduce unnecessary memory allocation for undeveloped cells, use better estimate (a585af9)

Module r.futures.pga will be renamed in FUTURES v3 to r.futures.simulation. This release adds a r.futures.pga wrapper called r.futures.simulation that can be used now (2b83d7f).

Full Changelog: v2.3.0...v2.4.0

FUTURES v2.3.0

What's Changed

• r.futures.potential: add random effect (random_column option) in #41
  • developed_column ~ predictor1 + predictor2 + (random_column | subregions_column)
• r.futures.devpressure: add nprocs parameter for parallel execution in #43 (requires GRASS GIS 8.2 to take advantage of that)
• r.futures.calib: parallelize patch library generation in #44

Full Changelog: v2.2.0...v2.3.0

FUTURES v2.2.0

This release brings new validation modules, speeds up r.futures.potential and includes better handling of some edge cases.

• add new modules for validation (#36)

  • includes allocation/quantity disagreement, cohen's kappa, Kappa simulation (Vliet et al, 2011)
  • r.futures.validation: global validation
  • r.futures.gridvalidation: validation on a grid (useful for large areas)

• r.futures.pga

  • fix for case when there are no derived patch sizes
  • fix reading patches when running on a subset of subregions (#29)
  • better verbose message showing progress (#28)
  • check all subregions in map are in potential file (#31)

• r.futures.potential: model selection improvements (#23)

  • speed up with parallelization of dredge function (requires R snow package)
  • option to output file summarizing all models
  • add option fixed_columns to specify predictors that should always be included

• r.futures.demand

  • fallback method for inverse relations that keeps density static (#24)
  • if there is zero in population data, discard data point (#22)
  • do not exit if using only 1 method which can't converge
  • fix matplotlib deprecation warning (#19)

• r.futures.parallelpga and r.futures.calibration - add memory option from r.futures.pga

FUTURES v2.1.1

This release includes important bug fixes in r.futures.pga, lower memory demands of r.futues.pga, and fixes for edge cases in r.futures.calib and r.futures.demand.

• r.futures.calib
  • fix error when null cells are present
  • fix errors when no new development found or when only 1 patch found
• r.futures.demand
  • better error handling for logarithmic edge cases
  • write columns only for subregions present in map
  • improved inverse cases (population decreases as developed area increases)
• r.futues.pga
  • fix potential_weight for w < 0
  • aggregate static predictors to save memory
  • fix segfault when growing patches - check boundaries to not add candidates outside of the extent

FUTURES v2.1.0

This release brings new improvements and several bug fixes. The main improvements are code testing with Continuous Integration and faster calibration procedure that can be run in parallel. Additionally, patch library can be now derived for each subregion to better reflect the patch size distribution for different (rural, urban) areas. See v2.0.0 release for installation instructions and other resources.

Changelog

General

• Added tests for all modules.
• Tests running in Continuous Integration (CI) using GitHub Actions (#6).
• CSV files standardized to use comma as separator by default (#5). Separator can be changed.
• Change Python shebang to Python 3

r.futures.pga

• Patch library is now a CSV file allowing separate library for each subregion.
• Fix bug in seed selection for 'random' method.
• Better reporting when there are not enough undeveloped sites (#4).
• Fix compilation warnings.

r.futures.calib:

• Replaced vector-based patch computations with raster based using r.object.geometry addon. It is faster and can be run in parallel easily. Needs new dependency, r.object.geometry addon, install with g.extension r.object.geometry.
• Improved output that automatically presents the best parameter selection.
• Fixed some missing/old interface options related to r.futures.pga.
• Added random_seed option for reproducibility.
• Optionally derives patch library for each subregion.

r.futures.parallelpga

• Fixed some missing/old interface options related to r.futures.pga.

FUTURES v2.0.0

We release new major version of FUTURES model to share improvements in r.futures.pga module, which went through major code refactoring bringing several improvements:

• r.futures.pga speed up (especially for large development pressure neighborhood)
• enable running r.futures.pga on large regions by enabling to manage memory on disk if needed
• allow to use different subregions for DEMAND and for POTENTIAL submodels
• Python 3 support for all r.futures.* Python modules

To quickly try FUTURES, follow online tutorial.

To install FUTURES, install and start GRASS GIS and then install r.futures addon from Settings -> Addons extensions -> Install extension from addons or using command line:

g.extension r.futures

See also official website with list of publications, manual pages and tutorials.

FUTURES v1.0.0

First official release of FUTURES model as GRASS GIS r.futures.* modules.

FUTURES is an urban growth model specifically designed to capture the spatial structure of future development.
Features include:

• stochastic patch-based spread to mimic structure of urban development
• scenario development (infill vs. sprawl, locally increase/decrease probability of development)
• positive spatial feedbacks
• spatial non-stationarity
• flexible predictor inputs
• runs on all major platforms
• graphical user interface

To install FUTURES, install and start GRASS GIS and then install r.futures addon from Settings -> Addons extensions -> Install extension from addons or using command line:

g.extension r.futures

See also official website with list of publications, manual pages and tutorials.