interpolate2: Interpolate 2

In pik-piam/luscale: PIK Landuse Group Data Scaling Tools

Interpolate 2

Description

Disaggregates a cellular MAgPIE output to 0.5 degree based on the given mapping and information about the initial 0.5 degree distribution.

Usage

interpolate2(x, x_ini, map, x_ini_lr = NULL)

Arguments

x	The object to be disaggregated. See details for further important information.
x_ini	The initial distribution of x in high resolution.
map	A relation map between low and high resolution
x_ini_lr	Low resolution version of x_ini. Will be calculated automatically, if not provided. Can speed up computation, if provided.

Details

There was a now deleted function called interpolate in this package before, hence the name interpolate2.

The function is based on the following assumption: x is an object in low resolution with more than one data dimension and the sum over the data dimensions is constant over time. Example: One column cropland, the other one (cell size - cropland). x_ini provides the same type of data as x but in high resolution and for the time step previous the initial time step of x (e.g. if x goes from t=a to t=a+10, x_ini must be provided for t=a-1). The function calculates the amount by which the individual data columns of x change in each timestep. The output is based on x_ini and only the differences in later timesteps to ths starting point are disaggregated by the given mapping. This assures that as little information as possible is lost from the original dataset x_ini.

The disaggregation procedure itself works as follows:

1. Differences in distribution between years are derived for the low resolution data set.

2. Based on these differences extension and reduction shares are calculated for the different pools. Reduction shares are calculated relative to the pool itself (e.g. a reduction in a cropland pool from 10ha to 6ha leads to a reduction share of (10ha-6ha)/10ha = 40%). At the same time extension shares are calculated relative to the pool which was made available by reductions of the other pools (e.g. cropland is reduced from 10ha to 6ha, forest area is reduced from 2ha to 1ha, but pastureland increases from 20ha to 25ha. In this case the extension share of pasture will be (25ha-20ha)/(10ha-6ha+2ha-1ha)=5ha/5ha=100%). This difference in calculation of reduction and extension share is crucial for the application at the high resolution level because otherwise the calculation will not add up.

3. Reduction and extension shares are disaggregated to the high resolution level by just assigning the same low resolution shares to all belonging cells at the higher resolution.

4.Starting with the provided high resolution pool data set for the initial year reduction shares are applied on all pools in all cells. The pool which is made available for expansions is calculated by summing up all values which were released by the pool reductions.

5. Pool expansions are calculated based on the pool made available in 4 for the first time step.

6. Steps 4 and 5 are repeated for all the following years based on the newly created high resolution data.

Applying this procedure makes sure that relative pool reductions are identical for the low resolution cell and all belonging high resolutions cells whereas the extension shares relative to the areas made available per cell are identical between low resolution cell and belonging high resolution cells.

Value

The disaggregated MAgPIE object containing x_ini as first timestep

Author(s)

Jan Philipp Dietrich

See Also

toolAggregate

pik-piam/luscale documentation built on April 27, 2024, 2:11 p.m.