EN Scope of application

This is an English translation of the Dutch page Toepassinggebied, which is leading. Translated on 8 November 2026 from the Dutch text committed together with this page, on top of wiki commit 428045e.

Status and use

RSopen is research software and under active development. The outcomes are explorations under assumptions about scenarios and policy, not predictions. The software comes without warranty and without liability, as sections 15 and 16 of the GPL v3 provide. Anyone who uses outcomes for advice or a decision checks them independently and states which version of the model was used: the branch and the commit or the tag.

Purpose and scope

The RuimteScanner is a spatial modelling instrument that simulates the distribution of the demand for and supply of space at national and regional level within policy and physical constraints, and computes the impacts of that distribution. The model is used for policy explorations and spatial analyses that require insight into:

the spatial consequences of programmes for housing and work locations, and of requirements for water storage, nature and agriculture;
the influence of planning restrictions and environmental zoning;
the distribution of extension and redevelopment locations across the Netherlands under different policy assumptions;
the impacts of all this, such as flood damage, greenery in the residential environment, accessibility, carbon and dwelling value.

The RuimteScanner is not intended for predictions at parcel level or for detailed planning, but for strategic explorations at levels such as provinces, COROP regions and municipalities.

Generic assumptions and limitations

This table is about the allocation. The assumptions and limitations of the impact modules are in the section Impact modules and indicators below.

Theme	Assumptions / Limitations
Purpose of the model	The model simulates new construction on extension locations and the redevelopment of existing dwellings: a redevelopment probability (in Dutch) per cadastral parcel, estimated on demolition and new construction in the BAG that adds dwellings, determines when a parcel is demolished (#869), and the allocation rebuilds the dwellings preferably on the parcel itself, with a package that is denser than what stood there (#860). A separate replacement flow replaces dwellings one for one on the same spot and does not change the state (#859). Limitations: redevelopment applies only to housing and only in a region that does not yet meet its housing claim, the model demolishes dwellings that are too old in the city, and what the model builds itself is not redeveloped again (#735).
Input dependencies	The demand for space for housing and employment comes from TIGRIS XL; the other sectors get their requirement from trends, policy targets, a delivered requirement or an imposed map. The initial state comes from the BAG of the Kadaster (dwellings and buildings), LISA (jobs) and the Land Use Statistics of Statistics Netherlands (land use), the dwelling prices from a hedonic estimation on NVM transactions and the parcels from the BRK. Errors and choices in those sources carry through directly.
Restrictions	Planning and physical constraints (such as Natura 2000, water and safety zones) come as fixed maps per variant from a separate project, see Restrictions (in Dutch). Policy therefore does not change endogenously. Per projection year only two things change: whether planning capacity still gives an exemption (`PlancapaciteitGeldigTotEnMet`), and the environmental zoning that grows around work to which the allocation itself gave industry or logistics.
Expert judgement	Part of the input rests on expert judgement: the division of restrictions by hardness (very hard, hard, medium, soft), the environmental category of work locations for which IBIS does not deliver one (`AanvullenBijOnbekend`) and the construction methods that a job subsector can use (#727).
Border effects	The study area is the Netherlands. What lies across the border does not count in the allocation nor in the measures of the surroundings, such as the nature and water fractions in the dwelling price. Near the national border, suitability and indicators therefore miss part of the surroundings.
Pull factors (public transport, amenities)	Accessibility (the travel time to 500,000 inhabitants and to a public transport hub) and amenities (the UAI) are part of the dwelling price, and similar measures are part of the suitability for employment, with coefficients from empirical estimations. Those maps are fixed and do not change with what the model builds, so a future public transport project is not included.
Policy and feedback	Policy responses to spatial pressure, such as adjusting restrictions or programming, are not dynamic: the model takes policy as given.
Interaction between sectors	Between sectors there is no price mechanism. The sectors are allocated one after another in the order of `ModelParameters/SectorAllocRegio`, and what one sector displaces of another comes back as a residual claim. Within a sector, the subsector with the highest empirical suitability wins a cell; for employment, priority rules for existing buildings, plans and stimuli come first. For housing that suitability is the development balance in euros per development package: the revenue from a hedonic price function against the construction, land production, acquisition and demolition costs, see Development balance housing (in Dutch). Employment and holiday accommodation choose on a probability from a logit model, without a cost side. Agriculture maximises the total suitability per region, the net margin per crop, within a lower and an upper bound per agricultural class. The revenues in the development balance do not respond to demand and supply; only the nature and water in the surroundings feed into the price per projection year.
Time dimension	The model computes a series of projection years, from a configurable first projection year at a fixed interval, and each projection year builds on the state of the previous one. The allocation is therefore path-dependent: computing straight to the final year (`AlleenEindjaar`) gives different outcomes than computing per projection year. Between the projection years there is limited feedback: the dwelling prices follow the nature and water at the start of the projection year, the suitability for employment follows the land use allocated earlier, and the redevelopment probability accumulates per projection year. The claims do not respond to the allocation and stay the same after `FromWhichDoWeKeepTIGRISconstant`. See Time dynamics.

Model behaviour compared with observations and system knowledge

The spatial patterns follow from mechanisms that rest on system knowledge and observations:

Housing ends up near existing settlements and public transport when the ladder of a variant gives those locations priority, as the BAU ladder does with the axes existing built-up area and public transport accessibility; see Rungs (in Dutch).
Growth stays away from areas with strict restrictions (such as Natura 2000, water and safety zones), because the sieve excludes those cells; see Availability.
The suitability for employment is estimated on the observed growth of the building footprint, with among other things the distance to motorway junctions, seaports and railway stations and the land use in and around the cell as explanatory variables; see Suitability.

The allocation as a whole is not calibrated on observed spatial development. Components are estimated on observations: the dwelling prices on NVM transactions from 2000–2023, the suitability for employment on the growth of the building footprint between 2012 and 2022, and the redevelopment probability on demolition and new construction in the BAG. A test harness (Diagnose.dms) sets a number of outcomes against a norm, such as the expected redevelopment per year against the demolition on profitable locations in the BAG. Such a comparison also shows where the model deviates: the level of redevelopment falls within the norm, but in the city the model demolishes dwellings that are too old; see Redevelopment probability (in Dutch).

Locally the outcomes can deviate because of:

local policy choices that are not modelled;
limited precision of input files (planning capacity, restrictions);
prices that do not respond to demand and supply.

Impact modules and indicators

Next to the allocation, the model computes a second layer: impact modules that turn the state after the allocation into indicators, per case and per projection year, and partly also for the base year. The index is on Impact modules and indicators. There are five themes:

land use and agriculture: the land-use map, land-use change, sealed surface, fertile agricultural land and yield loss;
allocation: claim realisation, overflow and the density of what has been built;
housing, value and costs: the value of new dwellings and the change in dwelling value due to changes in greenery, demolition and new construction, the dwelling stock by construction year, buy-out and demolition costs, the costs of housing construction and of work locations, the land development budget and foundation damage;
climate, water and soil: CO2 emissions from peat (with SOMERS), carbon storage, flood damage (with HIS-SSM), cloudburst storage, rainwater reuse and water storage;
greenery, accessibility and health: the accessibility of greenery and of jobs, and mortality due to changes in greenery.

Separate from the allocation, with a few exceptions

The impact modules run separately from the allocation: they read the state after the allocation, and the allocation reads no indicator back. A different ladder or a different density therefore changes the indicators, but an indicator does not change where the model builds. A number of calculations do exist on both sides, with the same formulas, and there an impact feeds into the suitability:

The nature and water fractions of Dwelling value due to changes in greenery (in Dutch) correct the dwelling prices in the allocation per projection year. It is the same template (Templates/Suitabilities/Groenfracties_T), fed with the state at the start of the projection year instead of at the end.
The foundation damage (in Dutch) lowers the acquisition costs of existing dwellings in the development balance, in a variant in which the damage is not repaired.
The construction of the measures for cloudburst storage (in Dutch), with the same fractions per variant as the indicator, and the additional costs of subsidence from Costs of housing construction (in Dutch), if the variant switches that item on, are items in the development balance.
The yield loss (in Dutch) goes directly into the suitability for agriculture.

No indicator feeds the sieve directly. The items above do carry through into availability, through the threshold on the development balance in the sieve per development package (MinimumExploitatieSaldo). The flood damage itself does not count in the allocation, but the sieve and the choice of construction method read the same depth maps.

What the indicators are and are not suitable for

The indicators are suitable for comparing variants and projection years with each other: the same calculation rule on a different state. For an absolute figure the assumptions of each indicator apply, and those are on its own page. A few limitations apply to many indicators at once:

An indicator sees what the model changes. What the model does not touch is included with the state or the observation of the base year; for example, the accessibility of greenery (in Dutch) does not see greening or densification within an existing neighbourhood that is not allocated again.
Part of the input is fixed, such as the travel times of the accessibility of jobs (in Dutch), which use the network of a fixed year.
Damage is gross: the flood damage (in Dutch) is computed without evacuation and without casualties.
Amounts are given nominally and as net present value in the base year, see Discounting (in Dutch). The dwelling values depend on the price level and the relationships of the hedonic estimation, and those relationships are correlational.
The tables count per regional division on the cells of 25 metres; a national figure from a table can differ slightly from the same sum over the map.
The claim realisation (in Dutch) is diagnostic: it shows whether the claims could be fitted spatially, and the regular measure can be misleadingly high where the claim lies below the state of the base year.

For a statement about a single cell, dwelling or parcel the indicators are not suitable: they inherit the limitations of the allocation, which distributes a regional demand and does not predict a place.

Main uncertainties

Policy development: future relaxation or tightening of restrictions (for example nitrogen or nature policy) can have a large influence, but is assumed to be exogenous.
Spatial data: plan files and policy boundaries are compiled from different sources and contain uncertainty of interpretation.
Transferability of estimations: the price function, the suitability for employment and the redevelopment probability are estimated on an observation period and applied to projection years long after it; the assumption is that the relationships are transferable. Restrictions do not work through an estimated sensitivity but as an exclusion in the sieve.
External models: uncertainties in the TIGRIS outcomes carry through directly into the RuimteScanner. After the last year for which the claims run, demand no longer grows, and differences between variants then come from the spatial steering and not from extra demand.
Temporality: the lack of dynamic feedback between demand, supply and policy increases the uncertainty over longer time horizons.

Suggested uses and limitations

The RuimteScanner is suitable for:

Policy explorations, for example with the WLO scenarios, for the Planmonitor NOVI and under climate scenarios;
Spatial impact analyses at regional or national level;
Comparison of policy variants in terms of land take, distribution and impacts.

Not suitable for:

Predictions at parcel or project level; which parcel is redeveloped follows from a fixed draw against the probability and is not a prediction;
Valuation of individual plan locations;
Valuation of land or real estate: the prices in the model are hedonic estimates per cell and dwelling type and not a valuation of an object.